In vitro investigations of the potential chemotherapeutic and ...

University of Wollongong University of Wollongong

Research Online Research Online

University of Wollongong Thesis Collection 2017+ University of Wollongong Thesis Collections

2018

In vitro investigations of the potential chemotherapeutic and In vitro investigations of the potential chemotherapeutic and

chemopreventive properties of bismuth(III) complexes of non-steroidal anti-chemopreventive properties of bismuth(III) complexes of non-steroidal anti-

inflammatory drugs inflammatory drugs

Emma Louise Hawksworth University of Wollongong Follow this and additional works at: https://ro.uow.edu.au/theses1

University of Wollongong University of Wollongong

Copyright Warning Copyright Warning

You may print or download ONE copy of this document for the purpose of your own research or study. The University

does not authorise you to copy, communicate or otherwise make available electronically to any other person any

copyright material contained on this site.

You are reminded of the following: This work is copyright. Apart from any use permitted under the Copyright Act

1968, no part of this work may be reproduced by any process, nor may any other exclusive right be exercised,

without the permission of the author. Copyright owners are entitled to take legal action against persons who infringe

their copyright. A reproduction of material that is protected by copyright may be a copyright infringement. A court

may impose penalties and award damages in relation to offences and infringements relating to copyright material.

Higher penalties may apply, and higher damages may be awarded, for offences and infringements involving the

conversion of material into digital or electronic form.

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily

represent the views of the University of Wollongong. represent the views of the University of Wollongong.

Recommended Citation Recommended Citation Hawksworth, Emma Louise, In vitro investigations of the potential chemotherapeutic and chemopreventive properties of bismuth(III) complexes of non-steroidal anti-inflammatory drugs, Doctor of Philosophy thesis, School of Chemistry, University of Wollongong, 2018. https://ro.uow.edu.au/theses1/429

Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected]

https://ro.uow.edu.au/

https://ro.uow.edu.au/theses1

https://ro.uow.edu.au/theses1

https://ro.uow.edu.au/thesesuow

https://ro.uow.edu.au/theses1?utm_source=ro.uow.edu.au%2Ftheses1%2F429&utm_medium=PDF&utm_campaign=PDFCoverPages

In vitro investigations of the potential chemotherapeutic and chemopreventive properties

of bismuth(III) complexes of non-steroidal anti-inflammatory drugs

Emma Louise Hawksworth Bachelor of Medicinal Chemistry (Honours Class I)

This thesis is presented as part of the requirement for the conferral of the degree: Doctor of Philosophy

The University of Wollongong School of Chemistry

March 2018

ii

ABSTRACT

Colorectal cancer (CRC) is the second most common cancer in both men and

women in Australia. The upregulation of cyclooxygenase-2 (COX-2), an enzyme

involved in the inflammation response, has been linked to a number of cancers,

including CRC. This has prompted a number of epidemiological and clinical studies

that suggest that prolonged administration of non-steroidal anti-inflammatory drugs

(NSAIDs) can reduce the incidence of cancer, in particular, CRC. However, side effects

such as gastrointestinal (GI) bleeding limit the prolonged daily use of (NSAIDs) for the

chemoprevention of cancer.

Bismuth, a group 15 post-transition metal, has been used for centuries to treat an

array of GI diseases. The Bi formulations exhibit an acceptable toxicity profile when

administered at low doses over extended periods. Thus, it is hypothesised that the

combination of Bi and NSAID in a single compound may potentially relieve the adverse

GI side effects of NSAIDs if used daily as a chemopreventive agent.

The aim of this Thesis was to determine the suitability of Bi complexes of NSAIDs

(BiNSAIDs) as potential chemotherapeutic or chemopreventive agents. The starting

point was to assess the ability of BiNSAIDs to interact with transformed cells (HCT-8

human ileocecal colon cancer cells) using a selection of BiNSAIDs of the general

formula [Bi(L)3]n, (where L = diflunisal (difl), mefenamate (mef) or tolfenamate (tolf)).

The hydrolytic stability of the BiNSAIDs in cell medium was investigated to

determine the biologically relevant structures that the cancer cells were exposed to in

the cell assays. NMR spectroscopy of high concentrations of BiNSAIDs in cell

medium (24 h, 37 °C) indicated that their structural stability and interactions with cell

medium components were specific to the complex. Specifically, Bi(tolf)3 appeared to

remain intact, whereas Bi(mef)3 and Bi(difl)3 were affected by interactions with the

cell medium.

The in vitro cytotoxicity, mode of cell death, and the ability of the BiNSAIDs to

inhibit COX in HCT-8 cells were investigated. Assessment of cell viability using the

[3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium] bromide (MTT) assay showed

that the toxicity ranking of the BiNSAIDs paralleled those of the respective free

NSAIDs: diflH < mefH < tolfH. While the IC50 values of the BiNSAIDs (ranging between

16 and 81 μM) were lower than the free NSAIDs (ranging between 37 and 403 μM),

it was apparent that the toxicity of the BiNSAIDs was in reasonable agreement with

the molar ratio of the three NSAIDs per BiNSAID, with the exception of Bi(difl)3.

iii

Phase contrast microscopy and Annexin-V/7AAD staining showed that treatment with

all BiNSAIDs and NSAIDs induced morphological changes indicative of cell death via

apoptosis. Recognising that over-expression of COX-2 is associated with CRC, the

ability of the BiNSAIDs to inhibit the production of PGE2 (and COX) in HCT-8 cells was

assessed. It was observed that the inhibition of PGE2 production by the BiNSAIDs

paralleled that of the respective free NSAIDs: diflH < mefH < tolfH. These results

indicated that the COX inhibition of the examined NSAIDs was not substantially

diminished by their coordination to Bi (in the BiNSAIDs).

Bismuth uptake was quantified using graphite furnace atomic absorption

spectroscopy (GFAAS) to establish whether BiNSAID stability and toxicity correlated

with cellular uptake. The highest cellular Bi content was observed following

treatment with Bi(tolf)3. Since NMR studies indicated that Bi(tolf)3 was the most

stable BiNSAID it appears that Bi uptake is assisted by the NSAID. Furthermore,

microprobe synchrotron radiation X-ray fluorescence (SRXRF) imaging was utilised to

determine the subcellular fate of the Bi contained in the BiNSAID complexes.

Microprobe SRXRF imaging showed that Bi generally accumulated in the cytoplasm

within 24-h exposure regardless of the BiNSAID treatment. The size and location of

the hot spots (0.3–5.8 μm2) were consistent with two intracellular fates. The first is

accumulation into lysosomes, which would be indicative of a cellular detoxification

mechanism. The second relates to the hot spots that are in close proximity to the

nuclear membrane (and potentially COX-2), which indicates the intracellular

movement of Bi. However, the specific targeting of COX-2 by the BiNSAID would

require further investigation.

Synchrotron radiation infrared microspectroscopy (SR-IRMS) was utilised in

order to assess the global biomolecular changes induced by BiNSAID treatment

compared to NSAIDs alone. Live HCT-8 cells were examined using a demountable cell

holder in order to maintain hydration. The results from 4-h experiments differed to

24-h experiments suggesting that the cells undergo time-dependent changes. In most

cases, it was observed that BiNSAID or NSAID treatment substantially affected the IR

bands associated with lipids.

Based on the results from the SR-IRMS study and literature precedence, the

effects of BiNSAID- and NSAID-induced cell death on the distribution and quantity of

three glycerophospholipids, phosphatidylcholine (PC), phosphatidylethanolamine

(PE) and phosphatidylserine (PS), in the HCT-8 cells was examined using electrospray

iv

ionisation tandem mass spectrometry (ESI-MS/MS). The results showed that

treatment with Bi(tolf)3 or Bi(difl)3 caused alterations to the profile and concentration

of individual PC, PE and PS molecules. Additionally, specific trends were identified,

whereby changes to the saturation levels of the fatty acids comprising the PC, PE and

PS were observed. Treatment with both BiNSAIDs or NSAIDs caused an increase in

the proportion of the ether-linked PC and PE molecules relative to the overall profile

of PC and PE in the HCT-8 cells, which was accompanied by a significant increase in

the relative concentration of ether-linked PE molecules (but not PC) in the

BiNSAID-treated cells.

Encouraged by the promising in vitro anti-cancer activity of the BiNSAIDs against

the HCT-8 cells, several Bi(III) derivatives of aminoarenesulfonates,

indole-carboxylates, hydroxamates were screened for anti-cancer activity using the

MTT assay. A small selection of complexes from each class showed promising in vitro

toxicity, that warrants further investigation into the chemotherapeutic potential of

non-NSAID Bi(III) complexes.

The results presented in this Thesis reveal that BiNSAIDs, Bi(tolf)3, Bi(mef)3 and

Bi(difl)3, exhibit promising anti-cancer action against CRC cells in vitro. These findings

set the precedence to establish if the chemotherapeutic and chemopreventive

potential of the BiNSAIDs translates to an in vivo model of CRC.

v

DEDICATION

To my dear Pop,

My entire desire to work in medical research has been inspired by you in the

hope that your premature death was not in vain and that other sufferers of this cruel

ailment may have access to, and benefit from, more effective cancer preventions and

treatments in the future. As my sincerest thank you, this Thesis is dedicated to you.

In loving memory:

Robert John McAvoy

22nd September 1938 – 6th February 2003

vi

ACKNOWLEDGEMENTS

This research has been conducted with the support of the Australian Government

Research Training Program Scholarship.

The use of the Advanced Photon Source was supported by the U.S. Department of

Energy (DOE) Office of Science under Contract No. DEAC02-06CH11357. Travel

funding was provided by the International Synchrotron Access Program (ISAP)

managed by the Australian Synchrotron and funded by the Australian Government.

The synchrotron radiation infrared microspectroscopy studies were undertaken

on the Infrared Microspectroscopy beamline at the Australian Synchrotron, part

of ANSTO.

First and foremost I would like to give the sincerest thank you to my PhD and

Honours Supervisor, Dr Carolyn Dillon. I really appreciate the many opportunities you

have provided me throughout my PhD project, including the chance to perform

experiments at the synchrotrons in Melbourne and Chicago, as I have learnt so much

from these experiences. Thank you for allowing me the opportunity to be independent

when exploring different scientific approaches to this project and supporting me

through the highs and lows. Your dedication to your research students and to your

academic work is truly admirable. Thank you for your feedback and support while I

completed writing my Thesis and for your encouragement and belief in me, as I would

not have reached this stage without it!

I offer a generous thank you to Professor Philip Andrews (School of Chemistry,

Monash University), for his collaboration, support and scientific input throughout the

course of this project. I also offer my sincerest thanks to the past PhD students and

post-docs of the Andrews group, particularly Dr Ish Kumar, Dr Amita Pathak and

Dr Madleen Busse, for synthesising the Bi compounds required for this project.

I offer my sincerest thank you to Associate Professor Ronald Sluyter (School of

Biological Sciences, UOW) for his collaboration on many aspects of this project. From

providing me with colon cell lines to screen the compounds, to teaching me the

techniques necessary for the progression of the project, you have always been

vii

generous with your time, resources and knowledge and you have been a fantastic (and

also very patient) mentor.

My thanks go to Dr Mark Tobin, Dr Lilijana Puskar, Dr Danielle Martin and

Dr Keith Bambery at the Australian Synchrotron IR beamline (Clayton, Australia) for

their technical support and the data analysis advice they have provided me. I would

like to especially thank Dr Bambery for helpful scientific discussions and feedback in

regards to SR-IRMS data analysis. I must also thank Associate Professor Bayden Wood

(School of Chemistry, Monash University) and Dr Elizabeth Carter (Vibrational

Spectroscopy Facility, The University of Sydney) for helpful scientific discussions.

My thanks go to Dr Barry Lai at the 2-ID-D beamline at the Advanced Photon Source

(APS, Lemont, USA) for assisting me during the beamtime and for converting the files

generated into the required format for analysis, and Dr Stefan Vogt for providing the

required software so that I could analyse the data. My fellow group members, Judith

Carrall, Dr Lloyd James, Renee Di Pietro, Sandra Spremo, Gerard Sansom, and my

Supervisor, must be recognised for their efforts as each synchrotron trip is very much

a team effort – thank you for taking over so I could get some much needed sleep

throughout the beamtimes.

For technical support, I wish to thank Tony Romeo at the UOW Electron

Microscopy Centre (EMC) for teaching Judith Carrall and myself the art of the

microtome and locating the necessary equipment for us in the fledgling days of the

EMC as we prepared our samples for analysis at APS. I offer my warmest thanks to

Dr Wilford Lie (School of Chemistry, UOW) for his generosity with his time and advice

about NMR spectroscopy and for his wonderful management of the NMR facility.

I offer a sincere thank you to Dr Jennifer Saville (School of Chemistry, UOW), my lipid

queen, for assisting me with learning techniques necessary for my project’s success

and for her willingness to answer all my queries as I analysed the data; thank you also

for your continued friendship and support following your move interstate. I would

like to thank Associate Professor Todd Mitchell (School of Medicine, UOW) for

providing a number of resources required for the lipid mass spectrometry

experiments and the subsequent data analysis. I would like to thank the Technical

Services team (School of Chemistry, UOW) for maintaining the instruments necessary

for my project and coming to my aid when things went awry with them. I must thank

Dr Phil Barker (School of Chemistry, UOW) for helpful scientific discussions.

viii

I would like to offer my warmest thanks to my fellow group member, Judith

Carrall, who has been a constant source of friendship, support and encouragement

throughout our time together in the Dillon group. I cannot express enough the

gratitude I have for the patience and understanding you have shown me throughout

our PhD years. I truly treasure the friendship I have formed with you and I wish you

all the best wherever life takes you next. I would also like to sincerely thank Dr Lloyd

James as you were a significant presence in the Dillon lab during the course of our

PhDs, and you too have become a wonderful friend. Thank you to my other amazing

friends, especially Dr Monica Birrento, Dr Kimberly Davies, and Dr Diane Ly (among

others), for many inspirational PhD pep talks over the years. There are many

wonderful students I have had the pleasure to work with, and garnered support from,

while I have been a member of the Dillon group including; Tara Brown, Jacob Lambert,

Sandra Spremo, Melanie Pearsall, Renee Di Pietro and Erica Minato.

There are many wonderful staff and academics in the UOW School of Chemistry

(honorable mentions to Louisa Wildin, Associate Professor Steven Ralph, Associate

Professor Dianne Jolley and Associate Professor Glennys O’Brien) who have provided

me with support in some of the more challenging times during my studies. The

Operations team at the Illawarra Health and Medical Research Institute (IHMRI) has

been a source of support throughout the final stages of my degree, and I have been

incredibly fortunate to work with such a fantastic team of people. Finally, everyone

can stop asking me how my Thesis is going...what a relief!

I would like to thank Mr Peter Horsley, my fantastic Year 12 Chemistry teacher,

who not only inspired me to pursue a Chemistry degree, but also taught me not to

settle for average and pushed me to achieve my best academically. I would like to

thank my family, especially my parents, siblings and Nanna, for their constant support

and encouragement throughout my PhD, and the entirety of my undergraduate degree,

even if they are still not quite sure what it was I was doing in the lab all these years!

To my partner, Dr Jacob Lewis, thank you for inspiring me to see this project through

to the end and supporting me in more ways than you know during my degree.

ix

CERTIFICATION

I, Emma Louise Hawksworth, declare that this thesis submitted in fulfilment of the

requirements for the conferral of the degree Doctor of Philosophy, from the University of

Wollongong, is wholly my own work unless otherwise referenced or acknowledged. This

document has not been submitted for qualifications at any other academic institution.

________________________________ Emma Louise Hawksworth 29th March 2018

x

ABBREVIATIONS Δψm mitochondrial transmembrane potential

2AN-SO3H 2-amino-1-naphthalensulfonic acid

4A3HN-SO3H 4-amino-3-hydroxy-1-naphthalenesulfonic acid

5AN-SO3H 5-amino-1-naphthalensulfonic acid

7AAD 7-aminoactinomycin D

AAS atomic absorption spectroscopy

AnnV fluorescein isothiocyanate-labelled Annexin-V

aspH aspirin, acetylsalicylic acid

Bcl-2 B-cell lymphoma 2

BHA-H benzohydroxamic acid

BiNSAID bismuth complex of non-steroidal anti-inflammatory drug

biyp 2,2’-bipyridine

BSA bovine serum albumin

BSS bismuth subsalicylate

CBS colloidal bismuth subcitrate

cell medium RPMI-1640 that does not contain foetal bovine serum,

penicillin/streptomycin or GlutaMAX™ solutions

CID collision induced dissociation

cisplatin cis-diamminedichloroplatinum(II)

COX cyclooxygenase

CRC colorectal cancer

CV cardiovascular

DMSO-d6 deuterated dimethyl sulphoxide

dicloH diclofenac

difl diflunisate ligand

diflH diflunisal

DMA N,N-dimethylacetamide

DMF N,N-dimethylformamide

DMSO dimethyl sulphoxide

ECACC European Collection of Cell Cultures

ESI-MS electrospray ionisation mass spectrometry

ESI-MS/MS electrospray ionisation tandem mass spectrometry

FAAS flame atomic absorption spectroscopy

xi

FAP familial adenomatous polyposis

FDA U.S. Food and Drug Administration

fluH flufenamic acid

FITC fluorescein isothiocyanate

FTIR Fourier transform infrared

GMB gemcitabine

GFAAS graphite furnace atomic absorption spectroscopy

GI gastrointestinal

growth medium RPMI-1640 medium supplemented with 10% fetal bovine

serum, GlutaMAX™ (2 mM), penicillin and streptomycin (final

concentration 200 IU/mL and 200 µg/mL, respectively)

HA hypocrellin A

HCT-8 human ileocecal colorectal cancer cell line

HEPES 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid

Hp haptoglobin

hPBMC human peripheral blood mononuclear cell

H. pylori Helicobacter pylori

ibuH ibuprofen

IC50 concentration required to cause 50% inhibition

ICP-MS inductively coupled plasma-atomic mass spectrometry

IGA-H indole-3-glyoxylic acid

IL interleukin

ILA-H indole-3-lactic acid

indoH indomethacin

IR infrared

IRMS infrared microspectroscopy

isox isoxicam

ketoH ketoprofen

LOD limit of detection

LOQ limit of quantification

mecloH meclofenamic acid

mef mefenamate ligand

mefH mefenamic acid

melox meloxicam

xii

MMP matrix metalloproteinase

MS/MS tandem mass spectrometry

MT metallothionein

MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

MTX methotrexate

napH naproxen

n.d. not determined

NMR nuclear magnetic resonance

ns not significant

NSAID non-steroidal anti-inflammatory drug

oAB-SO3H ortho-aminobenzenesulfonic acid

O-AcSHA-H O-acetylsalicylhydroxamic acid

pAB-SO3H para-aminobenzenesulfonic acid

PBS phosphate buffered saline

PC phosphatidylcholine

PCA principal component analysis

PE phosphatidylethanolamine

PGD2 prostaglandin D2

PGE2 prostaglandin E2

PGF2α prostaglandin F2α

PGG2 prostaglandin G2

PGH2 prostaglandin H2

PGI2 prostaglandin I2

phen 1,10-phenanthroline

pirox piroxicam

PL phospholipid

POX peroxidase

POPC 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

PS phosphatidylserine

py pyridine

QqQ triple quadrupole

RBC ranitidine bismuth citrate

RMieS resonant Mie scatter

RPMI-1640 Roswell Park Memorial Institute 1640 cell medium

xiii

ROI region of interest

ROS reactive oxygen species

SHA-H salicylhydroxamic acid

SR synchrotron radiation

sh shoulder

SRB sulforhodamine B

SR-IRMS synchrotron radiation infrared microspectroscopy

SRXRF synchrotron radiation X-ray fluorescence

tenox tenoxicam

tolf tolfenamate ligand

tolfH tolfenamic acid

TXA2 thromboxane A2

UV-Vis ultraviolet-visible

VEGF vascular endothelial growth factor

vehicle RPMI-1640 containing 2% v/v DMSO

xiv

LIST OF PUBLICATIONS

Hawksworth, E.L., Andrews, P.C., Lie, W., Lai, B., Dillon, C.T. Biological evaluation of

bismuth non-steroidal anti-inflammatory drugs (BiNSAIDs): Stability, toxicity and

uptake in HCT-8 colon cancer cells. J Inorg Biochem, 135 (2014) 28–39.

LIST OF CONFERENCE PRESENTATIONS

Hawksworth, E.L., Andrews, P.C., Sluyter, R., Saville, J.T., Spremo, S., Tobin, M.J.,

Martin, D., Dillon, C.T. (2014) Biomolecular studies of bismuth non-steroidal

anti-inflammatory drugs in human colon cancer cells. 7th Asian Biological Inorganic

Chemistry Conference, Gold Coast, Australia. Poster presentation.

Hawksworth, E.L., Andrews, P.C., Sluyter, R., Saville, J.T., Spremo, S., Tobin, M.J.,

Martin, D., Dillon, C.T. (2014) Biomolecular studies of bismuth non-steroidal

anti-inflammatory drugs in human colon cancer cells. RACI National Congress,

Adelaide, Australia. Poster presentation.

xv

TABLE OF CONTENTS

Abstract .................................................................................................................................................................... ii

Dedication ............................................................................................................................................................... v

Acknowledgements ............................................................................................................................................ vi

Certification ........................................................................................................................................................... ix

Abbreviations ......................................................................................................................................................... x

List of Publications ........................................................................................................................................... xiv

List of Conference Presentations ................................................................................................................ xiv

Table of Contents ............................................................................................................................................... xv

List of Figures ................................................................................................................................................... xxii

List of Tables ...................................................................................................................................................... xxx

Chapter 1. Introduction 1 1.1 Cancer ................................................................................................................................. 2

1.1.1 Colorectal cancer (CRC) ......................................................................................................... 2

1.2 Non-steroidal anti-inflammatory drugs (NSAIDs) ................................................ 4

1.2.1 Mechanism of action of NSAIDs .......................................................................................... 7

1.2.2 NSAIDs: Chemopreventive agents for CRC? ............................................................... 10

1.2.3 The role of COX-2 in CRC .................................................................................................... 12

1.2.4 NSAIDs mediate cell death via COX, and non-COX related mechanisms ....... 14

1.2.5 The drawbacks of prolonged NSAID use ..................................................................... 15

1.3 Metal complexes of NSAIDs (metal-NSAIDs) ......................................................... 17

1.3.1 Anti-inflammatory and anti-ulcerative properties of metal-NSAIDs ............. 18

1.3.2 Chemotherapeutic properties of metal-NSAIDs ...................................................... 21

1.3.3 Other biologically relevant properties of metal-NSAIDs ..................................... 24

1.4 Bismuth ............................................................................................................................ 25

1.4.1 Bismuth in medicine ............................................................................................................. 26

1.4.1.1 Absorption, transport and distribution of Bi(III) ........................................... 28

1.4.1.2 Bismuth as a GI protective agent ........................................................................... 29

1.4.1.3 Bismuth as an anti-bacterial agent ...................................................................... 30

1.4.1.4 Bismuth as an anti-cancer agent ........................................................................... 31

1.4.2 Bismuth toxicity .................................................................................................................... 35

xvi

1.5 BiNSAID complexes ....................................................................................................... 37

1.6 Project Aims .................................................................................................................... 39

1.7 References ....................................................................................................................... 41

Chapter 2. Influence of BiNSAID stability on toxicity and PGE2

inhibition in HCT-8 colon cancer cells 59 2.1 Introduction .................................................................................................................... 60

2.1.1 Solubility/Bioavailability ................................................................................................... 60

2.1.2 Drug stability ........................................................................................................................... 62

2.1.3 In vitro cytotoxicity towards colon cancer cells ....................................................... 63

2.1.3.1 Assessing in vitro drug toxicity: The MTT assay .............................................. 63

2.1.3.2 Modes of cell death ...................................................................................................... 64

2.1.4 Inhibition of COX: Reduction of PGE2 production ................................................... 67

2.1.5 Chapter aims ............................................................................................................................ 68

2.2 Materials and methods ................................................................................................ 69

2.2.1 Materials .................................................................................................................................... 69

2.2.2 Stability studies ...................................................................................................................... 70

2.2.3 General cell culture ............................................................................................................... 70

2.2.4 In vitro cytotoxicity towards cancer cells: MTT assay ........................................... 71

2.2.5 Cell death assays .................................................................................................................... 72

2.2.5.1 Morphological analysis by phase contrast microscopy ................................ 72

2.2.5.2 Measurement of cell death by cytofluorometric assay .................................. 72

2.2.6 COX inhibition: Prostaglandin assays .......................................................................... 73

2.2.7 Statistical analysis ................................................................................................................ 74

2.3 Results .............................................................................................................................. 75

2.3.1 Stability of BiNSAIDs in cell medium ............................................................................ 75

2.3.2 In vitro cytotoxicity towards cancer cells ................................................................... 84

2.3.3 Cell death assays .................................................................................................................... 86

2.3.3.1 Phase contrast microscopy ....................................................................................... 86

2.3.3.2 AnnV/7AAD assays ...................................................................................................... 87

2.3.4 The effect of BiNSAIDs on PGE2 production by HCT-8 cells ............................... 91

2.4 Discussion ........................................................................................................................ 94

2.4.1 NMR stability studies .......................................................................................................... 94

2.4.2 In vitro cytotoxicity towards cancer cells .................................................................. 95

xvii

2.4.3 Cell death assays ................................................................................................................... 98

2.4.4 PGE2 production ................................................................................................................. 100

2.5 Chapter summary ....................................................................................................... 103

2.6 References .................................................................................................................... 105

Chapter 3. Uptake and localisation of Bi by BiNSAID-treated HCT-8

colon cancer cells 111 3.1 Introduction ................................................................................................................. 112

3.1.1 Atomic absorption spectroscopy for studying cellular uptake of inorganic

complexes/drugs ................................................................................................................ 113

3.1.2 Microprobe synchrotron radiation X-ray fluorescence imaging for studying

intracellular metal localisation ..................................................................................... 114

3.1.3 Chapter aims ......................................................................................................................... 117

3.2 Materials and methods ............................................................................................. 118

3.2.1 Materials ................................................................................................................................. 118

3.2.2 Cell culture procedures .................................................................................................... 118

3.2.3 GFAAS uptake studies ....................................................................................................... 118

3.2.3.1 Preparation of cell digests ................................................................................... 118

3.2.3.2 Preparation of fixed/dehydrated cell digests ............................................. 119

3.2.3.3 Graphite furnace atomic absorption spectroscopy .................................. 120

3.2.4 Microprobe SRXRF analysis of Bi-treated cells ..................................................... 120

3.2.4.1 Preparation of thin-sectioned cell samples ................................................. 120

3.2.4.2 Microprobe SRXRF analysis of thin-sectioned cells ................................. 121

3.2.5 Statistical analysis .............................................................................................................. 122

3.3 Results ........................................................................................................................... 123

3.3.1 GFAAS analysis of cellular Bi ......................................................................................... 123

3.3.1.1 The dependence of cell preparation on the cellular uptake of Bi. ......... 125

3.3.2 Microprobe SRXRF spectroscopy ................................................................................ 126

3.4 Discussion ..................................................................................................................... 131

2.4.1 GFAAS uptake studies...................................................................................................... 131

2.4.2 Microprobe SRXRF spectroscopy of thin-sectioned cells ................................ 134

3.5 Chapter summary ....................................................................................................... 140

3.6 References .................................................................................................................... 141

xviii

Chapter 4. Synchrotron radiation infrared microspectroscopy

studies (SR-IRMS) of live HCT-8 cells treated with

BiNSAIDs or NSAIDs 145

4.1 Introduction ................................................................................................................. 146

4.1.1 FTIR spectroscopy of biological molecules ............................................................. 146

4.1.2 FTIR microscopy and synchrotron sources ............................................................ 149

4.1.3 Infrared microspectroscopy studies of drug effects on cancer cells ........... 150

4.1.4 SR-IRMS of live cells: advantages and disadvantages ........................................ 152

4.1.5 Chapter aims ......................................................................................................................... 156


4.2.1 Material ................................................................................................................................... 157

4.2.2 General cell culture ............................................................................................................ 157

4.2.3 SR-IRMS studies................................................................................................................... 157

4.2.3.1 Preparation of cell samples................................................................................. 157

4.2.2.2 SR-IRMS data collection ....................................................................................... 159

4.2.2.3 SR-IRMS spectral processing and chemometric analysis ...................... 159

4.3 Results ........................................................................................................................... 161

4.3.1 Comparison of spectral correction factors used for data analysis ............... 161

4.3.1.1 Water band correction using a spectral subtraction method ................. 161

4.3.1.2 Statistical analysis of data: Principal component analysis (PCA) ......... 164

4.3.1.3 Effect of vehicle on live HCT-8 cells (4 h study)............................................. 166

4.3.1.4 Effect of vehicle on live HCT-8 cells (24 h study) ......................................... 167

4.3.2 Comparison of the effects of BINSAIDs, and the respective NSAIDs, on

SR-IRMS spectra on live HCT-8 cells ......................................................................... 170

4.3.2.1 Bi(tolf)3- and tolfH-treated cells (4 h treatment) .......................................... 170

4.3.2.2 Bi(tolf)3- and tolfH-treated cells (24 h treatment) ...................................... 173

4.3.2.3 Bi(mef)3- and mefH-treated cells (4 h treatment) ....................................... 175

4.3.2.4 Bi(mef)3- and mefH-treated cells (24 h treatment) .................................... 177

4.3.2.5 Bi(difl)3- and diflH-treated cells (4 h treatment) ......................................... 179

4.3.2.6 Bi(difl)3- and diflH-treated cells (24 h treatment) ...................................... 181

4.3.2.7 Summary of the IR spectral and PCA differences following BiNSAID and

NSAID treatment ....................................................................................................... 181

4.4 Discussion ..................................................................................................................... 188

xix

4.4.1 Data analysis: Water absorption correction strategy .......................................... 188

4.4.2 The biochemical effects of treatment on live HCT-8 cells ................................. 190

4.4.2.1 Vehicle-treated cells ................................................................................................. 190

4.4.2.2 BiNSAID/NSAID-treated cells .............................................................................. 194

4.4.3 Limitations of the experimental approach and future improvements ........ 198

4.5 Chapter summary ....................................................................................................... 200

4.6 References .................................................................................................................... 202

Chapter 5. Lipidomic profiling of glycerophospholipids using

ESI-MS/MS of HCT-8 cells treated with BiNSAIDs

and NSAIDs 209

5.1 Introduction ................................................................................................................. 210

5.1.1 Shotgun lipidomics: Electrospray ionisation mass spectrometry (ESI-MS)

and tandem MS/MS for the examination of phospholipids ............................. 211

5.1.2 The structure of glycerophospholipids ..................................................................... 213

5.1.3 Nomenclature of glycerophospholipids ................................................................... 215

5.1.4 The biological importance of glycerophospholipids ........................................... 215

5.1.5 Chapter aims ......................................................................................................................... 217


5.2.1 Materials ................................................................................................................................. 218

5.2.2 Cell lines .................................................................................................................................. 218

5.2.3 Preparation of cell lipid extracts .................................................................................. 218

5.2.4 Mass spectrometry ............................................................................................................. 220

5.2.5 Data analysis .......................................................................................................................... 221

5.2.6 Statistical analysis ............................................................................................................... 221

5.3 Results ........................................................................................................................... 223

5.3.1 Phosphatidylcholine (PC) ................................................................................................ 223

5.3.2 Phosphatidylethanolamine (PE) .................................................................................. 229

5.3.3 Phosphatidylserine (PS) .................................................................................................. 235

5.3.4 Summary of results ............................................................................................................ 236

5.4 Discussion ..................................................................................................................... 239

5.4.1 Effect of BiNSAID and NSAID treatment on phospholipid concentration . 239

xx

5.4.2 The effect of BiNSAID and NSAID treatment on fatty acyl chain saturation ....

................................................................................................................................................................... 241

5.4.3 The effect of BiNSAID and NSAID treatment on ether-linked molecules .. 243

5.4.4 Overall summary of results and future directions ............................................... 243

5.5 Chapter summary ....................................................................................................... 247

5.6 References .................................................................................................................... 249

Chapter 6. Exploration of the in vitro toxicity of Bi(III)

aminoarenesulfonate, indole-carboxylate, and

hydroxamate complexes 253

6.1 Introduction ................................................................................................................. 254

6.1.1 Sulfonic acids and Bi(III) aminoarenesulfonates ................................................. 254

6.1.2 Indoles and Bi(III) indole-carboxylates .................................................................... 257

6.1.3 Hydroxamic acids and Bi(III) hydroxamates ......................................................... 258

6.1.5 Chapter aims ......................................................................................................................... 261


6.2.1 Materials ................................................................................................................................. 262

6.2.2 Cell lines .................................................................................................................................. 262

6.2.3 MTT assay .............................................................................................................................. 262

6.3 Results ........................................................................................................................... 263

6.4 Discussion ..................................................................................................................... 266

6.4.1 Bi(III) aminoarenesulfonates ........................................................................................ 266

6.4.2 Bi(III) indole-carboxylates ............................................................................................. 267

6.4.3 Bi(III) hydroxamates ......................................................................................................... 268

6.5 Chapter summary ....................................................................................................... 270

6.6 References .................................................................................................................... 271

Chapter 7. Conclusions and future directions 275 7.1 Conclusions .................................................................................................................. 276

7.2 Future directions ........................................................................................................ 282

Appendices 284

Appendix 1 Additional MTT and PGE2 assays .......................................................... 284

xxi

Appendix 2 Additional whole cell and thin-sectioned cell maps ....................... 288

Appendix 3 Additional SR-IRMS spectra .................................................................... 292

Appendix 4 Additional ESI-MS/MS results ................................................................ 304

Appendix 5 Additional MTT assay results ................................................................. 305

xxii

LIST OF FIGURES Figure 1.1: Age-standardised incidence (1982-2014) and mortality rates (1968-2014)

of colorectal cancers in Australia ............................................................................................................. 3

Figure 1.2: Chemical classes of NSAIDs ............................................................................................... 6

Figure 1.3: The conversion of arachidonic acid to prostaglandins via COX ........................ 8

Figure 1.4: The key structural residues of the COX-1 and COX-2 active sites .................... 9

Figure 1.5: Selected molecular pathways affected by the over-expression of COX-2 in

neoplasia .......................................................................................................................................................... 13

Figure 1.6: Chemical structures of bismuth compounds that have shown potential

anti-cancer activity in vitro ...................................................................................................................... 33

Figure 1.7: Representation of the chemical structures of (a) Bi(tolf)3 and Bi(mef)3, and

(b) Bi(difl)3 as inferred by elemental analysis, IR and NMR data .......................................... 39

Figure 2.1: The use of flow cytometry to distinguish modes of cell death using

AnnV and 7AAD ............................................................................................................................................ 67

Figure 2.2: 1H NMR spectra of the aromatic region (6.6-8.0 ppm) of: (a) tolfH; (b) tolfH

collected after incubation in cell medium (24 h, 37 °C); (c) Bi(tolf)3 and, (d) Bi(tolf)3

collected after incubation in cell medium (24 h, 37 °C) ............................................................. 76

Figure 2.3: 1H NMR spectra of the aromatic region (6.6-8.0 ppm) of: (a) mefH;

(b) mefH collected after incubation in cell medium (24 h, 37 °C); (c) Bi(mef)3 and,

(d) Bi(mef)3 collected after incubation in cell medium (24 h, 37 °C) ................................... 79

Figure 2.4: 1H NMR spectra of the aromatic region (6.6-8.0 ppm) of: (a) diflH; (c) diflH

collected after incubation in cell medium (24 h, 37 °C); (d) Bi(difl)3, and (f) Bi(difl)3

collected after incubation in cell medium (24 h, 37 °C) ................................................................... 82

Figure 2.5: 1H NMR spectra of: (a) unknown molecule present in the sample

containing Bi(difl)3 collected after incubation in cell medium (24 h, 37 °C); and

(b) D-glucose .................................................................................................................................................. 84

Figure 2.6: Concentration-response curves obtained from the MTT assays of HCT-8

cells treated with tolfH, Bi(tolf)3, mefH, Bi(mef)3, diflH, or Bi(difl)3 (24 h) ....................... 86

Figure 2.7: Morphological effects following 24-h treatment with: (a) vehicle (DMSO

2% v/v); (b) Bi(tolf)3 (15 μM); (c) tolfH (45 μM); (d) Bi(mef)3 (40 μM);

(e) mefH (120 μM); (f) Bi(difl)3 (75 μM); and (g) diflH (225 μM) ......................................... 88

xxiii

Figure 2.8: Flow cytometric analysis of cell death in HCT-8 cells treated for 24 h with

vehicle (RPMI-1640, 2% v/v DMSO; Control), BiNSAIDs (Bi(tolf)3 (15 µM),

Bi(mef)3 (40 µM) or Bi(difl)3 (75 µM)); or the respective equimolar free NSAIDs ........ 90

Figure 2.9: Concentration-response curve obtained from the PGE2 assays of HCT-8

cells treated (4 h) with: (a) tolfH, Bi(tolf)3, mefH, Bi(mef)3; and (b) diflH, Bi(difl)3, and

aspH, followed by arachidonic acid (20 µM, 30 min) .................................................................. 92

Figure 3.1: Bohr atom model illustrating the basic principle of X-ray fluorescence..115

Figure 3.2: Trypan blue exclusion assay and GFAAS analysis of HCT-8 cells treated

with Bi compounds .................................................................................................................................. 124

Figure 3.3: GFAAS detection of Bi in HCT-8 cells following exposure to the specified Bi

compounds (40 μM, 24 h) prepared by normal washing method or prepared following

fixation with glutaraldehyde (2% in PBS, 2 h) and dehydration with ethanol ............. 125

Figure 3.4: Microprobe SRXRF elemental maps for thin-sectioned HCT-8 cells

following 24-h treatment with: (a) vehicle only (DMSO 2% v/v); (b) Bi(difl)3 (40 μM);

(c) Bi(mef)3 (40 μM); (d) Bi(tolf)3 (40 μM); and (e) BSS (40 μM) ....................................... 127

Figure 3.5: Quantifications of: (a) P; (b) Zn; (c) Ca; and (d) Bi, in the cytoplasmic and

nuclear regions of the HCT-8 thin-sectioned cells following 24-h treatment with

vehicle (2% DMSO v/v), or the specified Bi compounds (40 μM) ...................................... 130

Figure 4.1: Fundamental modes of vibration of IR active molecules ............................... 147

Figure 4.2: Typical IR signatures of biological materials ...................................................... 148

Figure 4.3: A single fixed HCT-8 cell spectrum superimposed with a spectrum

collected from cell medium .................................................................................................................. 155

Figure 4.4: Construction of the compression cell holder components employed for

SR-IRMS measurements ......................................................................................................................... 158

Figure 4.5: Individual cell spectrum before and following water subtraction to a

correction value of −0.04 using OPUS 7.0 ...................................................................................... 160

Figure 4.6: (a) Averaged (n = 59), baseline corrected (rubberband baseline corrected,

64 baseline points), vector normalised HCT-8 cell spectra following 4-h incubation in

cell medium before, and after, water band subtraction. (b) Unit vector normalised,

second derivative of the spectra shown in (a) ............................................................................. 162

Figure 4.7: PCA scores plot (left) and corresponding PC-1 and PC-2 X-loadings plots

(right) of second derivative (Savitzky Golay, 13 smoothing points), unit vector

xxiv

normalised SR-IRMS spectra obtained from HCT-8 cells treated for 4 h with cell

medium, or vehicle (RPMI-1640, DMSO 2% v/v) ....................................................................... 165




medium, or vehicle (RPMI-1640, DMSO 2% v/v) ....................................................................... 166




medium alone, or vehicle (RPMI-1640, DMSO 2% v/v) .......................................................... 169

Figure 4.10: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots

generated from the second derivative (Savitzky Golay, 13 smoothing points), unit

vector normalised SR-IRMS spectra obtained from HCT-8 cells following 4 h treatment

with vehicle (RPMI-1640, DMSO 2% v/v), Bi(tolf)3, or tolfH (45 μM NSAID) ............... 172

Figure 4.11: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots of

second derivative (Savitzky Golay, 13 smoothing points), unit vector normalised

SR-IRMS spectra obtained from HCT-8 cells following 24 h treatment with vehicle

(RPMI-1640, DMSO 2% v/v), Bi(tolf)3, or tolfH (45 μM NSAID).......................................... 174




(RPMI-1640, DMSO 2% v/v), Bi(mef)3, or mefH (120 μM NSAID) ..................................... 177




(RPMI-1640, DMSO 2% v/v), Bi(mef)3, or mefH (120 μM NSAID) ..................................... 178




(RPMI-1640, DMSO 2% v/v), Bi(difl)3, or diflH (225 μM NSAID) ....................................... 180




(RPMI-1640, DMSO 2% v/v), Bi(difl)3, or diflH (120 μM NSAID) ....................................... 182

xxv

Figure 5.1: Schematic representation of the shotgun lipidomic profiling method by

tandem mass spectrometry using a triple quadrupole (QqQ) MS instrument .............. 212

Figure 5.2: Chemical structure of: (a) Glycerophospholipids; (b) Common

glycerophospholipid head groups; (c) A fatty acid (18 carbons in length) containing no

double bonds (saturated), a single double bond (monounsaturated), and two double

bonds (polyunsaturated) ....................................................................................................................... 214

Figure 5.3: Profile of phosphatidylcholine (PC) molecules in HCT-8 cells following

24-h treatment with vehicle (DMSO 2% v/v), Bi(tolf)3 (15 µM), or tolfH (45 µM),

calculated as a percentage of total PC .............................................................................................. 224

Figure 5.4: Profile of phosphatidylcholine (PC) molecules in HCT-8 cells following

24-h treatment with vehicle (DMSO 2% v/v), Bi(difl)3 (75 µM), or diflH (225 µM),

calculated as a percentage of total PC .............................................................................................. 225

Figure 5.5: The proportion of phosphatidylcholine (PC) molecules with fatty acid

chains containing saturated, monounsaturated, polyunsaturated or ether-linked fatty

acids calculated as a percentage of the total PC molecules identified in HCT-8 cell lipid

extracts .......................................................................................................................................................... 227

Figure 5.6: The normalised concentration of phosphatidylcholine (PC) molecules

containing saturated, monounsaturated, polyunsaturated or ether-linked fatty acids in

HCT-8 cell lipid extracts treated for 24 h with: (a) vehicle (RPMI-1640, DMSO 2% v/v),

Bi(tolf)3 (15 µM) or tolfH (45 µM); and (b) vehicle (RPMI-1640, DMSO 2% v/v),

Bi(difl)3 (75 µM) or diflH (225 µM) .................................................................................................. 228

Figure 5.7: Profile of phosphatidylethanolamine (PE) molecules in HCT-8 cells

following 24 h treatment with: (a) vehicle (DMSO 2% v/v), Bi(tolf)3 (15 µM), or tolfH

(45 µM); and (b) vehicle (DMSO 2% v/v), Bi(difl)3 (75 µM), or diflH (225 µM),

calculated as a percentage of total PE .............................................................................................. 230

Figure 5.8: The proportion of phosphatidylethanolamine (PE) molecules containing

monounsaturated, polyunsaturated or ether-linked fatty acids calculated as a

percentage of the total PE molecules identified in HCT-8 cell lipid extracts ................. 232

Figure 5.9: The normalised concentration of phosphatidylethanolamine (PE)

molecules containing monounsaturated, polyunsaturated or ether-linked fatty acids in



Bi(difl)3 (75 µM) or diflH (225 µM). ................................................................................................. 234

xxvi

Figure 5.10: Profile of phosphatidylserine (PS) molecules in HCT-8 cells following

24 h treatment with: (a) vehicle (RPMI-1640, DMSO 2% v/v), Bi(tolf) 3 (15 µM), or

tolfH (45 µM); and (b) vehicle (RPMI-1640, DMSO 2% v/v), Bi(difl)3 (75 µM), or

diflH (225 µM), calculated as a percentage of total PS ............................................................. 235

Figure 5.11: The normalised concentration of phosphatidylserine (PS) molecules in



Bi(difl) 3 (75 µM) or diflH (225 µM) .................................................................................................. 236

Figure 6.1: The chemical structures of biologically active sulfonic acids ...................... 255

Figure 6.2: The chemical structures of the aminoarenesulfonic acids used to synthesis

the five Bi(III) aminoarenesulfonate complexes investigated in the present study ... 256

Figure 6.3: The chemical structures of biologically active indoles, and the

indole-carboxylic acids used to synthesise the two Bi(III) indole-carboxylate

complexes investigated in this study ............................................................................................... 258

Figure 6.4: The chemical structures of a selection of biologically active hydroxamic

acids, and the hydroxamic acids used to synthesise the three Bi(III) hydroxamate

complexes investigated in this study ............................................................................................... 260

Figure 6.5: Concentration-response curves obtained from the MTT assays of HCT-8

cells treated with: (a) Bi(4A3HN-SO3)3 and 4A3HN-SO3H; (b) Bi(IGA)3, Bi(ILA)3, IGA-H

and ILA-H; (c) Bi(SHA)3, SHA-H, Bi(O-AcSHA)3 and O-AcSHA-H; and (d) Bi(BHA)3 and

BHA-H (24 h) ............................................................................................................................................... 264

Appendix 1.1: Concentration-response curve obtained from the MTT assay of DMSO

concentration (% v/v in RPMI-1640) in HCT-8 cells after 24-h treatment .................... 284

Appendix 1.2: Concentration-response curves obtained from the MTT assays of:

(a) tolfH and Bi(tolf)3; (b) mefH and Bi(mef)3; and (c) diflH and Bi(difl)3; in HCT-8 cells

after 4-h treatment ................................................................................................................................... 285

Appendix 1.3: Concentration-response curves obtained from the MTT assays of

cisplatin, BiCl3 and BSS after 24-h treatment ............................................................................... 286

Appendix 1.4: PGE2 production in HCT-8 cells after 4-h treatment with BSS ............. 286

Appendix 1.5: Concentration-response curve obtained from the MTT assays of aspH

in HCT-8 cells after 24-h treatment .................................................................................................. 287

xxvii

Appendix 2.2: Microprobe SRXRF elemental maps for toluidine-blue stained thin-

section of HCT-8 cells following 24-h treatment with: (a) vehicle only; (b) Bi(difl)3

(40 μM); (c) Bi(mef)3 (40 μM); (d) Bi(tolf)3 (40 μM), and (e) BSS (40 μM) .................... 289

Appendix 2.3: Microprobe SRXRF elemental maps for toluidine-blue stained

thin-section of HCT-8 cells following 4-h treatment with Bi(mef)3 (40 μM) ................. 290

Appendix 2.4: Correlative light micrograph of a toluidine-blue stained thin-section of

HCT-8 cells following 24-h treatment with: (a) Bi(difl)3 (40 μM); (b) Bi(mef)3 (40 μM);

and (c) Bi(tolf)3 (40 μM) and the corresponding microprobe SRXRF maps of P, Zn, Bi,

and the colocalisation (Col) of the three elements .................................................................... 291

Appendix 3.1: (a) Averaged, baseline corrected (rubberband baseline corrected, 64

baseline points), vector normalised HCT-8 cell spectra following 4-h incubation in

vehicle (RPMI-1640, DMSO 2% v/v) before and after water band compensation.

(b) Unit vector normalised, second derivative of the spectra shown in (a) ................... 292

Appendix 3.2: Averaged, baseline corrected (rubberband baseline corrected, 64

baseline points), vector normalised, HCT-8 cell spectra following 4-h incubation in cell

medium (RPMI-1640; n = 54), or vehicle (RPMI-1640, 2% DMSO v/v; n = 59):

(a) without the application of water band subtraction, and (b) following water band

subtraction ................................................................................................................................................... 293

Appendix 3.3: Averaged second derivative (Savitzky Golay, 13 smoothing points), unit

vector normalised, HCT-8 cell spectra following 4-h incubation in cell medium

(RPMI-1640; n = 54), or vehicle (RPMI-1640, 2% DMSO v/v; n = 59): (a) without the

application of water band correction, and (b) following water band correction ........ 294

Appendix 3.4: Averaged, baseline corrected (rubberband baseline corrected, 64


cell medium (RPMI-1640 alone, n = 39), or vehicle (RPMI-1640, 2% DMSO v/v;

n = 54): (a) without the application of water band correction, and (b) following water

band correction .......................................................................................................................................... 295

Appendix 3.5: Averaged second derivative (Savitzky Golay, 13 smoothing points), unit

vector normalised, HCT-8 cell spectra following 24-h incubation in RPMI-1640 alone

(n = 39), or vehicle (RPMI-1640, 2% DMSO v/v; n = 54): (a) without the application of

water band correction, and (b) following water band correction ...................................... 296



xxviii

vehicle (RPMI-1640, DMSO 2% v/v, n = 54), Bi(tolf)3 (15 µM, n = 45) or tolfH (45 µM, n

= 29). (b) Unit vector normalised, second derivative of the spectra shown in (a)... .. 297



vehicle (RPMI-1640, DMSO 2% v/v, n = 54), Bi(tolf)3 (15 µM, n = 33) or

tolfH (45 µM, n = 100). (b) Unit vector normalised, second derivative of the spectra

shown in (a) ............................................................................................................................................... ..298



vehicle (RPMI-1640, DMSO 2% v/v; n = 83), Bi(mef)3 (40 µM, n = 42) or

mefH (120 µM, n = 54). (b) Unit vector normalised, second derivative of the spectra

shown in (a) ................................................................................................................................................. 299



vehicle (RPMI-1640, DMSO 2% v/v; n = 54), Bi(mef)3 (40 µM, n = 95) or mefH

(120 µM, n = 90). (b) Unit vector normalised, second derivative of the spectra

shown in (a) ................................................................................................................................................. 300



vehicle (RPMI-1640, DMSO 2% v/v) (n = 59), Bi(difl)3 (75 µM, n = 34) or diflH

(225 µM, n = 41). (b) Unit vector normalised, second derivative of the spectra

shown in (a) ................................................................................................................................................. 301



vehicle (RPMI-1640, DMSO 2% v/v, n = 54), Bi(difl)3 (75 µM, n = 66) or

diflH (225 µM, n = 38). (b) Unit vector normalised, second derivative of the spectra

shown in (a) ................................................................................................................................................. 302

Appendix 3.12: Percentage of early apoptotic and late apoptotic cells based on flow

cytometric analysis of cell death in the adherent only HCT-8 cells treated for 24 h with

vehicle (RPMI-1640, 2% v/v DMSO; Control), BiNSAIDs (Bi(tolf)3 (15 µM), Bi(mef)3

(40 µM) or Bi(difl)3 (75 µM)); or the respective equimolar free NSAIDs ........................ 303

xxix

Appendix 5.1: Concentration-response curves obtained from MTT assays of HCT-8

cells treated (24 h) with the specified aminoarenesulfonic acids and corresponding

Bi(III) complexes ....................................................................................................................................... 305

xxx

LIST OF TABLES Table 1.1: The major COX-2 dependent or independent pro-apoptotic pathways

induced by NSAIDs ...................................................................................................................................... 16

Table 1.2: Results from selected in vivo studies comparing the anti-inflammatory,

analgesic, or anti-ulcerative activity of metal-NSAID complexes and uncomplexed

NSAIDs .............................................................................................................................................................. 19

Table 1.3: Results from in vitro studies comparing the anti-cancer activity of

metal-NSAID complexes and uncomplexed NSAIDs, or currently used anti-cancer

drugs .................................................................................................................................................................. 22

Table 1.4: A selection of historical and currently used medicinal Bi compounds ........ 26

Table 2.1: 1H and 13C NMR data of tolfH and Bi(tolf)3 in d6-DMSO, before and after

suspension in RPMI-1640 (24 h) .......................................................................................................... 77

Table 2.2: 1H and 13C NMR data of mefH and Bi(mef)3 in d6-DMSO, before and after


Table 2.3: 1H and 13C NMR data of diflH and Bi(difl)3 in d6-DMSO, before and after


Table 2.4: IC50 values determined by MTT toxicity assays following 24-h treatment of

HCT-8 colon cancer cells with the specified compounds ........................................................... 86

Table 2.5: Cell populations determined by flow cytometric analysis of HCT-8 cells

treated for 24h with: vehicle, BiNSAIDs, or the respective equimolar free NSAIDs ...... 91

Table 2.6: IC50 values determined by PGE2 assays following 4-h treatment of HCT-8

colon cancer cells with the specified compounds ......................................................................... 93

Table 4.1: Generalised assignments of characteristic IR bands of biological and

eukaryotic cell spectra ............................................................................................................................ 149

Table 4.2: Summary of some IRMS or SR-IRMS studies that examined the effects of

chemotherapeutic agents (either potentially or clinically-used) on cancer cells ........ 151

Table 4.3: Summary of average spectral trends and key PCA differences between

vehicle-treated cell samples versus NSAID/BiNSAID-treated cell samples. .................. 185

Table 5.1: Instrument parameters used for ESI-MS and targeted ion scan (MS/MS)

analysis on the Waters QuattroMicroTM triple quadrupole mass spectrometer .......... 220

Table 5.2: Targeted ion (MS/MS) scans ........................................................................................ 220

xxxi

Table 5.3: Summary of the overall trend in glycerophospholipid (PC and PE)

profile (% of total) following treatment of HCT-8 cells with BiNSAIDs of NSAIDs (24 h)

compared to vehicle-treated cells ..................................................................................................... 237

Table 5.4: Summary of the overall changes in glycerophospholipid (PC, PE or PS)

concentration following treatment of HCT-8 cells with BiNSAIDs of NSAIDs (24 h)

compared to vehicle-treated cells ..................................................................................................... 237

Table 6.1: IC50 values determined by MTT toxicity assays following 24 h treatment of

HCT-8 colon cancer cells with the specified compounds ........................................................ 265

Appendix 2.1: Graphite furnace operating conditions used for the analysis of Bi .... 288

Appendix 4.1: Analytes identified above the LOD in HCT-8 cell lipid extracts treated

with vehicle using ESI-MS/MS ............................................................................................................ 304

Chapter 1. Introduction

Colorectal cancer (CRC) is the second most common cancer in both men and women in

Australia. Since the incidence of CRC has not improved since the 1980s, prevention

strategies are needed to reduce the burden of this disease. The upregulation of

cyclooxygenase-2 (COX-2), an enzyme involved in the inflammation response, has

been linked to a number of cancers, including CRC. This has led to a large number of

studies that subsequently showed that several over-the-counter non-steroidal

anti-inflammatory drugs (NSAIDs) reduced the risk of CRC when taken regularly over

periods of months to years. This Chapter provides a review of the relevant literature

with a focus on the potential use of NSAIDs as chemopreventive and chemotherapeutic

agents for CRC, the advantages of coordinating NSAIDs to metals, and finally, the

medicinal use and anti-cancer activity of bismuth(III) complexes, including Bi(III)

complexes of NSAIDs (BiNSAIDs).


2

1.1 Cancer Cancer, in the simplest of terms, is a genetic disease that is characterised by the

uncontrolled propagation of cells that have acquired defects in regulatory circuits that

control normal cell proliferation and homeostasis [1]. The accumulation of sequential

chromosomal mutations allows a cancer cell to resist cell death, undergo limitless

cycles of cell division, resist anti-growth signals, maintain autonomous growth

signaling, sustain angiogenesis (the formation of blood vessels in the tumour), and

eventually acquire metastasis (the ability to invade tissues at distal locations) [1].

Collectively, these six features are described as the “hallmarks of cancer” [1]. Over 100

distinct types of cancer have been identified and the order of acquisition of the

aforementioned characteristics varies considerably between cell types [1].

The American Cancer Society estimates that cancer causes the death of one in seven

people worldwide [2]. In 2012, cancer represented the second leading, and third

leading, cause of death in economically developed, and economically developing

countries, respectively, highlighting the serious global nature of this disease [2].

1.1.1 Colorectal cancer

Worldwide, colorectal cancer (CRC) or bowel cancer, is the third most common

cancer in men (following lung and prostate cancers), and the second most common

cancer in women (following breast cancer) [2]. CRC commonly develops from

adenomatous polyps, which are non-malignant growths that propagate in the

epithelium of the digestive tract [3]. Owing to the slow accumulation of multiple

genetic mutations, conservative estimates suggest that adenomatous polyps can

remain non-malignant for periods of 5–10 years, or more [4]; this is reflected in the

diagnosis of over 90% of CRC cases in individuals over the age of 50 [5].

Alarmingly, Australia and New Zealand exhibit the highest incidence (per 100,000

people) of CRC in the world [6]. In Australia alone, CRC represents the second most

common cancer in both men and women [7]. Approximately 31% of individuals

diagnosed with CRC do not survive beyond five years [7]. In 2017, the Australian

Institute of Health and Welfare reported that a total of 4,144 deaths were directly

attributable to CRC [7]. Despite reductions in the age-standardised mortality rates

from CRC in Australia (Fig. 1.1) over the past three decades, the age-standardised

incidence of CRC has remained steady between 1982–2014 (Fig. 1.1) [7], with 16,682

new cases of CRC diagnosed in Australia in 2017 [7].


3

Figure 1.1: Age-standardised incidence and mortality rates (1982–2014) of colorectal cancers by sex in Australia. Source: Australian Institute of Health and Welfare [7].

Although the estimated lifetime risk of developing CRC in the general population

is 5.6% [8], 70–80% of CRC cases occur in individuals with no family history of the

disease [9, 10]. Largely preventable lifestyle factors, including high alcohol

consumption, diet (high fat, high intake of red and processed meats) and lack of

physical activity, are believed to contribute to the development of CRC [5].

Family history accounts for approximately 20% of CRC cases, while 5–10% of cases

are associated with the genetic syndromes, familial adenomatous polyposis (FAP) and

Lynch syndrome (also called hereditary non-polyposis CRC) [9, 10]. The estimated

lifetime risk of developing CRC in genetically susceptible subpopulations increases to

17% for individuals with two affected first-degree relatives, and up to 95% and 70%

for individuals with FAP and Lynch syndrome, respectively [11].

Successful methods of early detection such as faecal occult blood testing and

endoscopic surveillance (flexible sigmoidoscopy and colonoscopy), and the removal of

pre-carcinogenic lesions (polypectomy) are available in Australia [12-14].

However, primary prevention strategies have been suggested as an important

complement to established screening methods to reduce the incidence of CRC [15],

and an alternative approach to reducing mortality from cancer [16].

Chemoprevention, which is the use of a pharmacological or nutritional agent to

prevent, reverse, or delay the progression of carcinogenesis, represents one example

of a primary prevention strategy that has been widely studied in the context of

CRC [16]. Ideally, chemopreventive agents should be effective, easily administered


4

(i.e. oral administration), of low cost, involve a dosing schedule that enables

compliance (no more than once daily), and have no or limited side effects (i.e. the

benefit to risk ratio must be high) [16]. Therapeutic and nutritional agents that have

been shown to reduce CRC incidence in observational studies include non-steroidal

anti-inflammatory drugs (NSAIDs), folic acid, calcium (in conjunction with vitamin D in

some studies), antioxidants (including vitamins A, C, and E, beta-carotene or

selenium), statins and bisphosphonates [15]. Of the aforementioned agents, calcium

and vitamin D appear to exhibit modest clinical benefit, while NSAID therapies appear

to reduce the incidence of CRC substantially [15].

NSAIDs are readily available over-the-counter drugs that are commonly used to

relieve the symptoms of inflammation, pain, redness, swelling, and calor [17].

NSAIDs are commonly used for the treatment of pain and inflammation associated

with transient inflammatory conditions and tissue damage (including post-operative

pain, dysmenorrhoea, headaches, migraines, and pericarditis), and chronic conditions,

including osteoarthritis and rheumatoid arthritis [17]. The chemopreventive and

chemotherapeutic potential of NSAIDs, and the drawbacks of NSAID use, will be the

focus of discussion in Section 1.2.

1.2 Non-Steroidal Anti-Inflammatory Drugs The term NSAID describes several classes of organic molecules that exert their

quintessential anti-inflammatory, anti-pyretic and analgesic actions through the

inhibition of the cyclooxygenase-1/-2 (COX-1/-2) enzymes [17]. This subsequently

decreases the production of prostaglandins, the potent lipid autocrine and paracrine

factors involved in the mediation of the inflammatory response [18]. Over 30 million

people worldwide are estimated to use prescription NSAIDs on a daily basis [19],

making NSAIDs one of the most commonly used pharmaceutical class in the world. In

the USA alone, more than 111 million prescriptions are written for NSAIDs annually; in

addition, NSAIDs that are purchased over-the-counter account for approximately 60%

of all analgesics sold in the USA [20].

Historically, NSAID administration can be traced back thousands of years to the

use of bark from the white willow (Salix alba) by the physicians, Hippocrates (~460–

377 BC) and Dioscorides (~40–90 AD), for the alleviation of fever and musculoskeletal

pain [21]. Salicylic acid (2-hydroxybenzoic acid), a derivative of the active ingredient

salicin was first produced industrially in 1874 [22]; however, due to poor palatability,


5

acetylsalicylic acid (Fig. 1.2) was developed by Felix Hoffman in 1897, and marketed

by Frederick Bayer & Company as aspirin (aspH) in 1899 [23]. Non-salicylate NSAIDs,

indomethacin (indoH, Fig. 1.2) and ibuprofen (ibuH, Fig. 1.2), were first introduced to

the market in 1964 and 1969, respectively [24]. This was followed by the introduction

of diclofenac (dicloH, Fig 1.2) and ketoprofen (ketoH, Fig 1.2) in the mid-1970s [24].

As shown in Figure 1.2, arylalkanoic acids (of the general formula ArCRHCOOH,

where Ar = aryl or heteroaryl; R = H, CH3, alkyl), including salicylic acids (e.g. aspH,

diflunisal (diflH), and salsalate), indolic acids (e.g. indoH), acetic acids (e.g. sulindac),

propionic acids (e.g. ibuH, naproxen (napH), flurbiprofen, and ketoH), and fenamates

(e.g. mefenamic acid (mefH), tolfenamic acid (tolfH), flufenamic acid (fluH),

meclofenamic acid (mecloH), and dicloH), comprise the largest chemical group of

NSAIDs [25]. Notably, the arylalkanoic acid-based NSAIDs contain a carboxylate

moiety (Fig. 1.2). The selectivity preference of NSAIDs from these structural classes

for COX-1 and COX-2 varies depending on the individual NSAID [26, 27].

In the 1980s, Pfizer developed enolic acid NSAIDs, termed ‘oxicams’, which are

derivatives of 4-hydroxy-1,2-benzothiazine 3-carboxamides [28]. Piroxicam (pirox,

Fig. 1.2), marketed as Feldene® (Pfizer, New York, USA), was successfully introduced

into the market in 1982 [28]. Shortly following this other oxicams including

meloxicam (melox, Fig. 1.2), isoxicam (isox), lornoxicam, and tenoxicam (tenox) were

introduced to the market [28]. Oxicams are able to interact with both COX

isoforms [27, 29]; for instance, pirox has been shown to bind preferentially to

COX-1 [27], while melox has been shown to be moderately selective for COX-2 [29].

NSAIDs that are highly selective for COX-2 (termed ‘coxibs’, Fig. 1.2), celecoxib

(Celebrex®, Pfizer, New York, USA), rofecoxib (Vioxx®, Merck & Co. Inc., New Jersey,

USA), and parecoxib (Dynastat®, Pfizer, New York, USA), were introduced to the

market in the late 1990s [30]. The coxibs incorporate a sulfonyl moiety in their

chemical structures (Fig. 1.2), an important chemical feature that allows the molecules

to interact preferentially with the COX-2 active site [30]. The concept of COX-1 and

COX-2 selectivity, and the subsequent physiological consequences, will be discussed

further in Section 1.2.1.


6

Figure 1.2: Chemical classes of non-steroidal anti-inflammatory drugs (NSAIDs). The chemical structure, generic name and IUPAC name of selected NSAIDs representative of each class are presented to highlight the diversity of molecules that comprise NSAIDs [25, 31].


7

1.2.1 Mechanism of action of NSAIDs

Studies performed by Vane and colleagues in the 1970s established that the

primary mechanism through which NSAIDs exert their therapeutic action was through

the inhibition of prostaglandin endoperoxide G/H synthase, the enzyme commonly

referred to as COX [18, 32]. In the early 1990s, several studies established that there

were two distinct forms of the COX enzyme, COX-1 and COX-2 [33-35]. Importantly,

COX-1 and COX-2 differ in their subcellular location, constitutive levels of expression

in tissues throughout the body, and the structure of their active sites.

Immunocytofluorescence studies have established that both isoforms are present in

the endoplasmic reticulum [36]. COX-1 is also located close to the plasma membrane,

whereas COX-2 is specifically associated with the nuclear envelope [37]. COX-1 is

constitutively expressed in most human tissues and is responsible for regulating

essential physiological functions, including platelet aggregation, gastric protection,

and renal function [38]. Conversely, under normal physiological conditions, COX-2

expression is low to negligible in most tissues [39]. Growth factors, as well as

cytokines (including tumour necrosis factor-α, interferon-γ, and interleukins (ILs),

IL-1α and IL-1β) stimulate COX-2 expression [39]. As such, COX-2 is commonly

referred to as the ‘inducible’ form of the enzyme. It is notable that COX-2 is

constitutively expressed in specific regions of the brain [40] and renal tubular cells

[41]. The upregulation of COX-2 occurs primarily in response to inflammatory stimuli

and thus plays a crucial role in mediating pain and inflammation [39]. It is notable

that COX-3, a splice-variant of COX-1 has been identified [42]. Interestingly, some

NSAIDs including dicloH, and the non-NSAID acetaminophen, have been shown to

interact preferentially with COX-3 in vitro [42]; however, the physiological significance

of COX-3 in humans remains poorly understood [43].

COX-1 and COX-2 consist of two catalytic sites, one with COX activity, and the

other with peroxidase (POX) activity, which are located in two narrow channels on

opposite sides of the catalytic domain [44]. The COX catalytic site is responsible for

the oxidation of arachidonic acid to prostaglandin G2 (PGG2). Subsequently, this short-

lived intermediate is reduced to PGH2 by the POX catalytic site (Fig. 1.3) [38]. The

second intermediate, prostaglandin H2 (PGH2), is used as the substrate to form a

number of prostaglandins, namely prostaglandin D2 (PGD2), prostaglandin E2 (PGE2),

prostaglandin F2α (PGF2α), and prostaglandin I2 (PGI2) via prostaglandin-specific

synthases, and thromboxane A2 (TXA2) via thromboxane synthase (Fig. 1.3) [38].


8

Figure 1.3: The conversion of arachidonic acid to prostaglandins via the COX and POX active sites of COX-1/-2. Adapted from Vane et al. [38].

The eicosanoids (the collective term for prostaglandins, thromboxanes and

leukotrienes) are autocrine and paracrine factors that are responsible for the

mediation of specific physiological processes, including maintenance of gastric

mucosa, platelet aggregation, bronchoconstriction, and vascular patency [45].

NSAIDs explicitly interact with the active site in the COX channel, leaving the POX

catalytic site unaffected, preventing the conversion of arachidonic acid to


9

PGG2 [38]. The COX-selectivity of NSAIDs is based on the structure of the NSAID, and

the structure of the COX active sites. The homology of the COX active site is highly

conserved, except for the substitution of a few key residues, as demonstrated in

Figure 1.4 [30]. Specifically, the substitution of Ile523 (COX-1) for Val523 (COX-2)

allows unhindered access of bulkier substituents into a side pocket in the COX-2

channel due to the reduced size of the Val residue (Fig. 1.4) [30]. Additionally, the

substitution of Phe503 (COX-1) for Leu503 (COX-2) increases the space available at

the top of the COX-2 channel (Fig. 1.4) [30]. Non-selective NSAIDs are able to interact

with both the COX-1 and COX-2 active sites via interactions with Arg120 and

Tyr355 (Fig. 1.4), which are stabilised by the carboxylate moiety present in the NSAID

structures, while the phenyl group interacts with Tyr385 further up the hydrophobic

channel (Fig. 1.4) [30].

The ‘coxibs’ incorporate sulfonyl moieties by design to specifically interact (via

hydrogen bonding) with Arg513 in the accessible side pocket (Fig. 1.4), hence their

high selectivity for COX-2 [30]. The majority of NSAIDs interact reversibly with the

COX active site, thus competing with the natural substrate, arachidonic acid [46]. The

exception is aspH, which irreversibly inhibits COX activity via the acetylation of

Ser530 in COX-1 [47] and Ser516 in COX-2 [48]. The irreversible inhibition of COX by

aspH triggers de novo synthesis of COX by cells [49].

Figure 1.4: The key structural residues of the COX-1 and COX-2 active sites. Notably, substitution of Phe503 for Leu503, and Ile523 for Val523 increases the size of the COX-2 active site allowing for the incorporation of bulkier side groups present in COX-2 selective inhibitors. An equivalent numbering system for both isoforms was adopted for clarity. Adapted from Flower [30].


10

1.2.2 NSAIDs: Chemopreventive agents for CRC?

Numerous observational, epidemiological, and clinical studies collectively suggest

that daily use of NSAIDs reduces the occurrence of adenomatous polyps, and

decreases the overall incidence and mortality of CRC in both high- and average-risk

populations ([50], references within). Since Narisawa et al. [51] first described the

suppression of chemical-induced carcinoma in rats by indoH in 1981, over 40 other

rodent studies have shown that NSAIDs (including aspH, indoH, sulindac, ibuH, ketoH,

pirox and celecoxib) inhibit CRC or aberrant crypt formation induced by injection of

azoxymethane or other chemicals known to induce carcinogenesis ([50], references

within).

The chemopreventive effect of sulindac in humans was first recognised by

Waddell et al. [52, 53], whereby sulindac was effective in reducing the number of

polyps in patients with FAP in clinical studies. Given the inherent risk of developing

CRC by the third or fourth decade of life for FAP sufferers [54], a number of studies

focusing on the use of NSAIDs in this population have been performed. Several studies

in ApcMin mice and other murine models of human FAP have shown significant

inhibition of tumour development by NSAIDs including aspH [55], sulindac [56, 57],

and celecoxib [58]. One of the largest randomised, placebo-controlled trials conducted

in the setting of FAP (n = 133), the Colorectal Adenoma/Carcinoma Prevention

Programme [59], found no significant reduction in polyp number in the rectum and

sigmoid colon with daily aspH (600 mg) treatment over a period of one to seven years,

although a trend of reduced polyp size was observed. The suppression of the

formation of adenomatous polyps, and regression of existing polyps by sulindac has

been demonstrated in a number of randomised clinical trials in patients with

FAP [60-62], although one study [63] reported no significant reduction in the

formation of adenomas between the sulindac-treated and placebo groups. In the late

1990s, 77 patients participated in a pivotal trial that demonstrated that compared to

the placebo, a twice-daily dose of celecoxib (100 mg or 400 mg) over six months

significantly reduced the mean number of colorectal polyps and polyp burden [64].

The U.S. Food and Drug Administration (FDA) approved the use of celecoxib as an

adjunctive treatment for FAP (in conjunction with usual care procedures) under

subpart H (accelerated approval) in 2000; however, permanent FDA approval was not

pursued by the manufacturer and the preliminary approval was removed in 2011 [65].


11

The NSAID, aspH, has been shown to reduce the incidence and recurrence of CRC

in individuals with Lynch syndrome [66], and previous adenomatous polyps [67]. The

second Colorectal Adenoma/Carcinoma Prevention Programme trial [66] investigated

the effects of aspH on 1,000 individuals with Lynch syndrome, which was the first

large-scale genetically targeted chemoprevention trial of aspH with cancer as the

primary endpoint. The results suggested that following an average observation period

of 55.7 months cancer incidence was substantially reduced in hereditary CRC carriers

dosed with 600 mg of aspH once daily for a minimum of two years [66]. In a

randomised, double-blind study involving 635 patients that had previously suffered

from CRC [67], those dosed with aspH 325 mg once daily exhibited a significantly

lower risk of recurring colon adenomas (17% incidence of adenomas compared to

27% with the placebo). Furthermore, a randomised, placebo-controlled study, the

Adenoma Prevention with Celecoxib trial [68], evaluated the ability of celecoxib to

reduce the occurrence of endoscopically detected colorectal adenomas. Celecoxib

administration in patients who had previously had adenomatous polyps removed,

showed that following three years the incidence of recurrence of adenomas was

reduced in celecoxib-treated groups to 43.2% (200 mg twice daily) and 37.5%

(400 mg) compared to the placebo group (60.7%) [68]; however, the authors

concluded that due to cardiovascular (CV) risks celecoxib could not be recommended

for routine use for CRC prevention [68].

The results from 35 non-randomised epidemiological studies collectively suggest

that regular use of NSAIDs lowers the incidence of adenomatous polyps, and reduces

the incidence and mortality from CRC in individuals in the general population ([50],

references within). These observations are promising since the highest incidence of

CRC occurs in non-genetically predisposed populations. A recent meta-analysis [69]

compared data from observational studies to data from randomised controlled

trials. Encouragingly, the results from methodically rigorous observational studies

were consistent with the data obtained from randomised trials that regular use of

aspH reduces the long-term risk of CRC (and several other cancers) and the risk of

distant metastasis [69]. However, two large scale randomised placebo-controlled

studies, the Physicians’ Health Study (325 mg aspH, every other day) [70] and the

Women’s Health Study (100 mg of aspH, every other day) [71], showed no reduced

risk of CRC in individuals dosed with aspH compared to the placebo group. The lack of

effect has been debated in the literature; for instance, it is plausible that alternate day


12

dosing and relatively short follow-up periods may not produce a significant

result [72]. Importantly, a significant reduction in mortality from CRC was not

observed until 10 to 20 years post-intervention in a pooled analysis of three trials

examining the effect of daily aspH use [73].

Interestingly, the analysis of data from randomised trials that examined daily

aspH use for other endpoints, such as vascular events, have provided a large

population of aspH-dosed subjects in order to assess anti-cancer effects as a secondary

end point. For instance, pooled analysis of 34 randomised trials of daily aspH for

prevention of vascular events showed that risk of mortality from all cancers decreased

following a follow up period of more than five years [74]. Additionally, pooled analysis

of five large randomised trials (n = 17,285) of daily aspH (≥ 75 mg) versus placebo for

the prevention of vascular events suggested that daily aspH use could reduce the risk

of cancer with distant metastasis [75]. It should be noted that these analyses excluded

results from the Women’s Health Study [72].

In conclusion, the results from numerous epidemiological studies suggest there is

an inverse relationship between the regular use of NSAIDs and the incidence of CRC

and adenomas in a variety of populations. However, given the lack of strong clinical

evidence (from randomised placebo-controlled studies) to show an unequivocal

reduction in the CRC incidence of NSAID-treated subjects, combined with the risk of

adverse side effects, such as gastrointestinal (GI) bleeding, there are currently no

recommendations for the use of NSAIDs in CRC prevention [76, 77].

1.2.3 The role of COX-2 in CRC

In 1994, Eberhart et al. [78] determined that COX-2, but not COX-1, gene

expression was upregulated in approximately 86% of human colorectal carcinomas

and approximately 50% of adenomas compared to levels in adjacent healthy mucosal

cells. Other investigators [79-81] have confirmed the upregulation of COX-2

expression in CRC in human colon cancer tissues and animal models. Additionally,

elevated COX-2 expression has been identified in other types of cancer tissue including

non-small cell lung [81, 82], bladder [83], breast [81, 84], and head and neck [85]

cancers. A number of studies [86-88] have suggested that elevated COX-2 expression

in tumour tissue is associated with increased tumour invasiveness and poor prognosis

for the patient. As demonstrated in Figure 1.5, several molecular pathways are

impacted by the over-expression of COX-2 that ultimately promotes cancer cell growth


13

by: (i) mounting resistance to cell death, (ii) increasing cell proliferation,

(iii) increasing angiogenesis, and (iv) increasing the ability of the cells to metastasise.

Perhaps the most obvious effect of COX-2 over-expression is the increased

production of prostanoids. As demonstrated in Figure 1.5, the production of TXA2,

PGE2 and PGI2 collectively stimulate the normal growth of endothelial cells; however,

in the neoplastic environment, an over-production of prostaglandins can promote

cancer cell growth and proliferation [89]. In particular, an increase in PGE2

concentration triggers an increase in anti-apoptotic B-cell lymphoma 2 (Bcl-2) levels

resulting in the closing of mitochondrial pores, reducing the release of pro-apoptotic

cytochrome C, and thereby diminishing apoptosis [89]. Additionally, there is evidence

to suggest that the delayed progression of the cell cycle may also play a role in the

growth of COX-2 over-expressing cancer cells by increasing the resistance of cells to

apoptosis, thus increasing their tumourigenic potential [90]. The over-production of

PGE2 is believed to suppress dendritic cells, natural killer cells, T cells, and type-1

immunity, but promote type-2 immunity, which promotes tumour immune

evasion [91].

Figure 1.5: Selected molecular pathways affected by the over-expression of COX-2 in neoplasia [89-97]. Bcl-2 = B-cell lymphoma 2; Hp = haptoglobin; IL = interleukin; MMP = metalloproteinase; PGE2 = prostaglandin E2; PGI2 = prostaglandin I2, TXA2 = thromboxane A2; VEGF = vascular endothelial growth factor.


14

The over-expression of COX-2, and subsequently, the increased levels of TXA2, a critical intermediary of angiogenesis [92], and PGE2, which is associated with

increased expression of vascular endothelial growth factor (VEGF) in human tumour

mucosa [93], promotes the production of blood vessels (angiogenesis) necessary for

tumour growth and survival. Additionally, the increased production of PGE2 induces

IL-6, which subsequently induces haptoglobin (Hp), an important protein involved in

angiogenesis [89]. Matrix metalloproteinases (MMPs), which are enzymes that assist

in tumour and vascular cell invasion, have also been shown to be upregulated in COX-2

expressing tumours [91, 94]. A reduction in the expression of IL-12, a potent

anti-angiogenic cytokine, has also been observed in cells expressing COX-2 [95].

The pro-angiogenic effects of COX-2 were demonstrated in one study that employed

COX-2 over-expressing Caco-2 cells (versus low COX-2 expressing Caco-2 cells)

co-cultured with HUVEC endothelial cells [96]. The results showed that when

compared to the low COX-2 expressing Caco-2 cells, the endothelial cells were able to

penetrate a matrix eight-fold faster when co-cultured with COX-2 over-expressing

Caco-2 cells [96], suggesting COX-2 over-expression promotes a local environment

conducive to tumour growth.

The ability of cancer cells to invade distal tissue is an important feature of

metastatic cancers. Experimental evidence suggests that COX-2 over-expression is

associated with increased invasiveness of cancer cells [97]. The activation of MMP-2

and increased RNA levels for the membrane-type MMP-1 were observed in Caco-2

cells programmed to constitutively produce COX-2, compared to regular, low COX-2

producing Caco-2 cells [97]. The production of enzymes such as MMP-2 is associated

with the ability of cancer cells to invade basement membranes through the proteolysis

of aminin, fibronectin, type IV collagen, and proteoglycan [97].

1.2.4 NSAIDs mediate cell death via COX, and non-COX related mechanisms

Given the role that COX-2 plays in many cancers, it is implied that the

anti-neoplastic effects of NSAIDs are potentially associated with the inhibition of

COX-2 in cancer cells. The reversal of the pro-neoplastic effects of COX-2

over-expression has been demonstrated by NSAIDs in a number of studies [92, 96-99].

For instance, increased invasiveness and PGE2 production of Caco-2 cells programmed

to constitutively produce COX-2 was dose-dependently reversed by sulindac

treatment [97]. Additionally, treatment with aspH or the COX-2 inhibitor,


15

NS-398 (N-[2-(cyclohexyloxy)-4-nitrophenyl]methanesulfonamide), inhibited PGE2

production in a dose-dependent manner in the same COX-2-expressing Caco-2 cell

line [96]. Experiments performed by Chan et al. [98] suggest that indoH and sulindac

sulfide mediate colon cancer cell death by increasing arachidonic acid levels (rather

than a decrease in prostaglandin levels), which stimulates the conversion of

sphingomyelin to ceramide, a known mediator in apoptosis. In another study [92] the

selective COX-2 inhibitor, VU08 (N-(2-phenylethyl)indomethacin amide), decreased

TXA2 production which resulted in reduced endothelial migration and a decline in

fibroblast growth factor-induced corneal angiogenesis. Finally, the inhibition of COX-2

by celecoxib has been shown to reduce tumour growth and metastasis in xenograft

tumour models and suppress basic fibroblast growth factor 2-induced angiogenesis of

the rodent cornea [99].

The concept that inhibition of COX-2 in cancer cells contributes to the anti-cancer

activity of NSAIDs is somewhat oversimplified; an abundance of experimental

evidence suggests that not all NSAIDs induce cell death via COX-2 regulated pathways.

For instance, a number of studies have shown that NSAIDs are toxic towards tumour

cells that exhibit little to no COX-2 expression [100-102]. Furthermore, exogenously

applied prostaglandins have failed to prevent the inhibitory effects of NSAIDs on

cancer cell growth [103, 104]. The diversity of molecular pathways that NSAIDs affect

is presented in Table 1.1 (for more detailed reviews refer to [105] and [106]).

1.2.5 The drawbacks of prolonged NSAID use

The prolonged use of NSAIDs is associated with a number of side effects. These

side effects are mainly attributable to the mechanism of action of the NSAIDs

responsible for the desired therapeutic effects, and as such, the cells that produce

constitutive COX-1 (GI mucosa, kidneys, and platelets) can be affected adversely by

prolonged NSAID use. The decrease in the constitutive expression of protective

prostaglandins in the gastric mucosa as a result of COX-1 inhibition is understood to

be responsible for the manifestation of GI symptoms, including ulceration, perforation

and bleeding in the GI tract [107]. Of concern, the risk of GI bleeding associated with

aspH therapy has been shown to increase with age [108]. The GI side effects induced

by NSAID use encompass mild symptoms, including heartburn, nausea, dyspepsia and

abdominal pain, as well as endoscopically identified mucosal lesions and serious GI

complications [19]. NSAID-related GI problems are estimated to account for


16

Table 1.1: The major COX-2 dependent or independent pro-apoptotic pathways induced by NSAIDs. Apoptosis-inducing pathways NSAIDs COX-2-dependent COX-2 dependent accumulation of arachidonic acid and ceramide synthesis

sulindac [98], indoH [98]

Inhibition of IL-1β and phorbol 12-myristate 13-acetate induction preventing COX-2 mRNA and protein expression

aspH [109], sodium salicylate [109]

COX-2-independent

Oxidative stress and disruption of redox balance aspH [110], sulindac [111], indoH [112], tolfH [113]

Inhibition of Wnt/β-catenin signalling aspH [114], indoH [114], dicloH [115], sulindac [116]

Modulation of cyclins sulindac [117], tolfH [118]

Depletion of polyamines aspH [119], sulindac [120], indoH [121]

Inhibition of proteasomal function aspH [122] Increased expression of mda/IL24 sulindac [123] Depletion of survivin aspH [124], tolfH [125]

Modulation of nuclear factor kappa B signalling aspH [126], sulindac [127], ibuH [128], indoH [129], dicloH [115], tolfH [130]

Mitogen activated protein kinase modulation aspH [131], indoH [132] Caspase activation and Bcl-2 protein modulation aspH [133, 134], indoH [134] Inhibition of store-operated calcium entry sulindac [135], mefH [135] Increased activity of caspase-3 to cleave PARP1 mefH [136] Interactions with cell membrane phospholipids mefH [137], celecoxib [138] Suppression of specificity protein 1 tolfH [139] Activation of NSAID-activated gene-1 tolfH [113], sulindac [140]

Increased production of gastric glutathione-S-transferase aspH [141], ibuH [141], indoH [141]

100,000 hospitalisations per year in the USA [19]. Additionally, an estimated 16,500

deaths per year in the USA are attributable to NSAID use [142]. The production of

prostaglandins (particularly PGE2 and PGI2) by COX-1 in the kidneys is important for

renal function; inhibition can lead to dysfunction of normal renal regulation of blood

pressure [143]. The inhibition of TXA2 production by COX-1 in platelets can result in

prolonged bleeding time and a mild, systemic haemostatic defect [144].

In the case of selective COX-2 inhibition, PGI2 production is decreased and TXA2 is

left unchecked leading to the development of CV issues [145]. While a number of

clinical studies have shown that COX-2 selective inhibitors, celecoxib and rofecoxib,


17

reduce the risk of GI and renal toxicity, their long-term use has been associated with

an increased risk of CV events [146, 147]. In 2004, Merck announced the early

termination of the Adenomatous Polyp Prevention on Vioxx trial [148]. In this trial,

2,586 patients with a history of colorectal adenomas received either 25 mg of

rofecoxib daily (n = 1,257) or placebo (n = 1,299) [146]. The trial was terminated due

to an increased risk of CV events (including myocardial infarctions and ischemic

cerebrovascular events) in patients receiving the drug for more than 18 months [148].

Vioxx (rofecoxib) was subsequently withdrawn from the market [148]. This prompted

an investigation into the risk of CV events for patients prescribed celecoxib in the

Adenoma Prevention with Celecoxib trial [147]. Celecoxib (200 mg or 400 mg, twice

daily) was also determined to increase the risk of CV events 2.5-fold and 3.4-fold,

respectively, compared to the placebo group leading to the discontinuation of the trial

in late 2004 [147] and the FDA issuing a black box warning for celecoxib in

2005 [149]. One report suggests that the CV risk of the non-selective NSAID, dicloH, is

comparable to the COX-2 selective NSAID, rofecoxib [150]. Due to elevated risks of

hepatotoxicity, the COX-2 selective NSAID, lumiracoxib, was not approved for use in

the USA, and was withdrawn from the market in Australia and Europe in 2007 [151].

Additionally, a several non-selective NSAIDs (including dicloH, sulindac, and

nimesulide) are associated with increased risk of liver toxicity resulting in function

abnormalities and fatal liver injury [152, 153].

Given the current risk of side effects, it has been argued that the daily use of

NSAIDs for chemoprevention in the general population (where the risk of developing

CRC is less than 6%) would not be beneficial unless other health endpoints could be

achieved [50]. As such, the safety of NSAIDs needs to be improved. Strategies to

improve NSAID safety include: (i) the synthesis of novel modified NSAIDs, including

nitric oxide releasing NSAIDs [154], phospho-NSAIDs [155], and sulfide releasing

NSAIDs [156]; (ii) the development of dual COX and lipoxygenase

inhibitors [157-160]; and (iii) the coordination of NSAIDs to metal ions. The latter

approach is the focus of this Thesis.

1.3 Metal complexes of NSAIDs (metal-NSAIDs) The coordination of organic ligands to a metal ion(s), offers a number of potential

medicinal advantages over the administration of organic ligands alone.

These advantages include: (i) the alleviation of undesirable side effects; (ii) sustained


18

in vivo delivery of the organic ligand, i.e. increased/prolonged bioavailability (thus the

metal complex essentially acts as a ‘pro-drug’); and (iii) synergism between the metal

and the organic ligand resulting in the improvement of the original intended biological

effect. A wide variety of metal-NSAID combinations have displayed biological actions

including anti-inflammatory activity [161-169], analgesia [164], anti-ulcer

activity [161-165], anti-cancer activity [162], and anti-microbial activity [170-175]

(for concision, the latter will not be discussed). Of particular interest to the research

presented in this Thesis is the complexation of metals with phenylalkanonic acids, and

more specifically, fenamic acid NSAIDs (i.e. fluH, mecloH, mefH and tolfH) and

salicylates (i.e. aspH and diflH). For brevity, the following Sections will focus on the

discussion of data from studies pertaining to metal complexes with phenylalkanonic

acid NSAIDs, with brief mention of some studies examining metal-NSAIDs containing

oxicam ligands.

1.3.1 Anti-inflammatory and anti-ulcerative properties of metal-NSAIDs

A number of metal-NSAID complexes have shown improved (or similar) in vivo

anti-inflammatory and analgesic activities compared to their parent NSAID [161-169],

as summarised in Table 1.2. One of the perceived advantages of the employment of

metal-NSAIDs over uncomplexed NSAIDs is the alleviation of common dose-limiting

side effects, in particular, GI ulcerations. Encouragingly, in vivo studies performed

using rodent models [161, 162, 164, 165] have shown that gastric ulceration induced

by commonly used NSAIDs (including ibuH, indoH and napH) was reduced by

complexation with Cu(II), Ru(II), and Zn(II), as demonstrated by a decrease in lesion

index scores (Table 1.2). Additionally, [Cu2(indo)4(DMF)2] (where DMF = N,N-

dimethylformamide) and [Zn2(indo)4(DMA)2] (where DMA = N,N-dimethylacetamide)

have been shown to reduce intestinal ulceration in rats when compared to indoH

alone [163]. However, unlike [Cu2(indo)4(DMF)2], the [Zn2(indo)4(DMA)2] complex

induced the same degree of ulceration in the stomach as indoH alone [163]. As

demonstrated in Table 1.2, this result was in concordance with previous studies [169].

These results validate the importance of the choice of the metal centre on the

biological activity of the complexes. The Ru(III)-oxicam complexes,

[RuCl2(piroxH2)(piroxH)] and [RuCl2(tenoxH2)(tenoxH)], demonstrated significant

anti-inflammatory activity in the carrageenan-induced rat paw oedema assay, albeit a

smaller effect was achieved than equimolar pirox [176]. The authors


19

Tabl

e 1.

2: R

esul

ts fr

om se

lect

ed in

viv

o st

udie

s com

pari

ng th

e an

ti-in

flam

mat

ory,

ana

lges

ic, o

r ant

i-ulc

erat

ive

activ

ity o

f met

al-N

SAID

com

plex

es

and

unco

mpl

exed

NSA

IDs.

Ref

[167

]

[168

]

[165

]

[161

]

[164

]

Resu

lts

NSA

ID

76 ±

0.8

%

61.5

± 2

.3 %

1 h,

73.

48 %

6

h, 9

1.98

%

24 h

, 97.

04 %

45.2

± 1

.9 %

Neg

ativ

e co

ntro

l 65.

3 ±

1.4

%

597.

2 ±

43.5

Neg

ativ

e co

ntro

l 180

.9 ±

37.

4 5

mg/

kg, 2

6.3

± 0.

9 10

mg/

kg, 2

4.4

± 1.

0 20

mg/

kg, 1

6.8

± 0.

9 40

mg/

kg, 1

5.8

± 1.

2

66.2

%

A si

mila

r dos

e-de

pend

ent d

ecre

ase

in p

aw

volu

me

was

obs

erve

d be

twee

n Zn

-nap

and

na

pH o

n an

equ

ival

ent m

olar

bas

is

Met

al-N

SAID

(1) 8

2.6

± 0.

8 %

(2

) 93±

0.9

%

(1) 7

1 ±

1.1

%

(2) 1

7.5

± 0.

4 %

(3

) 81.

5 ±

1.3

%

1 h,

74.

91 %

6

h, 8

6.98

%

24 h

, 97.

04 %

(1) 4

4.9

± 3.

3 %

(2

) 46.

0 ±

2.7

%

(1) 2

90.5

± 3

1.3

(2) 3

25.5

± 1

4.4

5 m

g/kg

, 28.

2 ±

2.4

10 m

g/kg

, 19.

0 ±

1.5

20 m

g/kg

, 13.

2 ±

1.0

40 m

g/kg

, 10.

6 ±

1.6

86.5

%

Endp

oint

mea

sure

d

Inhi

bitio

n of

carr

agee

n-in

duce

d ra

t paw

oed

ema

(%

co

mpa

red

to u

ntre

ated

co

ntro

l paw

± S

.D.)

Inhi

bitio

n of

carr

agee

n-in

duce

d ra

t paw

oed

ema

(%

co

mpa

red

to u

ntre

ated

co

ntro

l paw

± S

.D.)

Inhi

bitio

n of

carr

agee

n-in

duce

d ra

t paw

oed

ema

(%

co

mpa

red

to sa

line-

trea

ted

cont

rol p

aw)

Incr

ease

in p

aw v

olum

e (m

ean

% ±

s.e.

)

Lesi

on i

ndex

(ul

cers

sco

red

base

d on

size

/sev

erity

)

Num

ber o

f abd

omin

al

cons

tric

tions

in 2

0 m

in

(mea

n ±

s.e.m

.)

Max

imum

pos

sibl

e ef

fect

of

anal

gesi

a (%

)

Incr

ease

in p

aw v

olum

e (%

)

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema.

Dos

e: 0

.1 m

mol

/kg

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema.

Dos

e: 0

.1 m

mol

/kg

(1)

and

(3),

(2) w

as te

sted

at

0.01

mm

ol/k

g du

e to

toxi

city

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema.

Dos

e: 1

0 m

g/kg

ace

cH

and

mol

ar e

quiv

alen

t [Zn

(ace

c)2]

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema.

Dos

e: 1

3.5

mg/

kg ib

uH,

15.6

mg/

kg (1

), 17

.3 m

g/kg

(2)

Ulce

ratio

n: G

astr

ic u

lcer

s wer

e sc

ored

in ra

ts re

stra

ined

for 1

7-h

at

30-m

in p

ost d

ose

Anal

gesia

: Ace

tic a

cid-

indu

ced

abdo

min

al co

nstr

ictio

n in

mic

e. O

ral

dosa

ge

Anal

gesia

: Tai

l-flic

k te

st in

mic

e

Dose

= 1

0 m

g/kg

, i.p

.

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema

Assa

y/M

odel

Met

al-N

SAID

*

(1) [

Co(t

olf)

2(H

2O) 2

] (2

) [Zn

(tol

f)2(

H2O

)]

(1) [

Co(m

ef) 2

(H2O

) 2]

(2) [

Ni(m

ef) 2

(H2O

) 2]

(3) [

Zn(m

ef) 2

]

[Zn(

acec

) 2]

(1) [

Cu2(

ibp)

4]

(2) [

Ru2C

l(ibp

) 4]

Zn-n

ap†


20

Tabl

e 1.

2 (c

ont.)

: Res

ults

from

sele

cted

in v

ivo

stud

ies c

ompa

ring

the

anti-

infla

mm

ator

y, a

nalg

esic

, or a

nti-u

lcer

ativ

e ac

tivity

of m

etal

-NSA

ID

com

plex

es a

nd u

ncom

plex

ed N

SAID

s. Re

f [1

64]

[163

]

[162

]

[163

]

[169

]

*Dep

roto

nate

d N

SAID

s: a

cec

= ac

eclo

fena

c; d

iclo

= d

iclo

fena

c; ib

p =

ibup

rofe

n; in

do =

indo

met

haci

n; m

ef =

mef

enam

ic a

cid;

nap

= n

apro

xen;

tolf

= to

lfena

mic

aci

d.

DMF

= N

,N-d

imet

hylfo

rmam

ide;

DM

A =

N,N

-dim

ethy

lace

tam

ide.

† Che

mic

al co

mpo

sitio

n not

def

ined

.

Resu

lts

NSA

ID

8 m

g/kg

, 4.5

± 0.

7 16

mg/

kg, 7

.0 ±

0.8

25 %

(d)/

11 %

(v)

Neg

ativ

e co

ntro

l: 48

% (d

)/60

% (v

)

36 m

m2

Neg

ativ

e co

ntro

l: 0

mm

2

140

mm

2

Neg

ativ

e co

ntro

l: 0

mm

2

25 %

(d)/

11 %

(v)

Neg

ativ

e co

ntro

l 48

% (d

)/60

% (v

)

39 m

m2

Neg

ativ

e co

ntro

l 3 m

m2

135

mm

2

Neg

ativ

e co

ntro

l 1 m

m2

9.83

mg/

kg, 6

9.26

Neg

ativ

e co

ntro

l, 2.

20

Met

al-N

SAID

8 m

g/kg

, 1.5

± 0

.0

16 m

g/kg

, 2.7

± 1.

5

30 %

(d)/

15 %

(v)

5 m

m2

5 m

m2

18 %

(d)/

10 %

(v)

37 m

m2

3 m

m2

9.83

mg/

kg, 7

0.80

3.

29 m

g/kg

, 0.7

5

Endp

oint

mea

sure

d

Lesi

on in

dex

(ulc

ers s

core

d ba

sed

on si

ze a

nd se

veri

ty)

(mea

n ±

s.e.m

.) In

crea

se in

paw

dia

met

er

(d) a

nd v

olum

e (v

) (%

com

pare

d to

bas

elin

e)

Mea

n ga

stri

c ulc

erat

ion

(sum

mat

ion

of th

e ar

ea o

f m

acro

scop

ic u

lcer

atio

n)

Mea

n in

test

inal

ulc

erat

ion

(sum

mat

ion

of th

e ar

ea o

f m

acro

scop

ic u

lcer

atio

n)

Incr

ease

in p

aw v

olum

e (d

) an

d vo

lum

e (v

) (%

com

pare

d to

bas

elin

e)

Mea

n ga

stri

c ulc

erat

ion

(sum

mat

ion

of th

e ar

ea o

f m

acro

scop

ic u

lcer

atio

n)

Mea

n in

test

inal

ulc

erat

ion

(sum

mat

ion

of th

e ar

ea o

f m

acro

scop

ic u

lcer

atio

n)

Lesi

on in

dex

(ulc

ers

scor

ed

base

d on

num

ber a

nd si

ze)

Ulce

ratio

n: G

astr

ic u

lcer

s in

rats

wer

e sc

ored

follo

win

g do

sing

ove

r 7 d

ays

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema.

Dos

e =

indo

H 1

0 m

g/kg

, [C

u 2(in

do) 4

(DM

F)2]

= 1

1.7

mg/

kg

Ulce

ratio

n: G

astr

ic u

lcer

s wer

e sc

ored

in

rats

sacr

ifice

d 3-

h po

st-d

ose;

inte

stin

al

ulce

rs w

ere

scor

ed in

rats

sacr

ifice

d 24

-h p

ost-

dose

Infla

mm

atio

n: C

arra

geen

-indu

ced

rat

paw

oed

ema.

Dos

e =

indo

H 1

0 m

g/kg

, [Z

n 2(in

do) 4

(DM

A)2]

= 1

3.3

mg/

kg

Ulce

ratio

n: G

astr

ic u

lcer

s wer

e sc

ored

in

rats

sacr

ifice

d 3-

h po

st-d

ose;

inte

stin

al

ulce

rs w

ere

scor

ed in

rats

sacr

ifice

d 24

-h

post

-dos

e. D

ose

= in

doH

10

mg/

kg

Ulce

ratio

n: G

astr

ic u

lcer

s wer

e sc

ored

in

rats

dos

ed 4

tim

es o

ver 2

day

s. D

ose

= in

doH

9.8

3 m

g/kg

, Zn-

Indo

3.7

5 m

g/kg

Assa

y/M

odel

Met

al-N

SAID

*

Zn-n

ap†

[Cu 2

(indo

) 4(D

MF)

2]

[Zn 2

(indo

) 4(D

MA)

2]


21

concluded that the two complexes were superior anti-inflammatory agents compared

to pirox due to the reduction in gastric damage in rats treated with the two complexes

compared to the damage induced by pirox alone.

The in vivo success of [Cu2(indo)4(DMF)2] has led to the development of

commercially available therapeutic products. For instance, these Cu-based NSAIDs

have been used in oral dosage forms for veterinary use, and for topical human

use [31]. An ethanolic gel-base of Cu-salicylate (marketed as Alcusal® by Australian

company Mentholatum) was formerly available for topical temporal relief of pain and

inflammation in humans [31]. Cu-Algesic® (containing [Cu2(indo)4(DMF)2]) is

available for veterinary use (dogs and horses) in Australasia, South East Asia and

South Africa due to its low toxicity profile compared to uncomplexed indoH [31].

1.3.2 Chemotherapeutic properties of metal-NSAIDs

Given the considerable evidence that NSAIDs possess chemopreventive and

chemotherapeutic properties (discussed in Section 1.2.2-1.2.4), it is reasonable to

hypothesise that metal-NSAID complexes may also possess this property. To this end,

a number of studies [167, 168, 175, 177-180] have assessed the toxicity of

metal-NSAIDs towards transformed (cancer) cells in vitro. As shown in Table 1.3, the

IC50 values (i.e. the concentration of drug required to produce 50% inhibition)

obtained from in vitro cancer cell studies allow a direct comparison of the toxicity of

the metal-NSAIDs to the analogous free NSAID, and the commonly used anti-cancer

drug, cisplatin (cis-diamminedichloroplatinum(II)). Encouragingly, a number of metal-

NSAID complexes possess improved (or similar) potency compared to the analogous

free NSAID (when considering the molar ratio of NSAID in the metal-complex) and

cisplatin against a number of cancer cell lines in vitro (Table 1.3). Many of these

studies varied the metal centre while employing the same NSAID ligands, revealing the

importance of the identity of the metal centre and the anti-cancer potential of the

complexes (Table 1.3). These studies also reveal that each of the metal-NSAID

complexes showed increased potency (i.e. selectivity) towards specific cancer cell lines

(Table 1.3).

Recently, a series of organometallic Ru(II)- and Os(II)-p-cymene complexes were

studied where the NSAIDs, indoH or dicloH, were attached to the organometallic

moieties via monodentate (pyridine/phosphine) or bidentate (bipyridine) ligands,

producing piano-stool Ru(II) and Os(II) arene complexes [181]. The modified NSAIDs


22

Table 1.3: Results from in vitro studies comparing the anti-cancer activity of metal-NSAID complexes with uncomplexed NSAIDs, and the anti-cancer agent, cisplatin (CisPt).

Metal-NSAID* Cell line† Assay‡ Time IC50 value (μM) Ref M+-NSAID NSAID CisPt [Co(tolf)2(H2O)2] T-24

MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

38.3 44.9 76.8 33.4

62.4 87.9 145 214

41.7 8.0 1.5 0.7

[167]

[Cu(tolf)2(H2O)]2 T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

13.9 26.1 31.4 5.3

62.4 87.9 145 214

41.7 8.0 1.5 0.7

[167]

[Mn(tolf)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

3.9 41.6 65.3 149

62.4 87.9 145 214

41.7 8.0 1.5 0.7

[167]

[Ni(tolf)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

35.3 54.5 54.1

100.1

62.4 87.9 145 214

41.7 8.0 1.5 0.7

[167]

[Zn(tolf)2(H2O)] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

41.0 41.3 57.9 123

62.4 87.9 145 214

41.7 8.0 1.5 0.7

[167]

[Co(mef)2(H2O)2]

T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

27.0 93.6

115.9 < 174.4

81.2 149.2 168.3 178.2

41.7 8.0 1.5 0.7

[168]

[Cu(mef)2(H2O)]2 T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

7.77 25.1

< 100.3 19.5

81.2 149.2 168.3 178.2

41.7 8.0 1.5 0.7

[168]

[Mn(mef)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

35.1 72.6

< 175.6 < 175.6

81.2 149.2 168.3 178.2

41.7 8.0 1.5 0.7

[168]

[Ni(mef)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

24.0 59.3 86.6 63.4

81.2 149.2 168.3 178.2

41.7 8.0 1.5 0.7

[168]

[Zn(mef)2]

T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

37.7 40.7

111.2 74.4

81.2 149.2 168.3 178.2

41.7 8.0 1.5 0.7

[168]

[Cd(mef)2(CH3OH)]n T-24 MCF-7

MTT MTT

72 h 72 h

0.86 0.12

81.2 149.2

40.9 1.84

[175]

[Ph3Sn(mef)] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

0.29 0.68 7.21 7.23

80.1 146.4 167.8 168.5

7.99 41.66 0.69 1.53

[177]


23

Table 1.3: (cont.) Results from in vitro studies comparing the anti-cancer activity of metal-NSAID complexes with uncomplexed NSAIDs, and the anti-cancer agent, cisplatin (CisPt). Metal-NSAID* Cell line† Assay Time IC50 value (μM) Ref M+-NSAID NSAID CisPt [Ph3Sn(flu)] T-24

MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

4.78 1.26 60.44 5.85

99.6 120.9 171.8 192.0

7.99 41.66 0.69 1.53

[177]

[Cd(meclo)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

3.3 5.1

29.8 22.7

70.2 63.1

139.1 133.0

41.7 8.0 1.5 0.7

[178]

[Cu(meclo)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

5.3 38.2 41.2 61.9

70.2 63.1

139.1 133.0

41.7 8.0 1.5 0.7

[178]

[Mn(meclo)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

3.8 39.8 27.4 25.3

70.2 63.1

139.1 133.0

41.7 8.0 1.5 0.7

[178]

[Zn(meclo)2(H2O)2] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

3.2 32.5 >244 24.2

70.2 63.1

139.1 133.0

41.7 8.0 1.5 0.7

[178]

[Ph3Sn(meclo)] T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

0.29 0.68 7.21 7.23

63.1 70.2 133

139.1

7.99 41.66 0.69 1.53

[177]

[Cd2(diclo)41.5(MeOH). 2(H2O)]n

T-24 MCF-7 A549 L-929

MTT SRB SRB SRB

72 h 72 h 72 h 72 h

4.4 5.8

30.2 24.6

71.2 62.8

138.9 132.8

40.9 7.8 1.5 0.7

[179]

cis-[Pt(NH3)2(melox)2]

CH1 HeLa SW480 HCT-15 HCT-11

MTT MTT MTT MTT MTT

96 h 96 h 96 h 96 h 96 h

0.6 92 96

194 119

>160 230

≥285 ≥300 >320

0.16 0.37 3.5 n.d. 2.7

[180]

cis-[Pt(NH3)2(isox)2] CH1 HeLa SW480 HCT-15 HCT-116

MTT MTT MTT MTT MTT

96 h 96 h 96 h 96 h 96 h

0.37 3.5 35

116 90

>80 >160 >320 >320 >320

0.16 0.37 3.5 n.d. 2.7

[180]

* In this instance the NSAID is deprotonated: tolf = tolfenamic acid; mef = mefenamic acid; flu = flufenamic acid; meclo = meclofenamic acid; diclo = diclofenac; tolm = tolmetin; indo = indomethacin; melox = meloxicam; isox = isoxicam. en = ethylenediamine. †A549 = human small cell lung cancer; C6 = rat glioma; CH1 = human ovarian carcinoma; HCT-116 = human colorectal cancer; HCT-15 = human colorectal cancer; HeLa = human cervical cancer; L-929 = murine fibroblast; MCF-7 = human breast cancer; SW480 = human colorectal cancer; T-24 = human bladder cancer; T-47D = human breast cancer. ‡MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, SRB = sulforhodamine B. “n.d.” = not determined. Value in bold text indicates the metal-complex IC50 value is comparable or superior to cisplatin.

https://en.wikipedia.org/wiki/Di-

https://en.wikipedia.org/wiki/Di-

https://en.wikipedia.org/wiki/Thiazole

https://en.wikipedia.org/wiki/Phenyl


24

in complexation with Ru(II)- and Os(II) showed promising in vitro anti-proliferative

activity against human ovarian cancer cells (A2780, and cisplatin-resistant

A2780cisR) [181].

The Pt-oxicam complexes, cis-[Pt(NH3)2(melox)2] and cis-[Pt(NH3)2(isox)2],

demonstrated greater cytotoxicity compared to melox and isox alone against a panel

of cancer cell lines ([180], Table 1.3). The complexes were particularly effective

against CH1 human ovarian carcinoma cells, with IC50 values almost equal to that of

cisplatin [180]. The Cu(II), Zn(II), Co(II), and Ni(II) complexes of melox were found to

have a more pronounced cytotoxic/cytostatic effect than melox alone against HeLa

human cervical cancer cells and 8MGBA glioblastoma multiforme cells [182].

Pirox complexes with transition metal ions, V(IV), Mo(VI), Mn(II), U(VI) and Fe(III),

induce apoptosis within 24 h in HL-60 leukaemia cells, where little activity was

possessed by pirox until treatment for 57 h [183].

While there have been a substantial number of studies that have evaluated the

toxicity of metal-NSAIDs towards cancer cell lines in vitro (Table 1.3), the in vivo

anti-cancer effects of these complexes have been marginally explored. The effects of

[Cu2(indo)4(DMF)2] or indoH on the prevention of aberrant crypt foci (ACF) formation

in the colon of rats has been evaluated [162]. In the study, male Sprague-Dawley rats

were dosed with [Cu2(indo)4(DMF)2] for a 4-week period in conjunction with

azoxymethane (an inducer of ACF formation). The results from the study showed that

[Cu2(indo)4(DMF)2] treatment (3.8 mg/kg) reduced the formation of ACF by 43%,

while indoH treatment (3 mg/kg) resulted in 73% reduction. Although indoH was

more effective at preventing ACF formation, the [Cu2(indo)4(DMF)2]-treated rats

suffered significantly less gastric and intestinal ulceration ([162], Table 1.2).

1.3.3 Other biologically relevant properties of metal-NSAIDs

It is notable that a variety of metal-NSAIDs have shown strong protein binding

in vitro, and exhibited the ability to intercalate with DNA and exacerbate free radical

scavenging properties. These properties have been particularly well-studied in

metal-NSAID complexes of Co(II) or Cu(II) containing dicloH [184], diflH [185],

fluH [186], indoH [187], mefH [188], napH [184, 189], or tolfH [190, 191] ligands, or

the complexes, [M(II)(NSAID)n(L)n], where L represents incorporated nitrogen donor

heterocyclic ligands such as 2,2’-bipyridine (bipy), 1,10-phenanthroline (phen), and

pyridine (py). Additionally, the DNA binding activity of Mn(II) or Ni(II) derivatives


25

containing dicloH [192, 193], mefH [194] or nifH have been reported [195].

For instance, the Co(II) derivative of mef, [Co(mef)2(MeOH)4] (MeOH = methanol), and

Co(II)-mef complexes incorporating nitrogen donor heterocyclic ligands,

[Co(mef)2(bipy)(MeOH)2], [Co(mef)2(phen)(MeOH)2] and [Co(mef)2(py)2(MeOH)2],

possess relatively strong binding affinities for calf thymus DNA [188]. The four Co(II)-

mef complexes exhibited relatively high DNA binding constants (Kb);

however, [Co(mef)2(py)(MeOH)2] was the only complex with increased DNA binding

affinity compared to mefH alone (Kb values of 3.32 × 105 M−1 and 1.05 × 105 M−1,

respectively) [188]. Furthermore, competitive studies with ethidium bromide

indicated mefH and the Co(II)-mef complexes most likely interact with calf thymus

DNA by the intercalative mode [188]. Additionally, all four Co(II)-mef complexes

exhibited superior superoxide and free radical scavenging ability, and similar or

superior hydroxyl free scavenging ability in comparison to mefH [188]. Given that

NSAIDs exhibit anionic charge in vivo (deprotonated at pH 7) their interaction with the

polyanionic backbone of DNA is unlikely [196]. Uncharged metal-NSAIDs, however,

might overcome these repulsion forces [196], depending on their intracellular

stability. The ability of the aforementioned metal-NSAIDs to intercalate with DNA

with high binding affinity may be an important chemotherapeutic property given that

many clinically established anti-cancer agents exert their anti-neoplastic effects

through DNA interactions.

1.4 Bismuth One vital consideration in the rational drug development of metal-NSAIDs is the

choice of metal centre, since many metals are inherently toxic at low doses. Bi has no

essential physiological role [197]. However, Bi induces therapeutic effects in humans

and has been used in healthcare for centuries, predominantly to treat GI disorders, and

as an anti-microbial agent ([198] and references within). Bi remains employed in

contemporary medicine primarily for its anti-ulcerative and anti-bacterial actions

(discussed in Sections 1.4.1.2 and 1.4.1.3, respectively). Importantly, the use of Bi in

medicinal formulations is associated with low systemic toxicity and transient side

effects [199]. Several novel Bi compounds show promising in vitro and in vivo

anti-cancer activity (discussed in Section 1.4.1.4). As such, complexation of Bi with

NSAIDs may prove to be a successful strategy to reduce NSAID-induced GI injury and

improve the desired chemopreventive and chemotherapeutic activity of the NSAID.


26

Located in group 15 of the periodic table, Bi (atomic number 83) is considered

metallic (or semi-metallic) in character, as opposed to other group 15 elements,

arsenic and antimony, which are classified as metalloids [200]. Bi is found in the

trivalent or pentavalent oxidation state (Bi(III) and Bi(V)), where the former is the

more stable and the most biologically relevant oxidation state [201]. Given Bi(V)

possesses powerful antioxidant properties in aqueous solution (Bi(V)/Bi(III) potential

E° = 2.03 V), Bi(V) is not believed to be stable in biological systems [201]. Bi(III) is

readily hydrolysed (pKa = 1.51) in aqueous solution, and has a high affinity for oxygen,

nitrogen and thiolate containing ligands [201]. It is thought that the Bi(III) ion has

high mobility within cells due to the kinetic lability of thiolate ligands and resultant

millisecond time scale required for exchange with free thiols [202].

Furthermore, there is a wealth of experimental evidence that shows that the

therapeutic and anti-bacterial actions of Bi arise primarily through interactions with

proteins (discussed further in Sections 1.4.1.1–1.4.1.3).

1.4.1 Bismuth in medicine

Louis Odier first described the use of bismuth subnitrate for medicinal purposes

in 1786 as a treatment for dyspepsia [203]. For over two centuries, Bi preparations

have been used medicinally to treat an assortment of diseases, primarily

gastroenterological and microbiological in nature, as demonstrated in Table 1.4.

Currently, three Bi drugs are marketed for the treatment of GI conditions and

Helicobacter pylori infection in combination therapies with antibiotics. These include:

bismuth subsalicylate (BSS; Pepto-Bismol®, the Procter & Gamble Company,

Cincinnati, Ohio, USA), colloidal bismuth subcitrate (CBS; De-Nol®, Gist Brocades, Delft,

The Netherlands), and ranitidine bismuth citrate (RBC; Tritec® and Pylorid®,

GlaxoWellcome, Hertfordshire, UK) [204]. RBC combines the mucosal protection of Bi

with the anti-secretory properties of ranitidine (N,N-dimethyl-5-(3-nitromethylene-7-

thia-2,4-diaaoctyl)-furan-2-meth-amine) [204]. The chemical structures of bismuth

salicylate (bismuth 2-hydroxybenzoate), bismuth citrate (bismuth 2-hydroxypropane-

1,2,3-tricarboxylate) and RBC are shown in Table 1.4. The ‘sub’ prefixes of BSS and

CBS refers to the poorly understood chemical structure of the formulations, although

recent studies have shed light on the possible conformation and polymerisation of BSS

[205] and CBS [206] in acidic environments.


27

Table 1.4: Historical and currently used medicinal bismuth compounds.

Compound; trade name/s

Chemical structure* Medical disorder

Bibrocathol; Noviform®, Posiformin®

Eye infections [207]

Bismuth potassium (or sodium) tartrate

Syphilis [208]

Bismuth subcarbonate

Antacid [208] Gastric/duodenal ulcer [209] Gastritis [208] Wounds [208]

Bismuth sodium thioglycollate; Thio-Bismol, Bistrimate®

Syphilis [208, 210] Malaria [211]

Bismuth subgallate; Devrom®

Gastritis [208] Haemostatic adjunct during tonsillectomy [212] Internal deodorant [213, 214] Wounds [208]

Bismuth subnitrate

Gastritis [208] Gastric/duodenal ulcer [215, 216] H. pylori infection [215-217] Hypertension [218] Wounds [208, 219]

Bismuth subsalicylate, BSS; Pepto-Bismol®

Diarrhea [220, 221] Gastric/duodenal ulcers [222] H. pylori infection [223]


28

Table 1.4: cont. Historical and currently used medicinal bismuth compounds. Compound; trade name/s

Chemical structure* Medical disorder

Colloidal bismuth subcitrate (CBS); De-Nol®

Dyspepsia [224] Gastric/duodenal ulcers [225, 226] H. pylori infection [227]

Bismuth tribromophen-ate; Xeroform® dressing gauze

Wounds [208, 228]

Ranitidine bismuth citrate (RBC); Tritec®, Pylorid®

Gastric/duodenal ulcers [229] H. pylori infection [223, 229]

*Basic chemical structure shown. The exact chemical structures of the compounds with the “sub” prefix are generally unknown or poorly defined.

The development of new Bi drugs with superior activity against H. pylori has been

an active area of research over the past decade (recently reviewed in [200] and [230]).

It is notable that recent in vitro studies have shown that Bi-containing complexes

possess anti-fungal (Saccharomyces cerevisiae) [231], anti-viral (severe acute

respiratory syndrome coronavirus) [232, 233] and anti-parasitic (Leishmania

major) [234-236] activities, highlighting the broad medicinal potential of Bi

compounds. The following Sections of this review will briefly describe the current

primary medicinal use of Bi as a GI protective (Section 1.4.1.2), and an anti-bacterial

agent (Section 1.4.1.3), and the potential use of Bi complexes as anti-cancer

agents (Section 1.4.1.4).

1.4.1.1 Absorption, transport and distribution of Bi(III)

The mechanisms of absorption, transport and systemic distribution of Bi(III) in

mammals are poorly understood. Studies have shown that up to 99% of ingested Bi is


29

passed through the feces without alteration [237]. However, it has also been reported

that Bi levels in the gastric mucosa significantly increase to a peak level of 30–60 μg/L

(0.14–0.28 μM) following a single oral dose of BSS, CBS or RBC [238]. Bi levels in the

blood have been reported to increase 51- to 1483-fold following ingestion [238], with

a relatively rapid rate of systemic uptake (15–60 min) [239]. Once absorbed, binding

to plasma proteins conceivably facilitates the systemic transport of Bi. Experimental

evidence suggests that Bi binds strongly to the iron transport enzyme, transferrin

[240, 241]. It has also been reported that Bi bound to lactoferrin (a member of the

transferrin family) was recognised by IEC-6 rat intestinal cells and blocked the uptake

of the native Fe–lactoferrin complex, leading to the suggestion that Bi is transported

into cells by transferrin receptor pathways [242]. Bi has also been shown to bind to

human serum albumin [243], one of the most abundant proteins in the bloodstream.

However, in the presence of a large excess of albumin (albumin/transferrin 13:1, at pH

7.4, 10 mM bicarbonate), 70% of Bi was bound to transferrin [243], suggesting

transferrin is the most likely target of Bi in the bloodstream.

In addition to iron transport proteins, Bi(III) has been shown to bind to other

cellular peptides and proteins that are believed to confer intracellular uptake

pathways. For instance, it has been reported [244] that Bi has a high affinity for

glutathione, a tripeptide that is present in cells in relatively high concentrations

(0.5–7 mM) and is believed to play a role in Bi transport in cells and biofluids.

Found in most tissues, metallothioneins (MTs), are low molecular mass proteins

(~ 60 amino acids) comprised of a cysteine-rich amino acid sequence [245].

The cysteine-rich character provides a high capacity for the metals ion (such as Zn(II),

Cu(I), Cd(II) and Bi(III)) to bind. As such, the primary functions of MTs are believed to

be metal sequestration, metal homeostasis and metabolism, and the detoxification of

metals [245]. Studies by Sun et al. [246] have shown that Bi binds strongly to the

cysteine residues of MT in a stoichiometry of Bi:MT = 7:1 and can readily replace

Zn(II) and Cd(II) ions. Additionally, Bi administration has been shown to increase MT

production in humans [247, 248]. A number of studies [249-253] suggest that

intracellular Bi accumulates in lysosomes, which are acidic organelles that are used to

remove debris from within the cell, most likely as part of cellular detoxification events.

1.4.1.2 Bismuth as a GI protective agent

The success of Bi as a GI protective agent is due in part to its anti-bacterial action

against H. pylori (discussed in more detail in Section 1.4.1.3). However, common Bi


30

compounds such as BSS and CBS have been shown to exert non-bactericidal effects in

the GI system, and have been employed as clinically effective agents for the treatment

of a variety of GI ailments (Table 1.4). It has been suggested that the protective GI

properties of Bi compounds, such as those produced by BSS and CBS, are due to the

interaction of colloidal particles of Bi(III) with the acidic conditions of the stomach

resulting in the formation of precipitated polymeric Bi(III) structures [254].

For instance, BSS can undergo ligand exchange and form insoluble bismuth

oxychloride (BiOCl) ions in the stomach [255]. Studies performed with CBS led to the

suggestion that [Bi(cit)2Bi]2– molecules form colloidal particles such as [Bi6O4(cit)4]6–

and [Bi12O8(cit)8]12– at neutral pH [256], but rearrange to form sheets and 3D polymers

under acidic conditions due to ligand exchange reactions [206]. It is likely that the

precipitation of BSS or CBS products and the formation of a polymeric coating on ulcer

craters prevents further erosion of the mucosa [254]. Additionally, it has been

suggested that the stabilisation of the mucous layer by CBS increases the resistance to

back-diffusion of hydrogen ions without significantly modifying the ion-exchange

properties of the mucous [257]. Other means by which Bi has been suggested to

attenuate GI protection include scavenging reactive oxygen species (ROS) [258, 259],

inactivating the digestive enzyme pepsin [260, 261], stimulating luminal release of

PGE2 [226], stimulating epidermal growth factor [262], restoration of

phosphoinositide specific phospholipase C activity in response to injury, thus

increasing mucosal cell proliferation [263], and the activation of the calcium-sensing

receptor and the intracellular increase in Ca2+ concentration, increasing mitogen-

activated protein kinase activity, which leads to the proliferation of normal human

gastric mucous epithelial cells [264].

1.4.1.3 Bismuth as an anti-bacterial agent

In 1984, Marshall and Warren discovered that H. pylori, a bacterium that localises

in human gastric mucosa, was present in the stomach lining of patients suffering from

gastritis and stomach ulcerations [265]. A decade later, the World Health Organisation

classified H. pylori as a class I carcinogen, since chronic infection can lead to the

formation of dysplasia and eventually gastric cancer if left untreated [266]. While only

a small proportion (< 3%) of individuals infected with H. pylori are expected to

develop gastric malignancies as a result of H. pylori colonisation, it is estimated that up

to 50% of the world’s population carry H. pylori [267]. Two Bi-based H. pylori


31

treatment regimens were approved by the FDA in the mid-1990s [223].

These consisted of: (i) a two week course of RBC (400 mg twice daily), and

clarithromycin (500 mg three times daily), followed by a further two week course of

RBC (400 mg twice daily), or (ii) a two week quadruple regime combining

BSS (525 mg four times daily), metronidazole (250 mg four times daily), and

tetracycline (500 mg four times daily), in combination with a histamine-2 receptor

antagonist (ulcer healing agent) as directed over 4 weeks [223]. However, Bi-based

therapies were often employed in relapsed forms of H. pylori treatment after primary

treatment with a triple-therapy of antibiotics had failed [268].

More recently, concerns have been raised over increasing antibiotic resistance,

particularly clarithromycin resistance in H. pylori [268]. Given there is no evidence

that H. pylori are able to develop resistance to Bi, it has been suggested that Bi-based

regimes (including the Bi-based quadruple therapy, comprising of a three-in-one

capsule of potassium bismuth subcitrate, metronidazole, and tetracycline (marketed

as Pylera®), and a proton pump inhibitor) may be considered a more suitable

first-line treatment for H. pylori infection over the coming decades [268].

While many Bi compounds are active against H. pylori, the exact mechanism of

action remains unclear. Experimental evidence from in vitro studies suggests that the

anti-bacterial activity of Bi may be multimodal; for instance, proposed mechanisms

include the disruption of the glycocalyx cell wall through the formation of Bi

aggregates [269], disruption of the adherence of the bacterium to gastric cells [255],

cessation of ATP synthesis [270], the inhibition of urease, an enzyme which protects

the bacterium from the acidic environment of the stomach [271], and inhibition of

other enzymes which are responsible for toxicity to gastric mucosal cells, such as

bacterial alcohol dehydrogenase [272]. Additionally, evidence shows that H. pylori

sequesters iron through the utilisation of host-specific lactoferrin, thus Bi(III) binding

to lactoferrin may disrupt bacterial iron acquisition [273]. In many cases the

interaction of Bi with the cells is physicochemical in nature, thus potentially explaining

the lack of bacterial resistance to Bi compounds [269].

1.4.1.4 Bismuth as an anti-cancer agent: Diagnostic and therapeutic roles

In addition to the GI protective and anti-bacterial properties of Bi, there is

growing substantiation that Bi compounds have the potential to play a significant and

diverse role in cancer diagnostics and anti-cancer therapies. For instance, isotopes of


32

Bi (212Bi or 213Bi) conjugated to various monoclonal antibodies have been investigated

as a method of targeted radiotherapy [274, 275]. Nanoparticles of Bi have shown

promise as contrast agents for computed tomography, providing equal or superior

quality imaging to currently used contrast agents, owing to the large atomic number of

Bi [276]. Bi has also displayed synergism when administered with chemotherapeutic

agents in vivo. For example, co-administration of Bi and cisplatin reduced the

nephrotoxic side effects induced by cisplatin [277-279]; Bi-induced MT synthesis was

determined as the primary mechanism of this effect [247, 248]. The induction of MT

synthesis by Bi has been shown to reduce the cardiotoxic side effects of adriamycin in

mice [280, 281].

A variety of novel Bi(III) complexes have shown promising in vitro

chemotherapeutic activity. Heterocyclic organobismuth(III) complexes, N-tert-butyl-

bi-chlorodibenzo[c,f][1,5]azabismocine (Fig. 1.6, 1), bi-chlorodibenzo[c,f][1,5]-

thiabismocine (Fig. 1.6, 2) and bi-chlorophenothiabismin-S,S-dioxide (Fig. 1.6, 3) have

shown potent cytotoxicity against leukaemia cell lines; Molt-4, U937, HL-60, NB4 and

K562 [282]. Apoptosis induced by 2 (Fig. 1.6) in HL-60 cells and was associated with

enhanced generation of intracellular ROS, loss of mitochondrial transmembrane

potential (Δψm), cytochrome C release from the mitochondria to the cytosol, and

activation of caspases -3, -8, and -9, suggesting the main mechanism of action of the

compound involves mitochondrial damage. In further work, 2 (Fig. 1.6) was shown to

cause G2/M cell cycle arrest in HeLa cells at low concentrations (~1.0 µM) leading to

apoptosis [283]. This was accompanied by the disruption of the microtubule network

in the cells, suggesting the compound may be a novel tubulin polymerisation

inhibitor [283].

The organobismuth complex, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl]-

pyrrolidine (Fig. 1.6, 4), showed anti-proliferative activity against several human

cancer cell lines (IC50 values 0.880−8.478 µM), with particularly high potency towards

the human acute promyelocytic leukaemia cell line, NB4 [284]. Treatment of the NB4

cells with 4 (Fig. 1.6) induced apoptosis, decreased Δψm and increased the production

of ROS, similar to the cellular effects induced by 2 (Fig. 1.6); however, 4 (Fig. 1.6) did

not show any disruption of tubulin polymerisation in vitro. It is notable that

substitution of the metal centre of 4 (Fig. 1.6) with Sb(III) resulted in no cytotoxicity in

NB4 cells (up to 10 μM) suggesting the combination of the Bi(III) centre and the

conjugated structure of the diazenylpyrrolidine moiety were essential for bioactivity.


33

Figure 1.6: Chemical structures of Bi compounds that have shown potential anti-cancer activity in vitro.


34

The [Bi(AcPhH)Cl2], [Bi(AcpClPhH)Cl2] and [Bi(AcpNO2PhH)Cl2] complexes

(Fig. 1.6, 5, where AcPhH2 = 2,6-diacetylpyridine bis(benzoylhydrazone), AcpClPhH2 =

2,6-diacetylpyridine bis(para-chlorobenzoylhydrazone), and AcpNO2PhH2 = 2,6-

diacetylpyridine bis(para-nitrobenzoylhydrazone)) showed enhanced toxicity

compared to the free ligands and analogous Sb(III) complexes, against a panel of

human cancer cell lines (IC50 values ranging from 0.09−10.56 µM). Importantly,

[Bi(AcpClPhH)Cl2] and [Bi(AcpNO2PhH)Cl2] showed relatively low potency towards

human peripheral blood mononuclear cells (IC50 values 117.0 µM and 22.94 µM,

respectively), demonstrating selectivity towards cancer cell lines [285]. More

recently, the same research group synthesised Bi(III) complexes with

2-acetylpyridine- and 2-benzoylpyridine-derived hydrazones [286], which also

showed promising cytotoxic activities against HL-60, Jurkat and THP-1 leukaemia

cells, and on MCF-7 and HCT-116 solid tumour cells. The IC50 values of the

AcpNO2PhH2 ligand and [Bi(2AcpNO2Ph)Cl2] complex (where AcpNO2PhH2 =

2-acetylpyridine-para-nitro-phenylhydrazone) were three-fold smaller when HCT-116

cells were cultured in soft-agar (3D) than when cells were cultured in monolayer (2D),

suggesting these compounds may show potential against solid tumours in vivo [286].

In other studies, the seven-coordinated Bi(III) complex [Bi(L)(NO3)2(CH3CH2OH)]

(HL = 2-acetylpyridine N(4)phenylthiosemicarbazone) (Fig. 1.6, 6) exhibited

comparable cytotoxicity to cisplatin (5.22 µM and 1.2 µM), and higher toxicity than the

free ligand (IC50 = 94.7 µM) [287]. The complex was also active against HCT-116, HeLa

and HepG2 cells [287]. Similar to other Bi(III) complexes discussed in this Section,

6 (Fig. 1.6) induced apoptotic cell death via an increase in ROS production and

reduced Δψm in HepG2 cells [287].

The Bi(III) complex, BiTPC (where TPC = 1,4,7,10-tetrakis(2-pyridylmethyl)-

1,4,7,10-tetraazacyclododecane) (Fig. 1.6, 7), a water soluble Bi compound, was found

to exhibit potent cytotoxicity towards melanoma B16-BL6 cells (IC50 = 41 nM) which

was 100 times more potent than cisplatin [288]. Although it is generally accepted that

the most likely targets of Bi compounds are proteins, DNA binding studies showed that

BiTPC could bind non-covalently to DNA under physiologically relevant

conditions [288].

While the aforementioned Bi complexes have demonstrated encouraging in vitro

activity, evaluation of a number of other Bi compounds have shown varied success in

animal models. For instance, the arylbismuth(III) oxinates, Bi(Ox)3, Bi(Ox)2I,


35

PhBi(Ox)2.EtOH, PhBi(MeOx)2, [NaPhBi(Ox)3] and [KPhBi(Ox)3] (where Ox =

quinolin-8-olate and MeOx = 2-methylquinolin-8-olate) showed promising anti-cancer

activity against murine leukaemia cells, L1210 (≤ 0.56 µM), the corresponding

cisplatin resistant cell line, L1210/DDP (≤ 0.37 µM), and human ovarian cancer cells,

SKOV-3 (≤ 2.9 µM) in vitro [289]. However, the compounds, PhBi(Ox)2.EtOH and

[KPhBi(Ox)3], showed no significant anti-tumour activity against murine

plasmacytoma, and [KPhBi(Ox)3] showed no significant anti-tumour activity against

P388 leukaemia in murine models [289]. The lack of in vivo activity was attributed to

problems with delivery of the compounds due to their poor solubility in biological

fluids [289].

Similarly, dithiocarbamate Bi complexes including, Bi(N,N-dimethyldithio-

carbamato)3, Bi(N,N-diethyldithiocarbamato)3, Bi(pyrrolidyldithio-carbamato)3 and

Bi(morpholidyldithiocarbamato)3 were found to be cytotoxic against several cancer

cell lines and, in many cases, more potent than established organic drugs [290].

Treatment with the Bi(N,N-diethyldithiocarbamato)3 complex slowed tumour growth

in mice inoculated with HT-29 cells relative to tumour growth in untreated mice [290].

The complex showed greater anti-tumour activity in OVCAR-3 inoculated mice

whereby tumour size began to reduce after 15 days of Bi(N,N-diethyldithio-

carbamato)3 treatment [290].

Finally, a nine-coordinate Bi(III) complex, [Bi(H2L)(NO3)2]NO3 (where H2L = 2,6-

diacetylpyridine bis(4N-methylthiosemicarbazone)) (Fig. 1.6, 8) has been synthesised

and its in vitro and in vivo anti-cancer activity evaluated [291]. The IC50 value of

8 (Fig. 1.6) was approximately three times lower than the ligand alone against the

K562 cell line [291]. In mice xenografted with H22 (mice hepatoma) cells, tumour

growth was inhibited by 61.6% and 43.5% following 7-day treatment with 8 (Fig. 1.6,

10 mg/kg), and the anti-cancer drug mitoxantrone (0.4 mg/kg), respectively,

compared to the control group [291]. Despite promising in vitro activity and some

preliminary success with in vivo studies, currently there are no Bi-containing

complexes in clinical trials or used in therapies for the treatment of cancer [292].

1.4.2 Bismuth toxicity

It is generally reported in the literature that Bi possesses uncharacteristically low

toxicity, particularly when compared to the group 15 counterparts, arsenic and

antimony. Compared to many medications, the side effect profile for Bi (taken at the


36

appropriate dose) is quite limited [237]; transient, harmless side effects such as

darkened stools [199] and blackening of the tongue [293] are the most commonly

reported. However, in the 1970s, concerns over the safety of Bi compounds (namely

bismuth subcarbonate, bismuth subgallate, and CBS) were raised due to cases of Bi

intoxication resulting in cases of encephalopathy in Australia [294, 295] and

France [296], which led to the withdrawal of some Bi medications from the market in

these countries. This severe side effect was reported in patients who consumed high

doses (3–20 g of Bi per day over a 6–36 month period) [296]. It is notable, however,

that Bi-induced encephalopathy can be reversed upon the cessation of Bi treatment,

with full restoration of neurological function over a few weeks [237].

A limited number of studies addressing the bioaccumulation of Bi in mammals

have been reported. Following a single dose of RBC in mice, Bi deposits have been

observed in the stomach, duodenum, ileum and kidney (2.5–10 h after dosing) [297].

Interestingly, Bi was evident in the lymph nodes, liver, spleen, kidney and

macrophages in the GI lamina propria 1–9 weeks following administration of RBC,

although Bi was absent from the GI epithelium (consistent with cellular

turnover) [297]. However, no signs of morphological changes in the Bi-treated tissues

or adverse effects in the behaviour of the mice were observed [297]. Although the link

between bioaccumulation and toxicity remains poorly understood, this study

highlights the potential for systemic accumulation of Bi, even with short-term

treatments. Another concern is the recent number of studies highlighting the

biotransformation of Bi into methylated species such as trimethylbismuth (Me3Bi) by

bacterial species present in intestinal flora in humans (recently reviewed in [255]).

Given evidence that Me3Bi species are more toxic to bacteria than inorganic

species [298], there may be implications for the regulation of gut flora in humans;

however, the health implications of the Me3Bi formation by human microflora requires

further investigation.

While the aforementioned toxicity of Bi raises some concern over the safety of

Bi-based medications, there is also evidence to suggest that when used periodically, Bi

drugs are generally safe and well tolerated. A systemic review and meta-analysis of

35 clinical studies concluded that Bi drugs (namely bismuth subnitrate, BSS, CBS, and

RBC, Bi dosage ranging from 400–1200 mg/day and duration from 7–56 days) used

for the treatment of H. pylori infection are safe and well-tolerated when administered

either alone, or in combination with antibiotics [199]. The only side effect that was


37

significantly more common in patients assigned Bi-based therapies was darkened

stools (not to be confused with GI bleeding) [199]. Accordingly, Pepto-Bismol is

currently available over-the-counter without a prescription in the USA, Canada and

UK. Given that the majority of regimes employing Bi for treatment of H. pylori

infection require less than two weeks of administration, the use of Bi in prolonged

chemopreventive therapies (months to years) would require careful monitoring to

determine if any adverse side effects or signs of toxicity negate the potential positive

therapeutic effects.

1.5 BiNSAID Complexes Despite experimental, epidemiological, and clinical evidence that indicates that

long-term use of NSAIDs is associated with reduced incidence of CRC (Section 1.2.2),

the prolonged use of NSAIDs for chemoprevention is not recommended for the general

population [77]. This is due, in part, to the risks that NSAIDs impose on the GI system.

It is, therefore, plausible to consider the combination of NSAIDs with GI protective Bi

(as a single therapeutic agent) in order to alleviate the GI side effects of NSAIDs [299].

Despite the specific GI protective advantage of Bi compared to other metals, and the

conceivable synergistic chemotherapeutic action of Bi coordinated to an organic

ligand, there are a limited number of reports [299, 300] detailing the synthesis and

biological activity of BiNSAID complexes, highlighting the potential for development in

this area.

The use of Bi(III) compounds for protection against NSAID-induced GI injury is

not without precedence. A number of studies have shown that pre-treatment or

co-administration with Bi compounds such as BSS [301, 302], CBS [303, 304], and

RBC [305] with NSAIDs reduces GI ulceration in vivo. In a double-blind randomised

three-way crossover study, healthy male volunteers (n = 24) received nine doses of

placebo, aspH (900 mg), or aspH and RBC (900 mg and 800 mg, respectively) at 12-h

intervals, with a two-week washout period between each treatment [305]. The

authors reported that co-administration with RBC significantly reduced the number of

aspH-induced erosions and micro-bleeding, suggesting RBC confers substantial

protection against aspH-related GI injury. In a more recent study, the GI-protective

effect of CBS, and other potential agents (misoprostol and omeprazole), against gastric

injury induced by indoH was evaluated in Wistar albino rats [306].

CBS (70 mg/kg/day and 15 mg/kg/day) was administered 30 min prior to indoH


38

(5 mg/kg/day) once daily for five consecutive days. The authors report that the mean

ulcer scores achieved by CBS treatment were not significantly reduced compared to

indoH alone, although the mean ulcer scores and histopathological examination

suggests that the higher dose CBS had a greater effect than the lower dose CBS.

Pre-treatment with misoprostol (100 g/kg and 10 g/kg) or omeprazole (5 mg/kg)

were both found to significantly reduce mean ulcer scores [306].

The in situ synthesis of a 1:1 Bi-diclo complex via the addition of diclofenac

sodium to CBS aqueous solution (3:1 molar ratio, dicloH to Bi) has been

reported [300]. The study concluded that diclo coordinated through the carboxylate

moiety to Bi, although the exact composition of the complex was not established.

Unexpectedly, oral administration of the Bi-diclo complex resulted in a higher stomach

lesion index than rats treated with diclofenac sodium alone, while a physical mixture

of diclofenac sodium and CBS induced the lowest lesion index; however, no statistical

analysis was performed to validate the significance of these results. The authors

hypothesised that the ulcerogenic activity of the Bi-diclo complex arose from the

decomposition of the complex into diclo particles, which were over three times

smaller than the diclofenac sodium particles, hence allowing diclo from the

decomposed Bi(III) complex to cover a greater surface area of the stomach mucosa. If

such an observation holds merit, the stability and resultant decomposition products of

other BiNSAID complexes in gastric fluid could be understatedly important and should

be established prior to in vivo studies.

Several homoleptic Bi(III) tris-carboxylate complexes derived from the

deprotonated acids of NSAIDs have been synthesised, with the general formula

[Bi(L)3]n, where L = deprotonated 5-chlorosalicylic acid, diflH, fenbufen, fluH, ibuH,

ketoH, mefH, napH, sulindac, or tolfH [299]. Elemental analysis determined that three

NSAID moieties are present for each Bi(III) ion, while complementary infrared (IR)

and nuclear magnetic resonance (NMR) spectroscopic data suggested that Bi(III)

adopts a bidentate coordination with the NSAID ligands through the deprotonation of

the carboxylic acid moiety. Crystal structures of the complexes could not be obtained,

hence the exact structural and polymeric propensities of each of the BiNSAID

complexes has not been definitively ascertained. As such, a representation of the

prospective structures of Bi(tolf)3, Bi(mef)3 and Bi(difl)3 is shown in Figure 1.7.

The anti-bacterial activity of all ten BiNSAIDs against three laboratory strains of

H. pylori, B128, 251 and 26695, compared favourably to the anti-bacterial activity of


39

BSS and [Bi(Hsal)3]n (≥ 6.25 µg/mL versus ≥ 12.5 µg/mL), while the NSAIDs alone

showed no inhibitory properties.

Figure 1.7: Representation of the chemical structures of: (a) Bi(tolf)3 and Bi(mef)3, and (b) Bi(difl)3, as inferred by elemental analysis, IR and NMR data [299].

The effect of the BiNSAIDs on cancer cell lines is an area that is yet to be explored,

and subsequently forms the basis of the investigations of this Thesis. Specifically, the

in vitro anti-cancer potential of three of the aforementioned BiNSAIDs, Bi(tolf)3,

Bi(mef)3 and Bi(difl)3, will be investigated.

1.6 Project Aims A number of epidemiological, observational, and clinical studies suggest that

NSAIDs are able to reduce the incidence of cancer, in particular, CRC; however, side

effects, such as GI bleeding, limit the prolonged daily use of NSAIDs. Bi has been used

to treat an array of GI diseases for centuries, with an acceptable toxicity profile at low

doses over extended periods of administration. Thus it is hypothesised that the

combination of Bi and an NSAID in a single compound may potentially prevent the

adverse GI side effects of NSAIDs, if used daily as a chemopreventive agent.

There are several in vitro aspects that need to be investigated before such claims

could be tested in vivo. The first of these is the recognition that when administering a

drug as a chemopreventive, it is expected that the patient is prone to developing

cancer (specifically adenomatous polyps). A starting point then is to look at the


40

interaction of BiNSAIDs with transformed (cancer) cells. As the primary proposed

mechanism for chemoprevention by NSAIDs is COX-2 inhibition.

Subsequently, the overarching aim of this Thesis was to investigate the in vitro

potency and biological effects of the BiNSAID complexes, Bi(tolf)3, Bi(mef)3 and

Bi(difl)3, against a human colon cancer cell line, HCT-8, and hence the suitability of the

complexes for in vivo studies.

The specific aims of the project were to:

1. Evaluate the hydrolytic stability of the complexes in cell medium using NMR

spectroscopy to ascertain the biologically available structures of the BiNSAIDs

(Chapter 2).

2. Evaluate the in vitro drug toxicity and mode of cell death of the BiNSAIDs

compared to uncomplexed NSAIDs in HCT-8 cells (Chapter 2).

3. Assess the ability of the BiNSAIDs to inhibit the production of PGE2 by HCT-8

cells in order to determine if the complexes exhibit COX inhibition (Chapter 2).

4. Determine the cellular uptake and subcellular localisation of the BiNSAIDs in

HCT-8 cells to establish whether stability and toxicity correlated with cellular

uptake and to determine the subcellular fate of the Bi contained in the BiNSAID

complexes (Chapter 3).

5. Assess the biomolecular changes induced by BiNSAID treatment compared to

NSAIDs alone in live HCT-8 cells using synchrotron radiation infrared

microspectroscopy (SR-IRMS) (Chapter 4).

6. Examine the effects of BiNSAID- and NSAID-induced cell death on the

distribution and concentration of abundant cellular glycerophospholipids (PC,

PE, and PS) using electrospray ionisation tandem mass spectrometry (ESI-

MS/MS) (Chapter 5).

7. Evaluate the in vitro toxicity of other previously untested Bi(III) complexes of

aminoarenesulfonic acids, indole-carboxylic acids, and hydroxamic acids

against HCT-8 colon cancer cells, and determine whether complexation with Bi

improves the potency of the ligands (Chapter 6).


41

1.7 References [1] D. Hanahan and R. A. Weinberg. (2000) The hallmarks of cancer, Cell, 100, 57-70. [2] American Cancer Society. (2015) Global cancer facts & figures, 3rd Edition, Atlanta:

American Cancer Society, 1-64. [3] D. L. Worthley and B. A. Leggett. (2010) Colorectal cancer: Molecular features and

clinical opportunities, Clin. Biochem. Rev., 31, 31-38. [4] S. Kozuka, M. Nogaki, T. Ozeki and S. Masumori. (1975) Premalignancy of the

mucosal polyp in the large intestine: II. Estimation of the periods required for malignant transformation of mucosal polyps, Dis. Colon Rectum, 18, 494-500.

[5] American Cancer Society. (2011) Global cancer facts & figures, 2nd Edition, Atlanta: American Cancer Society, 1-60.

[6] J. Ferlay, I. Soerjomataram, M. Ervik, R. Dikshit, S. Eser, C. Mathers, M. Rebelo, D. Parkin, D. Forman and F. Bray, GLOBOCAN 2012 v1.0, Cancer incidence and mortality worldwide: IARC CancerBase No. 11. Available from: http://globocan.iarc.fr, [accessed: Feb. 11, 2015]

[7] Australian Institute of Health and Welfare. (2017) Cancer in Australia 2017. Cancer series no.101. Cat. no. CAN 100. Canberra: AIHW.

[8] A. Jemal, R. Siegel, E. Ward, Y. Hao, J. Xu and M. J. Thun. (2009) Cancer statistics, 2009, CA Cancer J. Clin., 59, 225-249.

[9] H. T. Lynch and A. de la Chapelle. (2003) Hereditary colorectal cancer, N. Engl. J. Med., 348, 919-932.

[10] K. Cooper, H. Squires, C. Carroll, D. Papaioannou, A. Booth, R. F. Logan, C. Maguire, D. Hind and P. Tappenden. (2010) Chemoprevention of colorectal cancer: systematic review and economic evaluation, Health Technol. Assess., 14, doi:10.3310/hta14320.

[11] American Society of Clinical Oncology. (1998) Hereditary colorectal cancer syndromes, In Cancer genetics & cancer predisposition testing, Vol. 1 (American Society of Clinical Oncology, Ed.), Alexandria (VA).

[12] O. Holme, M. Bretthauer, A. Fretheim, J. Odgaard-Jensen and G. Hoff. (2013) Flexible sigmoidoscopy versus faecal occult blood testing for colorectal cancer screening in asymptomatic individuals, Cochrane Database Syst. Rev., 9, Cd009259.

[13] D. C. Rockey, E. Paulson, D. Niedzwiecki, W. Davis, H. B. Bosworth, L. Sanders, J. Yee, J. Henderson, P. Hatten, S. Burdick, A. Sanyal, D. T. Rubin, M. Sterling, G. Akerkar, M. S. Bhutani, K. Binmoeller, J. Garvie, E. J. Bini, K. McQuaid, W. L. Foster, W. M. Thompson, A. Dachman and R. Halvorsen. (2005) Analysis of air contrast barium enema, computed tomographic colonography, and colonoscopy: prospective comparison, Lancet, 365, 305-311.

[14] C. J. Fyock and P. V. Draganov. (2010) Colonoscopic polypectomy and associated techniques, World J. Gastroenterol., 16, 3630-3637.

[15] M. Crosara Teixeira, M. I. Braghiroli, J. Sabbaga and P. M. Hoff. (2014) Primary prevention of colorectal cancer: myth or reality?, World J. Gastroenterol., 20, 15060-15069.

[16] D. Yona and N. Arber. (2006) Coxibs and cancer prevention, J. Cardiovasc. Pharmacol., 47, 76-81.

[17] Australian Medicines Handbook 2017 (online) (2017) NSAIDs. Adelaide: Australian Medicines Handbook Pty Ltd, Available from: https://amhonline.amh.net.au/ [accessed: Nov. 4, 2017]

[18] J. R. Vane. (1971) Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs, Nat. New Biol., 231, 232-235.

[19] G. Singh. (2000) Gastrointestinal complications of prescription and over-the-counter nonsteroidal anti-inflammatory drugs: a view from the ARAMIS database. Arthritis, Rheumatism, and Aging Medical Information System, Am. J. Ther., 7, 115-121.


42

[20] Retail & Provider Perspective, National Prescription Audit, 1999 – 2000. Plymouth, PA: IMS Health, 2000.

[21] J. G. Mahdi. (2010) Medicinal potential of willow: A chemical perspective of aspirin discovery, J. Saudi Chem. Soc., 14, 317-322.

[22] J. Kolbe. (1874) Synthesis of salicylic acid, Prakt. Chem., 118, 107. [23] T. J. Rinsema. (1999) One hundred years of aspirin, Med. Hist., 43, 502-507. [24] K. Brune and B. Hinz. (2004) The discovery and development of antiinflammatory

drugs, Arthritis Rheum., 50, 2391-2399. [25] C. S. Boynton, C. F. Dick and G. H. Mayor. (1988) NSAIDs: an overview, J. Clin.

Pharmacol., 28, 512-517. [26] T. D. Warner, F. Giuliano, I. Vojnovic, A. Bukasa, J. A. Mitchell and J. R. Vane. (1999)

Nonsteroid drug selectivities for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal toxicity: a full in vitro analysis, Proc. Natl. Acad. Sci. U.S.A., 96, 7563-7568.

[27] J. A. Mitchell, P. Akarasereenont, C. Thiemermann, R. J. Flower and J. R. Vane. (1993) Selectivity of nonsteroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase, Proc. Natl. Acad. Sci. U.S.A., 90, 11693-11697.

[28] S. Xu, C. A. Rouzer and L. J. Marnett. (2014) Oxicams, a class of nonsteroidal anti-inflammatory drugs and beyond, IUBMB Life, 66, 803-811.

[29] L. Churchill, A. G. Graham, C.-K. Shih, D. Pauletti, P. R. Farina and P. M. Grob. (1996) Selective inhibition of human cyclo-oxygenase-2 by meloxicam, Inflammopharmacology, 4, 125-135.

[30] R. J. Flower. (2003) The development of COX2 inhibitors, Nat. Rev. Drug Discov., 2, 179-191.

[31] J. E. Weder, C. T. Dillon, T. W. Hambley, B. J. Kennedy, P. A. Lay, J. R. Biffin, H. L. Regtop and N. M. Davies. (2002) Copper complexes of non-steroidal anti-inflammatory drugs: an opportunity yet to be realized, Coord. Chem. Rev., 232, 95-126.

[32] S. H. Ferreira, S. Moncada and J. R. Vane. (1971) Indomethacin and aspirin abolish prostaglandin release from the spleen, Nat. New Biol., 231, 237-239.

[33] W. L. Xie, J. G. Chipman, D. L. Robertson, R. L. Erikson and D. L. Simmons. (1991) Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing, Proc. Natl. Acad. Sci. U.S.A., 88, 2692-2696.

[34] D. A. Kujubu, B. S. Fletcher, B. C. Varnum, R. W. Lim and H. R. Herschman. (1991) TIS10, a phorbol ester tumor promoter-inducible mRNA from Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue, J. Biol. Chem., 266, 12866-12872.

[35] B. S. Fletcher, D. A. Kujubu, D. M. Perrin and H. R. Herschman. (1992) Structure of the mitogen-inducible TIS10 gene and demonstration that the TIS10-encoded protein is a functional prostaglandin G/H synthase, J. Biol. Chem., 267, 4338-4344.

[36] J. C. Otto and W. L. Smith. (1994) The orientation of prostaglandin endoperoxide synthases-1 and -2 in the endoplasmic reticulum, J. Biol. Chem., 269, 19868-19875.

[37] I. Morita, M. Schindler, M. K. Regier, J. C. Otto, T. Hori, D. L. DeWitt and W. L. Smith. (1995) Different intracellular locations for prostaglandin endoperoxide H synthase-1 and -2, J. Biol. Chem., 270, 10902-10908.

[38] J. R. Vane, Y. S. Bakhle and R. M. Botting. (1998) Cycloxygenases 1 and 2, Ann. Rev. Pharmacol. Toxicol., 38, 97-120.

[39] W. L. Smith, R. M. Garavito and D. L. DeWitt. (1996) Prostaglandin endoperoxide H synthases (cyclooxygenases)-1 and -2, J. Biol. Chem., 271, 33157-33160.

[40] K. Yamagata, K. I. Andreasson, W. E. Kaufmann, C. A. Barnes and P. F. Worley. (1993) Expression of a mitogen-inducible cyclooxygenase in brain neurons: regulation by synaptic activity and glucocorticoids, Neuron, 11, 371-386.


43

[41] R. C. Harris, J. A. McKanna, Y. Akai, H. R. Jacobson, R. N. Dubois and M. D. Breyer. (1994) Cyclooxygenase-2 is associated with the macula densa of rat kidney and increases with salt restriction, J. Clin. Invest., 94, 2504-2510.

[42] N. V. Chandrasekharan, H. Dai, K. L. T. Roos, N. K. Evanson, J. Tomsik, T. S. Elton and D. L. Simmons. (2002) COX-3, a cyclooxygenase-1 variant inhibited by acetaminophen and other analgesic/antipyretic drugs: Cloning, structure, and expression, Proc. Natl. Acad. Sci. U.S.A., 99, 13926-13931.

[43] B. Kis, J. A. Snipes and D. W. Busija. (2005) Acetaminophen and the cyclooxygenase-3 puzzle: sorting out facts, fictions, and uncertainties, J. Pharmacol. Exp. Ther., 315, 1-7.

[44] A. L. Blobaum and L. J. Marnett. (2007) Structural and functional basis of cyclooxygenase inhibition, J. Med. Chem., 50, 1425-1441.

[45] E. Ricciotti and G. A. FitzGerald. (2011) Prostaglandins and inflammation, Arterioscler. Thromb. Vasc. Biol., 31, 986-1000.

[46] A. S. Kalgutkar and J. S. Daniels. (2010) Carboxylic acids and their bioisosteres, In Metabolism, pharmacokinetics, and toxicity of functional groups: impact of the building blocks of medicinal chemistry in ADMET (D. A. Smith, Ed.), pp 99-167, Royal Society of Chemistry.

[47] G. J. Roth and C. J. Siok. (1978) Acetylation of the NH2-terminal serine of prostaglandin synthetase by aspirin, J. Biol. Chem., 253, 3782-3784.

[48] L. P. Wennogle, H. Liang, J. C. Quintavalla, B. R. Bowen, J. Wasvary, D. B. Miller, A. Allentoff, W. Boyer, M. Kelly and P. Marshall. (1995) Comparison of recombinant cyclooxygenase-2 to native isoforms: aspirin labeling of the active site, FEBS Lett., 371, 315-320.

[49] V. Evangelista, S. Manarini, A. Di Santo, M. L. Capone, E. Ricciotti, L. Di Francesco, S. Tacconelli, A. Sacchetti, S. D’Angelo, A. Scilimati, M. G. Sciulli and P. Patrignani. (2006) De novo synthesis of cyclooxygenase-1 counteracts the suppression of platelet thromboxane biosynthesis by aspirin, Circ. Res., 98, 593-595.

[50] M. J. Thun, S. J. Henley and C. Patrono. (2002) Nonsteroidal anti-inflammatory drugs as anticancer agents: mechanistic, pharmacologic, and clinical issues, J. Natl. Cancer Inst., 94, 252-266.

[51] T. Narisawa, M. Sato, M. Tani, T. Kudo, T. Takahashi and A. Goto. (1981) Inhibition of development of methylnitrosourea-induced rat colon tumors by indomethacin treatment, Cancer Res., 41, 1954-1957.

[52] W. R. Waddell and R. W. Loughry. (1983) Sulindac for polyposis of the colon, J. Surg. Oncol., 24, 83-87.

[53] W. R. Waddell, G. F. Ganser, E. J. Cerise and R. W. Loughry. (1989) Sulindac for polyposis of the colon, Am. J. Surg., 157, 175-179.

[54] H. J. R. Bussey. (1975) Familial polyposis coli: family studies, histopathology, differential diagnosis, and results of treatment, 1-104, Johns Hopkins University Press.

[55] N. N. Mahmoud, A. J. Dannenberg, J. Mestre, R. T. Bilinski, M. R. Churchill, C. Martucci, H. Newmark and M. M. Bertagnolli. (1998) Aspirin prevents tumors in a murine model of familial adenomatous polyposis, Surgery, 124, 225-231.

[56] S. K. Boolbol, A. J. Dannenberg, A. Chadburn, C. Martucci, X. J. Guo, J. T. Ramonetti, M. Abreu-Goris, H. L. Newmark, M. L. Lipkin, J. J. DeCosse and M. M. Bertagnolli. (1996) Cyclooxygenase-2 overexpression and tumor formation are blocked by sulindac in a murine model of familial adenomatous polyposis, Cancer Res., 56, 2556-2560.

[57] C. H. Chiu, M. F. McEntee and J. Whelan. (1997) Sulindac causes rapid regression of preexisting tumors in Min/+ mice independent of prostaglandin biosynthesis, Cancer Res., 57, 4267-4273.


44

[58] R. F. Jacoby, K. Seibert, C. E. Cole, G. Kelloff and R. A. Lubet. (2000) The cyclooxygenase-2 inhibitor celecoxib is a potent preventive and therapeutic agent in the min mouse model of adenomatous polyposis, Cancer Res., 60, 5040-5044.

[59] J. Burn, D. T. Bishop, P. D. Chapman, F. Elliott, L. Bertario, M. G. Dunlop, D. Eccles, A. Ellis, D. G. Evans, R. Fodde, E. R. Maher, G. Möslein, H. F. A. Vasen, J. Coaker, R. K. S. Phillips, S. Bülow and J. C. Mathers. (2011) A randomized placebo-controlled prevention trial of aspirin and/or resistant starch in young people with familial adenomatous polyposis, Cancer Prev. Res., 4, 655-665.

[60] D. Labayle, D. Fischer, P. Vielh, F. Drouhin, A. Pariente, C. Bories, O. Duhamel, M. Trousset and P. Attali. (1991) Sulindac causes regression of rectal polyps in familial adenomatous polyposis, Gastroenterology, 101, 635-639.

[61] F. M. Giardiello, S. R. Hamilton, A. J. Krush, S. Piantadosi, L. M. Hylind, P. Celano, S. V. Booker, C. R. Robinson and G. J. Offerhaus. (1993) Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis, N. Engl. J. Med., 328, 1313-1316.

[62] K. P. Nugent, K. C. R. Farmer, A. D. Spigelman, C. B. Williams and R. K. S. Phillips. (1993) Randomized controlled trial of the effect of sulindac on duodenal and rectal polyposis and cell proliferation in patients with familial adenomatous polyposis, Br. J. Surg., 80, 1618-1619.

[63] F. M. Giardiello, V. W. Yang, L. M. Hylind, A. J. Krush, G. M. Petersen, J. D. Trimbath, S. Piantadosi, E. Garrett, D. E. Geiman, W. Hubbard, G. J. A. Offerhaus and S. R. Hamilton. (2002) Primary chemoprevention of familial adenomatous polyposis with sulindac, N. Engl. J. Med., 346, 1054-1059.

[64] G. Steinbach, P. M. Lynch, R. K. Phillips, M. H. Wallace, E. Hawk, G. B. Gordon, N. Wakabayashi, B. Saunders, Y. Shen, T. Fujimura, L. K. Su, B. Levin, L. Godio, S. Patterson, M. A. Rodriguez-Bigas, S. L. Jester, K. L. King, M. Schumacher, J. Abbruzzese, R. N. DuBois, W. N. Hittelman, S. Zimmerman, J. W. Sherman and G. Kelloff. (2000) The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis, N. Engl. J. Med., 342, 1946-1952.

[65] S. L. Patterson, K. Colbert Maresso and E. Hawk. (2013) Cancer chemoprevention: successes and failures, Clin. Chem., 59, 94-101.

[66] J. Burn, A.-M. Gerdes, F. Macrae, J.-P. Mecklin, G. Moeslein, S. Olschwang, D. Eccles, D. G. Evans, E. R. Maher, L. Bertario, M.-L. Bisgaard, M. G. Dunlop, J. W. C. Ho, S. V. Hodgson, A. Lindblom, J. Lubinski, P. J. Morrison, V. Murday, R. Ramesar, L. Side, R. J. Scott, H. J. W. Thomas, H. F. Vasen, G. Barker, G. Crawford, F. Elliott, M. Movahedi, K. Pylvanainen, J. T. Wijnen, R. Fodde, H. T. Lynch, J. C. Mathers and D. T. Bishop. (2011) Long-term effect of aspirin on cancer risk in carriers of hereditary colorectal cancer: an analysis from the CAPP2 randomised controlled trial, Lancet, 378, 2081-2087.

[67] R. S. Sandler, S. Halabi, J. A. Baron, S. Budinger, E. Paskett, R. Keresztes, N. Petrelli, J. M. Pipas, D. D. Karp, C. L. Loprinzi, G. Steinbach and R. Schilsky. (2003) A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer, N. Engl. J. Med., 348, 883-890.

[68] M. M. Bertagnolli, C. J. Eagle, A. G. Zauber, M. Redston, S. D. Solomon, K. Kim, J. Tang, R. B. Rosenstein, J. Wittes, D. Corle, T. M. Hess, G. M. Woloj, F. Boisserie, W. F. Anderson, J. L. Viner, D. Bagheri, J. Burn, D. C. Chung, T. Dewar, T. R. Foley, N. Hoffman, F. Macrae, R. E. Pruitt, J. R. Saltzman, B. Salzberg, T. Sylwestrowicz, G. B. Gordon, E. T. Hawk and A. P. C. S. Investigators. (2006) Celecoxib for the prevention of sporadic colorectal adenomas, N. Engl. J. Med., 355, 873-884.

[69] A. M. Algra and P. M. Rothwell. (2012) Effects of regular aspirin on long-term cancer incidence and metastasis: a systematic comparison of evidence from observational studies versus randomised trials, Lancet Oncol., 13, 518-527.


45

[70] P. H. Gann, J. E. Manson, R. J. Glynn, J. E. Buring and C. H. Hennekens. (1993) Low-dose aspirin and incidence of colorectal tumors in a randomized trial, J. Natl. Cancer Inst., 85, 1220-1224.

[71] N. R. Cook, I. M. Lee, J. M. Gaziano, D. Gordon, P. M. Ridker, J. E. Manson, C. H. Hennekens and J. E. Buring. (2005) Low-dose aspirin in the primary prevention of cancer: the Women's Health Study: a randomized controlled trial, JAMA, 294, 47-55.

[72] M. J. Thun, E. J. Jacobs and C. Patrono. (2012) The role of aspirin in cancer prevention, Nat. Rev. Clin. Oncol., 9, 259-267.

[73] P. M. Rothwell, F. G. Fowkes, J. F. Belch, H. Ogawa, C. P. Warlow and T. W. Meade. (2011) Effect of daily aspirin on long-term risk of death due to cancer: analysis of individual patient data from randomised trials, Lancet, 377, 31-41.

[74] P. M. Rothwell, J. F. Price, F. G. Fowkes, A. Zanchetti, M. C. Roncaglioni, G. Tognoni, R. Lee, J. F. Belch, M. Wilson, Z. Mehta and T. W. Meade. (2012) Short-term effects of daily aspirin on cancer incidence, mortality, and non-vascular death: analysis of the time course of risks and benefits in 51 randomised controlled trials, Lancet, 379, 1602-1612.

[75] P. M. Rothwell, M. Wilson, J. F. Price, J. F. Belch, T. W. Meade and Z. Mehta. (2012) Effect of daily aspirin on risk of cancer metastasis: a study of incident cancers during randomised controlled trials, Lancet, 379, 1591-1601.

[76] U.S. Preventive Services Task Force. (2007) Routine aspirin or nonsteroidal anti-inflammatory drugs for the primary prevention of colorectal cancer: Preventive Services Task Force Recommendation Statement, Ann. Intern. Med., 146, 361-364.

[77] U.S. Preventive Services Task Force (2014) Final recommendation statement: Aspirin/NSAIDs for prevention of colorectal cancer: Preventive medication, Available from: http://www.uspreventiveservicestaskforce.org/Page/Document/ RecommendationStatementFinal/aspirin-nsaids-for-prevention-of-colorectal-cancer-preventive-medication, [accessed: Aug. 15, 2016]

[78] C. E. Eberhart, R. J. Coffey, A. Radhika, F. M. Giardiello, S. Ferrenbach and R. N. DuBois. (1994) Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas, Gastroenterology, 107, 1183-1188.

[79] S. L. Kargman, G. P. O'Neill, P. J. Vickers, J. F. Evans, J. A. Mancini and S. Jothy. (1995) Expression of prostaglandin G/H synthase-1 and -2 protein in human colon cancer, Cancer Res., 55, 2556-2559.

[80] H. Sano, Y. Kawahito, R. L. Wilder, A. Hashiramoto, S. Mukai, K. Asai, S. Kimura, H. Kato, M. Kondo and T. Hla. (1995) Expression of cyclooxygenase-1 and -2 in human colorectal cancer, Cancer Res., 55, 3785-3789.

[81] R. A. Soslow, A. J. Dannenberg, D. Rush, B. M. Woerner, K. N. Khan, J. Masferrer and A. T. Koki. (2000) COX-2 is expressed in human pulmonary, colonic, and mammary tumors, Cancer, 89, 2637-2645.

[82] D. K. Petkova, C. Clelland, J. Ronan, L. Pang, J. M. Coulson, S. Lewis and A. J. Knox. (2004) Overexpression of cyclooxygenase-2 in non-small cell lung cancer, Respir. Med., 98, 164-172.

[83] R. Yoshimura, H. Sano, M. Mitsuhashi, M. Kohno, J. Chargui and S. Wada. (2001) Expression of cyclooxygenase-2 in patients with bladder carcinoma, J. Urol., 165, 1468-1472.

[84] C. Costa, R. Soares, J. S. Reis-Filho, D. Leitão, I. Amendoeira and F. C. Schmitt. (2002) Cyclo-oxygenase 2 expression is associated with angiogenesis and lymph node metastasis in human breast cancer, J. Clin. Pathol., 55, 429-434.

[85] B. S. Zweifel, T. W. Davis, R. L. Ornberg and J. L. Masferrer. (2002) Direct evidence for a role of cyclooxygenase 2-derived prostaglandin E2 in human head and neck xenograft tumors, Cancer Res., 62, 6706-6711.

[86] G. Ferrandina, L. Lauriola, G. F. Zannoni, A. Fagotti, F. Fanfani, F. Legge, N. Maggiano, M. Gessi, S. Mancuso, F. O. Ranelletti and G. Scambia. (2002) Increased


46

cyclooxygenase-2 (COX-2) expression is associated with chemotherapy resistance and outcome in ovarian cancer patients, Ann. Oncol., 13, 1205-1211.

[87] C. Bocca, M. Ievolella, R. Autelli, M. Motta, L. Mosso, B. Torchio, F. Bozzo, S. Cannito, C. Paternostro, S. Colombatto, M. Parola and A. Miglietta. (2014) Expression of Cox-2 in human breast cancer cells as a critical determinant of epithelial-to-mesenchymal transition and invasiveness, Expert Opin. Ther. Targets, 18, 121-135.

[88] A. Ristimaki, A. Sivula, J. Lundin, M. Lundin, T. Salminen, C. Haglund, H. Joensuu and J. Isola. (2002) Prognostic significance of elevated cyclooxygenase-2 expression in breast cancer, Cancer Res., 62, 632-635.

[89] E. Fosslien. (2000) Biochemistry of cyclooxygenase (COX)-2 inhibitors and molecular pathology of COX-2 in neoplasia, Crit. Rev. Clin. Lab. Sci., 37, 431-502.

[90] R. N. DuBois, J. Shao, M. Tsujii, H. Sheng and R. Daniel Beauchamp. (1996) G1 delay in cells overexpressing prostaglandin endoperoxide synthase-2, Cancer Res., 56, 733-737.

[91] B. Liu, L. Qu and S. Yan. (2015) Cyclooxygenase-2 promotes tumor growth and suppresses tumor immunity, Cancer Cell Int., 15, 106-111.

[92] T. O. Daniel, H. Liu, J. D. Morrow, B. C. Crews and L. J. Marnett. (1999) Thromboxane A2 is a mediator of cyclooxygenase-2-dependent endothelial migration and angiogenesis, Cancer Res., 59, 4574-4577.

[93] F. Cianchi, C. Cortesini, O. Fantappie, L. Messerini, I. Sardi, N. Lasagna, F. Perna, V. Fabbroni, A. Di Felice, G. Perigli, R. Mazzanti and E. Masini. (2004) Cyclooxygenase-2 activation mediates the proangiogenic effect of nitric oxide in colorectal cancer, Clin. Cancer Res., 10, 2694-2704.

[94] J.-H. Lee, M. S. Piao, J.-Y. Choi, S. J. Yun, J.-B. Lee and S.-C. Lee. (2013) Up-regulation of cyclooxygenase 2 and matrix metalloproteinases-2 and -9 in cutaneous squamous cell carcinoma: active role of inflammation and tissue remodeling in carcinogenesis, Ann. Dermatol., 25, 145-151.

[95] M. Stolina, S. Sharma, Y. Lin, M. Dohadwala, B. Gardner, J. Luo, L. Zhu, M. Kronenberg, P. W. Miller, J. Portanova, J. C. Lee and S. M. Dubinett. (2000) Specific inhibition of cyclooxygenase 2 restores antitumor reactivity by altering the balance of IL-10 and IL-12 synthesis, J. Immunol. Methods, 164, 361-370.

[96] M. Tsujii, S. Kawano, S. Tsuji, H. Sawaoka, M. Hori and R. N. DuBois. (1998) Cyclooxygenase regulates angiogenesis induced by colon cancer cells, Cell, 93, 705-716.

[97] M. Tsujii, S. Kawano and R. N. DuBois. (1997) Cyclooxygenase-2 expression in human colon cancer cells increases metastatic potential, Proc. Natl. Acad. Sci. U.S.A., 94, 3336-3340.

[98] T. A. Chan, P. J. Morin, B. Vogelstein and K. W. Kinzler. (1998) Mechanisms underlying nonsteroidal antiinflammatory drug-mediated apoptosis, Proc. Natl. Acad. Sci. U.S.A., 95, 681-686.

[99] K. M. Leahy, R. L. Ornberg, Y. Wang, B. S. Zweifel, A. T. Koki and J. L. Masferrer. (2002) Cyclooxygenase-2 inhibition by celecoxib reduces proliferation and induces apoptosis in angiogenic endothelial cells in vivo, Cancer Res., 62, 625-631.

[100] R. Hanif, A. Pittas, Y. Feng, M. I. Koutsos, L. Qiao, L. Staiano-Coico, S. I. Shiff and B. Rigas. (1996) Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway, Biochem. Pharmacol., 52, 237-245.

[101] D. J. Elder, D. E. Halton, A. Hague and C. Paraskeva. (1997) Induction of apoptotic cell death in human colorectal carcinoma cell lines by a cyclooxygenase-2 (COX-2)-selective nonsteroidal anti-inflammatory drug: independence from COX-2 protein expression, Clin. Cancer Res., 3, 1679-1683.

[102] S. Grösch, I. Tegeder, E. Niederberger, L. Bräutigam and G. Geisslinger. (2001) COX-2 independent induction of cell cycle arrest and apoptosis in colon cancer cells by the selective COX-2 inhibitor celecoxib, FASEB J, 15, 2742-2744.


47

[103] G. A. Piazza, A. L. K. Rahm, M. Krutzsch, G. Sperl, N. S. Paranka, P. H. Gross, K. Brendel, R. W. Burt, D. S. Alberts, R. Pamukcu and D. J. Ahnen. (1995) Antineoplastic drugs sulindac sulfide and sulfone inhibit cell growth by inducing apoptosis, Cancer Res., 55, 3110-3116.

[104] H. Kusuhara, H. Matsuyuki, M. Matsuura, T. Imayoshi, T. Okumoto and H. Matsui. (1998) Induction of apoptotic DNA fragmentation by nonsteroidal anti-inflammatory drugs in cultured rat gastric mucosal cells, Eur. J. Pharmacol., 360, 273-280.

[105] G. Farrugia and R. Balzan. (2013) The proapoptotic effect of traditional and novel nonsteroidal anti-inflammatory drugs in mammalian and yeast cells, Oxid. Med. Cell Longev., 2013, doi:10.1155/2013/504230.

[106] J. L. Liggett, X. Zhang, T. E. Eling and S. J. Baek. (2014) Anti-tumor activity of non-steroidal anti-inflammatory drugs: cyclooxygenase-independent targets, Cancer Lett., 346, 217-224.

[107] U. S. Akarca. (2005) Gastrointestinal effects of selective and non-selective non-steroidal anti-inflammatory drugs, Curr. Pharm. Des., 11, 1779-1793.

[108] C. Patrono , L. A. García Rodríguez , R. Landolfi and C. Baigent (2005) Low-dose aspirin for the prevention of atherothrombosis, N. Engl. J. Med., 353, 2373-2383.

[109] X.-M. Xu, L. Sansores-Garcia, X.-M. Chen, N. Matijevic-Aleksic, M. Du and K. K. Wu. (1999) Suppression of inducible cyclooxygenase 2 gene transcription by aspirin and sodium salicylate, Proc. Natl. Acad. Sci. U.S.A., 96, 5292-5297.

[110] H. Raza and A. John. (2012) Implications of altered glutathione metabolism in aspirin-induced oxidative stress and mitochondrial dysfunction in HepG2 cells, PLoS One, 7, e36325.

[111] M. Marchetti, L. Resnick, E. Gamliel, S. Kesaraju, H. Weissbach and D. Binninger. (2009) Sulindac enhances the killing of cancer cells exposed to oxidative stress, PLoS One, 4, e5804.

[112] H. Kusuhara, H. Komatsu, H. Sumichika and K. Sugahara. (1999) Reactive oxygen species are involved in the apoptosis induced by nonsteroidal anti-inflammatory drugs in cultured gastric cells, Eur. J. Pharmacol., 383, 331-337.

[113] J. W. Chang, S. U. Kang, J. W. Choi, Y. S. Shin, S. J. Baek, S. H. Lee and C. H. Kim. (2014) Tolfenamic acid induces apoptosis and growth inhibition in anaplastic thyroid cancer: Involvement of nonsteroidal anti-inflammatory drug-activated gene-1 expression and intracellular reactive oxygen species generation, Free Radic. Biol. Med., 67, 115-130.

[114] S. Dihlmann, A. Siermann and M. von Knebel Doeberitz. (2001) The nonsteroidal anti-inflammatory drugs aspirin and indomethacin attenuate beta-catenin/TCF-4 signaling, Oncogene, 20, 645-653.

[115] M. Cho, J. Gwak, S. Park, J. Won, D.-E. Kim, S. S. Yea, I.-J. Cha, T. K. Kim, J.-G. Shin and S. Oh. (2005) Diclofenac attenuates Wnt/β-catenin signaling in colon cancer cells by activation of NF-κB, FEBS Lett., 579, 4213-4218.

[116] M. F. McEntee, C. H. Chiu and J. Whelan. (1999) Relationship of beta-catenin and Bcl-2 expression to sulindac-induced regression of intestinal tumors in Min mice, Carcinogenesis, 20, 635-640.

[117] L. Qiao, S. J. Shiff and B. Rigas. (1998) Sulindac sulfide alters the expression of cyclin proteins in HT-29 colon adenocarcinoma cells, Int. J. Cancer, 76, 99-104.

[118] X. Zhang, S. H. Lee, K. W. Min, M. F. McEntee, J. B. Jeong, Q. Li and S. J. Baek. (2013) The involvement of endoplasmic reticulum stress in the suppression of colorectal tumorigenesis by tolfenamic acid, Cancer Prev. Res., 6, 1337-1347.

[119] N. Babbar, Eugene W. Gerner and Robert A. Casero. (2006) Induction of spermidine/spermine N(1)-acetyltransferase (SSAT) by aspirin in Caco-2 colon cancer cells, Biochem. J., 394, 317-324.


48

[120] N. Babbar, N. A. Ignatenko, R. A. Casero, Jr. and E. W. Gerner. (2003) Cyclooxygenase-independent induction of apoptosis by sulindac sulfone is mediated by polyamines in colon cancer, J. Biol. Chem., 278, 47762-47775.

[121] L. Turchanowa, N. Dauletbaev, V. Milovic and J. Stein. (2001) Nonsteroidal anti-inflammatory drugs stimulate spermidine/spermine acetyltransferase and deplete polyamine content in colon cancer cells, Eur. J. Clin. Invest., 31, 887-893.

[122] P. Dikshit, M. Chatterjee, A. Goswami, A. Mishra and N. R. Jana. (2006) Aspirin induces apoptosis through the inhibition of proteasome function, J. Biol. Chem., 281, 29228-29235.

[123] Y. Oida, B. Gopalan, R. Miyahara, S. Inoue, C. D. Branch, A. M. Mhashilkar, E. Lin, B. N. Bekele, J. A. Roth, S. Chada and R. Ramesh. (2005) Sulindac enhances adenoviral vector expressing mda-7/IL-24-mediated apoptosis in human lung cancer, Mol. Cancer Ther., 4, 291-304.

[124] M. Lu, A. Strohecker, F. Chen, T. Kwan, J. Bosman, V. C. Jordan and V. L. Cryns. (2008) Aspirin sensitizes cancer cells to TRAIL-induced apoptosis by reducing survivin levels, Clin. Cancer Res., 14, 3168-3176.

[125] U. T. Sankpal, M. Abdelrahim, S. F. Connelly, C. M. Lee, R. Madero-Visbal, J. Colon, J. Smith, S. Safe, P. Maliakal and R. Basha. (2012) Small molecule tolfenamic acid inhibits PC-3 cell proliferation and invasion in vitro, and tumor growth in orthotopic mouse model for prostate cancer, Prostate, 72, 1648-1658.

[126] E. Kopp and S. Ghosh. (1994) Inhibition of NF-kappa B by sodium salicylate and aspirin, Science, 265, 956-959.

[127] Y. Yamamoto, M. J. Yin, K. M. Lin and R. B. Gaynor. (1999) Sulindac inhibits activation of the NF-kappaB pathway, J. Biol. Chem., 274, 27307-27314.

[128] E. J. Greenspan, J. P. Madigan, L. A. Boardman and D. W. Rosenberg. (2011) Ibuprofen inhibits activation of nuclear β-catenin in human colon adenomas and induces the phosphorylation of GSK-3β, Cancer Prev. Res., 4, 161-171.

[129] S. Setia and S. N. Sanyal. (2012) Downregulation of NF-kappaB and PCNA in the regulatory pathways of apoptosis by cyclooxygenase-2 inhibitors in experimental lung cancer, Mol. Cell Biochem., 369, 75-86.

[130] J. B. Jeong, X. Yang, R. Clark, J. Choi, S. J. Baek and S. H. Lee. (2013) A mechanistic study of the proapoptotic effect of tolfenamic acid: involvement of NF-kappaB activation, Carcinogenesis, 34, 2350-2360.

[131] S. R. Im and Y. J. Jang. (2012) Aspirin enhances TRAIL-induced apoptosis via regulation of ERK1/2 activation in human cervical cancer cells, Biochem. Biophys. Res. Commun., 424, 65-70.

[132] Y.-C. Ou, C.-R. Yang, C.-L. Cheng, S.-L. Raung, Y.-Y. Hung and C.-J. Chen. (2007) Indomethacin induces apoptosis in 786-O renal cell carcinoma cells by activating mitogen-activated protein kinases and AKT, Eur. J. Pharmacol., 563, 49-60.

[133] B. Bellosillo, M. Pique, M. Barragan, E. Castano, N. Villamor, D. Colomer, E. Montserrat, G. Pons and J. Gil. (1998) Aspirin and salicylate induce apoptosis and activation of caspases in B-cell chronic lymphocytic leukemia cells, Blood, 92, 1406-1414.

[134] X. M. Zhou, B. C. Y. Wong, X. M. Fan, H. B. Zhang, M. C. M. Lin, H. F. Kung, D. M. Fan and S. K. Lam. (2001) Non-steroidal anti-inflammatory drugs induce apoptosis in gastric cancer cells through up-regulation of bax and bak, Carcinogenesis, 22, 1393-1397.

[135] H. Weiss, A. Amberger, M. Widschwendter, R. Margreiter, D. Ofner and P. Dietl. (2001) Inhibition of store-operated calcium entry contributes to the anti-proliferative effect of non-steroidal anti-inflammatory drugs in human colon cancer cells, Int. J. Cancer, 92, 877-882.

[136] D. H. Woo, I. S. Han and G. Jung. (2004) Mefenamic acid-induced apoptosis in human liver cancer cell-lines through caspase-3 pathway, Life Sci., 75, 2439-2449.


49

[137] M. Suwalsky, M. Manrique-Moreno, J. Howe, K. Brandenburg and F. Villena. (2011) Molecular interactions of mefenamic acid with lipid bilayers and red blood cells, J. Braz. Chem. Soc., 22, 2243-2249.

[138] A. Sade, S. Tunçay, İ. Çimen, F. Severcan and S. Banerjee. (2012) Celecoxib reduces fluidity and decreases metastatic potential of colon cancer cell lines irrespective of COX-2 expression, Biosci. Rep., 32, 35-44.

[139] E. S. Choi, J. H. Shim, J. Y. Jung, H. J. Kim, K. H. Choi, J. A. Shin, J. S. Nam, N. P. Cho and S. D. Cho. (2011) Apoptotic effect of tolfenamic acid in androgen receptor-independent prostate cancer cell and xenograft tumor through specificity protein 1, Cancer Sci., 102, 742-748.

[140] T. J. Jang, N. I. Kim and C. H. Lee. (2006) Proapoptotic activity of NAG-1 is cell type specific and not related to COX-2 expression, Apoptosis, 11, 1131-1138.

[141] E. M. Van Lieshout, D. M. Tiemessen, H. M. Roelofs and W. H. Peters. (1998) Nonsteroidal anti-inflammatory drugs enhance glutathione S-transferase theta levels in rat colon, Biochim. Biophys. Acta, 1381, 305-311.

[142] M. M. Wolfe, D. R. Lichtenstein and G. Singh. (1999) Gastrointestinal toxicity of nonsteroidal antiinflammatory drugs, N. Engl. J. Med., 340, 1888-1899.

[143] G.-H. Kim. (2008) Renal effects of prostaglandins and cyclooxygenase-2 inhibitors, Electrolytes Blood Press., 6, 35-41.

[144] T. D. Warner, S. Nylander and C. Whatling. (2011) Anti-platelet therapy: cyclo-oxygenase inhibition and the use of aspirin with particular regard to dual anti-platelet therapy, Br. J. Clin. Pharmacol., 72, 619-633.

[145] G. A. FitzGerald (2004) Coxibs and Cardiovascular Disease, N. Engl. J. Med., 351, 1709-1711.

[146] R. S. Bresalier, R. S. Sandler, H. Quan, J. A. Bolognese, B. Oxenius, K. Horgan, C. Lines, R. Riddell, D. Morton, A. Lanas, M. A. Konstam and J. A. Baron. (2005) Cardiovascular events associated with rofecoxib in a colorectal adenoma chemoprevention trial, N. Engl. J. Med., 352, 1092-1102.

[147] S. D. Solomon, J. J. McMurray, M. A. Pfeffer, J. Wittes, R. Fowler, P. Finn, W. F. Anderson, A. Zauber, E. Hawk and M. Bertagnolli. (2005) Cardiovascular risk associated with celecoxib in a clinical trial for colorectal adenoma prevention, N. Engl. J. Med., 352, 1071-1080.

[148] B. Sibbald. (2004) Rofecoxib (Vioxx) voluntarily withdrawn from market, CMAJ, 171, 1027-1028.

[149] U.S. Food and Drug Administration Information for Healthcare Professionals: Celecoxib (marketed as Celebrex). FDA Alert: 4/7/2005, Available from: http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm124655.htm, [accessed: Sept. 5, 2016]

[150] P. McGettigan and D. Henry. (2013) Use of non-steroidal anti-inflammatory drugs that elevate cardiovascular risk: an examination of sales and essential medicines lists in low-, middle-, and high-income countries, PLoS Med., 10, e1001388.

[151] P. Geusens and W. Lems. (2008) Efficacy and tolerability of lumiracoxib, a highly selective cyclo-oxygenase-2 (COX2) inhibitor, in the management of pain and osteoarthritis, Ther. Clin. Risk Manag., 4, 337-344.

[152] G. P. Aithal and C. P. Day. (2007) Nonsteroidal anti-inflammatory drug-induced hepatotoxicity, Clin. Liver Dis., 11, 563-575.

[153] F. Bessone. (2010) Non-steroidal anti-inflammatory drugs: what is the actual risk of liver damage?, World J. Gastroenterol., 16, 5651-5661.

[154] J. L. Williams, S. Borgo, I. Hasan, E. Castillo, F. Traganos and B. Rigas. (2001) Nitric oxide-releasing nonsteroidal anti-inflammatory drugs (NSAIDs) alter the kinetics of human colon cancer cell lines more effectively than traditional NSAIDs: implications for colon cancer chemoprevention, Cancer Res., 61, 3285-3289.

[155] L. Huang, G. G. Mackenzie, Y. Sun, N. Ouyang, G. Xie, K. Vrankova, D. Komninou and B. Rigas. (2011) Chemotherapeutic properties of phospho-nonsteroidal anti-


50

inflammatory drugs, a new class of anticancer compounds, Cancer Res., 71, 7617-7627.

[156] M. Chattopadhyay, R. Kodela, N. Nath, Y. M. Dastagirzada, C. A. Velázquez-Martínez, D. Boring and K. Kashfi. (2012) Hydrogen sulfide-releasing NSAIDs inhibit the growth of human cancer cells: a general property and evidence of a tissue type-independent effect, Biochem. Pharmacol., 83, 715-722.

[157] S. A. Laufer, J. Augustin, G. Dannhardt and W. Kiefer. (1994) (6,7-diaryldihydropyrrolizin-5-yl)acetic acids, a novel class of potent dual inhibitors of both cyclooxygenase and 5-lipoxygenase, J. Med. Chem., 37, 1894-1897.

[158] S. Barbey, L. Goossens, T. Taverne, J. Cornet, V. Choesmel, C. Rouaud, G. Gimeno, S. Yannic-Arnoult, C. Michaux, C. Charlier, R. Houssin and J. P. Henichart. (2002) Synthesis and activity of a new methoxytetrahydropyran derivative as dual cyclooxygenase-2/5-lipoxygenase inhibitor, Bioorg. Med. Chem. Lett., 12, 779-782.

[159] Q.-H. Chen, P. N. Praveen Rao and E. E. Knaus. (2006) Synthesis and biological evaluation of a novel class of rofecoxib analogues as dual inhibitors of cyclooxygenases (COXs) and lipoxygenases (LOXs), Bioorg. Med. Chem., 14, 7898-7909.

[160] M. V. Reddy, V. K. Billa, V. R. Pallela, M. R. Mallireddigari, R. Boominathan, J. L. Gabriel and E. P. Reddy. (2008) Design, synthesis, and biological evaluation of 1-(4-sulfamylphenyl)-3-trifluoromethyl-5-indolyl pyrazolines as cyclooxygenase-2 (COX-2) and lipoxygenase (LOX) inhibitors, Bioorg. Med. Chem., 16, 3907-3916.

[161] A. Andrade, S. F. Namora, R. G. Woisky, G. Wiezel, R. Najjar, J. A. A. Sertié and D. de Oliveira Silva. (2000) Synthesis and characterization of a diruthenium–ibuprofenato complex: Comparing its anti-inflammatory activity with that of a copper(II)–ibuprofenato complex, J. Inorg. Biochem., 81, 23-27.

[162] A. M. Bonin, J. A. Yanez, C. Fukuda, X. W. Teng, C. T. Dillon, T. W. Hambley, P. A. Lay and N. M. Davies. (2010) Inhibition of experimental colorectal cancer and reduction in renal and gastrointestinal toxicities by copper-indomethacin in rats, Cancer Chemother. Pharmacol., 66, 755-764.

[163] C. T. Dillon, T. W. Hambley, B. J. Kennedy, P. A. Lay, Q. Zhou, N. M. Davies, J. R. Biffin and H. L. Regtop. (2003) Gastrointestinal toxicity, antiinflammatory activity, and superoxide dismutase activity of copper and zinc complexes of the antiinflammatory drug indomethacin, Chem. Res. Toxicol., 16, 28-37.

[164] N. K. Jain, A. Singh and S. K. Kulkarni. (1999) Analgesic, anti-inflammatory and ulcerogenic activity of a zinc-naproxen complex in mice and rats, Pharm. Pharmacol. Commun., 5, 599-602.

[165] M. A. Kale, R. Shelke and R. B. Nawale. (2014) Zinc-aceclofenac complex: Synthesis, hydrolysis study and anti-inflammatory studies, Curr. Med. Chem. Anti Inflamm. Anti Allergy Agents, 13, 36-44.

[166] M. Konstandinidou, A. Kourounakis, M. Yiangou, L. Hadjipetrou, D. Kovala-Demertzi, S. Hadjikakou and M. Demertzis. (1998) Anti-inflammatory properties of diclofenac transition metalloelement complexes, J. Inorg. Biochem., 70, 63-69.

[167] D. Kovala-Demertzi, D. Hadjipavlou-Litina, A. Primikiri, M. Staninska, C. Kotoglou and M. A. Demertzis. (2009) Anti-inflammatory, antiproliferative, and radical-scavenging activities of tolfenamic acid and its metal complexes, Chem. Biodivers., 6, 948-960.

[168] D. Kovala-Demertzi, D. Hadjipavlou-Litina, M. Staninska, A. Primikiri, C. Kotoglou and M. A. Demertzis. (2009) Anti-oxidant, in vitro, in vivo anti-inflammatory activity and antiproliferative activity of mefenamic acid and its metal complexes with manganese(II), cobalt(II), nickel(II), copper(II) and zinc(II), J. Enzym. Inhib. Med. Ch., 24, 742-752.

[169] A. K. Singla and H. Wadhwa. (1995) Zinc-indomethacin complex: Synthesis, physicochemical and biological evaluation in the rat, Int. J. Pharm., 120, 145-155.


51

[170] C. Dendrinou-Samara, G. Tsotsou, L. V. Ekateriniadou, A. H. Kortsaris, C. P. Raptopoulou, A. Terzis, D. A. Kyriakidis and D. P. Kessissoglou. (1998) Anti-inflammatory drugs interacting with Zn(II), Cd(II) and Pt(II) metal ions, J. Inorg. Biochem., 71, 171-179.

[171] A. A. H. S. Al-Janabi. (2010) Elevation of antidermatophytic action of mefenamic acid by cobalt ions, Indian J. Pharmacol., 42, 351-353.

[172] H. Chiniforoshan, L. Tabrizi, M. Hadizade, M. R. Sabzalian, A. N. Chermahini and M. Rezapour. (2014) Anti-inflammatory drugs interacting with Zn (II) metal ion based on thiocyanate and azide ligands: synthesis, spectroscopic studies, DFT calculations and antibacterial assays, Spectrochim. Acta Mol. Biomol. Spectrosc., 128, 183-190.

[173] A. Jalal, S. Shahzadi, K. Shahid, S. Ali, A. Badshah, M. Mazhar and K. M. Khan. (2004) Preparation, spectroscopic studies and biological activity of mono-organotin (IV) derivatives of non-steroidal anti-inflammatory drugs, Turk. J. Chem., 28, 629-644.

[174] S. Shahzadi, K. Shahid, S. Ali, M. Mazhar, A. Badshah, E. Ahmed and A. Malik. (2005) Non-steroidal anti-inflammatory drugs (NSAIDs) as donor ligands in organotin(IV) derivatives: Synthesis, spectroscopic characterization and biological applications, Turk. J. Chem., 29, 273-287.

[175] L. Tabrizi, H. Chiniforoshan, P. McArdle, M. Ebrahimi and T. Khayamian. (2015) A novel bioactive Cd(II) polymeric complex with mefenamic acid: Synthesis, crystal structure and biological evaluations, Inorg. Chim. Acta, 432, 176-184.

[176] G. Tamasi, C. Bernini, G. Corbini, N. F. Owens, L. Messori, F. Scaletti, L. Massai, P. L. Giudice and R. Cini. (2014) Synthesis, spectroscopic and DFT structural characterization of two novel ruthenium(III) oxicam complexes. In vivo evaluation of anti-inflammatory and gastric damaging activities, J. Inorg. Biochem., 134, 25-35.

[177] V. Dokorou, A. Primikiri and D. Kovala-Demertzi. (2011) The triphenyltin(VI) complexes of NSAIDs and derivatives. Synthesis, crystal structure and antiproliferative activity. Potent anticancer agents, J. Inorg. Biochem., 105, 195-201.

[178] D. Kovala-Demertzi, M. Staninska, I. Garcia-Santos, A. Castineiras and M. A. Demertzis. (2011) Synthesis, crystal structures and spectroscopy of meclofenamic acid and its metal complexes with manganese(II), copper(II), zinc(II) and cadmium(II). Antiproliferative and superoxide dismutase activity, J. Inorg. Biochem., 105, 1187-1195.

[179] L. Tabrizi, H. Chiniforoshan and P. McArdle. (2015) Synthesis, crystal structure and spectroscopy of bioactive Cd(II) polymeric complex of the non-steroidal anti-inflammatory drug diclofenac sodium: Antiproliferative and biological activity, Spectrochim. Acta Mol. Biomol. Spectrosc., 136, 429-436.

[180] G. Tamasi, M. Casolaro, A. Magnani, A. Sega, L. Chiasserini, L. Messori, C. Gabbiani, S. M. Valiahdi, M. A. Jakupec, B. K. Keppler, M. B. Hursthouse and R. Cini. (2010) New platinum–oxicam complexes as anti-cancer drugs. Synthesis, characterization, release studies from smart hydrogels, evaluation of reactivity with selected proteins and cytotoxic activity in vitro, J. Inorg. Biochem., 104, 799-814.

[181] E. Păunescu, S. McArthur, M. Soudani, R. Scopelliti and P. J. Dyson. (2016) Nonsteroidal anti-inflammatory—organometallic anticancer compounds, Inorg. Chem., 55, 1788-1808.

[182] L. Dyakova, D.-C. Culita, T. Zhivkova, M. Georgieva, R. Kalfin, G. Miloshev, M. Alexandrov, G. Marinescu, L. Patron and R. Alexandrova. (2015) 3d metal complexes with meloxicam as therapeutic agents in the fight against human glioblastoma multiforme and cervical carcinoma, Biotechnol. Biotechnol. Equip., 29, 1190-1200.

[183] P. Christofis, M. Katsarou, A. Papakyriakou, Y. Sanakis, N. Katsaros and G. Psomas. (2005) Mononuclear metal complexes with piroxicam: synthesis, structure and biological activity, J. Inorg. Biochem., 99, 2197-2210.


52

[184] F. Dimiza, F. Perdih, V. Tangoulis, I. Turel, D. P. Kessissoglou and G. Psomas. (2011) Interaction of copper(II) with the non-steroidal anti-inflammatory drugs naproxen and diclofenac: Synthesis, structure, DNA- and albumin-binding, J. Inorg. Biochem., 105, 476-489.

[185] S. Fountoulaki, F. Perdih, I. Turel, D. P. Kessissoglou and G. Psomas. (2011) Non-steroidal anti-inflammatory drug diflunisal interacting with Cu(II). Structure and biological features, J. Inorg. Biochem., 105, 1645-1655.

[186] C. Tolia, A. N. Papadopoulos, C. P. Raptopoulou, V. Psycharis, C. Garino, L. Salassa and G. Psomas. (2013) Copper(II) interacting with the non-steroidal antiinflammatory drug flufenamic acid: Structure, antioxidant activity and binding to DNA and albumins, J. Inorg. Biochem., 123, 53-65.

[187] A. Tarushi, C. P. Raptopoulou, V. Psycharis, D. P. Kessissoglou, A. N. Papadopoulos and G. Psomas. (2014) Structure and biological perspectives of Cu(II)–indomethacin complexes, J. Inorg. Biochem., 140, 185-198.

[188] F. Dimiza, A. N. Papadopoulos, V. Tangoulis, V. Psycharis, C. P. Raptopoulou, D. P. Kessissoglou and G. Psomas. (2010) Biological evaluation of non-steroidal anti-inflammatory drugs-cobalt(II) complexes, Dalton Trans., 39, 4517-4528.

[189] F. Dimiza, A. N. Papadopoulos, V. Tangoulis, V. Psycharis, C. P. Raptopoulou, D. P. Kessissoglou and G. Psomas. (2012) Biological evaluation of cobalt(II) complexes with non-steroidal anti-inflammatory drug naproxen, J. Inorg. Biochem., 107, 54-64.

[190] S. Tsiliou, L.-A. Kefala, F. Perdih, I. Turel, D. P. Kessissoglou and G. Psomas. (2012) Cobalt(II) complexes with non-steroidal anti-inflammatory drug tolfenamic acid: Structure and biological evaluation, Eur. J. Med. Chem., 48, 132-142.

[191] A. Tarushi, S. Perontsis, A. G. Hatzidimitriou, A. N. Papadopoulos, D. P. Kessissoglou and G. Psomas. (2015) Copper(II) complexes with the non-steroidal anti-inflammatory drug tolfenamic acid: Structure and biological features, J. Inorg. Biochem., 149, 68-79.

[192] M. Kyropoulou, C. P. Raptopoulou, V. Psycharis and G. Psomas. (2013) Ni(II) complexes with non-steroidal anti-inflammatory drug diclofenac: Structure and interaction with DNA and albumins, Polyhedron, 61, 126-136.

[193] M. Zampakou, V. Tangoulis, C. P. Raptopoulou, V. Psycharis, A. N. Papadopoulos and G. Psomas. (2015) Structurally diverse manganese(II)–diclofenac complexes showing enhanced antioxidant activity and affinity to serum albumins in comparison to sodium diclofenac, Eur. J. Inorg. Chem., 2285-2294.

[194] X. Totta, A. A. Papadopoulou, A. G. Hatzidimitriou, A. Papadopoulos and G. Psomas. (2015) Synthesis, structure and biological activity of nickel(II) complexes with mefenamato and nitrogen-donor ligands, J. Inorg. Biochem., 145, 79-93.

[195] P. Tsiliki, F. Perdih, I. Turel and G. Psomas. (2013) Structure, DNA- and albumin-binding of the manganese(II) complex with the non-steroidal antiinflammatory drug niflumic acid, Polyhedron, 53, 215-222.

[196] S. Chakraborty, E. Sehanobish and M. Sarkar. (2012) Binding of Cu(II) complexes of oxicam NSAIDs to alternating AT and homopolymeric AT sequences: differential response to variation in backbone structure, J. Biol. Inorg. Chem., 17, 475-487.

[197] I. Pais and J. Benton Jones Jr. (1997) The handbook of trace elements, CRC Press, Boca Raton.

[198] G. G. Briand and N. Burford. (1999) Bismuth compounds and preparations with biological or medicinal relevance, Chem. Rev., 99, 2601-2658.

[199] A. C. Ford, P. Malfertheiner, M. Giguere, J. Santana, M. Khan and P. Moayyedi. (2008) Adverse events with bismuth salts for Helicobacter pylori eradication: Systematic review and meta-analysis, World J. Gastroenterol., 14, 7361-7370.

[200] D. M. Keogan and D. M. Griffith. (2014) Current and potential applications of bismuth-based drugs, Molecules, 19, 15258-15297.


53

[201] N. Yang and H. Sun. (2011) Biological chemistry of antimony and bismuth, In Biological chemistry of arsenic, antimony and bismuth (H. Sun, Ed.), pp 53-81, John Wiley & Sons, Ltd.

[202] P. J. Sadler, C. Munice and M. A. Shipman. (2007) Chapter Vll metals in medicine, In Biological Inorganic Chemistry Structure & Reactivity (I. Bertini, H. B. Gray, E. I. Stiefel and J. S. Valentine, Eds.), pp 95-135, University Science Books, California.

[203] L. Odier. (1786) J. Med. Chir. Pharm., 68, 49-56. [204] P. J. Sadler, H. Li and H. Sun. (1999) Coordination chemistry of metals in medicine:

target sites for bismuth, Coord. Chem. Rev., 185-186, 689-709. [205] P. C. Andrews, G. B. Deacon, C. M. Forsyth, P. C. Junk, I. Kumar and M. Maguire.

(2006) Towards a structural understanding of the anti-ulcer and anti-gastritis drug bismuth subsalicylate, Angew. Chem. Int. Ed., 45, 5638-5642.

[206] W. Li, L. Jin, N. Zhu, X. Hou, F. Deng and H. Sun. (2003) Structure of colloidal bismuth subcitrate (CBS) in dilute HCl: unique assembly of bismuth citrate dinuclear units ([Bi(cit)2Bi]2-), J. Am. Chem. Soc., 125, 12408-12409.

[207] P. A. Bezdetko, N. Sergienko, Y. Dyomin, A. Korol, N. Nikitin, M. Merzbacher, D. Groß and R. Kohnen. (2012) Successful treatment of blepharitis with bibrocathol (Posiformin(®) 2 %), Graefes Arch. Clin. Exp. Ophthalmol., 250, 1869-1875.

[208] W. Modell. (2013) Drugs in current use 1958, 22-23, Springer, Berlin Heidelberg. [209] B. W. Sippy. (1915) Gastric and duodenal ulcer: Medical cure by an efficient

removal of gastric juice corrosion, JAMA, 64, 1625–1630. [210] O. M. Gruhzit, E. Lyons and R. Perkins. (1927) Bismuth thioglycollate in

experimental and clinical treatment of syphilis, Arch. Derm. Syphilol., 15, 550-567. [211] W. F. Schwartz. (1939) The effect of Thio-Bismol on therapeutic malaria,

J. Pharmacol. Exp. Ther., 65, 175-184. [212] J. E. Fenton, A. W. Blayney and T. P. O'Dwyer. (1995) Bismuth subgallate – its role

in tonsillectomy, J. Laryngol. Otol., 109, 203-205. [213] The Parthenon Co. Inc. (2016) Devrom®, Available from: http://devrom.com,

[accessed: May 25, 2016] [214] J. R. Lambert and P. Midolo. (1997) The actions of bismuth in the treatment of

Helicobacter pylori infection, Aliment. Pharmacol. Ther., 11, 27-33. [215] I. Wilhelmsen, R. Weberg, K. Berstad, T. Hausken, O. Hundal and A. Berstad. (1994)

Helicobacter pylori eradication with bismuth subnitrate, oxytetracycline and metronidazole in patients with peptic ulcer disease, Hepatogastroenterology, 41, 43-47.

[216] A. F. Carvalho, L. A. Fiorelli, V. N. C. Jorge, C. M. F. Da Silva, G. De Nucci, J. G. P. Ferraz and J. Pedrazzoli. (1998) Addition of bismuth subnitrate to omeprazole plus amoxycillin improves eradication of Helicobacter pylori, Aliment. Pharmacol. Ther., 12, 557-561.

[217] S. Tefera, A. Berstad, C. Bang, G. Nesæter, J. Hatlebakk, S. Olafsson, L. Nesje, T. Hausken, K. Berstad and Ø. Hundal. (1996) Bismuth-based combination therapy for Helicobacter pylori-associated peptic ulcer disease (metronidazole for eradication, ranitidine for pain), Am. J. Gastroenterol., 91, 935-941.

[218] E. G. Stieglitz. (1930) Bismuth subnitrate in the treatment of arterial hypertension, JAMA, 95, 842-846.

[219] R. Morison. (1916) The treatment of infected suppurating war wounds, Lancet, 2, 268-272.

[220] D. Figueroa-Quintanilla, E. Salazar-Lindo, R. B. Sack, R. Leon-Barua, S. Sarabia-Arce, M. Campos-Sanchez and E. Eyzaguirre-Maccan. (1993) A controlled trial of bismuth subsalicylate in infants with acute watery diarrheal disease, N. Engl. J. Med., 328, 1653-1658.

[221] R. Steffen, H. L. DuPont, R. Heusser, A. Helminger, F. Witassek, M. D. Manhart and M. Schär. (1986) Prevention of traveler's diarrhea by the tablet form of bismuth subsalicylate, Antimicrob. Agents Chemother., 29, 625-627.


54

[222] D. Y. Graham, G. M. Lew, D. G. Evans, D. J. Evans, Jr. and P. D. Klein. (1991) Effect of triple therapy (antibiotics plus bismuth) on duodenal ulcer healing. A randomized controlled trial, Ann. Intern. Med., 115, 266-269.

[223] R. J. Hopkins. (1997) Current FDA-approved treatments for Helicobacter pylori and the FDA approval process, Gastroenterology, 113, 126-130.

[224] J. Y. Kang, H. H. Tay, A. Wee, R. Guan, M. V. Math and I. Yap. (1990) Effect of colloidal bismuth subcitrate on symptoms and gastric histology in non-ulcer dyspepsia. A double blind placebo controlled study, Gut, 31, 476-480.

[225] G. N. Tytgat. (1987) Colloidal bismuth subcitrate in peptic ulcer - a review, Digestion, 37, 31-41.

[226] S. J. Konturek, T. Radecki, I. Piastucki and D. Drozdowicz. (1987) Studies on the gastroprotective and ulcer-healing effects of colloidal bismuth subcitrate, Digestion, 37, 8-15.

[227] C. D. Giacomo, R. Fiocca, L. Villani, L. Lisato, G. Licardi, N. Diegoli, A. Donadini and G. Maggiore. (1990) Helicobacter pylori infection and chronic gastritis: Clinical, serological, and histologic correlations in children treated with amoxicillin and colloidal bismuth subcitrate, J. Pediatr. Gastroenterol. Nutr., 11, 310-316.

[228] DeRoyal Industries Inc. Xeroform petrolatum wound dressing, Available from: http://www.deroyal.com/medicalproducts/woundcare/product.aspx?id=wc-genwound-xeroform, [accessed: May 25, 2016]

[229] W. L. Peterson, A. A. Ciociola, D. L. Sykes, D. J. McSorley, D. D. Webb and The RBC H. Pylori Study Group. (1996) Ranitidine bismuth citrate plus clarithromycin is effective for healing duodenal ulcers, eradicating H. pylori and reducing ulcer recurrence, Aliment. Pharmacol. Ther., 10, 251-261.

[230] Y. Yang, R. Ouyang, L. Xu, N. Guo, W. Li, K. Feng, L. Ouyang, Z. Yang, S. Zhou and Y. Miao. (2015) Review: Bismuth complexes: synthesis and applications in biomedicine, J. Coord. Chem., 68, 379-397.

[231] T. Murafuji, Y. Miyoshi, M. Ishibashi, A. F. Mustafizur Rahman, Y. Sugihara, I. Miyakawa and H. Uno. (2004) Antifungal activity of organobismuth compounds against the yeast Saccharomyces cerevisiae: structure-activity relationship, J. Inorg. Biochem., 98, 547-552.

[232] N. Yang, J. A. Tanner, B. J. Zheng, R. M. Watt, M. L. He, L. Y. Lu, J. Q. Jiang, K. T. Shum, Y. P. Lin, K. L. Wong, M. C. Lin, H. F. Kung, H. Sun and J. D. Huang. (2007) Bismuth complexes inhibit the SARS coronavirus, Angew. Chem. Int. Ed., 46, 6464-6468.

[233] N. Yang, J. A. Tanner, Z. Wang, J. D. Huang, B. J. Zheng, N. Zhu and H. Sun. (2007) Inhibition of SARS coronavirus helicase by bismuth complexes, Chem. Commun., 4413-4415.

[234] P. C. Andrews, V. L. Blair, R. L. Ferrero, P. C. Junk, L. Kedzierski and R. M. Peiris. (2014) Bismuth(III) beta-thioxoketonates as antibiotics against Helicobacter pylori and as anti-leishmanial agents, Dalton Trans., 43, 1279-1291.

[235] P. C. Andrews, R. Frank, P. C. Junk, L. Kedzierski, I. Kumar and J. G. MacLellan. (2011) Anti-Leishmanial activity of homo- and heteroleptic bismuth(III) carboxylates, J. Inorg. Biochem., 105, 454-461.

[236] M. L. Gomes, G. DeFreitas-Silva, P. G. dos Reis, M. N. Melo, F. Frezard, C. Demicheli and Y. M. Idemori. (2015) Synthesis and characterization of bismuth(III) and antimony(V) porphyrins: high antileishmanial activity against antimony-resistant parasite, J. Biol. Inorg. Chem., 20, 771-779.

[237] P. T. Reynolds. (2012) Bismuth toxicity: A rare cause of neurologic dysfunction, Int. J. Clin. Med., 3, 46-48.

[238] H. Sun, H. Li and P. J. Sadler. (1997) The biological and medicinal chemistry of bismuth, Chem. Ber./Recueil, 130, 669-668.

[239] W. Hespe, H. J. M. Staal and D. W. R. Hall. (1988) Bismuth absorption from the colloidal subcitrate, Lancet, 332, 1258.


55

[240] H. Sun, H. Li, A. B. Mason, R. C. Woodworth and P. J. Sadler. (2001) Competitive binding of bismuth to transferrin and albumin in aqueous solution and in blood plasma, J. Biol. Chem., 276, 8829-8835.

[241] H. Li, P. J. Sadler and H. Sun. (1996) Unexpectedly strong binding of a large metal ion (Bi3+) to human serum transferrin, J. Biol. Chem., 271, 9483-9489.

[242] L. Zhang, K. Y. Szeto, W. B. Wong, T. T. Loh, P. J. Sadler and H. Sun. (2001) Interactions of bismuth with human lactoferrin and recognition of the BiIII-lactoferrin complex by intestinal cells, Biochemistry, 40, 13281-13287.

[243] H. Sun and K. Y. Szeto. (2003) Binding of bismuth to serum proteins: implication for targets of Bi(III) in blood plasma, J. Inorg. Biochem., 94, 114-120.

[244] N. Yang and H. Sun. (2007) Biocoordination chemistry of bismuth: recent advances, Coord. Chem. Rev., 251, 2354-2366.

[245] A. T. Miles, G. M. Hawksworth, J. H. Beattie and V. Rodilla. (2000) Induction, regulation, degradation, and biological significance of mammalian metallothioneins, Crit. Rev. Biochem. Mol. Biol., 35, 35-70.

[246] H. Sun, H. Li, I. Harvey and P. J. Sadler. (1999) Interactions of bismuth complexes with metallothionein(II), J. Biol. Chem., 274, 29094-29101.

[247] P. J. Boogaard, A. Slikkerveer, J. F. Nagelkerke and G. J. Mulder. (1991) The role of metallothionein in the reduction of cisplatin-induced nephrotoxicity by Bi3+-pretreatment in the rat in vivo and in vitro. Are antioxidant properties of metallothionein more relevant than platinum binding?, Biochem. Pharmacol., 41, 369-375.

[248] M. Satoh, Y. Aoki and C. Tohyama. (1997) Protective role of metallothionein in renal toxicity of cisplatinum, Cancer Chemother. Pharmacol., 40, 358-362.

[249] M. Stoltenberg, G. Danscher, R. Pamphlett, M. M. Christensen and J. Rungby. (2000) Histochemical tracing of bismuth in testis from rats exposed intraperitoneally to bismuth subnitrate, Reprod. Toxicol., 14, 65-71.

[250] M. Stoltenberg and G. Danscher. (2000) Histochemical differentiation of autometallographically traceable metals (Au, Ag, Hg, Bi, Zn): protocols for chemical removal of separate autometallographic metal clusters in epon sections, Histochem. J., 32, 645-652.

[251] M. Stoltenberg, J. A. Hogenhuis, J. J. Hauw and G. Danscher. (2001) Autometallographic tracing of bismuth in human brain autopsies, J. Neuropathol. Exp. Neurol., 60, 705-710.

[252] M. Stoltenberg, J. D. Schionning and G. Danscher. (2001) Retrograde axonal transport of bismuth: an autometallographic study, Acta Neuropathol., 101, 123-128.

[253] N. E. Magnusson, A. Larsen, J. Rungby, M. Kruhoffer, T. F. Orntoft and M. Stoltenberg. (2005) Gene expression changes induced by bismuth in a macrophage cell line, Cell Tissue Res., 321, 195-210.

[254] H. Sun, L. Zhang and K. Y. Szeto. (2004) Bismuth in medicine, Met. Ions Biol. Syst., 41, 333-378.

[255] F. Thomas, B. Bialek and R. Hensel. (2011) Medical use of bismuth: the two sides of the coin, J. Clin. Toxicol., S:3, doi:10.4172/2161-0495.s4173-4004.

[256] E. Asato, K. Katsura, M. Mikuriya, U. Turpeinen, I. Mutikainen and J. Reedijk. (1995) Synthesis, structure, and spectroscopic properties of bismuth citrate compounds and the bismuth-containing ulcer-healing agent colloidal bismuth subcitrate (CBS). 4. crystal structure and solution behavior of a unique dodecanuclear cluster (NH4)12[Bi12O8(cit)8](H2O)10, Inorg. Chem., 34, 2447-2454.

[257] C. Tasman-Jones, C. Maher, L. Thomsen, S. P. Lee and M. Vanderwee. (1987) Mucosal defences and gastroduodenal disease, Digestion, 37 Suppl 2, 1-7.

[258] D. Bagchi, O. R. Carryl, M. X. Tran, R. L. Krohn, D. J. Bagchi, A. Garg, M. Bagchi, S. Mitra and S. J. Stohs. (1998) Stress, diet and alcohol-induced oxidative


56

gastrointestinal mucosal injury in rats and protection by bismuth subsalicylate, J. Appl. Toxicol., 18, 3-13.

[259] D. Bagchi, T. R. McGinn, X. Ye, J. Balmoori, M. Bagchi, S. J. Stohs, C. A. Kuszynski, O. R. Carryl and S. Mitra. (1999) Mechanism of gastroprotection by bismuth subsalicylate against chemically induced oxidative stress in cultured human gastric mucosal cells, Dig. Dis. Sci., 44, 2419-2428.

[260] W. Beil, S. Bierbaum and K.-F. Sewing. (1993) Studies of the mechanism of action of collodial bismuth subcitrate, Pharmacology, 47, 141-144.

[261] J. H. Baron, J. Barr, J. Batten, R. Sidebotham and J. Spencer. (1986) Acid, pepsin, and mucus secretion in patients with gastric and duodenal ulcer before and after colloidal bismuth subcitrate (De-Nol), Gut, 27, 486-490.

[262] S. J. Konturek, A. Dembinski, Z. Warzecha, W. Bielanski, T. Brzozowski and D. Drozdowicz. (1988) Epidermal growth factor (EGF) in the gastroprotective and ulcer healing actions of colloidal bismuth subcitrate (De-Nol) in rats, Gut, 29, 894-902.

[263] B. L. Slomiany, X. Y. Wang, D. Palecz, K. Okazaki and A. Slomiany. (1990) Participation of phosphoinositides in gastric mucosal protection by colloidal bismuth subcitrate against ethanol-induced injury, Alcohol. Clin. Exp. Res., 14, 580-583.

[264] J. Gilster, K. Bacon, K. Marlink, B. Sheppard, C. Deveney and M. Rutten. (2004) Bismuth subsalicylate increases intracellular Ca2+, MAP-kinase activity, and cell proliferation in normal human gastric mucous epithelial cells, Dig. Dis. Sci., 49, 370-378.

[265] B. J. Marshall and J. R. Warren. (1984) Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration, Lancet, 1, 1311-1315.

[266] IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. (1994) Schistosomes, liver flukes and Helicobacter pylori, Volume 61. Lyon, France: IARC.

[267] R. M. Peek, Jr. and M. J. Blaser. (2002) Helicobacter pylori and gastrointestinal tract adenocarcinomas, Nat. Rev. Cancer, 2, 28-37.

[268] F. Mégraud. (2012) The challenge of Helicobacter pylori resistance to antibiotics: the comeback of bismuth-based quadruple therapy, Therap. Adv. Gastroenterol., 5, 103-109.

[269] C. W. Stratton, R. R. Warner, P. E. Coudron and N. A. Lilly. (1999) Bismuth-mediated disruption of the glycocalyx-cell wall of Helicobacter pylori: ultrastructural evidence for a mechanism of action for bismuth salts, J. Antimicrob. Chemother., 43, 659-666.

[270] T. E. Sox and C. A. Olson. (1989) Binding and killing of bacteria by bismuth subsalicylate, Antimicrob. Agents Chemother., 33, 2075-2082.

[271] L. Zhang, S. B. Mulrooney, A. F. Leung, Y. Zeng, B. B. Ko, R. P. Hausinger and H. Sun. (2006) Inhibition of urease by bismuth(III): implications for the mechanism of action of bismuth drugs, Biometals, 19, 503-511.

[272] L. Jin, K. Y. Szeto, L. Zhang, W. Du and H. Sun. (2004) Inhibition of alcohol dehydrogenase by bismuth, J. Inorg. Biochem., 98, 1331-1337.

[273] O. Senkovich, S. Ceaser, D. J. McGee and T. L. Testerman. (2010) Unique host iron utilization mechanisms of Helicobacter pylori revealed with iron-deficient chemically defined media, Infect. Immun., 78, 1841-1849.

[274] H. A. Song, C. S. Kang, K. E. Baidoo, D. E. Milenic, Y. Chen, A. Dai, M. W. Brechbiel and H.-S. Chong. (2011) An efficient bifunctional decadentate ligand 3p-C-DEPA for targeted alpha radioimmunotherapy applications, Bioconjugate Chem., 22, 1128-1135.

[275] M. Cherel, S. Gouard, J. Gaschet, C. Sai-Maurel, F. Bruchertseifer, A. Morgenstern, M. Bourgeois, J. F. Gestin, F. K. Bodere, J. Barbet, P. Moreau and F. Davodeau. (2013) 213Bi radioimmunotherapy with an anti-mCD138 monoclonal antibody in a murine model of multiple myeloma, J. Nucl. Med., 54, 1597-1604.


57

[276] J. M. Kinsella, R. E. Jimenez, P. P. Karmali, A. M. Rush, V. R. Kotamraju, N. C. Gianneschi, E. Ruoslahti, D. Stupack and M. J. Sailor. (2011) X-ray computed tomography imaging of breast cancer by using targeted peptide-labeled bismuth sulfide nanoparticles, Angew. Chem. Int. Ed., 50, 12308-12311.

[277] J. T. Chen, Y. Kurokawa, K. Nakayama, K. Saito, K. Sakamoto, K. Fukuda, Y. Hirai, T. Hamada, I. Fujimoto and K. Yamauchi. (1987) The effect of bismuth subnitrate on cisplatin toxicity, J. Jpn. Obstet. Gynecol. Soc., 39, 815-822.

[278] T. Morikawa and E. Kawamura. (1989) A method of cisplatin administration with the aid of high-dose bismuth subnitrate, and their pharmacokinetics, Jpn. J. Cancer Chemother., 16, 1094-1098.

[279] Y. Kondo, M. Satoh, N. Imura and M. Akimoto. (1992) Tissue-specific induction of metallothionein by bismuth as a promising protocol for chemotherapy with repeated administration of cis-diamminedichloroplatinum (II) against bladder tumor, Anticancer Res., 12, 2303-2307.

[280] A. Naganuma, M. Satoh and N. Imura. (1988) Specific reduction of toxic side effects of adriamycin by induction of metallothionein in mice, Jpn. J. Cancer Res., 79, 406-411.

[281] M. Satoh, A. Naganuma and N. Imura. (1988) Metallothionein induction prevents toxic side effects of cisplatin and adriamycin used in combination, Cancer Chemother. Pharmacol., 21, 176-178.

[282] K. Iuchi, Y. Hatano and T. Yagura. (2008) Heterocyclic organobismuth(III) induces apoptosis of human promyelocytic leukemic cells through activation of caspases and mitochondrial perturbation, Biochem. Pharmacol., 76, 974-986.

[283] K. Iuchi, K. Akagi and T. Yagura. (2009) Heterocyclic organobismuth(III) compound targets tubulin to induce G2/M arrest in HeLa cells, J. Pharmacol. Sci., 109, 573-582.

[284] K. Onishi, M. Douke, T. Nakamura, Y. Ochiai, N. Kakusawa, S. Yasuike, J. Kurita, C. Yamamoto, M. Kawahata, K. Yamaguchi and T. Yagura. (2012) A novel organobismuth compound, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl]-pyrrolidine, induces apoptosis in the human acute promyelocytic leukemia cell line NB4 via reactive oxygen species, J. Inorg. Biochem., 117, 77-84.

[285] K. S. O. Ferraz, N. F. Silva, J. G. da Silva, L. F. de Miranda, C. F. D. Romeiro, E. M. Souza-Fagundes, I. C. Mendes and H. Beraldo. (2012) Investigation on the pharmacological profile of 2,6-diacetylpyridine bis(benzoylhydrazone) derivatives and their antimony(III) and bismuth(III) complexes, Eur. J. Med. Chem., 53, 98-106.

[286] I. P. Ferreira, E. D. L. Piló, A. A. Recio-Despaigne, J. G. Da Silva, J. P. Ramos, L. B. Marques, P. H. D. M. Prazeres, J. A. Takahashi, E. M. Souza-Fagundes, W. Rocha and H. Beraldo. (2016) Bismuth(III) complexes with 2-acetylpyridine- and 2-benzoylpyridine-derived hydrazones: Antimicrobial and cytotoxic activities and effects on the clonogenic survival of human solid tumor cells, Bioorg. Med. Chem., 24, 2988-2998.

[287] Y.-K. Li, M. Yang, M.-X. Li, H. Yu, H.-C. Wu and S.-Q. Xie. (2013) Synthesis, crystal structure and biological evaluation of a main group seven-coordinated bismuth(III) complex with 2-acetylpyridine N4-phenylthiosemicarbazone, Bioorg. Med. Chem. Lett., 23, 2288-2292.

[288] X. Wang, X. Zhang, J. Lin, J. Chen, Q. Xu and Z. Guo. (2003) DNA-binding property and antitumor activity of bismuth(III) complex with 1,4,7,10-tetrakis(2-pyridylmethyl)-1,4,7,10-tetraazacyclododecane, Dalton Trans., 2379-2380.

[289] K. A. Smith, G. B. Deacon, W. R. Jackson, E. R. Tiekink, S. Rainone and L. K. Webster. (1998) Preparation and anti-tumour activity of some arylbismuth(III) oxine complexes, Met. Based Drugs, 5, 295-304.

[290] H. Li, C. S. Lai, J. Wu, P. C. Ho, D. de Vos and E. R. Tiekink. (2007) Cytotoxicity, qualitative structure-activity relationship (QSAR), and anti-tumor activity of bismuth dithiocarbamate complexes, J. Inorg. Biochem., 101, 809-816.


58

[291] M.-X. Li, M. Yang, J.-Y. Niu, L.-Z. Zhang and S.-Q. Xie. (2012) A nine-coordinated bismuth(III) complex derived from pentadentate 2,6-diacetylpyridine bis(4N-methylthiosemicarbazone): crystal structure and both in vitro and in vivo biological evaluation, Inorg. Chem., 51, 12521-12526.

[292] H. Li and H. Sun. (2012) Recent advances in bioinorganic chemistry of bismuth, Curr. Opin. Chem. Biol., 16, 74-83.

[293] M. D. Ioffreda, C. A. Gordon, D. R. Adams, S. J. Naides and J. J. Miller. (2001) Black tongue, Arch. Dermatol., 137, 968-969.

[294] R. Burns, D. W. Thomas and V. J. Barron. (1974) Reversible encephalopathy possibly associated with bismuth subgallate ingestion, BMJ, 1, 220-223.

[295] D. J. Lowe. (1974) Adverse effects of bismuth subgallate. A further report from the Australian Drug Evaluation Committee, Med. J. Aust., 2, 664-666.

[296] V. Supino-Viterbo, C. Sicard, M. Risvegliato, G. Rancurel and A. Buge. (1977) Toxic encephalopathy due to ingestion of bismuth salts: clinical and EEG studies of 45 patients, J. Neurol. Neurosur. Ps., 40, 748-752.

[297] A. Larsen, N. Martiny, M. Stoltenberg, G. Danscher and J. Rungby. (2003) Gastrointestinal and systemic uptake of bismuth in mice after oral exposure, Pharmacol. Toxicol., 93, 82-90.

[298] B. Bialek, R. A. Diaz-Bone, D. Pieper, M. Hollmann and R. Hensel. (2011) Toxicity of methylated bismuth compounds produced by intestinal microorganisms to bacteroides thetaiotaomicron, a member of the physiological intestinal microbiota, J. Toxicol., 608349, doi:10.1155/2011/608349.

[299] P. C. Andrews, R. L. Ferrero, P. C. Junk, I. Kumar, Q. Luu, K. Nguyen and J. W. Taylor. (2010) Bismuth(III) complexes derived from non-steroidal anti-inflammatory drugs and their activity against Helicobacter pylori, Dalton Trans., 39, 2861-2868.

[300] M. Abuznaid, A.-S. Sallam, I. Hamdan, M. Al-Hussaini and A. Bani-Jaber. (2008) Diclofenac-bismuth complex: synthesis, physicochemical, and biological evaluation, Drug Dev. Ind. Pharm., 34, 434-444.

[301] M. M. Goldenberg, L. J. Honkomp, S. E. Burrous and A. W. Castellion. (1975) Protective effect of Pepto-Bismol liquid on the gastric mucosa of rats, Gastroenterology, 69, 636-640.

[302] S. Tanaka, P. H. Guth, O. R. Carryl and J. D. Kaunitz. (1997) Cytoprotective effect of bismuth subsalicylate in indomethacin-treated rats is associated with enhanced mucus bismuth concentration, Aliment. Pharmacol. Ther., 11, 605-612.

[303] S. J. Konturek, N. Kwiecien, W. Obtulowicz, Z. Hebzda and J. Oleksy. (1988) Effects of colloidal bismuth subcitrate on aspirin-induced gastric microbleeding, DNA loss, and prostaglandin formation in humans, Scand. J. Gastroenterol., 23, 861-866.

[304] D. Bagchi, O. R. Carryl, M. X. Tran, M. Bagchi, P. J. Vuchetich, R. L. Krohn, S. D. Ray, S. Mitra and S. J. Stohs. (1997) Protection against chemically-induced oxidative gastrointestinal tissue injury in rats by bismuth salts, Dig. Dis. Sci., 42, 1890-1900.

[305] N. Hudson, F. E. Murray, A. T. Cole, G. M. Turnbull, S. Lettis and C. J. Hawkey. (1993) Ranitidine bismuth citrate and aspirin-induced gastric mucosal injury, Aliment. Pharmacol. Ther., 7, 515-521.

[306] F. V. Izzettin, M. Sancar, B. Okuyan, S. Apikoglu-Rabus and U. Cevikbas. (2012) Comparison of the protective effects of various antiulcer agents alone or in combination on indomethacin-induced gastric ulcers in rats, Exp. Toxicol. Pathol., 64, 339-343.

59

Chapter2. InfluenceofBiNSAIDstabilityonthetoxicityand

PGE2inhibitioninHCT-8coloncancercells

In order to determine whether BiNSAIDs are potential chemotherapeutics or

chemopreventives in vivo, the behaviour of the BiNSAIDs first needed to be studied

invitro.InthisChapter,thestabilityoftheBiNSAIDsincellmediumwasexaminedto

determine the biologically available species. The HCT-8 colon cancer cell line was

employed to assess the toxicity of the BiNSAIDs compared to uncomplexedNSAIDs,

and cisplatin, a clinicallyusedanti-cancerdrug. Upon confirming the toxicityof the

BiNSAIDs, cell death assays were employed to determine the mode of cell death

induced by the BiNSAIDs in the HCT-8 cells. Finally, the ability of the BiNSAIDs to

inhibit the production of PGE2 is presented with the respect to the proposed

mechanismofNSAIDchemoprevention,COXinhibition.

SectionsofthisChapterhavebeenpublishedin:

Hawksworth, E. L., Andrews, P. C., Lie, W., Lai, B. & Dillon, C. T. (2014) Biological

evaluation of bismuth non-steroidal anti-inflammatory drugs (BiNSAIDs): stability,

toxicity and uptake in HCT-8 colon cancer cells, J. Inorg. Biochem., 135, 28-39.

Chapter2.Stability,toxicityandPGE2inhibitionofBiNSAIDs

60

2.1 IntroductionAs discussed in Section 1.2.2, the results from a number of epidemiological,

observational and clinical studies indicate that long-term administration of NSAIDs,

suchasaspH,reducestheincidenceandreoccurrenceofCRC. Therearetwocriteria

that are important for this: (i)eradication of already transformed cells, and

(ii)preventingtheformationoftransformedcells.

Previously,NSAIDshavebeenextensivelystudiedinanumberofcancercelllines

(asdescribedinSection1.2.4)toexploreboththeirabilitytokillcancercellsandtheir

ability to prevent the transformation of normal cells to cancer cells. In addition, a

variety of Bi(III) complexes have also shown promising anti-cancer activity in vitro

and in vivo (discussed in Section 1.4.1.4) indicating their ability to kill transformed

cells.However,noworkhasbeenperformedtoestablishtheabilityofBiNSAIDstokill

cancerouscellsorpreventtheirformation.

Thereareanumberoffactorsthatneedtobeaddressedinordertodeterminethe

suitability of BiNSAIDs as potential chemotherapeutic or chemopreventive drug

candidates including general factors such as: (i) solubility/bioavailability, and

(ii)stability;andspecificfactorsrelatingtothechemopreventioncriteria:(iii)toxicity

towardscoloncancercells, (iv) inhibitionofCOX(andsubsequentreduction inPGE2

release), and (v) cellular uptake and intracellular distribution (discussed in

Chapter3). Factors (i) and (ii) are also important for in vitro testing whereby

compromisesneedtobeconsideredforthesepurposes.Assuch,Sections2.1.1–2.1.5

will present the current literature relevant to the optimal factors for BiNSAID

chemotherapeutic/chemopreventionsuitabilityandtherationaleforthemethodology

behindtheexperimentsdescribedinthisChapter.

2.1.1 Solubility/Bioavailability

Oral administration is themostdesirable routeof administrationof therapeutics

andpreventivesduetotheeaseofuseforthepatient[1]. Thedevelopmentoforally

administered drugs is often hampered by poor aqueous solubility and high

lipophilicity (logP > 3.5) which then leads to poor and variable absorption/

bioavailability [2]. Drugs with poor solubility can present in vivo issues including

reducedoralbioavailability, lackofefficacy,theneedtodevelopexpensiveandoften

problematic formulations, and the burden to patients through frequent dosing to

achieve therapeutic effectiveness [1]. Nonetheless, the requirement of the drug to


61

traverse the cellmembrane imposes a requirement for some degree of lipophilicity

suchthatabalancebetweenaqueoussolubilityandlipophilicityisessential[3].

NSAIDs,whichrepresentoneof themosthighlyprescribedclassesofdrug in the

world[4],aregenerallyadministeredorally[5];however,theydisplaypooraqueous

solubility and incomplete dissolution in the stomach [6]. In general, NSAIDs are

hydrophobicmolecules,aproperty that iscrucial for theirmechanismofactionas it

allowsthemoleculestoapproachthemembrane-boundtargetenzyme,COX,andenter

itshydrophobicarachidonate-bindingchannel [6]. Despite theseproperties,NSAIDs

generally display 80–95%oral absorption and 99%plasmaprotein binding (mainly

toalbumin)[6].

Similarly, commonly used bismuth drugs, BSS (Pepto-Bismol, Procter & Gamble

Company) and CBS (De-Nol®; Gist Brocades and Yamanouchi) exhibit poor aqueous

solubility and stability (discussed in more detail in Section 2.1.2). However, these

propertiescontributetothegastroprotectivenatureofthesedrugsasthedrugformsa

solidphysical coatingon theepithelial liningof the stomachsubsequentlyproviding

protectionviaaphysicalbarrier[7].

While the poor aqueous solubility of the BiNSAIDs may not necessarily hamper

invivo studiesandpotentialpatientadministration, itdoesposean issue for invitro

testing. For instance, lead compounds displaying poor aqueous solubility are often

considered less than ideal candidates for drug development as a number issues can

arise during in vitro testing, including; erratic assay results, artificially low potency,

anderroneousstructure−activityrelationships[1]. Assuch,theaqueoussolubilityof

theBiNSAIDswasafactorthatrequiredconsiderationfortheassaysreportedinthis

Chapter. Previous studies have shown that the BiNSAIDs, Bi(tolf)3, Bi(mef)3 and

Bi(difl)3,displaypoorsolubilityindeionizedwater[8]andthuswouldnotdissolvein

the cell culture medium required for cell testing. Furthermore, BiNSAIDs display

limited solubility in organic solvents such as toluene, acetone and ethanol [8];

however, all three BiNSAIDs showed appreciable solubility in dimethyl

sulphoxide(DMSO) [8]. DMSO is often employed as a solvent for biological in vitro

testing of compounds with low aqueous solubility; the high solvating efficiency of

DMSOstemsfromitsamphiphilicproperty,widetemperaturerangeoftheliquidstate,

and high polarity [9]. DMSO is generally not favoured for oral administration to

humans as the metabolism of DMSO to dimethyl sulfide produces a sulphuric-like

odour on the skin and a garlic-like after-taste and breath odour that can persist for


62

several days following DMSO administration[10]. In addition, DMSO is associated

withanumberofsideeffectsincludingsedation,nausea,headache,diarrhea,dizziness

and topical irritation [10]. However, DMSO is a commonly acceptable vehicle for

cell-basedscreeningassayswherebytheuseofsmallvolumesofDMSOiscommonly

reported for the purpose of testing metal complexes [11-16] and poorly soluble

organiccompounds[17-20].

2.1.2 Drugstability

Thereareanumberofproperties thatcanbeexploitedby themedicinalchemist

whendesigning/developingmetalcontainingdrugs. Forexample, theverynatureof

metal-ligandcomplexesprovidesascaffoldwherebyametalcentrecanbeconjugated

to an array of ligands with various geometries and dissociation rates. Inorder to

ensurethemetal-complexreachesitsbiologicaltarget,themetalcomplexshouldhave

asufficientlyhighthermodynamicstabilitytodeliverthemetaltothebiologicaltarget,

the metal-ligand binding should be hydrolytically stable, and understanding the

kineticswithwhichthemetalionundergoesligationordissociationreactionsshould

beconsidered[21]. Whilepresentingasignificantchallenge, it isnecessarytostudy

thestabilityofthemetal-complexesinthebiologicalenvironmenttoensurethemetal

complexeswill survive longenough in vivo to reach theirbiological target andexert

theintendedmedicinaleffects[22].

Generally, medicinal Bi compounds exhibit low aqueous solubility and lack

solutionstability;however,asdiscussedinSection1.4.1.2,lowaqueoussolubilitymay

beessentialforthegastroprotectiveeffectsofBi[7].Forinstance,anumberofstudies

have explored the formation of BSS, CBS and RBC polymeric products in acidic

environmentsthatmimicthestomach[23,24]. Thecellmediumusedtomaintain in

vitrocellculturesisformulatedtosimulatetheinvivogrowthenvironmentofthecells.

Buffersarepresent in themediumtomaintainphysiologicalpH7.4, the idealpHfor

cells to grow. Due to the nature of Bi(III) carboxylates (that are common to the

BiNSAIDs)itisessentialtodeterminethestabilityoftheBiNSAIDsinthecellmedium

inordertoassessthestructuresofthebiologicallyavailablespecies.Techniquesoften

used to study drug stability include ultraviolet-visible (UV-Vis) spectroscopy,

electrospray ionisationmass spectrometry (ESI-MS) andNMR spectroscopy. UV-Vis

spectroscopy is useful for determining if there is a change; however, the specific

chemicalinformationislimited.ESI-MSisoftenutilisedasatechniquetocharacterise


63

complexesbasedonthefragmentationpattern.However,ESI-MSisapoortechnique

fortheanalysisofhydrophobiccompoundsinDMSOduetothenon-volatilenatureof

this solvent. NMR spectroscopy is also a technique that is commonly used to

characterise compounds, providing structural information based on the chemical

environmentofthenuclei(commonly1Hand13C).Importantly,DMSO(intheformof

deuteratedDMSO,DMSO-d6) is a commonlyemployedsolvent inNMRspectroscopy,

particularly for compounds that have difficulty dissolving in other organic solvents

such as methanol and chloroform. The use of DMSO-d6 as a NMR solvent is

particularlyusefulforcompoundscontaining–OHand–NH2groupsduetothestrong

hydrogenbondingbetweentheseprotondonatinggroupsandaproticDMSO[9].

2.1.3 Invitrocytotoxicitytowardscoloncancercells

AsmentionedinSection2.1,oneofthecriteriaassociatedwiththeadministration

of aCRC chemopreventive agent is its ability to eradicate already transformed cells.

This is important given that the cohortwould generally include patients that are at

highriskofdevelopingbowelcancer.

2.1.3.1 Assessinginvitrodrugtoxicity:TheMTTassay

Inordertoassessthepotentialchemotherapeuticactivityofnovelcompounds,itis

necessarytoscreenthemusinginvitrooncology-relatedassays[25].Ideally,theassay

should be sensitive, reliable, inexpensive, have the capacity to test multiple

drugs/concentrationsat thesametime,andberelativelyeasytoperform[25]. First

describedbyMosmann in1983 [26], theMTTassay fits theaforementionedcriteria

and is commonly performed using cancer cell lines to screen and compare the

chemotherapeutic potential of new compounds [25]. While there are several

variations of the assay protocol [26-30], they generally employ a water-soluble

substrate, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), as a

colorimetric measure of cell viability, proliferation or activation [25]. The assay is

basedonthepremisethatonlyviable(ormetabolically-active)cellsareabletoreduce

yellow MTT to water-insoluble purple (E,Z)-5-(4,5-dimethylthiazol-2-yl)-1,3-

diphenylformazan (formazan) by cleavage of the tetrazolium ring via the enzyme,

mitochondrialreductase(Scheme2.1)[25].


64

Scheme2.1.ThereductionofMTTtoFormazanbythemitochondrialenzyme,mitochondrialreductase,inlivingcells.

The water-insoluble formazan can be dissolved using an organic solvent or

detergentandanalysedspectrophotometricallyat570nm[26-30]. Astheamountof

MTTreducedtoformazanisdirectlyproportionaltothemetabolicactivityofthecells,

a concentration-response curve canbe constructedbasedon thepercentageofMTT

converted by cells treated with increasing concentrations of a drug. This can be

compared to the control cells that are treated with the vehicle. As such, an

IC50value(i.e.theconcentrationofdrugrequiredtoinhibit50%enzymeactivity)can

be determined and used to compare the toxicity ofmultiple compounds against the

samecellline.

2.1.3.2 Modesofcelldeath

The identification of the mode of drug-induced cell death can be a useful for

determining the affected intracellular biochemical processes resulting from the

interactionofadrugwithinacell.Whilethereareseveraldistinctmodesofcelldeath

that can be distinguished due to distinct differences in morphological effects and

biochemical changes [31], apoptosis and necrosis are the two most common and

well-characterisedmodesof cell death that canbe triggered in cancer cellsdue to a

chemotherapeutic agent. As apoptosis and necrosis result in different anti-tumour

responses, it is important to determine which mode of cell death is triggered by

anti-cancertherapies.

Apoptosisisoftenconsidereda‘programmed’formofcelldeathandoccursduring

development and aging to maintain normal or healthy tissue populations of

cells[32,33].Apoptosiscanalsobetriggeredasadefencemechanismduetoimmune


65

responses, cell damage via disease, or noxious stimuli [32]. Notably, not all cells

respond to apoptotic stimuli in the same manner; for instance, some

chemotherapeutic drugs can initiate apoptosis in cancer cells, without affecting

normal cells [32]. Distinguishing features of early apoptosis include reduction of

cellularvolume (pyknosis)andcell rounding, retractionofpseudopods, reduction in

nuclearvolume,externalisationofphosphatidylserine(PS),andminormodificationof

cytoplasmicorganelles[31]. Laterstagesofapoptoticcelldeathcanbeidentifiedby

nuclearfragmentation(karyorrhexis),membraneblebbing,lossofplasmamembrane

integrity,andengulfmentbyresidentphagocytesinvivo[31].Apoptosisisoneofthe

mostcommonmechanismsofchemotherapy-inducedcelldeath[33]. Incancercells,

normalapoptoticpathwaysareoftencompromised[33],whichisadvantageoustothe

fortificationof cancer cells against cell death. As such, chemotherapeutics areoften

designedtotargetaspecificpartoftheapoptoticpathwayinordertocircumventthe

defunctsignalingpathwaysandrestoreapoptosis[33].

In contrast, necrosis is a form of cell death that is morphologically and

biochemically distinct from apoptosis. While it is often considered an ‘accidental’,

uncontrolled form of cell death, accumulating evidence suggests that transduction

pathwaysandcatabolicmechanismsareinvolvedinregulatingnecrosis[31]. Clearly

distinct key features of necrotic cell death (as opposed to apoptosis) include

cytoplasmic swelling (oncosis), rupture of plasma membrane (occurs early, as

opposed to a late stage step in apoptosis), swelling of cytoplasmic organelles, and

moderate chromatin condensation[31]. Unlike apoptosis, necrosis releases the

intracellularcontentand tumourantigens,whichattracts immunecells to thesiteof

the tumour to help drive anti-tumour immune responses [33]. While several

anti-cancer compounds have been shown to induce cell death via necrosis [34-36],

triggering apoptosis is generally described a safer mechanism for the death of the

cancercells[37].

Themostcommonmethodofassessingthemodeofcelldeathisviaflowcytometry

coupledwithstainingformarkersspecifictotheeventsassociatedwithapoptosisor

necrosis[31]. Two fluorescent probes, fluorescein isothiocyanate (FITC)-labeled

Annexin V (referred to herein as AnnV) and 7-aminoactinomycinD (7AAD), are

commonly used to distinguish between apoptotic cells and necrotic cells. AnnV is

utilised as it is a Ca2+ dependent phospholipid-bindingproteinwith high affinity for

PS[38].Duringtheearlystagesofapoptosis,PStranslocatesfromtheinnerleafletof


66

theplasmamembranetotheouterleaflet[39].AnnVbindingtoPSisnotexclusiveto

apoptosis. During the later stages of necrosis the cells increase in permeability

allowingentryofAnnVwhichcanbindtoPSontheinnerleaflet[38]. Incontrastto

this, during the early stages of necrosis, cells lose their membrane integrity and

become permeable, while this event does not occur until the later stages of

apoptosis[31]. Hence, the assessment of cell death is performed with AnnV in

combination with another fluorescent dye that is excluded by an intact plasma

membrane,butisabletodiffusethroughporesofacompromisedcellmembrane[38].

Traditionally, propidium iodide or 7AAD are employed to this end; thesemolecules

increase in fluorescence intensity when they intercalate with DNA following entry

throughthecompromisedcellmembrane,andprovideameasureoflossofmembrane

integrity[40,41].

AnexampleoftheflowcytometrydataobtainedfromAnnVand7AADco-stained

cellsisshowninFigure2.1.ViablecellsmaintainanintactplasmamembraneandPS

remains on the inner leaflet, thus viable cells do not bind AnnV and exclude 7AAD,

such that these cells exhibit low fluorescent intensity (AnnV−/7AAD−); and can be

identifiedasthepopulationinthebottom-leftquadrantofthebivariateplot(Fig.2.1).

In the early stagesof apoptosis, PS translocates from the inner leaflet of theplasma

membranetothecellsurfaceallowingAnnVtobind;however,theplasmamembrane

remainsintactexcluding7AADfromthecell(AnnV+/7AAD−).AsshowninFigure2.1,

thesecellsarelocatedinthebottom-rightquadrantofthebivariateplot. Inthelater

stagesofapoptosisporesappear in thecellmembraneallowing7AADtoenter. The

population in the top-right corner is indicative of late apoptosis or late necrosis

(AnnV+/7AAD+, as shown in Fig. 2.1). In necrosis the loss of plasma membrane

integrity occurs before the translocation of AnnV (AnnV−/7AAD+). As such, early

necroticcellsappearinthetop-leftquadrantinFigure2.1.


67

Figure2.1:TheuseofflowcytometrytodistinguishmodesofcelldeathusingAnnVand7AAD.ThepresenceofAnnV(depictedingreen)boundtoPS(depictedinblue)and7AAD(depictedinred)boundtoDNAinthecellnucleus(depictedinlightgrey)inthecellisdependentonthemodeofcelldeaththecellshaveundergone.

2.1.4 InhibitionofCOX:ReductionofPGE2production

Since one of the proposed mechanisms of chemoprevention by NSAIDs is the

inhibitionofCOX-2, it is important toestablishwhetherBiNSAIDscan inhibitCOXat

comparable concentrations. This information can provide preliminary evidence to

determinetheirpotentialusefulnessaschemopreventiveagentsandformaguidefor

futurestudies.AsdiscussedinSection1.2.3,theupregulationofCOX-2productionhas

beenestablishedinanumberofcancercelllines.TheupregulationofCOX-2,inturn,

results in the overproduction of PGE2, a PG that is associatedwith fortifying cancer

cells against cell death and increasing the invasiveness of cancer cells (as

demonstratedinFig.1.5).WhileitisofinteresttostudyCOXinhibitionbyincubation

of the BiNSAIDswith purified COX-1 and COX-2 enzymes, these experimentswould

not take into account the conditions that the NSAISDs and BiNSAIDs would be

subjectedtoinacellularenvironment(i.e.cellpermeabilityandmetabolicpathways).

ThedisadvantageofutilisingacellsystemtodetermineCOXinhibitionisthatCOX-1

versus COX-2 inhibition cannot be distinguished. However, this study will directly


68

compare BiNSAID versus NSAID, as it is important to establish if the BiNSAIDs can

provideequalorsuperiorCOXinhibitioninthecellularenvironment.Thustheability

ofBiNSAIDstoinhibitPGE2production(andsubsequentlytheabilitytoinhibitCOX)in

atransformedcelllinewillbeassessedinthisChapter.

2.1.5 Chapteraims

The establishment of the in vitro activity is the first step to determine whether

BiNSAIDs will be suitable candidates for in vivo studies as chemotherapeutic or

chemopreventive agents for CRC, and ultimately, whether BiNSAIDs should be

investigatedasapotentialmedication for thereductionof theaccompanyingGIside

effects of long-term NSAID use. The experiments described in this Chapter can be

dividedintotwopurposes:(i)toestablishthechemotherapeuticpotentialofBiNSAIDs

bydeterminingwhetherBiNSAIDsareabletokilltransformedcells(toxicityandcell

deathassays), and (ii) toestablishwhetherBiNSAIDsdemonstratechemopreventive

potentialbydeterminingwhetherBiNSAIDsareabletoinhibittheproductionofPGE2

(andthusinhibitCOX)intransformedcells.Ahumanepithelialileocecalcoloncancer

cell line, HCT-8, was employed as a model CRC cell line for the evaluation of the

BiNSAIDs, Bi(tolf)3, Bi(mef)3 and Bi(difl)3 and the respective uncomplexed NSAIDs,

tolfH,mefHanddiflH.

Thespecificaimsofthischapterwereto:

1. Investigate the stability of Bi(tolf)3, Bi(mef)3 and Bi(difl)3 in cell medium

usingNMRspectroscopyinordertoelucidatethebiologicallyactiveformof

thecomplexesintheinvitroassays.

2. Assess the in vitro toxicity of theBiNSAIDs,Bi(tolf)3, Bi(mef)3 andBi(difl)3,

and compare to the uncomplexed NSAIDs against human colon cancer cell

line,HCT-8.

3. DeterminewhethertheBiNSAIDs,Bi(tolf)3,Bi(mef)3andBi(difl)3,inducethe

same typeof celldeath (apoptosisornecrosis) as the respectiveNSAIDs in

HCT-8cells.

4. Ascertain whether complexation with Bi affects the ability of the NSAIDs,

tolfH, mefH and diflH, to interact with COX in vitro, by quantifying PGE2

productionbyHCT-8cells.


69

2.2 Materialsandmethods2.2.1 Materials

The Bi complexes of tolfH, mefH and diflH (Bi(tolf)3, Bi(mef)3 and Bi(difl)3,

respectively)werechosenforthisstudy. TolfHhasbeenextensivelystudied invitro

against a number of different types of human cancer cell lines [42-50]. MefH is

structurally similar to tolfH (both belonging to the structural class called fenamic

acids, Fig. 1.2) and was tested as a direct comparison to tolfH; additionally, mefH

exhibitstoxicity incancercell lines invitro [51,52]. DiflHwas investigatedas it isa

salicylate which is the same chemical class that aspH belongs

to(Fig.1.2). Additionally, reports suggest that diflH exhibits gastroprotective

properties[53,54]. TheBiNSAIDs,Bi(tolf)3,Bi(mef)3andBi(difl)3,weresynthesised,

purifiedandcharacterisedbyDrIshKumar(SchoolofChemistry,MonashUniversity)

and Dr Amita Pathak (School of Chemistry, Monash University) as previously

published by Andrews et al [8]. The NSAIDs, diflH (≥98%), mefH (≥98%), and

tolfH(≥98%), were purchased from Sigma-Aldrich(StLouis, USA). In addition,

bismuth salicylate was chosen to represent the marketed Bi drug, BSS. Bismuth

chloride(BiCl3)waschosentorepresentasimpleBicompoundwithnoexpectedanti-

canceractivityor toxicitystemming fromtheanion(Cl−). Cisplatinwasemployedto

provide a comparison of the BiNSAIDs to a standard chemotherapeutic drug. The

chemicals,BSS (99.9%),BiCl3(≥98%)andcisplatin (99.999%)werepurchased from

Sigma-Aldrich(StLouis,USA).

MilliQ™(18.2MΩ/cmresistivity,Millipore)waterwasusedtoprepareallaqueous

solutions.DMSO(99.8%)wasobtainedfromThermofisherScientific(Waltham,USA).

The chemicals, acetylsalicylic acid (aspH, ≥99.0%), deuterated DMSO (DMSO-d6,

99.9%),D-(+)-glucose(≥99.5%),4-(2-hydroxyethyl)-piperazine-1-ethanesulfonicacid

(HEPES, ≥99.5%), MTT (98%), and trypan blue (0.5% in phosphate buffered saline

(PBS)solution)werepurchasedfromSigma-Aldrich(StLouis,USA).Absoluteethanol

was sourced from Ajax Finechem (Seven Hills, Australia). PBS tablets (1tablet per

100mL of MilliQ water contains 0.8 g sodium chloride; 0.02 g potassium chloride;

0.115gdi-sodiumhydrogenphosphate;0.02gpotassiumdihydrogenphosphate)were

obtained from Oxoid (Basingstoke Hampshire, England). Annexin V binding buffer

andAnnVwereacquiredfromBioLegend(SanDiego,USA),and7AADwaspurchased

fromEnzoLifeSciences(PlymouthMeeting,USA).Arachidonicacid(0.1Minethanol,

Item No. 460103), potassium hydroxide (0.1M, Item No. 460105), and the


70

Prostaglandin E2 EIA kit – Monoclonal kits (Item No. 514010) were obtained from

CaymanChemical(AnnArbor,USA).

The materials used in the cell culture assays including, Roswell Park

Memorial Institute (RPMI)-1640 medium, penicillin/streptomycin (10,000IU/mL,

10,000μg/mL), trypsin (2.5%w/v in PBS) and GlutaMAX™ (200 mM L-alanyl-L-

glutaminedipeptidein0.85%NaCl)wereacquiredfromLifeTechnologies(NewYork,

USA). Australian certified fetal bovine serum was purchased from Thermofisher

Scientific (Waltham, USA) and was heat-inactivated before use (60 °C, 1 h).

CELLSTAR® 96-well cell culture plates, CELLSTAR® 60-mm culture dishes, and

CELLSTAR® filter cap cell culture flasks (75cm2) were sourced from Greiner

Bio-One(Kremsmünster,Austria).

2.2.2 Stabilitystudies

The stability ofBiNSAIDs inRPMI-1640 (referred toherein as cellmedium)was

assessedbycomparisonoftheNMRspectrawiththosefortheuncomplexedNSAIDsto

determine whether the NSAID dissociated from Bi. These studies were initially

performed by dissolving theNSAIDs or BiNSAIDs in accordancewith the procedure

and concentrations used for the cell assays (i.e.<100µM, 2% d6-DMSO/98% D2O);

however,onlyveryweak1HNMRsignalsoftheNSAIDsweredetected.Subsequently,

theconcentrationsoftheNSAIDsorBiNSAIDswereincreasedsuchthat10–40mgwas

dissolved in d6-DMSO (200 µL) and the resultant solution was added to

RPMI-1640(1.8mL). Theresultantsuspensionswerevortexedand incubated(37°C,

24 h). Following incubation, the suspended compounds were collected by

centrifugation(300g,5min)anddriedunderanitrogenatmosphere.The1Hand13C

NMR spectra of the BiNSAIDs, NSAIDs and the redissolved products resulting from

incubation in medium were recorded using a 500 MHz Varian spectrometer in

d6-DMSO at 25°C, referenced to DMSO (δH2.49, δC39.5). NMRassignments were

confirmedwith 2D-NMR, specifically homonuclear correlation spectroscopy (COSY),

heteronuclear single quantum coherence spectroscopy (HSQC), and heteronuclear

multiple-bondcorrelationspectroscopy(HMBC).

2.2.3 Cellculture

The human ileocecal colon cancer cell line, HCT-8, was chosen to represent a

humanbowelcancercelllineandsinceanumberofpublicationsreportthatCOX-2is


71

expressed in the HCT-8 cell line [55-57]. The cell line was kindly provided by

Associate Professor Ronald Sluyter (School of Biological Sciences, University of

Wollongong), and was originally sourced from the European Collection of

AuthenticatedCellCultures(ECACC).Allexperimentalworkinvolvingthecultureand

treatmentofHCT-8cells,andthesubsequenthandlingandpreparationofHCT-8cell

sampleswasperformedundersterileconditionsinabiologicalsafetycabinet(ClassII,

Email Westinghouse Pty Ltd). HCT-8 cells were cultured in growth medium that

consisted of RPMI-1640 medium supplemented with 10% fetal bovine serum,

GlutaMAX™(2mM),penicillinandstreptomycin (finalconcentration200 IU/mLand

200µg/mL,respectively).Cellswereincubatedina5%CO2atmosphereat37°Cand

passagedevery3–4days(whenconfluent)afterharvestingwithtrypsin(0.25%inPBS

solution).

2.2.4 Invitrocytotoxicitytowardscancercells:MTTassay

The toxicities of Bi(tolf)3, Bi(mef)3, Bi(difl)3, tolfH,mefH, diflH, BSS and cisplatin

were assessedusing adaptationsof theMTTassaydescribedbyMosmann [26], and

Carmichael et al [29]. Briefly, cells were seeded in 96-well plates at a density of

4×104cellsperwellingrowthmedium(100μL)andwereincubatedfor24htoallow

cell adhesion to occur. Stock solutions (typically 50µM to 100mM)of theNSAIDs,

BiNSAIDs, BiCl3 and BSS, were prepared in DMSO. These solutions (20 μL) were

pipetted into RPMI-1640 (980 μL), vortexed, and the resultant solutions (100 μL

treatment solution, final concentration of drug typically 1 to 2000 µM)were added

into the wells of the 96-well plate. Control cells were incubated with the vehicle

(DMSO(2μL) and RPMI-1640 (98 μL)) containing no test compounds. In order to

avoidanypotentialligandexchangereactionsbetweenDMSOandthelabileCl−groups

resulting in the deactivation of the drug in the cell medium [58], cisplatin was

prepared inRPMI-1640(typicalstockconcentration2−4mM)andseriallydiluted in

RPMI-1640. The 96-well plate was incubated (37°C) for 24h, after which the

treatment mediumwas removed and the cells were washed twice with sterile PBS

solution.FreshRPMI-1640(100μL)wasaddedtoeachwell,followedbytheaddition

ofMTTsolution(50μL,2mg/mLinPBS).The96-wellplatewasincubated(37°C)for

4h,afterwhichthemediumwasremovedandDMSO(200μL)wasaddedtoeachwell.

The96-wellplatewasagitatedfor25sandtheabsorbanceofeachwellwasrecorded

at 570 nm and 630nm using a POLARstar Omega microplate reader with Omega


72

software (Version1.1.2,BMGLabtekGmbH,Ortenberg,Germany). Theresultswere

expressedaspercentMTTconversion,whichisproportionaltothenumberofhealthy

cells(Equation1):

MTT conversion % = !!"#!!!"# !"#$!#% !"##$!!"#!!!"# !"#$%&#%' !"##$

× 100(1)

Thecalculatedcellviabilitywasplottedagainstthetreatmentconcentrationsusing

Microsoft EXCEL™ (Version 12 for Windows 2007, Microsoft, Redmond, USA) to

determine the IC50 value of each compound. The MTT assays were performed in

triplicateforeachcompound.

2.2.5 Celldeathassays

2.2.5.1 Morphologicalanalysisbyphasecontrastmicroscopy

HCT-8 cells were plated in 60-mm cell culture dishes (2 × 106 cells, 5mL) in

growthmediumandmaintainedat37°C,5%CO2overnight. Thecellswerewashed

twicewith PBS solution and the treatment solutions (3mL)were then added. The

treatmentconcentrationsweretheapproximateIC50dosesofeachBiNSAIDdissolved

intreatmentmedium(2%DMSOv/v)asdeterminedbyMTTassays(Section2.3.2),or

the equimolar concentration of the corresponding free NSAID, as denoted in the

figuresand text. Aminimumof three fieldsofviewwereexamined foreachsample

usingaMoticBinocularAE20Invertedlightmicroscopeequippedwitha20×objective

lens. Imageswere recorded using aMoticam 2000 2.0MP Live Resolution camera

withMoticImagesPlusVersion2.0forWindows(MoticChinaGroupCo.,Ltd.,Xiamen,

China).

2.2.5.2 Measurementofcelldeathbycytofluorometricassay

HCT-8 cells were plated in 60-mm cell culture dishes (1 × 106 cells, 5mL) in

growthmediumandmaintainedat37°C,5%CO2overnight. Thecellswerewashed

twice with PBS solution and treatment solutions were added (3 mL RPMI-1640,

2%DMSO v/v). Cellswere treatedwith the approximate IC50 doses of theBiNSAID

(Bi(tolf)3, Bi(mef)3 or Bi(difl)3) as determined by MTT assays (Section 2.3.2), or

equimolar concentrations of the corresponding free NSAID (tolfH, mefH or diflH).

Following incubation(37°C,24h), the treatmentsolutionsandPBSsolutionwashes

were collected and the remaining cells were detached using trypsin (0.25% in PBS

solution). Thedetachedcellswerecombinedwiththecellsinthetreatmentsolution

washings and the resultant suspension was centrifuged (300g, 5min) to ensure


73

collection of all cells (live and dead). Cell death was determined by quantitative

measurement using flow cytometry. The collected HCT-8 cells were washed twice

withHEPES-bufferedsalinesolution(145mMNaCl,5mMKCl,10mMHEPES,pH7.4,

1mL),followedbyAnnexinVbindingmedium(1mL).Thecellswereincubatedwith

AnnV (5µL) and 7AAD (1mg/mL in PBS solution, 1µL) at room temperature

protected from light for15min. Following staining,AnnVbindingmedium (400µL)

wasaddedtoeachtubeandevents(20,000persample)werecollectedusingaLSRII

flow cytometer (BD Biosciences, San Diego, CA) using an excitation wavelength of

488nm and emission collectedwith 515/20 and 695/40 bandpass filters forAnnV

and7AAD,respectively.Themeanfluorescentintensity(MFI)ofAnnVand7AADwas

determined using FloJo software (Tree Star, Ashland, OR). Quadrantmarkers were

used to determine the percentage of AnnV‒/7AAD‒ (viable), AnnV+/7AAD‒ (early

apoptotic), AnnV+/7AAD+ (late apoptotic/late necrotic), and AnnV‒/7AAD+ (early

necrotic)cells.

2.2.6 COXinhibition:Prostaglandinassays

HCT-8cells(1×106cells,5mL)wereseededinto60-mmcellculturedishesand

were incubated for24h,afterwhichthegrowthmediumwasremovedandthecells

werewashedtwicewithPBSsolution. TheBiNSAID(Bi(difl)3,Bi(mef)3,orBi(tolf)3),

NSAID(diflH,mefHor tolfH),aspHorBSS treatments inDMSO(5µMto50,000µM)

were added to each dish as a solution in RPMI-1640 (2mL, 2% DMSO v/v; final

concentration of drug 0.1−1000 µM),while the control cellswere incubated (37°C)

with the vehicle (2mL RPMI-1640, 2% DMSO v/v) for 4h. AspH was chosen as a

positivecontrolasitisawell-establishedCOXinhibitor,andclinicalevidencesuggests

that aspHmay possess chemopreventive properties (as discussed in Section 1.2.2).

BSS was chosen to represent a marketed anti-ulcer Bi medication. Following the

incubation period, the treatment medium was removed and the cells were washed

twicewithPBS solution. Fresharachidonic acid solutionwaspreparedaccording to

the manufacturer’s instructions. Arachidonic acid (0.1 M in ethanol, 20 µL) was

diluted by the addition of potassium hydroxide (0.1M, 20 µL) and the resultant

solutionwasaddedtoRPMI-1640(960µL).FreshRPMI-1640medium(1980µL)and

arachidonic acid solution (20 µL; final concentration 20 µM)was added to the cells

and the 60-mm dishes were incubated (37°C) for 30min. The cell medium was

collectedandcentrifuged(11000g,30sec) toremoveanycellularmaterial,andthe


74

supernatantwasstoredat−80°Cuntiltheassaywasperformed.PGE2productionwas

assayed using a Prostaglandin E2 EIA Kit in accordance with the manufacturer’s

instructions. Briefly, the absorbance of each well was recorded at 420 nm using a

POLARstar Omega microplate reader with Omega software (Version 1.1.2,

BMGLabtek). The PGE2 concentration(pg/mL) of each samplewas calculated using

GraphPad Prism Version 5.04 forWindows (GraphPad Software, La Jolla, USA), and

PGE2 release was expressed as a percentage of control PGE2 production. The

sensitivityoftheassay,asreportedbythemanufacturer,was15.6pgPGE2/mL.

2.2.7 Statisticalanalysis

Statistical analysis of the results of the AnnV/7AAD assays was performed by

one-way ANOVA, followed by the Tukey-Kramer Multiple Comparison Test, using

GraphPad Prism, Version 5.04 for Windows (GraphPad Software, La Jolla, USA).

TheresultswereanalysedasvehicleversusBiNSAIDversusthecorrespondingNSAID,

inordertodrawastatisticalcomparisonatthesameNSAIDconcentration.Avalueof

P≤0.05wasconsideredstatisticallysignificant.


75

2.3 Results2.3.1 StabilityofBiNSAIDsincellmedium

Featuresofthe1HNMRspectra(aromaticregiononly,presentedforclarity;Figure

2.2a–d)oftolfHandBi(tolf)3asobtainedinDMSO-d6priortoandafterincubationin

cellmediumaresummarisedinTable2.1.Itisapparentthatthereisnoprotonsignal

corresponding to the –COOH in the Bi(tolf)3 spectrum, which is consistent with

previous studies [8], and corresponds to the expected absence of the proton

subsequent to coordination to Bi. As shown in Table 2.1, the 13C NMR spectra of

Bi(tolf)3 also confirm the complexation to Bi, with downfield shifts observed from

170.1ppmto174.0ppmfortheresonancecorrespondingtothecarboxylatecarbon,

C14(C–OOBi),incomparisonwithuncomplexedtolfH;additionally,asubstantialshift

in the resonance of the quaternary carbon C1 (C–COOBi) was observed from

112.3ppmto118.6ppm.Themajorityofthesignalshiftsrelatingtotheintroduction

of thed6-DMSOsolutionof tolfHandBi(tolf)3 to cellmediumwereobserved for the

aromaticprotons,asshowninFigure2.2bandd.Smalldownfieldshiftsinthe1HNMR

signalsof tolfH incubated incellmediumoccurred forH5,andH8whichoverlapped

with the signal of H4 (Fig. 2.2b), indicating that there were some changes in the

chemicalenvironmentsurroundingtheseatomsfollowingincubationoftheNSAIDin

cellmediumcomparedtotheoriginalNSAID(Fig.2.2a).Inaddition,theprotonsignal

at 13.11 ppm related to –COOH disappeared and there was broadening and a

downfieldshiftofthesignalassociatedwith–NHinthe1HNMRspectraobtainedfrom

the precipitated tolfH (and Bi(tolf)3). These observations are attributed to the

presenceofasmallamountofwaterinthesampleandsubsequentexchangewithH2O

protons.Following incubation in cell medium, the 1H NMR spectrum of

Bi(tolf)3(Fig.2.2d) showed some differences from the initial spectrum of

Bi(tolf)3(Fig.2.2c), most substantially for the 1H signals of H4, H8, H9 and H10.

However,the13CNMRspectrum(Table2.1)obtainedforBi(tolf)3after incubationin

cellmediumshowedminimalshift in theC14(C–OOBi)signal (174.0 to173.8ppm).

Additionally, there was minimal shift in the C1 signal (Table 2.1) suggesting that

following incubation of the complex in cell medium, the interaction between the

carboxylategroupandBiremainsintact.


76

Figure 2.2: 1H NMR spectra of the aromatic region (6.6–8.0 ppm) of: (a) tolfH; (b) tolfHcollectedafterincubationincellmedium(24h,37°C);(c)Bi(tolf)3and,(d)Bi(tolf)3collectedafterincubationincellmedium(24h,37°C).Allsampleswereanalysedind6-DMSOat25°C(referencedtoDMSO,2.49ppm).


77

Table2.1:1 H

and

13CNMRdata

a,b,c,d

,e oftolfHand

Bi(tolf)

3ind 6-DMSO

,beforeandaftersuspensionincellm

edium(2

4h).

Cellmedium(2

4h)

13C11

8.8

131.9

117.4

132.5

113.8

145.9

141.5

123.2

127.3

118.9

134.3

127.9

14.4

173.8

a Chemicalshift(δ)inppm

;b M

ultip

licity

isgivenass,singlet;d,doublet;t,triplet;q

,quartet;m

,multip

let;c C

ouplingconstantsareinparentheses3 J(1H,1 H

);

d bris“broadsignal”;e n.ois“n

otobserved”.

DMSO

1 H

7.86

d(7

.5)

6.72

t(7)

7.23

m(n

.o)

6.93

d(7

.5)

7.23

d(8

)7.12

t(8)

7.07

m(8

)

2.14

s

n.o

10.20brs

13C

118.6

131.9

117.5

133.4

114.0

146.0

141.0

123.7

127.3

119.6

134.3

128.1

14.2

174.0

Bi(tolf)3

None

DMSO

1 H

7.84

d(8

.5)

6.76

t(6.75)

7.30

t(7)

6.92

d(8

)

7.21

d(7

)7.13

m(7

.5)

7.13

m(7

)

2.07

s

n.o

9.64

s

Cellm

edium(2

4h)

13C

113.0

131.8

117.4

134.0

113.7

147.4

140.7

124.7

127.6

121.7

134.4

129.6

14.7

170.2

DMSO

1 H

7.90

d(7

.5)

6.75

t(7.5)

7.33

m(8

)6.84

d(8

)

7.30

m(7

)7.20

m(6

.75)

7.20

m(6

.75)

2.24

s

n.o

9.90

brs

13C

112.3

131.8

117.3

134.2

113.6

147.6

140.5

124.9

127.5

122.1

134.5

129.8

14.7

170.1

tolfH

None

DMSO

1 H

7.90

d(8

)6.75

t(7.25)

7.33

t(7.75)

6.80

d(8

) 7.28

d(6

.75)

7.21

m(7

.5)

7.21

m(7

.5)

2.23

s

13.11s

9.56

s

Compoun

dTreatm

ent

Solvent

Atom

1 2 3 4 5 6 7 8 9 10

11

12

13

14

OH

NH


78

The 1H NMR spectra of mefH and Bi(mef)3 (aromatic region only presented for

clarity, Figure 2.3a–d), as obtained in d6-DMSO prior to and after incubation in cell

medium, are summarised in Table 2.2. Prior to incubation in cell medium it is

apparentthatthereisnosignalcorrespondingtotheprotonof–COOHintheBi(mef)3

spectrum,whichisconsistentwiththeabsenceofcoordinationoftheNSAIDtoBi[8].

AsshowninTable2.2,the13CNMRspectraofBi(mef)3alsoconfirmthecomplexation

toBi,withdownfieldshiftsobservedfrom170.2ppmto175.5ppmfortheresonance

corresponding to the carboxylate carbon, C15 (C–OOBi), in comparison with

uncomplexedmefH. The1HNMRsignalsofBi(mef)3(Table2.2)wereshiftedupfield

in comparison to freemefH (with the exception of H3 and NH). Themost obvious

differencebetweenthe1HNMRspectraofmefHandBi(mef)3 isthechemicalshiftof

H5 (Fig. 2.3 a and c). After incubation in cell medium, there appeared to be little

changetothe1HNMRspectrumofmefH(Fig.2.3b,Table2.3).Comparatively,Figure

2.3dshowsthattheH5signaloverlapswiththeH3signalafterincubationofBi(mef)3

incellmediumwhichwasaccompaniedbyadditionaldownfieldshiftsinthemajority

of theother 1HNMRsignals (Table2.3). Furthermore, the 13CNMRsignalsrevealed

thataftertheincubationofBi(mef)3incellmedium,theC–OOBisignalshiftedupfield

to 170.8 ppm (Table 2.3).Overall, these results suggest that following incubation of

Bi(mef)3incellmediumthereisliberationofmeffromtheBicomplex.

The 1H NMR spectra of diflH and Bi(difl)3 (aromatic region only, presented for

clarity; Fig. 2.4a–d), as obtained in d6-DMSO prior to and after incubation in cell

medium, are summarised in Table 2.3. Interestingly, reference to the spectrum of

diflHpriortoincubationincellmediumindicatestheabsenceoftheprotonsignalof

the –COOH (of C13) and –OH (of C10); however, solid IR analysis (data not shown)

confirmedthepresenceof–OHinthediflHstructure. Additionally,theC10,C11and

C13 (–COOH) signals of Bi(difl)3 were not observed (Table 2.3, Fig. 2.4a). This

phenomenon ismost likely a result of the slower tumblingof a largermolecule and

hencethebroadeningofsomeofthe13Csignals,therebydecreasingtheirdetectability;

thisisfurtherevidencedbythebroadeningoftheH8,H9andH12signalsobservedin

the 1H NMR spectrum of Bi(difl)3 (Fig. 2.4c). The 24-h incubation of diflH in cell

mediumresultedinupfieldshiftsoftheprotonsignalsassociatedwiththephenylring

containingthe–COOHmoiety(Fig.2.4d).FollowingtheincubationofBi(difl)3


79

Figure 2.3: 1H NMR spectra of the aromatic region (6.6–8.0 ppm) of: (a) mefH; (b) mefHcollectedafterincubationincellmedium(24h,37°C);(c)Bi(mef)3,and(d)Bi(mef)3collectedafterincubationincellmedium(24h,37°C).Allsampleswereanalysedind6-DMSOat25°C(referencedtoDMSO,2.49ppm).


80

Table2.2:1 H

and

13CNMRdata

a,b,c,d

,e ofm

efHand

Bi(mef) 3ind

6-DMSO


edium(2

4h).

Cellmedium(2

4h)

13C11

2.7

131.7

116.2

133.8

113.1

148.5

131.0

121.7

126.1

126.0

137.8

138.6

20.2

13.7

170.8


;b M

ultip

licity


,quartet;m

,multip

let;c C


);


otobserved”.

DMSO

1 H

7.87

d(8

)6.69

m(7

.5)

7.27

t(7)

6.65

m(7

)

6.99

d(5

)7.10

m(n

.o)

7.10

m(n

.o)

2.26

s2.08

s

n.o

9.65

brs

13C

n.o

131.9

116.3

133.4

113.2

147.5

130.0

120.4

125.7

125.4

137.6

138.8

20.1

13.2

175.5

Bi(m

ef) 3

None

DMSO

1 H

7.83

d(8

)6.68

t(7.25)

7.24

t(7.5)

6.78

d(8

.5)

6.90

d(5

.5)

7.03

m(7

.25)

7.00

m(7

.25)

2.13

s1.91

s

n.o

9.45

s

Cellm

edium(2

4h)

13C

111.6

131.8

116.3

134.1

113.1

148.8

131.2

122.1

126.3

126.1

137.9

138.4

20.3

13.7

170.4

DMSO

1 H

7.88

d(7

.5)

6.67

m(8

.5)

7.26

t(7.5)

6.64

m(7

)

6.98

d(6

)7.08

m(7

.25)

7.06

m(7

.25)

2.24

s2.07

s

n.o

9.50

brs

13C

111.2

131.7

116.2

134.2

113.1

148.7

131.2

122.2

126.4

126.0

137.9

138.3

20.2

13.6

170.2

mefH

None

DMSO

1 H

7.87

d(7

)6.67

m(7

)7.29

t(7.25)

6.69

m(8

) 7.06

d(5

)7.11

m(6

.75)

7.09

m(6

.75)

2.27

s2.08

s

12.9s

9.43

s

Compoun

dTreatm

ent

Solvent

Atom

1 2 3 4 5 6 7 8 9 10

11

12

13

14

15

OH

NH


81

incellmedium,apronouncedupfieldshiftintheH8,H9andH121HNMRsignals

was observed (Fig.2.4d), in comparison to freediflH (Fig.2.4b). Additionally, the

carbonsignals forC10,C11andthe–COOHwereobserved(Table2.4)suggestingan

alterationinstructure.However,theC11signaloftheBi(difl)3differssubstantiallyto

thatofdiflHafterincubation(120.4ppmand115.9ppm)whichsuggeststhatdiflmay

remaincoordinatedtoBi.

Interestingly, weak yet observable signals were detected in the 1H NMR from

anothermolecule(s) present in the Bi(difl)3 sample after incubation in cellmedium

which were not detected in any other sample studied. Complete characterisation

couldnotbeachievedduetothe lowconcentrationof themolecule(s). Nonetheless,

the appearanceofweak 1HNMRsignals of theunknownmolecule(s) suggested that

onlymoleculeswithconcentrationcomparabletotheconcentrationofBi(difl)3inthe

cellmediumwouldbedetected,inparticularD-glucoseorHEPES.Referencetothe1H

NMRspectraofD-glucoseorHEPES(alone)ind6-DMSO,suggestedthatthepotential

identity of the molecule(s) in the incubated Bi(difl)3 sample consisted of

D-glucose(Fig.2.5). Overall, the changesobserved in the 1Hand 13CNMRsignalsof

theincubatedBi(difl)3complexcomparedtotheoriginalcompoundmaysuggestthat

thecoordinationofBitothediflligandsisalteredfollowingincubationincellmedium

due,possibly,tosubsequentinteractionswithD-glucose.

AswellastheDMSO-solublespeciesdescribedabove(Fig.2.2–2.5,Table2.1–2.3),

it was noted that a DMSO-insoluble species was present upon preparation of the

samples for NMR analysis (after collection and drying following incubation in cell

medium).Theformationofthewhiteprecipitatewasnotobserveduponincubationof

the free NSAID in cell medium and consequently implied that at least some of the

BiNSAID must be interacting with other species present in the solution. High

concentrations of anions such as phosphate, chloride, sulphate and carbonate are

presentintheinorganicsaltsfoundinRPMI-1640.GiventhatBi(III)hasaparticularly

high affinity for phosphate ions, it is envisaged that the precipitation could be

insoluble Bi(III) phosphate. Identification of the species was attempted using IR

spectroscopy;however,thedominatingsignalsinthespectrumoriginatedfromDMSO

thatproveddifficulttoremoveduetoitsnon-volatilenature.Attemptstorepeatedly

wash the precipitate with MilliQ water were performed; however, due to the

lowquantityofthesolidproduced,isolationofasufficientquantityofdried


82

Figure 2.4: 1H NMR spectra of the aromatic region (6.6–8.0 ppm) of: (a) diflH; (c) diflHcollectedafterincubationincellmedium(24h,37°C);(d)Bi(difl)3,and(f)Bi(difl)3collectedafterincubationincellmedium(24h,37°C).Allsampleswereanalysedind6-DMSOat25°C(referencedtoDMSO,2.49ppm).


83

Table2.3:1 H

and

13CNMRdata

a,b,c,d

,e ofdiflHand

Bi(difl)

3ind 6-DMSO


edium(2

4h).

Cellmedium(2

4h)

13C13

1.2

111.8

161.7

104.3

159.8

126.2

121.6

131.7

116.2

163.2

120.4

130.5

171.0


;b M

ultip

licity


,quartet;m

,multip

let;c C


);


otobserved”;

f 13 Cdoubletdueto13C-

19Fcoup

ling.

DMSO

1 H

7.47

q(8

)7.10

t(7.75)

7.25

t(10)

7.29

d(8

)6.70

d(8

.5)

7.82

s n.o

13C

131.3

111.9

161.2

104.3

160.0

124.2

125.0

134.4

117.2

n.o

n.o

130.4

n.o

Bi(difl) 3

None

DMSO

1 H

7.46

m(7

)7.10

t(8)

7.27

t(10)

7.51

m(1

2)

6.93

brs

7.85

brs

n.o

Cellm

edium(2

4h)

13C

131.7

112.3

161.4

104.6

159.4

124.4

124.5

134.8

117.5

161.5

115.9

130.6

172.0

DMSO

1 H

7.48

m(8

.5)

7.10

td(8

.5)

7.23

t(10)

7.52

m(7

.25)

6.94

d(9

)

7.87

s n.o

13C

131.5

112.0

161.5

104.4

159.3

123.7

125.1

135.8

117.6

160.7

113.2

130.3

171.6

diflH

None

DMSO

1 H

7.54

q(8

.2)

7.13

t(8.5)

7.29

t(10.25

)

7.64

d(8

.5)

7.04

d(8

)

7.90

s n.o

Compoun

dTreatm

ent

Solvent

Atom

1 2 3f 4 5f 6 7 8 9 10

11

12

13

OH


84

Figure 2.5: 1H NMR spectra of: (a) unknown molecule present in the sample containingBi(difl)3 collected after incubation in cell medium (24 h, 37 °C) (l indicates signalscorresponding to D-glucose, × indicates unassigned signals); and (b) D-glucose. All sampleswereanalysedind6-DMSO,25°C,referencedtoDMSO(2.49ppm).

compoundtoproducegoodquality IRspectrawasnotpossible. Whetherphosphate

(oranyotherion)hasaneffectontheBiNSAIDsatlowconcentrationsincellmedium

(i.e.biologicallyrelevantconcentrations)willneedtobeconfirmedinfurtherstudies.

2.3.2 Invitrocytotoxicitytowardscancercells

To evaluate the effects of theBiNSAIDs andNSAIDs on cell viability,HCT-8 cells

were exposed to the compounds for 4 h or 24h, followingwhich the conversion of

MTT by functioning mitochondria was measured. As DMSO was necessary to

solubilisetheBiNSAIDsandNSAIDs, theMTTassaywasperformedtodeterminethe

tolerance of HCT-8 cells to increasing concentrations of DMSO in the cell medium.

Arepresentative concentration-response curve is shown inAppendix 1.1. Following

24-hexposuretoDMSO,theviabilityoftheHCT-8cellsdecreasedinaconcentration-

dependent manner. At DMSO concentration of 1% v/v approximately 95% of


85

mitochondrialactivitywasmaintainedcomparedtothecontrolcells(treatedwithcell

medium in the absenceofDMSO). Increasing the concentrationofDMSO to2%v/v

resulted in a decrease in viability to approximately 78% of mitochondrial activity

comparedtothecontrolcells; furtherincreasingtheconcentrationofDMSOresulted

in a further reduction in mitochondrial activity. Despite a minor impact on

mitochondrial activity, a concentration of 2%v/v DMSO was chosen for compound

deliveryinordertomaximizethedissolutionofthehighlywater-insolubleBiNSAIDs

andNSAIDsinthecellmedium.

The concentration-response curves shown in Appendix 1.2 and Figure 2.6 are

graphedfortheNSAIDconcentration(acknowledgingthatthereare3moleofNSAID

present in each BiNSAID complex) following 4-h or 24-h treatment, respectively.

AsshowninAppendix1.2,noappreciablelossofmitochondrialactivitywasobserved

inHCT-8cellsfollowing4-htreatmentwithanyoftheBiNSAIDsorNSAIDsuptothe

highest tested concentrations (300 µM or 1000µM). However, following 24-h

exposure to BiNSAIDs the viability of HCT-8 cells decreased in a concentration-

dependentmanner (Figure 2.6). The IC50 values determined after 24-h exposure to

the BiNSAID complexes and reference compounds (cisplatin, BiCl3, and BSS) are

summarised in Table 2.4. Notably, the compounds, Bi(tolf)3, Bi(mef)3, and tolfH

(Table2.4,Fig.2.6),weremoretoxictoHCT-8cellsthantheanti-cancerdrug,cisplatin

(Table2.4,Appendix1.3).Bismuthsalicylate,whichwasusedtomimictheanti-ulcer

drug,BSS,showedlittleeffectontheviabilityoftheHCT-8cellsupto2000µM(Table

2.4, Appendix 1.3). However, the Bi compound, BiCl3, showed moderate toxicity

(IC50=383μM)towardstheHCT-8cells (Table2.4,Appendix1.3). Followingclose

inspection of the results obtained from the BiNSAIDs it is apparent that their

IC50values are lower than the respective free NSAIDs (Table 2.4), ranging from

16–81μM(BiNSAIDs) to 37–403μM (NSAIDs). Importantly, the toxicity of the

BiNSAIDs increased in the order: Bi(difl)3 < Bi(mef)3 < Bi(tolf)3 paralleling the

responseexhibitedbythefreeNSAIDs(diflH<mefH<tolfH).Figure2.6demonstrates

that the toxicity profiles of theBiNSAIDs, Bi(tolf)3 andBi(mef)3,were similar to the

respective freeNSAIDs(tolfHandmefH)whereby theBiNSAIDswereapproximately

three timesmore toxic than the respective freeNSAIDs (representativeof themolar

ratio of NSAID contained in the BiNSAID complexes). The results of the assay for

Bi(difl)3differ,however,wherebythebindingofdifltobismuthenhancedthetoxicity

oftheNSAID(Fig.2.6).


86

Figure 2.6: Concentration-response curves obtained from the MTT assays of HCT-8 cellstreated(24h)withtolfH,Bi(tolf)3,mefH,Bi(mef)3,diflH,orBi(difl)3.Errorbarsrepresentthestandard deviation from the mean (n = 6). One representative of three independentexperimentsisshown.

Table2.4:IC50valuesdeterminedbyMTTtoxicityassaysfollowingthe24-htreatmentofHCT-8coloncancercellswiththespecifiedcompounds.

IC50(µM)aLigand(LH)Complex[Bi(L)3]

IC50ratioLH:Bi(L)3

tolf 37±2 16±1 2.3mef 118±2 44±5 2.7difl 403±6 81±6 5.0cisplatin 50±5 BiCl3 383±38 BSS >2000

aIC50valuewasdeterminedusingatleastthreeindependentMTTassays.


2.3.3.1 Phasecontrastmicroscopy

In order to determine whether themode of cell death of the BiNSAIDs was the

sameastherespectiveNSAIDs,phasecontrastmicroscopywasinitiallyperformedto

assessthemorphologicalchangesoftheHCT-8cells. AsobservedinFigure2.7a,the

vehicle-treatedcellsmaintaintheirnormalfibroblasticintactshapeandclosecontact

withneighboringcells. Following24-htreatmentwiththeapproximateIC50dosesof

theBiNSAIDs,ortheequivalentmolarconcentrationoftheNSAIDs(asdeterminedby


87

MTTassays,Section2.3.2),anincreasedproportionofHCT-8cellshaddetachedfrom

the dish surfaces and were observed floating in the medium of all drug-treated

samples.Thefloatingdebriscouldbeidentifiedasintactcellsandapoptoticbodiesas

indicated by the variation in size of thematerial. As observed in Figure 2.7b–g, an

increasedproportionofBiNSAID-andNSAID-treatedcellpopulationsappearedtobe

rounded and smaller in size, in contrast to the cells treated with vehicle alone

(Fig.2.7a).Morphologicalchangesofthecellmembraneareconsistentwithapoptosis

and include the formation of blebs, whichwere clearly evident in HCT-8 cells after

treatmentwithBiNSAIDsorNSAIDs(Fig.2.7b–g).Thesewereseldomobservedinthe

vehicle-treated dishes (Fig.2.7a). The diflH-treated (Fig.2.7g) cells showed less

evidenceofcellmembraneblebbingcomparedtotheothercompoundstested,which

is reflective of the lower toxicity of diflH towards theHCT-8 cells determined using

MTT assays. Furthermore, a proportion of the diflH-treated cell population,

maintainedtheirfibroblasticshape(Fig.2.7g),anexpectedobservationgiventhatthe

diflH-treated cells were treated at a concentration lower than the IC50value

determinedbyMTTassay(225μMand401μM,respectively).

2.3.3.2 AnnV/7AADassays

Flow cytometric analysis employing the use of the fluorescent probes AnnV and

7AADwasusedtodeterminewhetherBiNSAIDsandNSAIDsinducethesamemodeof

cell death in HCT-8 cells. As demonstrated in Figure 2.8a, 24-h treatment with

Bi(tolf)3, Bi(mef)3, Bi(difl)3 or the respective free NSAIDs resulted in a substantial

increase in the proportion of cells in the AnnV+/7AAD− population (bottom right

quadrant), and AnnV+/7AAD+ population (top right quadrant), but no appreciable

increaseinAnnV-/7AAD+population(topleftquadrant).Asanappreciableincreasein

the population of AnnV+/7AAD− cells was observed, with little increase in the

AnnV−/7AAD+population,itappearedthatcellswereproceedingthroughanapoptotic

pathwayand thusAnnV+/7AAD+ couldbedefinedas lateapoptotic (rather than late

apoptotic/necrotic). The overall proportions of dying cells (encompassing early

apoptotic,lateapoptoticandearlynecroticcellpopulations)weresignificantlyhigher

(P ≤ 0.001) in all BiNSAID- or NSAID-treated cells compared to the control

cells(Fig.2.8b). When the total population of dying cells induced by each BiNSAID

wascomparedtothatfortherespectivefreeNSAID,onlyBi(difl)3-treatedcellsshowed


88

Figure2.7:Morphologicaleffectsfollowing24-htreatmentwith:(a)vehicle(DMSO2%v/v);(b)Bi(tolf)3(15μM);(c)tolfH(45μM);(d)Bi(mef)3(40μM);(e)mefH(120μM);(f)Bi(difl)3(75μM);and(g)diflH(225μM).Imageswereviewedusingthe20×objectivelensonalightmicroscopeequippedwithaMoticam2000,2.0MPLiveResolutioncamerasystem.


89

asignificantlyhigherproportionofcelldeath(P≤0.01)comparedtothediflH-treated

cells.Following24-htreatmentwiththevehicle,themajorityoftheHCT-8cellswere

viable(Fig.2.8bandTable2.5).Incomparison,treatmentwithBi(tolf)3(15μM,24h)

resulted in significant increases in the populations of early apoptotic, and late

apoptotic cells (P ≤ 0.001 andP ≤ 0.05 respectively, Table 2.5). TolfH-treated cells

showed a significant increase in early apoptotic and late apoptotic populations

compared to vehicle-treated cells (P≤ 0.001 and P ≤ 0.01, respectively, Table 2.5).

While no statistical differencewas found between the proportion of early apoptotic

cellsinducedinBi(tolf)3-treatedcellscomparedtotolfH-treatedcells(Table2.5),tolfH

treatmentinducedasignificantlygreaternumberoflateapoptoticcells(P≤0.05).The

total percentage of dying cells following treatment with Bi(mef)3 and mefH were

similar(Fig.2.8b,Table2.5,P>0.05).Asignificantincreaseintheproportionofearly

apoptotic cells was observed in the Bi(mef)3- and mefH-treated cells compared to

vehicle-treatedcells(bothP≤0.01,Table2.5),althoughmefH-treatedcellsinduceda

higherproportionof early apoptotic cells compared toBi(mef)3 (P≤0.01). Notably,

therewas a 3-fold increase in thepercentage of late apoptotic cells in theBi(mef)3-

treated cell population compared to the mefH-treated cell population (P ≤ 0.01,

Table2.5). While both treatments increased the proportion of late apoptotic cells,

onlytheBi(mef)3-treatedcellpopulationwasfoundtobesignificantcomparedtothe

control (P≤ 0.001, Table 2.5). Following 24-h treatment with Bi(difl)3 or diflH,

significant increases in thepercentageof early apoptotic cells (bothP≤0.001)were

observed(Table2.5);however,treatmentwithBi(difl)3induceda2.4-foldincreasein

the proportion of late apoptotic cells (P ≤ 0.001 compared to the control; Fig.2.8b)

compared to treatment with diflH (P ≤ 0.001 compared to Bi(difl)3; Fig.2.8b).

Importantly, no significant increase in the proportion of early necrotic cells was

observedinanyoftheBiNSAID-orNSAID-treatedcellpopulations(Table2.5).Overall

theseresultsindicatethatBi(tolf)3,Bi(mef)3,Bi(difl)3,andtherespectivefreeNSAIDs,

inducecelldeathviaapoptosis.


90

Figure 2.8: Flow cytometric analysis of cell death in HCT-8 cells treated for 24 h withvehicle(RPMI-1640,2%v/vDMSO;Control),BiNSAIDs(Bi(tolf)3(15µM),Bi(mef)3(40µM)orBi(difl)3(75µM));ortherespectiveequimolarfreeNSAIDs.(a)Bivariateflowcytometricplotsof AnnV fluorescence versus 7AAD fluorescence. Bottom left quadrant represents viablecells(AnnV−/7AAD−);bottomrightquadrantrepresentsearlyapoptoticcells(AnnV+/7AAD−);top right quadrant represents late apoptotic/late necrotic cells (AnnV+/7AAD+); top leftquadrant represents early necrotic cells (AnnV−/7AAD+). Data are representative of n = 3replicates.(b)Percentageofearlyapoptoticandlateapoptoticcellsbasedonflowcytometryanalysis. Each segment represents themean ± s.d. (n = 3 replicates from one experiment).Statistical significance of the population of dying cells (encompassing early apoptotic, lateapoptotic and necrotic cell populations) compared to the control is indicated by***=P≤0.001. Statistical significance of the BiNSAID-treated samples compared to therespectiveNSAID-treatedsamplesisindicatedby‡=P≤0.001.Theviable(AnnV−/7AAD−)andearlynecrotic(AnnV−/7AAD+)cellpopulationshavebeenomittedfromthegraphforclarity.


91

Table2.5:CellpopulationsdeterminedbyflowcytometricanalysisofHCT-8cellstreatedfor 24 h with: vehicle (RPMI-1640, 2% v/v DMSO), BiNSAIDs (Bi(tolf)3 (15 µM),Bi(mef)3(40µM),orBi(difl)3(75µM)),ortherespectiveequimolarfreeNSAIDs.

Percentageofcells(%) Treatment

ViableAnnV–/7AAD–

EarlyapoptoticAnnV+/7AAD–

LateapoptoticAnnV+/7AAD+

EarlynecroticAnnV–/7AAD+

Vehicle 86.7±1.0 7.8±0.5 4.8±0.8 0.7±0.2Bi(tolf)3,15µM 46.4±12.2 41.0±6.9*** 12.3±5.8* 0.3±0.2tolfH,45µM 29.5±3.1 48.7±2.2*** 21.5±1.4**†† 0.3±0.2Bi(mef)3,40µM 40.9±3.8 39.6±4.6*** 18.8±0.8*** 0.8±0.4mefH,120µM 41.9±7.0 51.5±6.1***†† 6.15±1.4‡ 0.6±0.3Bi(difl)3,75µM 53.2±3.1 26.2±2.0*** 19.5±1.7*** 1.2±0.3diflH,225µM 63.3±2.8 27.5±1.5*** 8.20±2.4‡ 1.1±0.7

Eachvaluerepresentsmean±s.d.(n=3replicatesfromoneexperiment).Statisticalsignificancecomparedtothecontrolisindicatedby*=P≤0.05,**=P≤0.01,***=P≤0.001.Statistical significance of the BiNSAID compared to the corresponding free NSAID is indicated by†† =P≤0.05,or‡=P≤0.001.

2.3.4 TheeffectofBiNSAIDsonPGE2productionbyHCT-8cells

The production of PGE2 by HCT-8 cells was determined in order to establish

whetherBiNSAIDswereabletoinhibitarachidonicacidconversiontoPGE2byCOXat

a similar extent to the respective free NSAIDs. As shown in Figure 2.9, the

concentration-response curves are presented for the NSAID concentration

(recognisingthatthereare3moleofNSAIDpermoleofBiNSAID).Figure2.9ashows

that the Bi(tolf)3 and Bi(mef)3 complexes inhibited PGE2 production in a

concentration-dependentmanner,whichparalleled the inhibitiondisplayedby tolfH

and mefH alone, respectively. The concentration-response curve of Bi(difl)3 also

mirrored the inhibition of diflH (Fig. 2.9b). However, PGE2 production increased

comparative to control cells at treatment concentrations of 1–100 µM difl, but

decreasedathigherconcentrations(300–1000µMdifl). Althoughtheconcentration-

responsecurveobtainedforBi(difl)3mirroredtheeffectofdiflH,PGE2productionwas

less effectively inhibited at higher concentrations of Bi(difl)3 compared to

diflH(Fig.2.9b).Theconcentration-responsecurveofaspHexhibitedaconcentration-

dependentdecreaseinPGE2production(Fig.2.9b)similartotolfHandmefH.

InspectionoftheIC50values(50%reductionofPGE2productioncomparedtothe

control)determinedfromthePGE2assays(Table2.6)showedthattheCOXinhibition

oftheBiNSAIDs/NSAIDsinHCT-8cellsincreasedintheorder:difl<mef<tolf,which

paralleled the toxicity response determined by the MTT assays (Section2.3.2).


92

ThePGE2 production IC50 values for tolfH and Bi(tolf)3 were lower than that of

aspH(Table2.6), while mefH and Bi(mef)3 showed comparable inhibitory activity

toaspH(Table2.6). SubstantiallyhigherPGE2productionIC50valueswereobserved

fordiflHandBi(difl)3comparedtoaspH(Table2.6).TreatmentwithBSS(1–300µM)

caused no significant reduction in PGE2 production compared to control

cells(Appendix1.4).

Figure 2.9: Concentration-response curve obtained from the PGE2 assays of HCT-8 cellstreated (4 h) with: (a) tolfH, Bi(tolf)3, mefH, Bi(mef)3; and (b) diflH, Bi(difl)3, and aspH,followedbyarachidonicacid(20µM,30min).Resultsareexpressedasthemean±s.d.(n=3fromthreeindependentexperiments,withtheexceptionofaspH,n=2fromtwoindependentexperiments).


93

Table2.6: IC50valuesdeterminedbyPGE2assaysfollowing4-htreatmentofHCT-8cellswiththespecifiedcompounds.

Treatment

IC50(µM)aLigand(LH)Complex[Bi(L)3]

IC50ratioLH:Bi(L)3

tolf 1.3±0.7 0.5±0.3 2.6mef 23±16 10±8 2.3difl 340±180 203±32 1.7asp 13±4 BSS n.d.

aIC50valuewasdeterminedfromthreeindependentexperiments(withtheexceptionofaspH,whichwasdeterminedfromtwoindependentexperiments).n.d.=notdetermined.


94

2.4 Discussion2.4.1 NMRstabilitystudies

NMRspectroscopywasused to investigate thestructural stabilityofBiNSAIDs in

RPMI-1640 cellmedium to gain an understanding of the possible biologically active

species and interactions that give rise to the observed toxicity (examined in

Section2.3.2), cell uptake and localisation of the Bi (discussed in Chapter 3). The

sparingsolubilityofBiNSAIDsinaqueoussolutionsandanumberofcommonorganic

solvents (such as methanol, acetonitrile and chloroform) limited the analytical

techniqueswhichcouldbeusedforthestabilitystudies;however,thehighsolubility

of the compounds in DMSO meant that NMR spectroscopy was suitable.

Unfortunately, the low sensitivity of 1HNMR spectroscopy for the complexes at the

lowconcentrationsemployedincellassays(40μM)meantthatitwasnecessarytouse

higherconcentrationsoftheBiNSAIDs.Itshouldbenotedthatlittletonoprecipitate

wasobserved in the cellmediumat theBiNSAIDconcentrationsused for cell assays

andassuchtheNMRstudy(~600μMBi)representsthemostextremecasescenario

performedinordertoexplorewhatpotentialproductsmightform.

The NMR studies were limited to studying the chemical shifts of the 1H and13Csignals.Theoretically,the209Bisignal(chemicalshiftrangeof~5000ppm)could

beusedtostudytheenvironmentalsymmetryoftheBinuclei,andhenceanychanges

in the complexation of the NSAIDs. Bi has ideal properties for NMR studies due to

100% natural abundance of 209Bi and its high receptivity (0.144 relative to 1H, and

8.48×102relativeto13C)[59,60].Anattemptwasmadetostudythe209Binucleiin

thepresentstudy;however, thisprovedunsuccessfulduetodifficultiestuningto209Bi nuclear resonance frequency with the broadband probe of the NMR

spectrometerused for theNMRstudies (DrWilfordLie, personal communication).

In theabsenceof instrument limitations,arecentreview[60]commentedonthe

difficultnatureofrecording209Bispectraduetothebroadsignalthatitproduces,

andasaresulttherearefewBiNMRstudiesreportedintheliterature.

TheNMRresults(Section2.3.1)indicatedthatthestabilityoftheBiNSAIDsincell

mediummight be influenced directly by the structure of the NSAID ligands, since a

marked effect in the stability of the BiNSAIDs was observed by the substitution of

–Cl(tolfH)for–CH3(mefH). ThedifferenceinthestabilityofBi(tolf)3andBi(mef)3is

unlikely to be related to physicochemical properties of the free NSAIDs because

similar values are reported for both the pKa [61-63] and logP [64] values of tolfH


95

andmefH.ItisnotablethatthetolfHandmefHmoleculesdifferonlybythepresence

of an electron-withdrawing group (–Cl) and an electron-donating group (–CH3).

However,thesegroupsarelocatedontheoppositephenylringofmoleculetotheBi–

carboxylatebond;assuchtheatomicdistanceislikelytobetoofarawaytohaveany

through-bond influence that affects the stability of the Bi–carboxylate bond.

Ultimately, further studies would be required to determine the exact nature of the

interactionofBi(tolf)3andBi(mef)3incellmedium.

TheNMR studies suggested that following incubationofBi(difl)3 in cellmedium,

ligandexchangereactionsmaybe facilitatedby thehighconcentrationsofD-glucose

present in RPMI-1640(~11 mM [65]). It is postulated that partial substitution by

glucosylontotheBicomplexcouldpotentiallyresultintheformationofBi(difl)2gluor

Bi(difl)glu2 (where glu = glucosyl) complexes; however, further studies would be

required to confirm this. The involvement of glucose should be further studied to

explore the likelihood of this occurring in bloodwherein the normal blood glucose

concentrationisapproximately5mM[66].

While the NMR studies performed in the current study provide some useful

insights into thestabilityof theBiNSAIDs incellmedium, ithasbeensuggested that

theuseofsolid-stateNMRcouldbeanothersuitabletechniquetostudythestabilityof

the BiNSAIDs following incubation of the complexes in various biological

media(ProfessorPhilipAndrews,personalcommunication).Oneofthedisadvantages

ofsolid-stateNMRisthatitcantakeuptothreedaystorunonesomesample,and3-h

warm-up time is required prior to replacing the cryo-probe with an ambient

temperature solid-state probe (Dr Wilford Lie, personal communication).

Additionally,giventhatinorganicsamplestypicallyrequire100–200mgofcompound

foreachmeasurement[67]andthelimitedquantitiesofBiNSAIDssynthesisedforthe

studiesdescribedinthisThesis,theuseofsolidstateNMRwasunfeasible.

2.4.2 Invitrocytotoxicitytowardscancercells

Inordertodetermine ifa leadcompoundhasanypotential fordevelopmentasa

chemotherapeutic drug, the in vitro toxicity towards tumour cell lines needs to be

determined first [68]. As discussed in Section 2.1.2, the BiNSAIDs exhibit limited

solubility in aqueous solutions (including biological medium) and some organic

solvents.FollowingrangefindingstudiestodeterminefeasibleDMSOconcentrations,

the BiNSAIDswere solubilised in DMSO (final concentration 2% v/v) prior to their


96

additiontocellmediumfor invitro testingagainst theHCT-8cells. While theuseof

DMSO should be revisited for in vivo testing, the solubility of the BiNSAIDs was

substantiallyimprovedinordertointroducethecomplexesintothecellmediumforin

vitro testing. In addition, the range finding ensured the optimal balance between

minimal DMSO toxicity and maximal solubilisation of the BiNSAIDs and NSAIDs,

allowingforthesuccessfuldeterminationofIC50values(Table2.4)andcomparisonof

thetoxicityoftheBiNSAIDstoeachotherandtherespectiveNSAIDs.

TheMTTassays(Section2.3.2)showedthattheorderoftoxicityoftheBiNSAIDs

paralleledtheresponseexhibitedbythefreeNSAIDs(tolfH>mefH>diflH),suggesting

thatNSAIDsareprimarilyresponsibleforthetoxicityoftheBiNSAIDs.Additionally,if

the IC50 ratio LH:BiL3 was close to 3, it may be assumed that the NSAIDs are

predominantlyresponsibleforthetoxicityoftheBiNSAIDs(acknowledgingthereare

3moleofNSAIDpermoleofBi). TheIC50ratiosoftolfH:Bi(tolf)3andmefH:Bi(mef)3

were found tobe2.3 and2.7, respectively (Table2.4). WhileBi(tolf)3 exhibited the

highestIC50valueofthethreeBiNSAIDstested,theIC50ratiotolfH:Bi(tolf)3potentially

indicated that tolf is slightly less toxic to HCT-8 cells when complexed to Bi.

Alternatively, the IC50 ratio of mefH:Bi(mef)3 was closer to 3, potentially indicating

that complexation of mef to Bi had little effect on the toxicity of mefH. Although

Bi(difl)3wastheleasttoxicofthethreeBiNSAIDswhentheratiooftheIC50valueswas

considered (diflH:Bi(difl)3 = 5.0, Table 2.4), Bi(difl)3 showed substantially higher

toxicity than that expected for diflH, indicating that difl was more toxic when

complexedtoBi.Thereareanumberoffactorsthatmayinfluencethetoxicityofthe

BiNSAIDs towards the HCT-8 colon cancer cells and the observed

differences/similarities to the NSAIDs: (i) stability in cell medium (Sections2.3.1.1

and2.4.1), (ii)intrinsic physicochemical properties, and (iii) cellular uptake (to be

discussedinChapter3).

TheNMRstabilitystudiessuggestedthatBi(tolf)3displayedgreaterstabilitythan

Bi(mef)3 following incubation in cellmedium (Section 2.3.1.1); however, the overall

coordination(orlackthereof)ofBitotheNSAIDappearedtohavelittle influenceon

the toxicity of Bi(tolf)3 and Bi(mef)3 compared to the corresponding free NSAIDs.

AsthetoxicityofBi(tolf)3wasslightlylowerthanthecorrespondingconcentrationof

tolfH, it is plausible that the increased stability of Bi(tolf)3 may be influencing the

toxicityofthetolfligands.


97

Incontrast,Bi(difl)3wassubstantiallymoretoxicthantheanalogousdiflH. There

aretwopossiblereasonsforthis;thefirstexplanationisthecontributingtoxicityofBi.

Reference to Table 2.4 shows that diflH exhibits an IC50 value of 403 µM. When

consideringBiCl3,whichpossessesaconceivablynon-toxicanion(Cl−),itisnotedthe

IC50value(IC50=383µM)iscomparabletodiflHthatindicatesthatanadditiveeffect

wouldbemorenoticeable.Thesecondpossibleexplanationfortheincreasedtoxicity

ofBi(difl)3(versusdiflH)towardsHCT-8cells istheNMRevidenceforglucoseinthe

sample obtained from the incubation of Bi(difl)3 in cell medium. This raises the

questionofwhether glucose ispotentially coordinatingwithBi(difl)3 and increasing

its toxicity compared to free diflH. For instance, it might be possible that the

coordination of glucosyl ligands to Bi assists the uptake of the complex via glucose

uptakemediated transporters. Thiswouldbe expected tobegreater in cancer cells

sincetheyhaveahighaffinityforglucose[69,70].InotherstudiesbyDillonandco-

workers[71],Bi(difl)3showed1.2-foldlowertoxicityagainstHepG2livercancercells

compared to HCT-8 cells, as assessed by theMTT assay. Additionally, proportional

enhancementoftoxicity(Bi:NSAIDIC50ratio3.6)wasobserved[71]. Onehypothesis

put forth to rationalise this observation was the content of glucose in the cell

medium[71].Interestingly,theculturemediumthatwasusedtomaintaintheHepG2

duringtheMTTassayscontainedalowerconcentrationofD-glucosethanRPMI-1640

used in the present study (5.5mMand11.1mM, respectively). Therefore itmaybe

possible that the increased concentration of D-glucose in RPMI-1640 increases the

propensity of Bi(difl)3 to undergo ligand exchange reactions with glucose[71].

Infuture experiments,NMR studies could be performed to validate this hypothesis.

ThiswouldinvolvestudiesofBi(difl)3inDMEMforcomparisontotheresultobtained

inthisThesis.

TheorderoftoxicityoftolfH>mefH>diflH,mayberelatedtothephysicochemical

properties of themolecules. The reported logP values of tolfH, mefH and diflH are

very similar (4.94, 4.79 and5.20, respectively) [64]. Theseminor variations in logP

andthelackofcorrelationwiththetoxicitysuggestthelogPisnotaccountableforthe

observed differences in toxicity or cellular uptake.Additionally, the NSAIDs have

reported pKa values in a similar range (3.5–4.3)[61], all of which lie well below

pH7.4(thepHofRPMI-1640). ThemaindifferencesintheNSAIDstructuresinclude

the structural incorporation of the diphenylamine (tolfH andmefH) or the diphenyl

(diflH) [72-74]. Diphenylamine NSAIDs (such as tolfH and mefH) were shown to


98

induce lactate dehydrogenase leakage in primary cells obtained from isolated rat

hepatocytes, more significantly than diflH, at the same concentration [72]. The

absenceof an amine in the structureof diflHmight, therefore, result in the reduced

toxicity observed in this study. The current study showed that the substitution of

–CH3(mef)for–Cl(tolf)resultedinalmosta3-foldincreaseintoxicityassociatedwith

boththefreeNSAIDandBiNSAIDs. ThisresultparallelsotherstudieswherebytolfH

reportedly exhibited higher activity than mefH against BT474 and SKBR3 breast

cancercells[52].

The toxicity of the BiNSAIDswas substantially higher than the two reference Bi

compounds, BiCl3 and BSS. Notably, the IC50 values obtained for the BiNSAIDs are

comparabletothatoftherangeoftheanti-cancerdrug,cisplatin,althoughitshouldbe

recognisedthatcisplatinhasachievedgreatestclinicaleffectivenessagainsttesticular

and ovarian cancers [75]. Althoughnanomolar toxicity is desirablewhen screening

forleadcompoundsforanti-cancerdrugdevelopment,themoderatetoxicitydisplayed

by the BiNSAIDs and the similar toxicity compared to cisplatin is an encouraging

result. Ultimately, the BiNSAIDs will need to be evaluated for chemotherapeutic,

chemopreventiveandgastroprotectiveactivityinanimalmodelsinordertodetermine

whethertheyhaveanypotentialuseaschemotherapeuticorchemopreventivedrugs

inhumans.However,theinvitrocytotoxicityoftheBiNSAIDsdeterminedinthisstudy

suggests that further study is warranted into the interactions of the BiNSAIDs in

biologicalsystems.


The results from the phase contrast microscopy and the AnnV/7AAD assays

indicated that apoptosis is the primary mode of cell death induced by exposure of

HCT-8 cells toBi(tolf)3,Bi(mef)3, orBi(difl)3, and the respective freeNSAIDs. These

resultsare importantas theyshowthatall threeBiNSAIDs inducethesamemodeof

cell death as the respective uncomplexed NSAIDs. These results were also in

agreementwith a number of published studies that have reported apoptosis as the

mode of cell death induced by NSAIDs, including tolfH [42, 44, 45, 48, 50, 76-79],

mefH[51, 80] and diflH[80, 81] in other cancer cell lines in vitro. Additionally, a

numberofBi compoundshavebeenreported to inducecelldeathviaapoptosis ina

numberofcancercelllinesinvitro[82-86]. TheresultsfromtheAnnV/7AADassays

were in agreement with the toxicity determined by the MTT assays, whereby the


99

Bi(tolf)3 or tolfH, Bi(mef)3 or mefH, and Bi(difl)3 treatments induced apoptosis in

approximately 47–70% of the HCT-8 cell populations when treated with the

approximateIC50concentration(asdeterminedfromtheMTTassay).Additionally,the

diflH treatment resulted in a smaller increase in the proportion of apoptotic cells

(approximately 37%of the cell population),which is also consistentwith the lower

toxicityastheconcentrationofdiflHappliedintheassay(225μM)waslowerthanits

IC50value(403μM),andthehealthierphysicalappearanceofthecellsobservedusing

phasecontrastmicroscopy(presentedinSection2.3.3.1).

Thepresent studyconfirmed thatapoptosis is themodeof celldeath inducedby

BiNSAID and NSAID treatment in HCT-8 cells. As discussed in Section 2.1.3.2,

apoptosisisoftendescribedasa‘programmed’formofcelldeath.Apoptosiscanoccur

viatwodistinctpathwaystermed‘extrinsic’and‘intrinsic’.Apoptosisviaanextrinsic

signaling pathway results from transmembrane receptor-mediated interactions

involvingdeathreceptorsthataremembersofthetumornecrosisfactorreceptorgene

superfamily [32]. Alternatively, apoptosis initiated via the intrinsic signaling

pathwaysaremitochondrial-initiatedeventsinvolvingnon-receptor-mediatedstimuli

thatproduce intracellular signals that actdirectlyon targetswithin the cell (further

details on these mechanisms are reviewed in [32]). The MTT assay is a useful

indicatoroftoxicity,whichreliesonfunctioningmitochondriainmetabolicallyactive

cellstoconvertMTTtoformazan(discussedinSection1.2.3.1).Whenconsideredwith

the results from the cell death studies, the reduction in mitochondria function

observed in the MTT assays may suggest that the mitochondria are involved in

BiNSAID- andNSAID-inducedapoptosis, and thus theBiNSAIDs/NSAIDsmay trigger

celldeaththroughtheintrinsicapoptoticpathway. WhileitisrecognisedthatCOX-2

contributes to the anti-apoptotic effects and proliferation of cancer cells, there are

reports that suggest that NSAID-induced cell death is also mediated via non-COX-2

regulatedpathways (discussed inSection1.2.4). In thepresent study, theBiNSAIDs

andNSAIDs inhibited the production of PGE2, suggesting the compounds inhibit the

COX enzymes (discussed further in the next Section 2.4.4). However, it cannot be

ignored that numerous studies have reported that diphenylamine NSAIDs (such as

tolfH and mefH) can also cause uncoupling of oxidative phosphorylation in

mitochondria[73,74,87-89].Furthermore,mefHanddiflHshoweduncouplingeffects

onoxidativephosphorylationinisolatedratmitochondria[90],andprimarycellsfrom

isolated rat hepatocytes [72] resulting in depletion of ATP. Overall, the exact


100

mechanismsbywhichtheNSAIDsandBiNSAIDsinduceapoptosisandwhythethree

NSAIDsinthisstudyshowvariationintoxicitycannotbedrawnconclusivelywithout

further investigation. Further investigation could be performed in the future to

determine the detailed molecular pathway involved in BiNSAID/NSAID induced

celldeath.

2.4.4 PGE2production

As discussed in Section 2.1.4, the increased expression of COX-2 has been

established in many cancer cell types and the subsequent production of

prostaglandins, including PGE2, is advantageous for neoplastic cell survival and

proliferation.Assuch,studyingtheabilityoftheBiNSAIDstoinhibitPGE2production

isimportantfordeterminingthechemopreventivepotentialofthecomplexes.

ThePGE2concentration-responsecurvesshowedthat theBiNSAIDsmirrored the

effectsofthefreeNSAIDsafter4hexposuretotheHCT-8cells. Thisisanimportant

resultasitdemonstratesthatadministrationoftheBiNSAID(whereintheNSAIDdoes

not initially posses a free carboxylate) does not hinder the ability of the NSAID to

target the COX enzymes. As no appreciable loss of cell viability was

observed(Appendix 1.2a–c), it could be concluded that the reduction of PGE2

production at increasing concentrations of BiNSAIDs/NSAIDs was not due to cell

toxicity. Additionally,noappreciabledecreaseintheviabilityoftheHCT-8cellswas

observedfollowing24-htreatmentwithaspH(Appendix1.5);assuchnodecreasein

viabilitywasexpected followingonly4-h treatment, confirming toxicityofaspHwas

notassociatedwithreducedPGE2production.

When the PGE2 production concentration-response curves (Fig. 2.9a) are

compared to theMTT assay concentration-response curves (Fig 2.6), it is clear that

Bi(tolf)3 and Bi(mef)3 paralleled the respective free NSAIDs whereby the BiNSAIDs

were approximately three times more toxic than the respective free NSAIDs

(representative of themolar ratio of NSAID per BiNSAID). The results of the PGE2

productionconcentration-responsecurvesofBi(difl)3differed(Figure2.9b),however,

whereby diflH appeared to have slightly enhanced PGE2 inhibition compared to

Bi(difl)3 (at higher concentrations of NSAID). Additionally, the potency of the

reduction inPGE2 occurred in the sameorder as the toxicity of theNSAIDs (tolfH<

mefH < diflH), which may suggest a correlation between PGE2 production and cell

health. It was observed that the IC50 values associated with the PGE2


101

inhibition(Table2.6),were lower than the IC50 values required to induce toxicity in

theHCT-8cells(Table2.5).

ThetimeframeofthePGE2assayscomparedtotheMTTassaysusedtodetermine

the IC50 values of the compounds (4 h versus 24 h) reveals some insight into the

intracellularstabilityofthecomplexes.IthasbeenestablishedthatNSAIDsmusthave

a freecarboxylate inorder to interactwith theCOX-1/-2activesite [91]. Giventhat

the BiNSAIDs decreased PGE2 production in a similar manner compared to the

analogousNSAIDs,itisreasonabletoassumethattheNSAIDligandmustbeliberated

fromtheBiNSAIDcomplex,allowingafreecarboxylatetointeractwiththeCOX.This

is particularly significant in the case of Bi(tolf)3 and Bi(difl)3 as the stability results

describedinSection2.3.1suggestthatthesecomplexesremainintactinthemedium;

however, they can undergo hydrolysis/dissociation once inside the cells following

interactionwith enzymes and competing ligands. Subsequently, it seems likely that

the Bi(tolf)3 and Bi(difl)3 complexes are metabolised by intracellular processes

relativelyquickly(i.e.<4h).

TheabilityoftheBiNSAIDsandNSAIDstoinhibitPGE2intheHCT-8cell linewas

studiedinordertoensurethatBiNSAIDsshowedcomparableactivitywithinacancer

cellline,andthusprovideinformationontheabilitytoreducePGE2inamalignantcell

for chemopreventive potential. It is important to note that HCT-8 cells have been

showntoexpresshigherlevelsofCOX-1thanCOX-2[55,56,92];hencethepreference

forBiNSAIDsorNSAIDsforCOX-2overCOX-1inHCT-8cellscannotbeevaluatedfrom

thepresentstudy. Furtherworkneedstobeperformedtodeterminetheexpression

of COX-1 and COX-2 in the HCT-8 cells used in this study, especially as the level of

expression of COX-1/-2 by HCT-8 cells varies between reports [55-57, 92, 93].

ItwouldalsobeofvaluetostudytheinhibitoryactivityoftheBiNSAIDscomparedto

theNSAIDswithpurifiedCOXenzymesinanacellularenvironmenttodetermineifthe

selectivityforCOX-1/-2isaffectedbythecomplexationoftheNSAIDtoBi.However,

thisassaywouldnottakeintoaccounttheconditionsexperiencedbytheBiNSAIDsin

thecellenvironment. Forinstance,thestabilityofthecomplexesincellmediumand

the intracellular fluids, cell uptake and any alterations in structure following cell

uptakeandsubsequentintracellularmetabolismareimportantfactorsthataffectCOX

inhibition. As such, the cell-based assay as performed in the present study, and an

enzyme-basedassayutilizingpurifiedCOX-1/-2wouldprovidecomplementaryresults

andshouldbeevaluated inconjunction. Theresults fromthepresentstudyutilising


102

theHCT-8cellmodelshouldbecomparedtoothercelllinesthatdisplayhighlevelsof

COX-2,andcellswithessentiallynoCOX-2expression(currentlybeingperformedby

Brown,DillonandRanson,personalcommunication).

OffinalnotewasthefindingthatdiflHincreasedtheamountofPGE2productionby

HCT-8 cells at low concentrations ranging from1−100 μM (Fig. 2.9). The increased

production may be related to the gastroprotective activity reported by other

authors[53,54].ThisresultmayindicatethattheconcentrationofBi(difl)3needsto

be considered carefully if the complex is tested in vivo for cancer prevention as

treatmentwith a dose that increases theproductionof PGE2mayhave the opposite

result to the intended therapeutic effect. This is clearly an observation that will

requirefurtherinvestigation.


103

2.5 ChaptersummaryThemainaimofthischapterwastouse invitroassaystodeterminewhetherthe

BiNSAIDs, Bi(tolf)3, Bi(mef)3, and Bi(difl)3, possess any chemotherapeutic or

chemopreventive potential in a colon cancer (HCT-8) cell line. To this end, the

followingstudieswereperformedwiththesubsequentoutcomes:

i. NMR spectroscopic determination of the BiNSAID stability in RPMI-1640

cellmediumshowedthatBi(tolf)3isrelativelystable,whilstmefappeared

to dissociate from Bi. The 1H NMR spectrum of Bi(difl)3 suggested that

D-glucoseinthecellmediummayinteractwiththecomplexresultinginthe

formationofanewspecies.

ii. Examinationof in vitro toxicity towards cancer cells using theMTTassay

demonstrated the ability of the complexes to eradicate transformed cells.

Specifically,theBiNSAIDsdisplayedmoderatetoxicitytowardsHCT-8cells

withtheIC50valuesrangingbetween16–81µM.Theorderoftoxicityofthe

BiNSAIDswasBi(difl)3<Bi(mef)3<Bi(tolf)3,whichparalleledtheorderof

toxicity of the free NSAIDs. When considering the molar ratio of NSAID

containedintheBiNSAIDcomplexes,Bi(tolf)3andBi(mef)3showedsimilar

activity to the free NSAIDs; the activity of diflH was improved when

complexedtoBi.

iii. Themode of cell deathwas determinedusing phase contractmicroscopy

andflowcytometricexperiments.Phasecontrastmicroscopyshowedthat

themorphologicalchangesinHCT-8cellsindicativeofapoptosis(including

cell rounding, shrinking and membrane blebbing) were induced by the

BiNSAIDs and the respective uncomplexed NSAIDs. Flow cytometric

analysisofco-stained(AnnVand7AAD)cellsconfirmedthatbothBiNSAIDs

and their respective NSAIDs induced cell death through apoptosis rather

thannecrosis.

iv. The ability of BiNSAIDs to inhibit COX and reduce PGE2 production in

HCT-8 cells was determined. The BiNSAIDs inhibited the production of

PGE2bytheHCT-8cellsattreatmentconcentrationscomparabletothefree

NSAIDs.TheorderofPGE2inhibitionwasBi(difl)3/diflH<Bi(mef)3/mefH<

Bi(tolf)3/tolfH.


104

AlthoughtheresultsfromtheNMRstudiessuggestthattheBiNSAIDshavelimited

stability inRPMI-1640 cell culturemedium, theBiNSAIDs (or the biologically active

forms of these complexes) elicited micromolar toxicity towards HCT-8 cells (as

demonstratedintheMTTassays)thatwerecomparableorbetterthantherespective

NSAIDs. Importantly, the presence of Bi did not adversely affect the toxicity or the

ability of the NSAIDs to induce apoptosis or inhibit PGE2 production. The superior

stability in cell medium, increased toxicity, and highest PGE2 inhibition of Bi(tolf)3

compared to the Bi(mef)3 and Bi(difl)3, suggests that Bi(tolf)3 might be the best

candidate for further testing using in vivo systems. However, the diversity in the

resultswarrants further investigation intohowthe threeBiNSAIDs interactwith the

HCT-8 cells. Further work will be required to determine whether the molecular

mechanisms involved in apoptosis are analogous between each BiNSAID and

respectivefreeNSAID.


105

2.6 References[1] E. Kerns and L. Di. (2008) Drug-like properties: concepts, structure design and

methods:fromADMEtotoxicityoptimization,1sted.,56-84,AcademicPress.[2] V.H.Thomas,S.Bhattachar,L.Hitchingham,P.Zocharski,M.Naath,N.Surendran,

C. L. Stoner and A. El-Kattan. (2006) The road map to oral bioavailability: anindustrialperspective,ExpertOpin.DrugMetab.Toxicol.,2,591-608.

[3] A. Dahan and J. M. Miller. (2012) The solubility–permeability interplay and itsimplications in formulation design and development for poorly soluble drugs,AAPSJ.,14,244-251.

[4] G. Singh. (2000) Gastrointestinal complications of prescription and over-the-counternonsteroidalanti-inflammatorydrugs:aviewfromtheARAMISdatabase.Arthritis, Rheumatism, and Aging Medical Information System, Am. J. Ther., 7,115-121.

[5] L. Laine. (2001) Approaches to nonsteroidal anti-inflammatory drug use in thehigh-riskpatient,Gastroenterology,120,594-606.

[6] T. Grosser. (2006) Chapter 29: Non-steroidal and anti-inflammatory drugs andcoxibs, In Applied Pharmacokinetics & Pharmacodynamics: Principles ofTherapeuticDrugMonitoring (M. E. Burton, Ed.) 4th ed., pp 752-780, LippincottWilliams&Wilkins,UnitedStatesofAmerica.

[7] H.Sun,L.ZhangandK.Y.Szeto.(2004)Bismuthinmedicine,Met. IonsBiol.Syst.,41,333-378.

[8] P.C.Andrews,R.L.Ferrero,P.C.Junk,I.Kumar,Q.Luu,K.NguyenandJ.W.Taylor.(2010) Bismuth(III) complexes derived from non-steroidal anti-inflammatorydrugsandtheiractivityagainstHelicobacterpylori,DaltonTrans.,39,2861-2868.

[9] Z.W.YuandP.J.Quinn.(1994)Dimethylsulphoxide:Areviewofitsapplicationsincellbiology,Biosci.Rep.,14,259-281.

[10] R.D.Brobyn.(1975)Thehumantoxicologyofdimethylsulfoxide,Ann.N.Y.Acad.Sci.,243,497-506.

[11] A.M.Bonin,J.A.Yanez,C.Fukuda,X.W.Teng,C.T.Dillon,T.W.Hambley,P.A.Layand N. M. Davies. (2010) Inhibition of experimental colorectal cancer andreduction in renalandgastrointestinal toxicitiesby copper-indomethacin in rats,CancerChemother.Pharmacol.,66,755-764.

[12] D.Kovala-Demertzi,D.Hadjipavlou-Litina,A. Primikiri,M. Staninska, C. Kotoglouand M. A. Demertzis. (2009) Anti-inflammatory, antiproliferative, and radical-scavengingactivitiesoftolfenamicacidanditsmetalcomplexes,Chem.Biodivers.,6,948-960.

[13] D.Kovala-Demertzi,D.Hadjipavlou-Litina,M. Staninska,A. Primikiri, C. Kotoglouand M. A. Demertzis. (2009) Anti-oxidant, in vitro, in vivo anti-inflammatoryactivity and antiproliferative activity ofmefenamic acid and itsmetal complexeswithmanganese(II), cobalt(II), nickel(II), copper(II) and zinc(II), J. Enzym. Inhib.Med.Ch.,24,742-752.

[14] G.Tamasi,C.Bernini,G.Corbini,N.F.Owens,L.Messori,F.Scaletti,L.Massai,P.L.Giudice and R. Cini. (2014) Synthesis, spectroscopic and DFT structuralcharacterizationoftwonovelruthenium(III)oxicamcomplexes.Invivoevaluationofanti-inflammatoryandgastricdamagingactivities,J.Inorg.Biochem.,134,25-35.

[15] M. Çeşme, A. Gölcü and I. Demirtaş. (2015) New metal based drugs: spectral,electrochemical, DNA-binding, surface morphology and anticancer activityproperties,Spectrochim.ActaA,135,887-906.

[16] P.Christofis,M.Katsarou,A.Papakyriakou,Y.Sanakis,N.KatsarosandG.Psomas.(2005) Mononuclear metal complexes with piroxicam: synthesis, structure andbiologicalactivity,J.Inorg.Biochem.,99,2197-2210.

[17] Cell Signaling Technology Inc. (2015) Paclitaxel #9807, Available from:https://media.cellsignal.com/pdf/9807.pdf,[accessed:Mar.112018]


106

[18] Cell Signaling Technology Inc. (2018) Vinblastine #14255, Available from:https://www.cellsignal.com/products/activators-inhibitors/vinblastine/14255,[accessed:Mar.112018]

[19] C. Griffin, A. Karnik, J. McNulty and S. Pandey. (2011) Pancratistatin selectivelytargets cancer cell mitochondria and reduces growth of human colon tumorxenografts,Mol.CancerTher.,10,57-68.

[20] K. L. Vine, J. M. Locke, M. Ranson, S. G. Pyne and J. B. Bremner. (2007) In vitrocytotoxicityevaluationofsomesubstituted isatinderivatives,Bioorg.Med.Chem.,15,931-938.

[21] S.Ahmad,A.A. Isab,S.AliandA.R.Al-Arfaj. (2006)Perspectives inbioinorganicchemistryofsomemetalbasedtherapeuticagents,Polyhedron,25,1633-1645.

[22] T.W.Hambley. (2007)Developingnewmetal-basedtherapeutics:challengesandopportunities,DaltonTrans.,4929-4937.

[23] W. Li, L. Jin, N. Zhu, X. Hou, F. Deng and H. Sun. (2003) Structure of colloidalbismuth subcitrate (CBS) in dilute HCl: unique assembly of bismuth citratedinuclearunits([Bi(cit)2Bi]2-),J.Am.Chem.Soc.,125,12408-12409.

[24] F.Thomas,B.BialekandR.Hensel.(2011)Medicaluseofbismuth:thetwosidesofthecoin,J.Clin.Toxicol.,S:3,doi:10.4172/2161-0495.s4173-4004.

[25] E.Vega-AvilaandM.K.Pugsley.(2011)Anoverviewofcolorimetricassaymethodsusedtoassesssurvivalorproliferationofmammaliancells,Proc.West.Pharmacol.Soc.,54,10-14.

[26] T. Mosmann. (1983) Rapid colorimetric assay for cellular growth and survival:Applicationtoproliferationandcytotoxicityassays,J.Immunol.Methods,65,55-63.

[27] F. Denizot and R. Lang. (1986) Rapid colorimetric assay for cell growth andsurvival. Modifications to the tetrazolium dye procedure giving improvedsensitivityandreliability,J.Immunol.Methods,89,271-277.

[28] H. Tada, O. Shiho, K. Kuroshima, M. Koyama and K. Tsukamoto. (1986) Animprovedcolorimetricassayforinterleukin2,J.Immunol.Methods,93,157-165.

[29] J. Carmichael, W. G. DeGraff, A. F. Gazdar, J. D. Minna and J. B. Mitchell. (1987)Evaluationofa tetrazolium-basedsemiautomatedcolorimetricassay:assessmentofchemosensitivitytesting,CancerRes.,47,936-942.

[30] M. B. Hansen, S. E. Nielsen and K. Berg. (1989) Re-examination and furtherdevelopmentofapreciseandrapiddyemethodformeasuringcellgrowth/cellkill,J.Immunol.Methods,119,203-210.

[31] G.Kroemer,L.Galluzzi,P.Vandenabeele,J.Abrams,E.S.Alnemri,E.H.Baehrecke,M. V. Blagosklonny,W. S. El-Deiry, P. Golstein, D. R. Green,M. Hengartner, R. A.Knight,S.Kumar,S.A.Lipton,W.Malorni,G.Nunez,M.E.Peter,J.Tschopp,J.Yuan,M. Piacentini, B. Zhivotovsky, G. Melino and Nomenclature Committee on CellDeath. (2009)Classificationof celldeath: recommendationsof theNomenclatureCommitteeonCellDeath2009,CellDeathDiffer.,16,3-11.

[32] S.Elmore. (2007)Apoptosis:a reviewofprogrammedcelldeath,Toxicol.Pathol.,35,495-516.

[33] M.S.RicciandW.-X.Zong.(2006)Chemotherapeuticapproachesfortargetingcelldeathpathways,Oncologist,11,342-357.

[34] Y. Z. Li, C. J. Li, A. V. Pinto and A. B. Pardee. (1999) Release of mitochondrialcytochromeCinbothapoptosisandnecrosisinducedbybeta-lapachoneinhumancarcinomacells,Mol.Med.,5,232-239.

[35] A. R. Salomon, D. W. Voehringer, L. A. Herzenberg and C. Khosla. (2000)Understanding and exploiting themechanistic basis for selectivity of polyketideinhibitorsofF(0)F(1)-ATPase,Proc.Natl.Acad.Sci.U.S.A.,97,14766-14771.

[36] X.Bai,F.Cerimele,M.Ushio-Fukai,M.Waqas,P.M.Campbell,B.Govindarajan,C.J.Der, T. Battle, D. A. Frank, K. Ye, E. Murad, W. Dubiel, G. Soff and J. L. Arbiser.(2003)Honokiol,asmallmolecularweightnaturalproduct, inhibitsangiogenesisinvitroandtumorgrowthinvivo,J.Biol.Chem.,278,35501-35507.


107

[37] G. Pistritto,D. Trisciuoglio, C. Ceci, A.Garufi andG.D'Orazi. (2016)Apoptosis asanticancermechanism: Function and dysfunction of itsmodulators and targetedtherapeuticstrategies,Aging,8,603-619.

[38] I.Vermes,C.Haanen,H.Steffens-NakkenandC.Reutelingsperger.(1995)Anovelassay for apoptosis. Flow cytometric detection of phosphatidylserine expressiononearlyapoptoticcellsusingfluoresceinlabelledAnnexinV,J.Immunol.Methods,184,39-51.

[39] E. M. Bevers, P. Comfurius and R. F. Zwaal. (1996) Regulatory mechanisms inmaintenance and modulation of transmembrane lipid asymmetry:pathophysiologicalimplications,Lupus,5,480-487.

[40] N. J. Philpott, A. J. Turner, J. Scopes, M.Westby, J. C. Marsh, E. C. Gordon-Smith,A.G.Dalgleish and F. M. Gibson. (1996) The use of 7-amino actinomycin D inidentifying apoptosis: simplicity of use and broad spectrum of applicationcomparedwithothertechniques,Blood,87,2244-2251.

[41] I.Nicoletti,G.Migliorati,M.C.Pagliacci,F.GrignaniandC.Riccardi.(1991)Arapidand simple method for measuring thymocyte apoptosis by propidium iodidestainingandflowcytometry,J.Immunol.Methods,139,271-279.

[42] E.S.Choi,J.H.Shim,J.Y.Jung,H.J.Kim,K.H.Choi,J.A.Shin,J.S.Nam,N.P.ChoandS. D. Cho. (2011) Apoptotic effect of tolfenamic acid in androgen receptor-independentprostatecancercellandxenograft tumorthroughspecificityprotein1,CancerSci.,102,742-748.

[43] U.T.Sankpal,M.Abdelrahim,S.F.Connelly,C.M.Lee,R.Madero-Visbal,J.Colon,J.Smith, S. Safe, P. Maliakal and R. Basha. (2012) Small molecule tolfenamic acidinhibits PC-3 cell proliferation and invasion in vitro, and tumor growth inorthotopicmousemodelforprostatecancer,Prostate,72,1648-1658.

[44] J.B.Jeong,J.Choi,S.J.BaekandS.-H.Lee.(2013)Reactiveoxygenspeciesmediatetolfenamicacid-inducedapoptosisinhumancolorectalcancercells,Arch.Biochem.Biophys.,537,168-175.

[45] J.B. Jeong,X.Yang,R.Clark, J.Choi,S. J.BaekandS.H.Lee.(2013)Amechanisticstudy of the proapoptotic effect of tolfenamic acid: involvement of NF-kappaBactivation,Carcinogenesis,34,2350-2360.

[46] D.Eslin,C.Lee,U.T.Sankpal,P.Maliakal,R.M.Sutphin,L.AbrahamandR.Basha.(2013) Anticancer activity of tolfenamic acid in medulloblastoma: a preclinicalstudy,TumorBiol.,34,2781-2789.

[47] D.Eslin,U.T. Sankpal,C.Lee,R.M.Sutphin,P.Maliakal,E.Currier,G. Sholler,M.Khan and R. Basha. (2013) Tolfenamic acid inhibits neuroblastoma cellproliferationandinducesapoptosis:anoveltherapeuticagentforneuroblastoma,Mol.Carcinog.,52,377-386.

[48] X.Zhang,S.H.Lee,K.W.Min,M.F.McEntee,J.B.Jeong,Q.LiandS.J.Baek.(2013)Theinvolvementofendoplasmicreticulumstressinthesuppressionofcolorectaltumorigenesisbytolfenamicacid,CancerPrev.Res.,6,1337-1347.

[49] U.T.Sankpal,C.M.Lee,S.F.Connelly,O.Kayaleh,D.Eslin,R.Sutphin,S.Goodison,L. Adwan, N. H. Zawia, L. M. Lichtenberger and R. Basha. (2013) Cellular andorganismal toxicity of the anti-cancer small molecule, tolfenamic acid: a pre-clinicalevaluation,CellPhysiol.Biochem.,32,675-686.

[50] J.W. Chang, S. U. Kang, J.W. Choi, Y. S. Shin, S. J. Baek, S. H. Lee and C. H. Kim.(2014) Tolfenamic acid induces apoptosis and growth inhibition in anaplasticthyroid cancer: Involvement of nonsteroidal anti-inflammatory drug-activatedgene-1expressionandintracellularreactiveoxygenspeciesgeneration,FreeRadic.Biol.Med.,67,115-130.

[51] D. H. Woo, I. S. Han and G. Jung. (2004) Mefenamic acid-induced apoptosis inhumanlivercancercell-linesthroughcaspase-3pathway,LifeSci.,75,2439-2449.


108

[52] X. Liu, M. Abdelrahim, S. Abudayyeh, P. Lei and S. Safe. (2009) The NSAIDtolfenamicacidinhibitsBT474andSKBR3breastcancercellandtumorgrowthbyrepressingerbB2expression,Mol.CancerTher.,8,1207-1217.

[53] M.M.Cohen.(1983)Diflunisalprotectshumangastricmucosaagainstdamagebyindomethacin,Dig.Dis.Sci.,28,1070-1077.

[54] M.Trautmann,B.M.PeskarandB.A.Peskar. (1991)Aspirin-likedrugs, ethanol-inducedratgastricinjuryandmucosaleicosanoidrelease,Eur.J.Pharmacol.,201,53-58.

[55] T.-B. Lee,K.-J.M.Kim,Y-D., S.-I.Kang,K.-R. Jung, J.-U. Lee andC.-H. Choi. (2005)Adjuvant effect of NSAIDs on the cytotoxicity of colon cancer cells to 5-FU,J.KoreanSoc.Coloproctol.,21,121-128.

[56] H. Sui, S. Zhou, Y. Wang, X. Liu, L. Zhou, P. Yin, Z. Fan and Q. Li. (2011) COX-2contributes toP-glycoprotein-mediatedmultidrugresistanceviaphosphorylationofc-JunatSer63/73incolorectalcancer,Carcinogenesis,32,667-675.

[57] S. Dinicola, A. Pasqualato, S. Proietti, M. G. Masiello, A. Palombo, P. Coluccia, R.Canipari,A.Catizone,G.Ricci,A.H.Harrath,S.H.Alwasel,A.CucinaandM.Bizzarri.(2016)ParadoxicalE-cadherinincreasein5FU-resistantcoloncancerisunaffectedduringmesenchymal-epithelialreversioninducedbygamma-secretase inhibition,LifeSci.,145,174-183.

[58] M.D.Hall,K.A.Telma,K.-E.Chang,T.D.Lee,J.P.Madigan,J.R.Lloyd,I.S.Goldlust,J.D.HoescheleandM.M.Gottesman. (2014)Sayno toDMSO:Dimethylsulfoxideinactivates cisplatin, carboplatin, and other platinum complexes,Cancer Res., 74,3913-3922.

[59] R.K.Harris,E.D.Becker,S.M.Cabral,S.M.CabraldeMenezes,R.GoodfellowandP.Granger.(2001)NMRnomenclature.Nuclearspinpropertiesandconventionsforchemicalshifts(IUPACRecommendations2001),PureAppl.Chem.,73,1795–1818.

[60] L. Ronconi and P. J. Sadler. (2008) Applications of heteronuclear NMRspectroscopy in biological andmedicinal inorganic chemistry,Coord. Chem. Rev.,252,2239-2277.

[61] J. O. Boison, F. J. Ramos and A. Chicoine. (2016) Nonsteroidal antiinflammatorydrugs, InChemicalanalysisofnon-antimicrobialveterinarydrugresiduesinfood(J.F.Kay,J.D.MacNeilandJ.Wang,Eds.),pp427-496,Wiley,NewYork.

[62] T.Velkov, J.Horne,A.Laguerre,E. Jones,M. J. ScanlonandC. J.H.Porter. (2007)Examinationoftheroleofintestinalfattyacid-bindingproteinindrugabsorptionusingaparallelartificialmembranepermeabilityassay,Chem.Biol.,14,453-465.

[63] M. Polasek, M. Pospisilova and M. Urbanek. (2000) Capillary isotachophoreticdetermination of flufenamic, mefenamic, niflumic and tolfenamic acid inpharmaceuticals,J.Pharm.Biomed.Anal.,23,135-142.

[64] A. G. Siraki, T. Chevaldina and P. J. O’Brien. (2005) Application of quantitativestructure–toxicity relationships for acute NSAID cytotoxicity in rat hepatocytes,Chem.Biol.Interact.,151,177-191.

[65] ThermoFisherScientificTechnicalresources22400-RPMI1640,HEPES,Availablefrom:https://www.thermofisher.com/au/en/home/technical-resources/media-formulation.117.html,[accessed:Feb.12,2018]

[66] B. Thorens. (1992) Molecular and cellular physiology of GLUT-2, a high-Kmfacilitated glucose diffusion transporter, In Molecular Biology of Receptors andTransporters: Bacterial and Glucose Transporters (M. Friedlander and M.Mueckler,Eds.),pp209-235,AcademicPress,SanDiego.

[67] S.A.CurranandR.Brambilla.(1996)Magneticresonancespectroscopy,InAguideto materials characterization and chemical analysis (J. P. Sibilia, Ed.), pp 39-60,Wiley,NewYork.

[68] U.ReinholdandW.Tilgen. (2012)Chemosensitivity testing inoncology,SpringerBerlinHeidelberg.


109

[69] O.WarburgandF.Kubowitz.(1927)Themetabolismofgrowingcells(fibroblasts,heart,chorion).,Biochem.Z.,189,242-248.

[70] E. K. Pauwels, E. J. Sturm, E. Bombardieri, F. J. Cleton andM. P. Stokkel. (2000)Positron-emission tomography with [18F]fluorodeoxyglucose. Part I. Biochemicaluptakemechanismanditsimplicationforclinicalstudies,J.CancerRes.Clin.Oncol.,126,549-559.

[71] S. Spremo. (2014) Novel bismuth complexes as potential anti-cancer drugs,BachelorofMedicinalChemistry(Advanced),UniversityofWollongong,Australia.

[72] Y. Masubuchi, H. Saito and T. Horie. (1998) Structural requirements for thehepatotoxicityofnonsteroidalanti-inflammatorydrugsinisolatedrathepatocytes,J.Pharmacol.Exp.Ther.,287,208-213.

[73] Y. Masubuchi, S. Yamada and T. Horie. (1999) Diphenylamine as an importantstructure of nonsteroidal anti-inflammatory drugs to uncouple mitochondrialoxidativephosphorylation,Biochem.Pharmacol.,58,861-865.

[74] Y.Masubuchi, S.YamadaandT.Horie. (2000)Possiblemechanismofhepatocyteinjury induced by diphenylamine and its structurally related nonsteroidal anti-inflammatorydrugs,J.Pharmacol.Exp.Ther.,292,982-987.

[75] L. Kelland. (2007) The resurgence of platinum-based cancer chemotherapy,Nat.Rev.Cancer,7,573-584.

[76] M. Abdelrahim, C. H. Baker, J. L. Abbruzzese and S. Safe. (2006) Tolfenamic acidand pancreatic cancer growth, angiogenesis, and Sp protein degradation, J. Natl.CancerInst.,98,855-868.

[77] R.Basha,S.B. Ingersoll,U.T.Sankpal,S.Ahmad,C.H.Baker, J.R.Edwards,R.W.Holloway, S. Kaja and M. Abdelrahim. (2011) Tolfenamic acid inhibits ovariancancer cell growth and decreases the expression of c-Met and survivin throughsuppressingspecificityproteintranscriptionfactors,Gynecol.Oncol.,122,163-170.

[78] S. U. Kang, Y. S. Shin, H. S. Hwang, S. J. Baek, S.-H. Lee and C.-H. Kim. (2012)Tolfenamicacidinducesapoptosisandgrowthinhibitioninheadandneckcancer:involvementofNAG-1expression,PLoSOne,7,e34988.

[79] H.-J.Kim,S.-D.Cho,J.Kim,S.-J.Kim,C.Choi,J.-S.Kim,J.-S.Nam,K.HanKwon,K.-S.Kang and J.-Y. Jung. (2013) Apoptotic effect of tolfenamic acid onMDA-MB-231breastcancercellsandxenografttumors,J.Clin.Biochem.Nutr.,53,21-26.

[80] X. Lu,W. Xie, D. Reed,W. S. Bradshaw and D. L. Simmons. (1995) Nonsteroidalantiinflammatory drugs cause apoptosis and induce cyclooxygenases in chickenembryofibroblasts,Proc.Natl.Acad.Sci.U.S.A.,92,7961-7965.

[81] K.Shirakawa,L.Wang,N.Man,J.Maksimoska,A.W.Sorum,H.W.Lim,I.S.Lee,T.Shimazu,J.C.Newman,S.Schröder,M.Ott,R.Marmorstein,J.Meier,S.NimerandE.Verdin.(2016)Salicylate,diflunisalandtheirmetabolitesinhibitCBP/p300andexhibitanticanceractivity,eLife,5,e11156.

[82] K. S. O. Ferraz,N. F. Silva, J. G. da Silva, L. F. deMiranda, C. F. D. Romeiro, E.M.Souza-Fagundes, I. C. Mendes and H. Beraldo. (2012) Investigation on thepharmacologicalprofileof2,6-diacetylpyridinebis(benzoylhydrazone)derivativesandtheirantimony(III)andbismuth(III)complexes,Eur.J.Med.Chem.,53,98-106.

[83] D.H.Ishak,K.K.Ooi,K.P.Ang,A.M.Akim,Y.K.Cheah,N.Nordin,S.N.Halim,H.L.Seng and E. R. Tiekink. (2014) A bismuth diethyldithiocarbamate compoundpromotesapoptosisinHepG2carcinoma,cellcyclearrestandinhibitscellinvasionthrough modulation of the NF-κB activation pathway, J. Inorg. Biochem., 130,38-51.

[84] A. Islam, B. L. Rodrigues, I. M. Marzano, E. C. Perreira-Maia, D. Dittz, M. T. PazLopes,M. Ishfaq, F. Frézard and C. Demicheli. (2016) Cytotoxicity and apoptoticactivityofnovelorganobismuth(V)andorganoantimony(V)complexesindifferentcancercelllines,Eur.J.Med.Chem.,109,254-267.


110

[85] K.Iuchi,Y.HatanoandT.Yagura.(2008)Heterocyclicorganobismuth(III)inducesapoptosis of human promyelocytic leukemic cells through activation of caspasesandmitochondrialperturbation,Biochem.Pharmacol.,76,974-986.

[86] K.Onishi,M.Douke,T.Nakamura,Y.Ochiai,N.Kakusawa,S.Yasuike, J.Kurita,C.Yamamoto, M. Kawahata, K. Yamaguchi and T. Yagura. (2012) A novelorganobismuth compound, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl]-pyrrolidine,inducesapoptosisinthehumanacutepromyelocyticleukemiacelllineNB4viareactiveoxygenspecies,J.Inorg.Biochem.,117,77-84.

[87] S.Somasundaram,G.Sigthorsson,R.J.Simpson,J.Watts,M.Jacob,I.A.Tavares,S.Rafi,A.Roseth,R.Foster,A.B.Price,J.M.WrigglesworthandI.Bjarnason.(2000)UncouplingofintestinalmitochondrialoxidativephosphorylationandinhibitionofcyclooxygenasearerequiredforthedevelopmentofNSAID-enteropathyintherat,Aliment.Pharmacol.Ther.,14,639-650.

[88] G.Galati,S.Tafazoli,O.Sabzevari,T.S.ChanandP.J.O'Brien.(2002)IdiosyncraticNSAIDdruginducedoxidativestress,Chem.Biol.Interact.,142,25-41.

[89] Y. Li, X.M.Qi, X. Xue, X. F.Wu, Y. F.Wu,M. Chen, G. Z. Xing, Y. Luan and J. Ren.(2009) The relationship between diphenylamine structure and NSAIDs-inducedhepatocytesinjury,Toxicol.Lett.,186,111-114.

[90] F.E.Mingatto,A.C.Santos,S.A.Uyemura,M.C.JordaniandC.Curti.(1996)Invitrointeractionofnonsteroidalanti-inflammatorydrugsonoxidativephosphorylationofratkidneymitochondria:respirationandATPsynthesis,Arch.Biochem.Biophys.,334,303-308.

[91] R.J.Flower.(2003)ThedevelopmentofCOX2inhibitors,Nat.Rev.DrugDiscov.,2,179-191.

[92] T.J.Jang,K.H.JeonandK.H.Jung.(2009)Cyclooxygenase-2expressionisrelatedtotheepithelial-to-mesenchymaltransitioninhumancoloncancers,YonseiMed.J.,50,818-824.

[93] G.J.Du,H.H.Lin,Q.T.XuandM.W.Wang.(2005)Thalidomideinhibitsgrowthoftumors through COX-2 degradation independent of antiangiogenesis, Vascul.Pharmacol.,43,112-119.

111

Chapter3. UptakeandlocalisationofBibyBiNSAID-treated

HCT-8cells

HavingestablishedthestabilityoftheBiNSAIDsincellmediumandthetoxicityofthe

BiNSAIDstowardsHCT-8cellsinChapter2,itwasofinteresttoinvestigatetheuptake

of the complexes. The focus of this work was to quantify the cellular Bi following

treatment with BiNSAIDs to determine if there were any correlations between Bi

uptake and BiNSAID stability or toxicity. Additionally, studies of the subcellular

distributionofBiwereperformedonHCT-8cellsandtheresultswerediscussedwith

respect to the stability results (Chapter 2) and potential intracellular targets of the

BiNSAIDs.

PartsofthisChapterhavebeenpublishedin:

Hawksworth, E. L., Andrews, P. C., Lie, W., Lai, B. & Dillon, C. T. (2014) Biological

evaluation of bismuth non-steroidal anti-inflammatory drugs (BiNSAIDs): Stability,

toxicityanduptakeinHCT-8coloncancercells,J.Inorg.Biochem.,135,28-39.

Chapter3.UptakeandlocalisationofBiNSAIDs

112

3.1 IntroductionAs introduced in Section 2.1, there are a number of factors that need to be

addressed in order to determine the suitability of BiNSAIDs as potential

chemotherapeutic or chemopreventive candidates. The cellular uptake and

intracellulardistributionofadrugisanimportantcriteriathatrequiresexamination,

particularly in the contextof chemopreventionas it is important toestablish: (i)the

quantity of thedrug that can enter the cell (i.e. howmuch is required to induce the

desired therapeutic effect); and (ii) where the drug distributes in the cell (i.e. if it

localisesatthedesiredbiologicaltargetorindiscriminatelythroughoutthecell).

HavinginvestigatedthestabilityoftheBiNSAIDsincellmedium,andthetoxicityof

theBiNSAIDsandPGE2inhibitiontowardsHCT-8cellsinChapter2,itwasofinterest

to quantify the uptake of the BiNSAIDs to determine if therewere any correlations

betweenuptakeandthestabilityandtoxicityoftheBiNSAIDs.Additionally,assessing

the intracellulardistributionofBi inHCT-8 cellswasof interest inorder toprovide

insightintotheintracellulartargetsofBiandpotentiallytheBiNSAIDs(intheinstance

ofthecomplexremainingintact).

Beneficially,BiNSAIDscontaintwocomponentsthatcanbeusedtodeterminedrug

uptake and intracellular distribution, the inorganic Bi atom and the organic NSAID

molecule. Thestudiesdescribedherein focusedon theuptakeanddistributionofBi

withintheHCT-8cellsfollowingtreatmentwithBiNSAIDs.Whileitwillbeimportant

infutureworktofocusontheuptakeanddistributionoftheNSAIDs,therewereafew

reasonswhyitwasinitiallyimportanttostudytheuptakeandintracellularfateofBi.

Firstly, given the potential for Bi accumulation within mammalian cells and the

potential forBi toxicity (discussed inSection1.4.2), it is important todetermine the

quantity of Bi that enters cells at the BiNSAID concentration required to cause

cytotoxicity.Secondly,organicmolecules(includingNSAIDs)aregenerallycomprised

ofendogenouselementsthatarefoundwithincells(C,H,O,N)thatmakeitdifficultto

detect drug molecules inside cells without using labeling techniques. Radioactive

labeling (e.g. 3H) overcomes this issue, but ismore advantageous forwholebodyor

tissueimagingratherthanthemicronresolutionrequiredforintracellularimaging[1].

Thirdly, chemical labeling, such as the use of fluorophores, involves altering the

chemicalstructureandoften increasing themolecularweightof themolecule,which

can potentially alter the uptake and distribution of the drug [2]. In contrast,

metal/metalloid complexes offer an advantage in this instance, as manymetals are


113

often found endogenously within cells in trace quantities. The presence of a

metal/metalloidinthecompound,therefore,negatestheneedtolabelthedrug. The

limitationofthisapproachisthatitisnotpossibletotellwhetherthemetal/metalloid

isstillattachedtotheorganicligands.

There are a number of analytical techniques that can be used to detect trace

quantities of metals/metalloids in cell samples including atomic absorption

spectroscopy(AAS;discussedinSection3.1.1.)andmicroprobesynchrotronradiation

X-rayfluorescence(SRXRF)spectroscopy(discussedinSection3.1.2).

3.1.1 Atomicabsorptionspectroscopyforstudyingcellularuptakeofinorganic

complexes/drugs

Establishing the uptake of drugs entering cancer cells is important to determine

theireffectiveness,wherebythefocusofbowelcancerchemopreventivesistoprovide

insightintotheirtoxicitytowardsbowelcancercells(Section2.3.2)andtheirabilityto

inhibit PG production through the inhibition of COX-2 (Section 2.3.4). The cellular

uptake of inorganic complexes/drugs containing non-endogenous elements (such as

As,Bi,Co,Cr,Ni,Pd,Pt,Rd,RuandRh)canbemonitoredbyanumberofanalytical

techniques.

Acommonchoice isAAS,whichuseseithera flame(FAAS),oranelectrothermal

source, also known as graphite furnace atomic absorption spectroscopy (GFAAS), to

analyse metals/metalloids [3]. The sample is atomised using high temperatures

(2,000−3,000K) toproduceground state freeatoms in thevapour state [3].A light

source (generally a hollow cathode lamp), which produces the characteristic

wavelengths of light specific to the analyte, is used to quantify the element in the

sample [3]. The techniques known as inductively coupled plasma-atomic emission

spectroscopy(ICP-AES)andinductivelycoupledplasma-massspectrometry(ICP-MS),

employ a plasma source as the sample atomiser. A plasma source reaches

temperatures that exceed twice the temperature of a combustion flame

(6,000−10,000K)[3].

A number of factors are important when determining the suitability of these

techniquesforanalysingmetalsincellsamples,including;sensitivity,samplevolumes

andchemicalmatrix/interference.Thetypicaldetectionlimitsoftheaforementioned

techniques vary depending on the element of interest, and the particular brand of

instrument employed for the analysis [4]. The typical detection limits of Bi are


114

50parts per billion (ppb), 0.3 ppb, 18 ppb, and 0.01-0.1parts per trillion (ppt) for

FAAS,GFAAS,ICP-AES,andICP-MS,respectively[4].GFAAShasadetectionlimitupto

1000 times greater than FAAS [4]. The greater density of atoms generated by the

graphitetubeandimprovedresidencytimeoftheanalyteintheopticalpath(<1sin

FAASversus several seconds inGFAAS) contribute to the superiordetection limitof

GFAAS[3, 4]. As such, GFAAS [3-9] andmulti-element GFAAS [10] have been used

successfullytodeterminetheconcentrationofBiinbiologicalsampleswithadequate

detectionlimits(≤1μg/L).

Additionally,GFAASonlyrequiressmallsamplevolumes(approximately20μLper

detection)foraccurateelementaldetection,whereas1−2mLofsampleisrequiredfor

FAAS[3].Asatomisationofthesampleoccurs indiscretesteps,selectiveremovalof

matrixcomponentscanbeperformedwiththeaidofchemicalmodifiers. Theuseof

chemical modifiers reduces the effect of chemical interferences arising from

components of the sample matrix, which would otherwise cause issues with

determiningtheconcentrationoftheelementofinterest[4].

ThemaindisadvantagesofGFAASare its lowdynamic range (102),whichmeans

that the concentrations of the analytemust be keptwithin a narrow range, and the

amount of time that it takes to analyse the samples (> 5 min per sample) [4]; the

lengthof time taken toanalyseonesample isaparticulardisadvantage ifmore than

one element is to be studied. ICP-MS has the advantage, over GFAAS, of greater

sensitivity(partspertrillionforsomeelements),agreaterdynamicrange,andrapid

multi-elementdetection;however, thereareanumberof limitationsassociatedwith

ICP-MS. These includehigh instrumentcost, spectral interferences,andhighsample

volume; the latter being a crucial determining factor for the analysis of cell

samples.Withtheseadvantagesanddisadvantagsinmind,GFAASwasselectedasthe

mostappropriatemethodbywhichtoexaminetheuptakeofBibyHCT-8cellsinthe

presentstudy.

3.1.2 Microprobe synchrotron radiation X-ray fluorescence imaging for

studyingintracellularmetallocalisation

WhileSection3.1.1describedhighlysensitivetechniquesforthequantificationof

metal/metalloidswithinbiologicalsamples,theaforementionedtechniquescannotbe

used to determine the cellular localisation and consequently potential intracellular

targets of the metal/metalloid. Microprobe SRXRF spectroscopy (also known as


115

microprobe synchrotron radiation-induced X-ray emission (SRIXE) spectroscopy) is

ananalyticaltechniquethatcanprovideimagesofintracellularmetallocalisationand

colocalisation information with respect to other elements. Techniques such as

microprobe SRXRF spectroscopy remove the need for labeling the drug, due to the

direct analysis of the element of interest [5-8]. A number of studies have used

traditional staining and microscopy techniques such as autometallography (which

involves silver enhancement of bismuth sulfide/selenide clusters) to determine the

subcellular location of Bi following the treatment of various cell lines [9, 10],

animal[11-15] and human tissues [16]. However, microprobe SRXRF offers the

advantageofmulti-elementdetectionmappingandquantification [5-8]. This allows

researchers to study the changes in distribution and semi-quantitatively determine

fluctuations in concentrations of endogenous elements (such as Ca, Fe, P, S and Zn)

thatareinvolvedinimportantcellularprocesses,andhencetheeffectsofthedrugon

thedistributionoftheseelements[5-8].

Figure 3.1 illustrates the basic principle of XRF with respect to the Bohr atom

model[17].Ejectionofacoreshellelectron(generallyaKorLshellelectron)froman

atomoccurswhenphotonsofanX-raybeampossessahighenoughenergy(i.e.greater

thanthebindingenergy[18];inthehardX-rayregion,>1keV)toexcitetheelectron.

Theejectionof theelectron leavesavacancy in theshell. Anelectron fromanouter

orbitalcanthenfillthevacancyandrestorestabilitytotheatom.Theprocessresults

Figure 3.1: Bohr atom model illustrating the basic principle of X-ray fluorescence.(a)AnX-ray possessing energy that is equal to or greater than the binding energy of theelectronresultsinejectionofthatelectronintothecontinuum.(b)Anelectronfromanoutershell fills the vacancy left by the ejected electron resulting in the emissionof a fluorescencephotonwithanenergythatisequaltothedifferenceintheenergyofthetwoshellsassociatedwiththetransition.AdaptedfromFahrni[17].


116

intheemissionofaphotonwithenergyequaltothedifferenceinbindingenergiesof

the two shells involved in the transition. The binding energy of each electron is

proportional to the squarednuclear charge; for example, the analysis of Bi requires

energy of 10.7keV (the binding of the Lα2 electrons). In addition, every element

possessesuniquebindingenergiesandcharacteristicphotonemissions(fluorescence)

enablingdetectionof therelevantelements inthesample. Furthermore, therelative

quantities of the elements can be determined since the fluorescence emission is

directlyproportionaltotheamountofanelementpresentinasample.

TheuseofX-raysproducedatasynchrotronoffersanumberofadvantagesover

conventional bench-top X-ray sources. These include the capability of generating

X-raysup to12ordersofmagnitudebrighter thana conventionalX-raysource [19],

andfollowingtheadvancesinmodernX-rayoptics,theabilitytofocustheX-raystoa

submicronspotonasamplewiththeuseofdevicessuchasaKirkpatrick-Baezmirror

[20]orFresnelzoneplate[21].Bothofthesefactorshavefacilitatedthedevelopment

ofmicroprobe SRXRF spectroscopy. Detection sensitivities in the ppb range can be

routinely obtained from using microprobe SRXRF, in comparison to the ppm

sensitivities of alternative microprobe (particle induced X-ray emission) and

nanoprobe(electrondispersiveX-rayanalysis)techniques. Theimprovedsensitivity

ofmicroprobe XRF stems from the absence of the continuum background radiation

that isa featureof thespectraobtained fromtheuseofchargedparticlebeams,and

the increased X-ray flux on the sample associated with the use of third generation

synchrotronfacilities([22]andreferenceswithin).

TheachievementofsubmicronresolutionwithmicroprobeSRXRFcombinedwith

thehighsensitivityandpenetrationdepthofhardX-raysmakesthetechniquehighly

amenable to the study of single mammalian cells [22]. Raster scanning produces

quantitative information froman entire scanned area of a specimen resulting in the

generation of elemental maps. Depending on the particular element, detection

sensitivities have been reported in the range of 10−17−10−14 g for mammalian cell

studies([22]andreferenceswithin).MicroprobeSRXRFhasbeenappliedtothestudy

oftheintracellularfateofbothmarketedandpotentialmetal-baseddrugscontaining

Pt [6,23,24],Cr [5],Ru [8,25], Se [26,27],Gd [28,29], andAs[30] inanumberof

cancer cell lines. However, to date there are few, if any, reports assessing the

distribution of Bi-based drugs in human cells using microprobe SRXRF

microspectroscopy.


117

3.1.3 Chapteraims

HavingestablishedthestabilityoftheBiNSAIDs,Bi(difl)3,Bi(mef)3andBi(tolf)3,in

cellmedium,andthetoxicityoftheBiNSAIDstowardsHCT-8cells,itwasofinterestto

quantify the cellular Bi content following treatment with BiNSAIDs to determine if

therewereanycorrelationsbetweencellularuptakeversusstabilityand/or toxicity.

AssessingthesubcellulardistributionofBiwasofinterestinordertoprovideinsight

intotheintracellulartargetsoftheBiNSAIDs,Bi(difl)3,Bi(mef)3andBi(tolf)3,inHCT-8

cells.ThespecificaimsofthisChapterwereto:

1. Quantify Bi uptake by the HCT-8 cells using GFAAS in order to determine

whetherBiuptakecorrelateswithBiNSAIDstabilityandtoxicity.

2. Ascertain the subcellular location of Bi in HCT-8 cells following treatment

with BiNSAIDs using microprobe SRXRF imaging to assess how this might

influencetoxicitytowardsthecancercellsandhowthismight influencethe

targetingofCOX.


118

3.2 Materialsandmethods3.2.1 Materials

ThematerialsusedforthecellcultureexperimentsdescribedinthisChapterhave

been previously described in Section 2.2.1. The chemicals, ammonium acetate

(≥98%),sodiumchloride(99.999%tracemetalbasis),andtrypanblue(0.5%inPBS)

werepurchasedfromSigma-Aldrich(StLouis,USA). Pyronegdetergentwassourced

fromDiverseyLeverAustraliaPtyLtd(Smithfield,Australia).Analyticalgradeethanol

(absolute, ≥99.5%) and nitric acid (68−70%) were sourced from Ajax

Finechem(Seven Hills, Australia). Acertified Bi standard solution (1004±3mg/mL,

2% HNO3 v/v) was acquired from PerkinElmer Instruments (Massachusetts, USA),

whilethemagnesium(Mg(NO3)2,10mg/mL)andpalladiummatrixmodifiersolutions

(Pd(NO3)2, 5mg/mL)were procured from Astral Scientific (Taren Point, Australia).

Analytical volumetric flasks (Class A) were purchased from GlassCo (Dandenong

South, Australia). Tracepur nitric acid (69%)was sourced fromMerck (Darmstadt,

Germany). A Bi hollow cathode lamp and standard transversely heated graphite

atomizer tubes(insertedpyrolyticL’vovplatform)wereobtained fromPerkinElmer

Instruments (Victoria, Australia). Glutaraldehyde (25% solution), Spurr's low

viscosityembeddingresinandtoluidineblueweresourcedfromProscitech(Kirwan,

Australia). Siliconnitride (Si3N4)membranes (frame size, 5×5mm2; sample surface

area,1.5×1.5mm2;membranethickness,500nm;andframethickness,200nm)were

manufacturedbySilsonLtd(Northampton,England).

3.2.2 Cellcultureprocedures

HCT-8cellsweresourcedandemployedasdescribedinSection2.2.3.

3.2.3 GFAASuptakestudies

3.2.3.1 Preparationofcelldigests

Cell digestswere prepared for GFAAS analysis using a similar procedure to that

documentedpreviously [23]. Briefly,HCT-8 cells (2 × 106 cells, 5mL)were seeded

into60-mmcellculturedishesandincubatedfor24h,afterwhichthegrowthmedium

was removed and the cells were washed twice with PBS. The control cells were

incubatedwith vehicle (5mLRPMI-1640, 2%DMSO v/v) for 24 h,while treatment

solutionsofBi(difl)3,Bi(mef)3,Bi(tolf)3orBSSwereaddedtotheotherdishes(15µM

or 40 µM Bi in 5mL RPMI-1640, 2% DMSO v/v) for 24 h. The two treatment


119

concentrationswereapproximatelyequaltothetwolowestIC50valuesobtainedfrom

the toxicity studies (Section 2.2.4). Triplicate disheswere prepared for the control

andeachtreatment.Followingthetreatmentperiod,thesolutionswereremovedand

thecellswerewashed twicewithPBS(2mL),harvestedwith trypsin(0.25%,2mL)

andcentrifuged(300g,5min).ThecellswereresuspendedinRPMI-1640(1mL)and

trypan blue exclusion assays were performed by removing the resultant cell

solution(50µL) fromeachsampleandmixingwith trypanblue(50µL,0.5%w/v in

PBS). The populations of intact cells and disrupted cells were scored using a

haemocytometer(100×magnification),wherebycellsthatappearedbluerepresented

uptake of trypan blue and loss ofmembrane integrity (dead cells) while those that

werecolourlessrepresentedcellswithanintactmembrane(healthycells). Thecells

inthemainsuspensionswerewashedtwicewithsterilePBSsolution(2mL),followed

by twowashes with sterile saline solution (0.9%, 2mL). The final saline solutions

were drained from the pellets and the cells were freeze-dried for 2h. Nitric acid

(Tracepur,69%v/v)wasaddedtoeachEppendorftube(43µL)andthepelletswere

digested overnight. The digested pellets were transferred to volumetric flasks

(prewashedwithasolutionofPyronegdetergent(72h),followedbynitricacid(72h),

andMilliQ water(72 h) and rinsed 5times withMilliQ water) and the volumewas

adjusted to5.00mL (usingMilliQwater)or25.00mL (usingMilliQwaterandnitric

acid(177µL,Tracepur,69%v/v)),toobtainafinalHNO3concentrationof0.6%v/v.

3.2.3.2 Preparationoffixed/dehydratedcelldigests

It has been previously reported that formalin fixation can causemetal and trace

element leaching from tissues [31]. In this study, the integrity/effect of the

glutaraldehyde fixation and ethanol dehydration process used for preparation of

thin-sectioned cell samples for microprobe SRXRF analysis was investigated to

determine if there was any change in Bi uptake. As such, the cell samples were

preparedasoutlinedintheprevioussectionbutwiththefollowingchanges.Afterthe

final PBS wash of the treated cells, the cells were resuspended in glutaraldehyde

(1mL,2%v/vinPBS)andfixedfor2hatroomtemperature.Dehydrationofthecells

was performed incrementally using increasing ethanol concentrations (in MilliQ

water; 30% ×2, 50% ×2, 70% ×2, 80% ×2, 90% ×2, 95%×2, 100% ×4) with

centrifugation,andremovalofthesupernatantandcellresuspensionaftereachstep.

The pellets were then freeze-dried after the final ethanol wash and digested as


120

describedinSection3.2.3.1.

3.2.3.3 Graphitefurnaceatomicabsorptionspectroscopy

Digested cell solutions were analysed for Bi using a PerkinElmer AAnalyst 600

atomic absorption spectrometer equippedwith theWinlab 32 AA Furnace program

incorporating Zeeman-effect background correction, an AS-800 autosampler and an

AAAccessory cooling system. ABihollowcathode lamp (PerkinElmer)wasusedat

223nm, using a slit width of 0.7nm, to give an instrument energy reading of

approximately 35W. A certified Bi standard solutionwas employed to prepare the

stock Bi standard (100µg/L, 0.6% HNO3v/v) for the calibration curve (correlation

coefficients: 0.993−0.997). The autosampler volumes consisted of the chemical

modifier(5µL,0.005mg/mLPd(NO3)2and0.003mg/mLMg(NO3)2), followedbythe

standard or sample (20µL). Bi was atomised from the surface of the pyrolytic

graphite-coated tubes using the furnace operating conditions described in

Appendix2.1. Theanalytereadingwasperformedduringatomisation(1700°C)and

Argaswassuppliedtotheinstrumentduringanalysis.Afteranalysisofeachdigested

cell sample, a blank (containing 0.6% HNO3v/v) was analysed to minimise

interference from previous samples and ensure that Bi detection returned to

backgroundlevels.

3.2.4 MicroprobeSRXRFanalysisofBi-treatedcells

3.2.4.1 Preparationofthin-sectionedcellsamples

Thin-sectioned cells were prepared for microprobe SRXRF analysis using

procedures documented previously [5, 6, 32, 33]. Briefly, cells were seeded in cell

cultureflasks(75cm2)andgrownto80%confluence.TreatmentsolutionsofBi(difl)3,

Bi(mef)3,Bi(tolf)3orBSS(40µMBi,10mLRPMI-1640,2%DMSOv/v)wereappliedto

the cell culture flasks, while control cells were incubated with vehicle (10mL

RPMI-1640, 2%DMSO v/v) for 4 h or 24 h. The treatment concentration of the

complexes was chosen based on the uptake results obtained from the toxicity and

GFAASexperiments(Sections2.3.2and3.3.1)toensurethatdetectableconcentrations

of Bi (suitable for microprobe SRXRF imaging) were present in the cells following

treatment.ThecellswerewashedtwicewithPBSandharvestedusingtrypsin(0.25%

inPBS).TheharvestedcellswerewashedtwicewithsterilePBSusingcentrifugation

andthe finalcellpelletwas fixed inglutaraldehyde(1mL,2%v/v inPBS) for2hat


121

room temperature; and dehydrated in ethanol (30% ×2, 50% ×2, 70%×2, 80% ×2,

90% ×2, 95% ×2, 100% ×4). Infiltration was achieved by rocking the cells in

50%ethanol/50% Spurr’s resin overnight and then replacing with 100% Spurr’s

resin(×2over2days)toensuretheremovalofall tracesofwater. Onthe finalday,

thecellswereembeddedin100%Spurr’sresin,centrifuged(8000g,1h)andcuredat

60 °Covernight in aBeemcapsule. Thin sections (1µm)of the cell pelletwere cut

using an ultramicrotome (Leica EM UC7). The thin-sections were stained with

toluidinebluebeforemountingonsiliconnitridemembranes.

3.2.4.2 MicroprobeSRXRFanalysisofthin-sectionedcells

Siliconnitridemembraneswereattachedtokinematicmountsdesignedtofitboth

thelightmicroscope(LeicaDMXREEpi-fluorescence/visualmicroscope)andtheX-ray

microprobe(2-ID-D).OpticalmicrographsandXYcoordinateswereobtainedinorder

tolocatesuitablecellsforanalysis.Thosecellsthatexhibitedanintactappearance,a

clearly defined nucleus and were located away from other cells and cellular debris

werechosenforanalysis.

Microprobe SRXRF was performed at beamline 2-ID-D at the Advanced Photon

Source,ArgonneNational Laboratory (Lemont, IL,USA). The13.45keVX-raybeam

wasfocusedtoaspotsizeof0.3×0.3µm2usingtwozoneplates. Thesampleswere

housedinaheliumatmosphereandthedistributionsoftheelementsfromP(2.1keV)

toBi(13.45keV)weremappedbyrasterscanningthesamplein0.3µmsteps,withthe

aidofaXYZmotorisedstage.FluorescentX-raysweredetectedusingadwelltimeof

2.5s/pt and an energy dispersive Vortex EM (Hitachi High-Technologies Science

America, Northridge, CA) silicon drift detector. Elemental distribution maps were

processedusingMAPSVersion1.7.3.02softwarepackageprovidedbyDrStefanVogt

(AdvancedPhotonSource) [34]. Conversionof elemental fluorescence intensities to

absolute densities in microgram per square centimetre(µg/cm2) was performed by

comparing X-ray fluorescence intensities with those from thin film standards,

NBS-1832andNBS-1833(NIST,Gaithersburg,MD,USA).Theelementalquantification

ofspecificregionsoftheelementaldistributionmaps(i.e.thenucleusandcytoplasm)

was performed by selecting the region of interest (ROI) and extracting the

quantificationinformationfortheROIusingMAPSVersion1.7.3.11softwarepackage

providedbyDrStefanVogt(AdvancedPhotonSource)[34].


122


Statistical analysis of the results of the trypan blue exclusion assays, the GFAAS

assays and the microprobe SRXRF elemental quantification were performed using

one-way ANOVA, followed by the Tukey-Kramer Multiple Comparison Test, using

GraphPadPrism,Version5.04 forWindows(GraphPadSoftware,La Jolla,USA). The

trypan blue exclusion assay and GFAAS results were analysed as two treatment

sets(i.e.controland15µM,orcontroland40µMtreatmentsamples)inordertodraw

astatisticalcomparisonatthesameBitreatmentconcentration.Resultsarepresented

asmean± standarddeviation (s.d.). A valueofP≤0.05was considered statistically

significant.


123

3.3 Results3.3.1 GFAASanalysisofcellularBi

Figure 3.2 shows the results of the trypan blue assay and the corresponding Bi

detected in cells treated with the vehicle, BiNSAIDs or BSS (15 µM and 40 µM Bi).

Thetrypanbluepositivecells(Fig.3.2a)formedarelativelylowproportion(<1%)of

the totalnumberofcontrolandBiNSAID-treatedcells, indicating that themembrane

integrityofthemajorityofthecellswasuncompromised.Thisisanimportantresult

becauseitestablishesthatanycellularBidetected(Fig.3.2b)isnottheresultofrapid

uptakeduetolossofcellmembraneintegrity.Notably,significantdecreasesinthecell

populationswereobservedfollowingtreatmentwithBi(mef)3(40µMBi,P≤0.01)and

Bi(tolf)3 (15 µM Bi, P ≤ 0.01; 40 µM Bi, P ≤ 0.001, Fig. 3.2a). The number of

Bi(tolf)3-treated cells (15 µM Bi) was approximately 45% of the number of control

cells and further decreased to approximately 20% with increased Bi treatment

concentration(40µMBi,Fig.3.2a).Thecellpopulationobservedfollowingtreatment

with Bi(mef)3 (40 µM Bi, 24 h) was approximately 43% of the population of

vehicle-treatedcells(Fig.3.2a).ThelowercellpopulationsobservedforBi(mef)3-and

Bi(tolf)3-treatedcells(40µM,Figure3.2a)wereconsistentwiththeresultsoftheMTT

assay(Fig.2.6,Table2.4)thatshowedthatBi(mef)3andBi(tolf)3werethemoretoxic

complexes.

In the treatmentgroupexposed to15-µMBi, thecellularBi (Fig.3.2b)wasmore

than three times higher in Bi(tolf)3-treated cells than in the Bi(difl)3- and

Bi(mef)3-treatedcells.Interestingly,BSS-treatedcellsexhibitedthehighestamountof

cellular Bi (three times higher than Bi(tolf)3), and displayed the only statistically

significant (P≤ 0.001) increase above the control for the 15-µM treatment group.

Increasingthetreatmentconcentrationfrom15µMto40µMresultedinanincreasein

cellular Bi content in all Bi-treated cells (Fig. 3.2b). Treatmentwith themost toxic

BiNSAID, Bi(tolf)3 (40 µM, 24 h), showed the highest Bi concentration per cell

(P≤0.001 compared to the control, and the other Bi treatments), which was

approximately 70 times higher than the cellular Bi associated with Bi(difl)3- and

Bi(mef)3-treated cells (40 µM, 24 h) and approximately 340 times higher than the

cellular uptake of Bi(tolf)3-treated cells at 15µM. The cellular uptake of Bi(difl)3-,

Bi(mef)3-, and BSS-treated cells increased only approximately 9, 18 and 5 times,

respectively, incomparisontocellstreatedwith15-µMofthecompounds(P>0.05).

Interestingly, despite the differences in the toxicities of the 40-µM treatments with


124

Bi(difl)3orBi(mef)3(whereby40-µMrepresentsasub-toxicconcentrationandtheIC50

value,respectively;Fig.2.6,Table2.4),Bi(mef)3-treatedcellsonlyexhibited1.5times

thecellularBiconcentrationoftheBi(difl)3-treatedcells.

Figure 3.2: Trypan blue exclusion assay and GFAAS analysis of HCT-8 cells treatedwith Bicompounds. (a)Cellpopulationsobtained fromtrypanblueexclusionassaysofvehicle-andBi-treatedcells(15and40µM,24h). Resultsareexpressedasmean± s.d. (n=3replicatesfrom one experiment). Error bars represent the standard deviation from the mean of thenegative and positive trypan blue cells. (b) GFAAS detection of Bi in HCT-8 cells followingexposure to vehicle (2%DMSOv/v) and specified Bi compounds (15 μM and 40 µM) for24h.Resultsareexpressedasmean±s.d.(n=3replicatesfromoneexperiment).Statisticalsignificancewithrespecttothecontrolisindicatedby**P≤0.01or***P≤0.001.


125

3.3.2 ThedependenceofcellpreparationonthecellularuptakeofBi

Figure3.3shows theresultsof theBicelluptakestudiesperformedoncells that

werepreparedusingthefixationanddehydrationprotocolemployedforpreparation

ofthin-sectionedcellsforthemicroprobeSRXRFanalysescomparedwiththeprotocol

typicallyusedforGFAASanalysis. Therewerenosignificantdifferences(P>0.05)in

the amount of cellular Bi detected after fixation and dehydration compared to the

analogous non-fixed and non-dehydrated cells. This is an important result as it

establishes that there is no loss of intracellular Bi due to leaching following

preparation of thin-sectioned cells for microprobe SRXRF analysis. Conversely,

samples treatedwithBSS showedadecrease in the amount ofBi containedper cell

compared to the normally washed cells; however, this result was not statistically

significant(P>0.05).

Figure 3.3: GFAAS detection of Bi in HCT-8 cells following exposure to the specified Bicompounds(40μM,24h)preparedbynormalwashingmethodorpreparedfollowingfixationwithglutaraldehyde(2%inPBS,2h)anddehydrationwithethanol.Resultsareexpressedasmean±s.d.(n=3replicatesfromoneexperiment).


126

3.3.3 MicroprobeSRXRFspectroscopy

Figure 3.4a shows the elemental maps for P, S, Zn, Ca, and Bi generated from

microprobeSRXRF imagingof a thin-sectionprepared froma typical vehicle-treated

HCT-8 cell. The concentration range of each element detected within the scanned

region is specified above each elemental map (Fig. 3.4 and Appendix 2.2). High

concentrations of the endogenous elements, P, S, Zn, and Ca, were detected in the

vehicle-treated cells (Fig.3.4a). Inspection of the micrograph (top panel, Fig. 3.4a)

reveals theoverallsizeandshapeof thecell,whichcorrelateswellwithendogenous

elementsS,P,ZnandCa.ProteinsfoundinboththenucleusandcytoplasmcontainS;

additionally, S is also associated with the toluidine blue stain (which targets DNA),

whichwasnecessarytolocatethecellsforthemicroprobeSRXRFimagingduetotheir

translucent appearance. The position of the nucleus is also defined by the high

presenceofPandZn;PisafundamentalelementinthebackboneofDNA[35],while

Znplaysan importantrole inDNAbindingproteins(knownaszinc fingers) thatare

localisedinthenucleus[36].DistinctlydenseareasofP(observedasthedenseregion

of green/red/yellow pixels) inside the nuclear membrane can be attributed to the

densely packed DNA regions known as the heterochromatin [37]. In addition, the

dense spherical structure in the top left of the nucleus (Fig. 3.4a) is the nucleolus

(observable in the S andZnmaps as yellow/redpixels). Both, the heterochromatin

andthenucleoluscorrelatewiththedarkstainedregionsinthemicrographimagethat

representdenseDNAmaterial.Asobservedinthevehicle-treatedcells,Caappearsto

beevenlydistributedthroughoutthecytoplasmandthenucleus(Fig.3.4a).Onlyvery

lowconcentrationsofBiwerepresentwithinthescannedregionofthevehicle-treated

cells (maximum Bi 0.00431μg/cm2, Fig. 3.4a, and 0.00469μg/cm2, Appendix 2.2a),

andarenotdistinctlyassociatedwiththecellsagainstthebackgroundregions.

Figure 3.4b shows the elemental map of typical Bi(difl)3-treated cells following

24-h treatment. The Bi(difl)3-treated cell samples showed evidence of intracellular

accumulation of Bi in ‘hot spots’ (areas of high Bi concentration) distributed

throughout the cytoplasm (Fig. 3.4b, bottom panel). The Bi(difl)3-treated cells

(Fig.3.4bandAppendix2.2b)containedseveralhotspotsrangingfromapproximately

0.3−1.5μm2insize.ThemaximumcellularBidetectionwas0.0578μg/cm2(Fig.3.4b)

and0.106μg/cm2(Appendix2.2b).


127

Figure3.4:MicroprobeSRXRFelementalmapsforthin-sectionedHCT-8cells following24-htreatmentwith:(a)vehicleonly(DMSO2%v/v);(b)Bi(difl)3(40μM);(c)Bi(mef)3(40μM);(d)Bi(tolf)3(40μM);and(e)BSS(40μM). Thedetectedelementsarespecifiedtotheleftofeachrow.Operatingconditionsinclude:beamenergy=13.45keV;beamsize=0.3×0.3μm2;step size = 0.3 μm; dwell time = 2.5 s/pt; and scan dimensions (H × V) = (a) 20 × 17μm2;(b)19×16μm2;(c)22×15μm2;(d)31×21μm2and(e)14×12μm2.


128

Figure 3.4c shows the elemental map of typical Bi(mef)3-treated cells following

24-htreatment. Thereweresubstantialdifferencesobservedinthedistributionand

maximumBidetectedbetweenCellAandCellB(Fig.3.4c).ThemaximumBidetected

in CellA was 0.128μg/cm2, twice the amount detected in CellB (maximum

0.06136μg/cm2).Additionally,BiwasdistributedthroughoutCellA,withthehighest

concentration of Bi localised in the nucleus and considerable Bi also present in the

cytoplasm(Fig.3.4c). It shouldbenotedthat thiswas theonlyBiNSAID-treatedcell

where theBi localisationwasobserved inboth thenuclearandcytoplasmic regions.

ThemajorityoftheBipresentinthecytoplasmofCellBappearstobeconcentratedin

one large hot spot, approximately 5.8μm2 in size (Fig. 3.4c). A similar result was

observed in a third Bi(mef)3-treated cell (24h), whereby one large hot spot was

observed in thecytoplasm(Appendix2.2c). Notably, therewas littleevidenceofBi

accumulationwithintheBi(mef)3-treatedcells following4-htreatment(maximumBi

0.0065μg/cm2and0.0134μg/cm2,Appendix2.3).

The 24-h Bi(tolf)3-treated cell samples also showed evidence of intracellular

accumulationofBiinhotspotsdistributedthroughoutthecytoplasm.Themaximum

Bi detected in Bi(tolf)3-treated cells (Fig. 3.4d) was 0.0496μg/cm2 (Cell A) and

0.0408μg/cm2 (Cell B), 10times greater than the amount of Bi detected in vehicle-

treated cells (Fig. 3.4a). A third Bi(tolf)3-treated cell (Appendix 2.2d) contained the

largest amount of Bi of all the BiNSAID-treated cells analysed (1.01 μg/cm2),which

appearedtobelocalisedinonecytoplasmichotspot.TheBihotspotsrangedinsize

fromapproximately0.7μm2to2.8μm2intheBi(tolf)3-treatedcells.

Figure3.4eshowsthattherewasnoevidenceofaccumulationofBiincellstreated

withBSS(24h).ThemaximumBicountsdetectedwere0.005μg/cm2(Fig.3.4e)and

0.006μg/cm2 (Appendix 2.2e), similar to the Bi content of the vehicle-treated cells

(Fig.3.4aandAppendix2.2a).

AcrossallBiNSAID-treatedcells,manyoftheBihotspotsshowedcloseproximity

to the nucleus; the colocalisation maps with P and Zn, confirmed the spots were

located close to thenuclearmembrane in a numberof cells (Appendix2.4 a(i), b(i),

b(ii)andc(i)). Analysisof themicroprobeSRXRFelementalmapsalsorevealedthat

theCadistributionofalltheBi-treatedHCT-8cells(Fig.3.4b-eandAppendix2.2b-e)

differed compared to the vehicle-treated cells (Fig.3.4a and Appendix 2.2a).

Forinstance, in the case of all Bi-treated cells (Fig. 3.4b-e and Appendix 2.2b-e),

cellularCaconcentrationappearedhigherthanthevehicle-treatedcells(Fig.3.4aand


129

Appendix 2.2a) as evident in the Ca range (maximum quantity detected in vehicle-

treated cells ranged from 0.027−0.0344 μg/cm2, Bi-treated cells 0.0507−0.371

μg/cm2)andthehigherdefinitionintheCamaps.

Theelementalquantificationsof thecytoplasmicandnuclear regionsof the thin-

sectionsobtainedfromvehicle-,BiNSAID-andBSS-treatedHCT-8cellsarepresented

inFigure3.5.Sincequantificationoftheentirecellularcontentsisnotpossiblefroma

thin-section, the quantifications are represented per unit area (μm2). Statistical

analysisshowsthattherewerenosignificantdifferencesinthecytoplasmicornuclear

P concentrations in the vehicle-treated cells compared to the Bi-treated cells

(Fig.3.5a).TheconcentrationofZnwassignificantlyelevatedinthecytoplasmofthe

Bi(difl)3-treated cells, and the nucleus of the Bi(mef)3-treated cells compared to the

vehicle-treated cells (P ≤ 0.05 and P≤0.01, respectively); however, no significant

changeswereobservedintheZncontentoftheBi(tolf)3-treatedcellscomparedtothe

vehicle-treatedcells(Fig.3.5b).

TheconcentrationofCawassignificantly increased in thecytoplasmandnucleus

of the Bi(mef)3-treated cells (both P ≤ 0.001) compared to the vehicle-treated cells

(Fig.3.5c).TherewerenosignificantfluctuationsinCainthecytoplasmofanyofthe

otherBi-treatedcells(Fig.3.5c).AnincreaseinCa(1.4–1.8fold)wasobservedinthe

nucleusoftheotherBi-treatedcellgroups(Fig.3.5c),althoughthisincreasewasonly

significantintheBi(difl)3-treatedcells(P≤0.05). NegligiblelevelsofintracellularBi

weredetectedinthecytoplasmandnucleusofthevehicle-andBSS-treatedcells(Fig.

3.5d); however, the BiNSAID-treated cells showed large fluctuations in the

intracellularlevelsofBiinboththecytoplasmandthenucleus.


130

Figure3.5:Quantificationsof:(a)P;(b)Zn;(c)Ca;and(d)Bi,inthecytoplasmicandnuclearregions of the HCT-8 thin-sectioned cells following 24-h treatment with vehicle(2%DMSOv/v),orthespecifiedBicompounds(40μM).Alldatawerecollectedusinga0.3μmstepsizeanda2.5sdwelltime.Dataarepresentedasmean±s.d.wheren=3(exceptBi(tolf)3,n=4,andBSS,n=2). Significancewithrespecttothevehicle-treatedsampleisindicatedby*P≤0.05, **P ≤0.01, ***P ≤0.001. The cytoplasmandnucleuswereanalysedas separategroups.


131

3.4 Discussion3.4.1 GFAASuptakestudies

QuantitativestudiesofthecellularuptakeofBiresultingfromtreatmentofHCT-8

cellswithBiNSAIDswereperformedtodeterminewhethertherewasanycorrelation

between Bi uptake, and the stability and/or toxicity of the BiNSAIDs (results

presentedinChapter2).Itisimportanttonotethatthetrypanblueexclusionassays

(Fig.3.2a)showedthatthemembraneintegrityremainedintactinthemajorityofthe

cells;therefore,uptakeoftheBiNSAIDsmostlikelyoccursthroughanuncompromised

cellmembrane. Inorder to rationalisehow theBiNSAIDsenter theHCT-8 cells, the

documentedmechanisms bywhich Bi(III) complexes and NSAIDs entermammalian

cellsmustalsobeconsideredinthefollowingdiscussion.

The results from the GFAAS studies showed that 24-h treatment of HCT-8 cells

with 15-μM or 40-μM Bi(tolf)3 resulted in substantially higher Bi uptake compared

withtheanalogousconcentrationsoftheotherBicomplexes(Fig.3.2b).Importantly,

the results from the NMR stability studies (Section 2.3.1) of Bi(tolf)3 also indicated

thatBi remains coordinated to tolf in the cellmedium for the duration of the 24-h

treatment. Accordingly, the Bi uptake associated with Bi(tolf)3 treatment must be

facilitatedbytheNSAIDortheoveralllipophilicityofthecomplex.Thereareatleast

twoknownmechanismsofNSAIDdiffusionacross themembranesof thecells in the

upper and lower intestines. The first is the monocarboxylate/H+ transporters [38]

while the second is a pH-dependent, but non-carrier mediated absorption of

NSAIDs[39]. ThefirstmechanismishighlyunlikelyfortheintactBi(tolf)3asthetolf

carboxylate necessary for interaction with monocarboxylate/H+ transporters is

unavailableduetothecoordinationtoBi.Giventhatthecomplexlikelyremainsintact

inthecellmedium,itismorelikelythatthelipophilicityofthecomplex,impartedby

the coordinated tolf, is facilitating interaction and diffusion through the cell

membrane. Apassive transportmechanism forBi compoundshasbeenobserved in

recentstudiesperformedbyotherresearchers.Hongetal.[40]haveshownthatHK-2

humanproximaltubularcellstreatedwithCBS(0.5mM)over24h,exhibitedapattern

of Bi saturation that was consistent with the profile of passive transport. More

recently, neutron reflectometry experiments performed by Brown and Dillon ([41]

andunpublishedwork)showedthatBi(tolf)3interactedwiththeheadgroupsoflipid

bilayer membrane mimics and partitioned into the tail region of the membranes

(comprised of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) molecules).


132

While these experiments were performed in a non-cell system, they showed that

Bi(tolf)3 was able to displace lipid molecules, which provides credence to the

hypothesis that passive diffusion is a potential uptake mechanism employed by

Bi(tolf)3topassthroughthecellmembraneandaccumulateinHCT-8cells.

The much lower Bi uptake associated with the Bi(mef)3- and Bi(difl)3-treated

cells(Fig.3.2)canprobablybeattributedtothedifferingstabilityandreactionsofthe

complexes in the cell medium as observed by NMR spectroscopic

investigations(Section2.3.1).Giventhestructuralsimilaritiesinthetolfandmef,one

would expect that Bi uptake would be similar for both Bi(tolf)3 and Bi(mef)3

treatment.However,thelowerBiuptakeisconsistentwiththeBi(mef)3NMRstability

studies (Section 2.3.1) that indicated that mef dissociates from Bi. As such, the Bi

uptakewouldnotbefacilitatedbythelipophilicitythatwouldbeimpartedbymefand

itssubsequentinteractionswiththemembrane.

Interestingly, Bi(mef)3-treated cells only exhibited 1.5 times greater Bi uptake

compared with the Bi(difl)3-treated cells despite the more substantial difference in

toxicityof 40-μMBi(mef)3-andBi(difl)3-(approximately50%versus20%reduction

of MTT conversion, respectively (Section 2.3.2). The NMR results indicated that

glucosemayinteractwiththeBi(difl)3complex.Itmightbepossiblethatglucosemay

facilitate uptake of any Bi(difl)2(glu) or Bi(difl)(glu)2 species that are formed.

Unfortunately,giventhattheglucosyl1HNMRsignalsonlyrepresentedapproximately

1/6 that of the difl signals, one would not expect a significant increase in the

detectableBifoundinthecells.

AnotherobservationfromthisstudywasthehighBidetectedfortheBSS-treated

cells. For instance, the cellular Bi was approximately 4−10 times higher than all

BiNSAID-treatedcellsforthe15μMtreatment,andapproximately4timeshigherthan

Bi(mef)3 and Bi(difl)3 for the 40 μM treatment. It is proposed that the increased

cellularBiassociatedwiththeBSS-treatedcells(ascomparedtotheBiNSAID-treated

cells) was due to extracellular bound Bi that could not be completely removed by

washing the cells using the protocol employed for GFAAS cell sample preparation

(Section 3.2.3.1). Inspection of the treatment dishes revealed that BSSwas far less

solublethantheBiNSAIDsasnotedbytheprecipitationinthetreatmentmedium.The

adherenceofmetalcomplexestotheexteriorofthecellmembranehasbeenobserved

in other studies [42]. Specifically, it was observed that during the preparation of

Cr(NO3)3.9H2O-treated A549 lung cancer cell samples for GFAAS analysis, the cell


133

pellets appeared blue in colour after several washes with PBS, followed by sterile

saline[42].WhenGFAASanalysiswasperformedononlytheintracellularcontentsof

theA549cells, theCr contentof theCr(NO3)3.9H2O-treated cellswas reduced to the

samequantityastheotherCr(III)complexestested[42], indicatingthatthemajority

oftheCrwasassociatedwiththecellmembrane.Unfortunately,BSSiswhitesuchthat

itsappearanceinthecellpelletisnotaseasilydiscerned. Itisnoteworthy,however,

that the GFAAS analysis of the BSS-treated cells differed when the more extensive

washing, fixationanddehydrationprotocolused formicroprobeSRXRFanalysiswas

employed (Fig.3.3). In contrast, no changewasobserved in the cellularBi following

treatmentwith the solubleBiNSAIDs (Fig.3.3). In addition, themaximumBi counts

detected in BSS-treated thin-sectioned cells were similar to the Bi content of the

vehicle-treatedcells(Fig.3.4)indicatingtherewasnegligibleintracellularBi.

It is importanttonotethattheGFAASanalysesdonotprovideinformationabout

the mechanisms by which “unbound” Bi enters and accumulates in cells. While

beyondthescopeofthisstudy,someassumptionscanbemadebasedontheliterature.

Importantly, the mechanisms of uptake, accumulation and metabolism of Bi by

mammaliancellsarepoorlyunderstood.ThereisevidencethatsuggeststhatBibinds

stronglytotheenzymetransferrin(discussedinSection1.4.1.1),whichcycles inand

out of cells as a method of iron transport [43]. Acknowledging that transferrin

receptors are overexpressed in cancer cells [44-46] it is feasible that in the serum-

deficient (and Fe-deficient) cell medium used during treatment that Bi may be

substitutedintotransferrinforuptakeintothecells.Additionally,ithasbeenreported

that Bi has a high affinity for glutathione, a tripeptide that is present in cells in

relativelyhighconcentrations(0.5–7mM)andisbelievedtoplayaroleinBitransport

incellsandbiofluids [47]. Recently, the importanceof the roleofglutathione in the

intracellulartransportofBiwasexaminedinHK-2cellstreatedwithCBS,wherebyit

was apparent that glutathione depletion adversely affected the ability of the cell to

detoxify itselfofBi [40]. Assuch, it isplausible thatglutathionepresent in thecells

may assist in transporting unbound Bi (in the case of Bi(mef)3) in the HCT-8 cells.

Furthermore, if Bi(tolf)3 enters the cell intact, as indicated by the NMR stability

studies,itcouldbepostulatedthatcellularprocessesresultinthereleaseoftolffrom

Bi,wherebythemechanismofBireleasemaybearesultofthestrongaffinityofBifor

other more abundant intracellular ligands such as glutathione, metallothionein, or

other metal-binding proteins that are responsible for intracellular transport of


134

endogenousmetalions(suchasZn(II)andFe(III)).Theexactuptakeandmetabolism

pathwaysoftheBiNSAIDswillneedtobeexaminedinfuturestudies.

OnefinalconsiderationstemmingfromtheGFAASanalysisoftheBi-treatedcellsis

the impact resulting from the Bi accumulation in the cells. At the lower treatment

concentration (15 μM Bi, which reflects the IC50 value of Bi(tolf)3), the Bi uptake

followingBi(tolf)3 treatmentwasonly three timeshigher than thatobserved for the

other BiNSAID treatments. Treatment with the higher concentration (40 μM),

however,resultedinsubstantialBiaccumulationinthecellsinhighquantities.These

resultshighlightthequestionastowheretheBiislocalisinginthecellswithrespectto

themechanismsofactionoftheBiNSAIDs,asaddressedinSection3.4.2.

3.4.2 MicroprobeSRXRFspectroscopyofthin-sectionedcells

MicroprobeSRXRFwasutilisedasa sensitive technique todeterminesubcellular

location of Bi in thin-sectioned HCT-8 cells to glean some information regarding

mechanisms of action of theBiNSAIDs. The results from theGFAAS cellular uptake

studies(Section3.3.2)demonstratedthatthetotalcellularBiconcentrationremained

essentially thesame followingacomparisonof thecellpreparationproceduresused

for thin-sectioning versus normal washing procedures (with the exception of BSS-

treated samples; Section 3.2.3.1 and 3.4.1), suggesting that therewas no significant

difference in Bi leaching from the cells using the fixation procedures employed for

samplepreparationforthemicroprobeSRXRFimaging. However,itshouldbenoted

that it can be assumed that under those conditions Bi does not leach into any of

solutions used during the washing steps in the normal GFAAS sample preparation

method(Section3.2.3.1).

The consistent trend observed from the microprobe SRXRF imaging of thin-

sectionedcellswasthattheintracellularfateofBiwasessentiallythesame,regardless

oftheBiNSAIDadministered.InthemajorityofBiNSAID-treatedcellsdiscreteBihot

spots were detected in the cytoplasm; however, there was some evidence that Bi

mightaccumulate in (oraround) thenucleus. These findingsneed tobe considered

with respect to a number of potential chemotherapeutic targets including (i) the

nucleus,(ii)COX,and(iii)mitochondria,andpotentialdetoxificationmechanismsvia

accumulationin(iv)lysosomes.

Thefirstpotentialtargetfordiscussionisthecellnucleus,whichisrelevantdueto

thefactthatitisacommontargetforanti-canceragents.Itwasnotedthattherewas


135

disperse distribution of Bi throughout the cell nucleus in one BiNSAID-treated cell,

(Bi(mef)3-treated Cell A, Fig. 3.4c). Of the approximately twenty cells studied,

however, thisresultwasatypicalandraises thequestionofwhether thisoccurs ina

greater number of cells (if a larger sample size was studied). In other microprobe

SRXRF studies of the clinically-used drugs, arsenite [30] and cisplatin [23],

colocalisation of As and Pt (respectively) was evident in the nucleus of a high

proportionofthedrug-treatedcells,whichisconsistentwiththemainmechanismof

anti-canceractionofthesecompounds,wherebyDNAistherelevanttarget.Giventhat

thereislittleevidenceofBiaccumulationinthenucleusoftheotherBiNSAID-treated

cells (very low proportion of Bi-treated cells compared to every cell analysed

followingtreatmentwithPtdrugs[23]andAscompounds[30]),itappearedunlikely

thatthemaintargetofBiNSAIDswasDNAinHCT-8cells.

As described in Sections 1.2.3 and1.2.4, the chemopreventive/chemotherapeutic

natureofNSAIDs isperceived fromtheirability to inhibitCOX-2. TheCOXenzymes

are membrane-bound enzymes, specifically located on the endoplasmic reticulum

(COX-1andCOX-2),plasmamembrane(COX-1andCOX-2)or thenuclearmembrane

(COX-2only) [48,49]. The colocalisationmaps (Appendix2.4) indicated thatBihot

spots were not located in close proximity to the plasma membrane. However, a

numberofBihotspotswerelocatedincloseproximitytothenucleusasshowninthe

colocalisationmaps(Appendix2.4a(i),a(ii),b(i),b(ii),andc(i)). Whilethisdoesnot

provide indisputable evidence of BiNSAID localisation at the site of COX, it does

provide evidence that places Bi in the vicinity of COX-2. In the instance that the

BiNSAIDremains intact, thissuggeststhatthemobility issuchthatthere ispotential

fortheinteractionwithCOX. It isimportanttonotethatit is likelythattheBiNSAID

alsodissociatesinsidethecellasimpliedbythePGE2productionassays(Section2.3.4)

asafreecarboxylateisrequiredtointeractwiththeCOX-2activesite.Consequently,it

wouldbebeneficialtostudythelocationoftheNSAID.Unfortunately,directprobing

oftheNSAIDusingmicroprobeSRXRFisnotfeasibleasthechemicalelementsofthe

NSAIDs are endogenous to the cell and cannot be discerned. An alternative study

worth consideringmight be the substitution of an element such as Br for Cl (in the

instanceoftolf).Additionally,theuseofCOX-2-specificantibodywithAunanoparticle

labelling and microprobe SRXRF could be used to image Au and Bi and determine

colocalisationakintoanexperimentreportedbyYuanetal.[50].


136

Duetobeamtimerestrictions,theamountofsamplesthatcouldbeanalysedwas

limited. Samples for microprobe SRXRF were prepared following 4-h or 24-h

treatmentwith theBi compounds,however,only4-hBi(mef)3-treatedsampleswere

analysed(Appendix2.3).NegligibleBiwasobserved in the4-hBi(mef)3-treated cell

microprobeSRXRFelementalmaps(Appendix2.3),indicatingthatanyBitakenupby

thecells is lower than thedetection limit. Thereare twonotableobservations from

these results. Firstly, given that the NMR studies (Section 2.3.1) suggested that

Bi(mef)3 isunstable inthecellmedium,theuptakeand localisationofBimaynotbe

indicativeofCOX. Secondly, itwouldappearthata longer incubationtime(>4h) is

requiredforsignificantuptakeofBi(Fig.3.5andAppendix2.2)ifitisnotboundtoan

NSAID. Furtherbeamtimewouldberequired tostudy theelementaldistributionof

the 4-h Bi(tolf)3- and Bi(difl)3-treated cells in order to determine the concentration

andlocationofBiofthestablecomplexes.

The thirdpotentialmechanismof actionofBiNSAIDs that isworth addressing is

targeting the mitochondria. The validity of this mechanism stems from results

reported inChapter 2 (toxicity and cell death assays) since the intrinsic pathwayof

apoptosis is inextricably linkedwithmitochondrial health [51]. The sizeof someof

the hot spots detected in the microprobe SRXRF studies are similar to the size

reported formitochondria (0.5−10μm) [52, 53]. Ameans for determining if Biwas

located in themitochondria is toexaminewhether it iscolocalisedwithFesince the

mitochondriaprimarilyconsumeFeforthesynthesisofhemeandFe–Sclustersandas

such serve as the centre for cellular Fe homeostasis [54]. A study by

Morrisonetal.[28] showed that Gd(III) from several Gd(III) complexes colocalised

with Fe that was contained in the mitochondria of T98G human glioblastoma

multiforme cells. Unfortunately, examination of the thin-sectioned cell maps in the

current study failed to discriminate Fe from the background (data not shown), thus

the colocalisation of Fe with other elements (such as Bi) could not be definitively

established. Alternatively, Ca has been shown to localise in the euchromatin and

mitochondria[55]. MitochondriaarevitalformaintainingCa2+homeostasisandplay

animportantroleinCa2+storageinthecell[56].However,therewaslittleevidenceto

showcolocalisationofCaandBiinthecytoplasmwheremitochondriawouldbefound.

For example, only one Bi(mef)3-treated cell showed any colocalisation with Ca

(Appendix 2.4, Fig. 3.4, CellA). Coincidentally, thiswas the onlyBi-treated cell that


137

displayed nuclear Bi localisation, which was an exception compared to the other

BiNSAID-treatedcells.

ThefinalsiteofpossibleBiintracellularlocalisationthatneedstobeconsideredis

the lysosomes. Lysosomes are membrane-bound, acidic organelles that play an

importantroleinremovingcellularwaste.Lysosomescontainacidhydrolases,which

areresponsibleforthedegradationofunwantedcellularbiomolecules[57].Previous

reports following autometallography studies have shown that Bi accumulates in

lysosomesofvariouscell types, includingkidneycells,macrophages,motorneurons,

ganglion,LeydigandSertolicellsinratsaspartofadetoxificationprocess[10-12,16].

Morerecently,Hongetal.[40]haveperformedfluorescencemicroscopyexperiments

that showed thatBi colocalised in lysosomesofCBS-treated (0.5mM)HK-2 (human

proximal tubular) cells. The microprobe SRXRF results are consistent with these

observationssincethesmallhotspotsoftheBiaresimilar insizeto lysosomes(50–

500nmdiameter) [58,59]. Thereare twopossibleexplanations for thediversity in

the size and number of theBi hot spots thatwere observed in theBiNSAID-treated

cells in this study; firstly, since the thin-section represents a 1-μm slice of the cell,

theremaybeheterogeneousdistributionthroughoutthecell. Secondly,thevariation

in the size of lysosomes could be the result of certain cellular events; for instance

events leadingtothe fusionofmultiple lysosomeswithautophagosomesresulting in

autolysosomeswithincreasedsizeandreducedlysosomenumber[60].Itisunknown

whether these lysosomes are excreted from the cells over time, which may also

accountforthediversityinthedistributionandthenumberofthelysosomespresent.

Unfortunately, due to the limitation in spatial resolution (beam dimensions (0.3 ×

0.3μm2)andstepsize(0.3μm)associatedwiththistechnique,anylysosomessmaller

than300nmwouldnot be easily detected. Given the evidence in the literature that

showsthatBilocalisesinlysosomesofmanyothercelltypes,andwithoutevidenceof

colocalisation with mitochondrion-related elements (Fe & Ca), it is rational to

conclude that the Bi hot spots observed in the present study are indicative of

lysosomes rather than mitochondria. Further experiments such as transmission

electronmicroscopycouldbeperformedtounequivocallyconfirmthisisthecase.

TherewereafewinterestingobservationsemanatingfromthemicroprobeSRXRF

imagingwithrespecttothedistributionandchangesinCaconcentrationthatwarrant

discussion. The first observationwas the increase in nuclear Ca localisation of the

Bi-treatedcells(Fig.3.5). ThisobservationhasbeenreportedbyLuiandHuang[61]


138

andMunroetal.[30]inarsenite-treatedCHO-K1andHep-G2cells,respectively.Both

studies [30, 61] concluded that themechanism of toxicity of arsenitemight involve

disturbancesinintracellularCahomeostasis,aknownindicatorofapoptosis[62].The

resultsobtainedfromthetoxicityassaysinChapter2indicatedthatthehealthofthe

cells was affected in all BiNSAID treatments at the treatment concentration and

treatmenttimechosenforthemicroprobeSRXRFcellsampletreatment(40μM,24h).

Additionally, the redistribution of Ca as indicated by themicroprobe SRXRF studies

reported herein, coupled with the cell death studies in Chapter 2, suggests that

BiNSAID-inducedapoptosisisassociatedwithadisruptionofCahomeostasis.Similar

to the conclusions drawn byMunro et al. [30], it is unclear from the present study

whether the redistribution of nuclear Ca is related to drug-induced mitochondrial

damage or another intracellular source of Ca such as disruption to the endoplasmic

reticulumleadingtothereleaseofCaintothecytosol[63].

The increase in Ca concentration in the nucleus of the BSS-treated cellswas the

second interestingobservation, since low toxic effectswereobserved inHCT-8 cells

followingtreatmentwithBSS(40μM,24h)inthetoxicityassaysreportedinChapter

2. It has been observed in normal human gastric mucous epithelial cells that

treatmentwithBSSincreasedintracellularCa2+[64]. Theincreasedconcentrationof

Ca2+promotedp44/p42andp38MAP-kinaseactivation,andsubsequentlystimulated

the growth of the gastricmucous epithelial cells [64]. Given that themetabolism of

normal cells differs to that of cancer cells [65], further studieswould be needed to

determine the significance of intracellular Ca2+ flux in BSS-treated HCT-8 cells.

Additionally, it cannot be ruled out that the Ca redistribution within BSS-treated

HCT-8cellsmayalternativelybeduetosalicylateratherthanBi.

Thirdly, elementalquantificationof the thin-sectionedcells showed thatBi(mef)3

significantly increased cytosolic and nuclear levels of Ca compared to the

vehicle-treatedcells.Assuch,thismightsuggestthatBi(mef)3inducesapoptosisviaa

pathwaywhichisattenuatedtotheeffectsofCa2+comparedtoBi(tolf)3andBi(difl)3.

GiventhatBi(mef)3causedsuchasignificantincreasein[Ca2+]icomparedtotheother

Bi compounds at the same treatment concentration, it seems unlikely that Bi alone

wouldbethecauseoftheincrease.Thisraisesthequestionastowhethertheeffecton

[Ca2+]iisduetotheligandaloneorthecombinationofBiandligand.Assuch,further

experiments could be performed to shed more light on the effects of Bi- and

NSAID-induced toxicity on intracellular Ca concentration and distribution. One


139

approach couldbe thepreparationofNSAID-treated samples formicroprobeSRXRF

analysis to compare to the BiNSAID-treated samples (Fig. 3.4 and Appendix 2.2).

Alternatively, bulk cell assays (with the advantage of increased statistical power)

couldbeperformedtoquantify[Ca2+]ilevelsusingatechniquesuchasflowcytometry

and a Ca fluorescent probe such as Fluo-4-AM [63]. Additionally, co-stained cells

containingacalciumfluorescentprobeandanucleardyesuchasHoechst33342could

beanalysedusingfluorescencemicroscopy[66].


140

3.5 ChaptersummaryTheaimof thisChapterwas toquantifyBi cellularuptakeand the localisation in

HCT-8 cells following treatment with BiNSAIDs to further examine potential

mechanismsofcelltoxicity.Specifically:

i. The most striking result from the GFAAS study was that Bi uptake was

influenced by the treatment concentration, whereby a higher treatment

concentration(40µM)resultedinincreasedBiuptakebyBiNSAID-treatedcells

(ranging from5 to340times)compared to the lowerconcentration(15µM).

ThemosttoxicBiNSAID,Bi(tolf)3,showedthehighestconcentrationofcellular

Bi at both treatment concentrations 15 µM and 40µM (24h), while the

Bi(mef)3-andBi(difl)3-treatedcellsshowedsimilarquantitiesofcellularBiat

bothtreatmentconcentrationstested. TheuptakeofBimaybeinfluencedby

stability,wherebyBi(tolf)3showedsubstantiallygreaterBiuptakeandwasthe

moststableoftheBiNSAIDsasdeterminedbyNMRspectroscopy(Chapter2).

ii. The microprobe SRXRF imaging of thin-sectioned cells showed that Bi

generallyaccumulatesindiscretehotspotsfoundinthecytoplasm.Thesizeof

thehotspotsisconsistentwithtwopossibleintracellularfates.Thefirstisthe

accumulationofBiwithinthatofthecellsdetoxificationorganelles,lysosomes.

Additionally, thecloseproximityof theBihotspotstothenuclearmembrane

(and location of COX-2), may suggest that intracellular movement of Bi can

occur.However,thissecondpropositionrequiresfurtherinvestigation.

While the GFAAS results indicated substantially greater Bi uptake for Bi(tolf)3

(versus Bi(mef)3 and Bi(difl)3), which paralleled the greater toxicity and greater

inhibition of PGE2 production of the former complex, themicroprobe SRXRF results

showed less compelling differenceswith respect to the toxicities. Of course greater

insight into the disparity between the cellular quantities of Bi associated with the

different BiNSAID treatments could be obtained following studies of NSAID uptake,

whichischallengingworkandhasbeencommencedbySpremoandDillon([67]and

unpublished results) and is continued Brown and Dillon ([41], and unpublished

results).


141

3.6 References[1] C.T.Dillon.(2011)Synchrotronradiationx-rayspectroscopyforinvestigationsof

intracellular metallointercalators: X-ray fluorescence imaging and X-rayabsorption spectroscopy, In Metallointercalators : synthesis and techniques toprobe their interactions with biomolecules (J. Aldrich-Wright, Ed.), pp 273-298,Springer-Verlag/Wien,NewYork.

[2] N.Zheng,H.N.Tsai,X.Zhang,K.SheddenandG.R.Rosania.(2011)Thesubcellulardistributionofsmallmolecules:ameta-analysis,Mol.Pharm.,8,1611-1618.

[3] D. C. Harris. (2011) Chapter 20: Atomic spectroscopy, In Quantitative ChemicalAnalysis8thEditioned.,Freeman,W.H.&Company.

[4] ThermoElemental.(2001)AAS,GFAAS,ICPorICP-MS?WhichtechniqueshouldIuse?Anelementaryoverviewofelementalanalysis.,1-20.

[5] C.T.Dillon,P.A.Lay,B.J.Kennedy,A.P.Stampfl,Z.Cai,P.Ilinski,W.Rodrigues,D.G. Legnini, B. Lai and J. Maser. (2002) Hard X-ray microprobe studies ofchromium(VI)-treated V79 Chinese hamster lung cells: intracellular mapping ofthebiotransformationproductsofachromiumcarcinogen,J.Biol. Inorg.Chem.,7,640-645.

[6] M.D.Hall,C.T.Dillon,M.Zhang,P.Beale,Z.Cai,B.Lai,A.P. J. StampflandT.W.Hambley.(2003)Thecellulardistributionandoxidationstateofplatinum(II)andplatinum(IV)antitumourcomplexesincancercells,J.Biol.Inorg.Chem.,8,726-732.

[7] C.M.Weekley,J.B.Aitken,L.Finney,S.Vogt,P.K.WittingandH.H.Harris.(2013)Seleniummetabolism in cancer cells: the combined application of XAS and XFMtechniquestotheproblemofseleniumspeciationinbiologicalsystems,Nutrients,5,1734-1756.

[8] S.Antony,J.B.Aitken,S.Vogt,B.Lai,T.Brown,L.SpicciaandH.H.Harris.(2013)X-ray fluorescence imaging of single human cancer cells reveals that theN-heterocyclic ligands of iodinated analogues of ruthenium anticancer drugsremaincoordinatedaftercellularuptake,J.Biol.Inorg.Chem.,18,845-853.

[9] M. Stoltenberg, A. Larsen, M. Zhao, G. Danscher and U. T. Brunk. (2002)Bismuth-inducedlysosomalruptureinJ774cells,APMIS,110,396-402.

[10] N. E. Magnusson, A. Larsen, J. Rungby, M. Kruhoffer, T. F. Orntoft and M.Stoltenberg.(2005)Geneexpressionchangesinducedbybismuthinamacrophagecellline,CellTissueRes.,321,195-210.

[11] M.Stoltenberg,G.Danscher,R.Pamphlett,M.M.ChristensenandJ.Rungby.(2000)Histochemical tracingof bismuth in testis from rats exposed intraperitoneally tobismuthsubnitrate,Reprod.Toxicol.,14,65-71.

[12] M. Stoltenberg, J. D. Schionning and G. Danscher. (2001) Retrograde axonaltransport of bismuth: an autometallographic study, Acta Neuropathol., 101,123-128.

[13] A. Larsen, N. Martiny, M. Stoltenberg, G. Danscher and J. Rungby. (2003)Gastrointestinal and systemic uptake of bismuth in mice after oral exposure,Pharmacol.Toxicol.,93,82-90.

[14] L.H.Pedersen,M.Stoltenberg,E.ErnstandM.J.West.(2003)Leydigcelldeathinratsexposedtobismuthsubnitrate,J.Appl.Toxicol.,23,235-238.

[15] A.Larsen,M.Stoltenberg,M.J.WestandG.Danscher.(2005)Influenceofbismuthon the number of neurons in cerebellum and hippocampus of normal andhypoxia-exposedmousebrain:astereologicalstudy,J.Appl.Toxicol.,25,383-392.

[16] M. Stoltenberg, J. A. Hogenhuis, J. J. Hauw and G. Danscher. (2001)Autometallographic tracing of bismuth in humanbrain autopsies, J. Neuropathol.Exp.Neurol.,60,705-710.

[17] C. J. Fahrni. (2007) Biological applications of X-ray fluorescence microscopy:exploring the subcellular topography and speciation of transition metals,Curr.Opin.Chem.Biol.,11,121-127.


142

[18] A.C.Thompson,D.T.Attwood,E.M.Gullikson,M.R.Howells, J.B.Kortright,A.L.Robinson,J.H.Underwood,K.-J.Kim,J.Kirz,I.Lindau,P.Pianetta,H.Winick,G.P.Williams and J. H. Scofield. (2001) Center for X-ray Optics and Advanced LightSourceX-rayDataBooklet(2ndedition),LawrenceBerkeleyNationalLaboratory.

[19] R. Meuli, Y. Hwu, J. H. Je and G. Margaritondo. (2004) Synchrotron radiation inradiology: radiology techniques based on synchrotron sources, Eur. Radiol., 14,1550-1560.

[20] S.Matsuyama,H.Mimura,H.Yumoto,Y.Sano,K.Yamamura,M.Yabashi,Y.Nishino,K.Tamasaku,T.IshikawaandK.Yamauchi.(2006)Developmentofscanningx-rayfluorescencemicroscopewithspatialresolutionof30 nmusingKirkpatrick-Baezmirroroptics,Rev.Sci.Instrum.,77,103102.

[21] E.DiFabrizio,F.Romanato,M.Gentil,S.Cabrini,B.Kaulich,J.SusiniandR.Barrett.(1999)High-efficiencymultilevelzoneplatesforkeVX-rays,Nature,401,895-898.

[22] C.T.Dillon.(2012)Synchrotronradiationspectroscopictechniquesastoolsforthemedicinal chemist: microprobe X-ray fluorescence imaging, X-ray absorptionspectroscopy,andinfraredmicrospectroscopy,Aust.J.Chem.,65,204-217

[23] K. J. Davis, J. A. Carrall, B. Lai, J. R. Aldrich-Wright, S. F. Ralph and C. T. Dillon.(2012)DoescytotoxicityofmetallointercalatorscorrelatewithcellularuptakeorDNAaffinity?,DaltonTrans.,41,9417-9426.

[24] M. Shimura, A. Saito, S. Matsuyama, T. Sakuma, Y. Terui, K. Ueno, H. Yumoto, K.Yamauchi,K.Yamamura,H.Mimura,Y.Sano,M.Yabashi,K.Tamasaku,K.Nishio,Y.Nishino,K.Endo,K.Hatake,Y.Mori,Y. IshizakaandT. Ishikawa. (2005)Elementarray by scanning X-ray fluorescence microscopy after cis-diamminedichloro-platinum(II)treatment,CancerRes.,65,4998-5002.

[25] J. B. Aitken, S. Antony, C. M.Weekley, B. Lai, L. Spiccia and H. H. Harris. (2012)Distinct cellular fates forKP1019andNAMI-AdeterminedbyX-ray fluorescenceimagingofsinglecells,Metallomics,4,1051-1056,1007.

[26] J.B.Aitken,P.A.Lay,T.T.Duong,R.Aran,P.K.Witting,H.H.Harris,B.Lai,S.VogtandG.I.Giles.(2012)SynchrotronradiationinducedX-rayemissionstudiesoftheantioxidantmechanismof theorganoseleniumdrug ebselen, J. Biol. Inorg. Chem.,17,589-598.

[27] C.M.Weekley,G.Jeong,M.E.Tierney,F.Hossain,A.M.Maw,A.Shanu,H.H.Harrisand P. K. Witting. (2014) Selenite-mediated production of superoxide radicalanions in A549 cancer cells is accompanied by a selective increase in SOD1concentration, enhanced apoptosis and Se-Cu bonding, J. Biol. Inorg. Chem., 19,813-828.

[28] D.E.Morrison, J.B.Aitken,M.D.deJonge,F. Issa,H.H.HarrisandL.M.Rendina.(2014) Synthesis andbiological evaluationof a class ofmitochondrially-targetedgadolinium(III)agents,Chem.Eur.J.,20,16602-16612.

[29] M.Busse,M.S.A.Windsor,A.J.Tefay,M.Kardashinsky,J.M.Fenton,D.E.Morrison,H. H. Harris and L. M. Rendina. (2017) Tumor cell uptake and selectivity ofgadolinium(III)-phosphonium complexes: the role of delocalisation at thephosphoniumcentre,J.Inorg.Biochem.,177,313-321.

[30] K.L.Munro,A.Mariana,A.I.Klavins,A.J.Foster,B.Lai,S.Vogt,Z.Cai,H.H.Harrisand C. T. Dillon. (2008)Microprobe XRFmapping and XAS investigations of theintracellular metabolism of arsenic for understanding arsenic-induced toxicity,Chem.Res.Toxicol.,21,1760-1769.

[31] M. J. Hackett, J. A. McQuillan, F. El-Assaad, J. B. Aitken, A. Levina, D. D. Cohen,R.Siegele, E. A. Carter, G. E. Grau, N. H. Hunt and P. A. Lay. (2011) Chemicalalterations tomurine brain tissue induced by formalin fixation: implications forbiospectroscopic imaging and mapping studies of disease pathogenesis, Analyst,136,2941-2952.


143

[32] J. B. Waern, H. H. Harris, B. Lai, Z. Cai, M. M. Harding and C. T. Dillon. (2005)Intracellular mapping of the distribution of metals derived from the antitumormetallocenes,J.Biol.Inorg.Chem.,10,443-452.

[33] E.L.Hawksworth,P.C.Andrews,W.Lie,B.LaiandC.T.Dillon.(2014)Biologicalevaluationofbismuthnon-steroidalanti-inflammatorydrugs(BiNSAIDs):stability,toxicityanduptakeinHCT-8coloncancercells,J.Inorg.Biochem.,135,28-39.

[34] S.Vogt.(2003)MAPS:Asetofsoftwaretoolsforanalysisandvisualizationof3DX-rayfluorescencedatasets,J.Phys.IVFrance,104,635-638.

[35] W.M.McIndoeandJ.N.Davidson.(1952)Thephosphoruscompoundsofthecellnucleus,Br.J.Cancer,6,200-214.

[36] C.Matheny,M. L.Day and J.Milbrandt. (1994)Thenuclear localization signal ofNGFI-A Is locatedwithin the zinc fingerDNAbindingdomain, J.Biol. Chem., 269,8176-8181.

[37] C. Quintana. (1991) X-raymicroanalysis of cell nuclei, J. Elec. Microsc. Tech., 18,411-423.

[38] J.S.Choi,M.J.JinandH.K.Han.(2005)RoleofmonocarboxylicacidtransportersinthecellularuptakeofNSAIDs,J.Pharm.Pharmacol.,57,1185-1189.

[39] I. Legen andA. Kristl. (2003) pH and energy dependent transport of ketoprofenacrossratjejunuminvitro,Eur.J.Pharm.Biopharm.,56,87-94.

[40] Y. Hong, Y.-T. Lai, G. C.-F. Chan and H. Sun. (2015) Glutathione and multidrugresistance protein transporter mediate a self-propelled disposal of bismuth inhumancells,Proc.Natl.Acad.Sci.U.S.A.,112,3211-3216.

[41] T.Brown.(2016)Bismuth-NSAIDsascolorectalcancerchemopreventives:Insightsinto theirmembrane interactionsanduptakemechanisms,BachelorofMedicinalChemistry(Advanced)withHonours,UniversityofWollongong,Australia.

[42] E.L.Hawksworth.(2010)Cellularstudiesofphotoactivechromiumintercalators,BachelorofMedicinalChemistry,UniversityofWollongong,Australia.

[43] B. D. Grant and J. G. Donaldson. (2009) Pathways andmechanisms of endocyticrecycling,Naturereviews.Molecularcellbiology,10,597-608.

[44] M. Prutki, M. Poljak-Blazi, M. Jakopovic, D. Tomas, I. Stipancic and N. Zarkovic.(2006) Altered iron metabolism, transferrin receptor 1 and ferritin in patientswithcoloncancer,CancerLett.,238,188-196.

[45] P. G. Cavanaugh, L. Jia, Y. Zou and G. L. Nicolson. (1999) Transferrin receptoroverexpressionenhancestransferrinresponsivenessandthemetastaticgrowthofaratmammaryadenocarcinomacellline,BreastCancerRes.Treat.,56,203-217.

[46] K.T.Chan,M.Y.Choi,K.K.Lai,W.Tan,L.N.Tung,H.Y.Lam,D.K.Tong,N.P.Leeand S. Law. (2014) Overexpression of transferrin receptor CD71 and itstumorigenic properties in esophageal squamous cell carcinoma, Oncol. Rep., 31,1296-1304.

[47] N. Yang and H. Sun. (2007) Biocoordination chemistry of bismuth: recentadvances,Coord.Chem.Rev.,251,2354-2366.

[48] I. Morita, M. Kawamoto, M. Hattori, K. Eguchi, K. Sekiba and H. Yoshida. (1990)Determinationof tryptophanand itsmetabolites inhumanplasmaandserumbyhigh-performance liquid chromatography with automated sample clean-upsystem,J.Chromatogr.B,526,367-374.

[49] J.C.OttoandW.L. Smith. (1994)Theorientationofprostaglandinendoperoxidesynthases-1and-2intheendoplasmicreticulum,J.Biol.Chem.,269,19868-19875.

[50] Y. Yuan, T. Paunesku, H. Arora, J. Ward, S. Vogt and G. Woloschak. (2011)Interrogation of EGFR targeted uptake of TiO2 nanoconjugates by X-rayfluorescencemicroscopy,AIPConf.Proc.,1365,423-426.

[51] S.Elmore. (2007)Apoptosis:a reviewofprogrammedcelldeath,Toxicol.Pathol.,35,495-516.

[52] National Institute of General Medical Sciences. (2005) Inside the cell, 1-80, U.S.DepartmentofHealthandHumanServices.


144

[53] P.Khanna.(2010)Cellandmolecularbiology, I.K. InternationalPublishingHousePvt.Limited,Delhi,India.

[54] C. Chen and B. H. Paw. (2012) Cellular and mitochondrial iron homeostasis invertebrates,BBAMol.CellRes.,1823,1459-1467.

[55] A. Kitamura, K. Takano and T. Inokuchi. (1994) Study on intracellular calciumlocalization by potassium antimonate technique in rabbit VX2 carcinoma, Med.ElectronMicrosc.,27,333-336.

[56] L.Contreras,I.Drago,E.ZampeseandT.Pozzan.(2010)Mitochondria:thecalciumconnection,Biochim.Biophys.Acta,1797,607-618.

[57] J.P.Luzio,P.R.PryorandN.A.Bright.(2007)Lysosomes:fusionandfunction,Nat.Rev.Mol.CellBiol.,8,622-632.

[58] M.A.Sukhai,S.Prabha,R.Hurren,A.C.Rutledge,A.Y.Lee,S.Sriskanthadevan,H.Sun, X. Wang, M. Skrtic, A. Seneviratne, M. Cusimano, B. Jhas, M. Gronda, N.MacLean,E.E.Cho,P.A.Spagnuolo,S.Sharmeen,M.Gebbia,M.Urbanus,K.Eppert,D.Dissanayake,A.Jonet,A.Dassonville-Klimpt,X.Li,A.Datti,P.S.Ohashi,J.Wrana,I.Rogers,P.Sonnet,W.Y.Ellis,S.J.Corey,C.Eaves,M.D.Minden,J.C.Y.Wang,J.E.Dick, C. Nislow, G. Giaever and A. D. Schimmer. (2013) Lysosomal disruptionpreferentiallytargetsacutemyeloidleukemiacellsandprogenitors,J.Clin.Invest.,123,315-328.

[59] D. J.PippinandK.D.Pippin. (2003)Sizeofazurophil lysosomesofhumanPMNsdeterminedbystereologicalmorphometry,Ultrastruct.Pathol.,27,261-263.

[60] L.Yu,C.K.McPhee,L.Zheng,G.A.Mardones,Y.Rong,J.Peng,N.Mi,Y.Zhao,Z.Liu,F. Wan, D. W. Hailey, V. Oorschot, J. Klumperman, E. H. Baehrecke and M. J.Lenardo. (2010) Termination of autophagy and reformation of lysosomesregulatedbymTOR,Nature,465,942-946.

[61] Y. C. Liu and H. Huang. (1996) Lowering extracellular calcium content protectscells from arsenite-induced killing and micronuclei formation, Mutagenesis, 11,75-78.

[62] S. Orrenius, B. Zhivotovsky and P. Nicotera. (2003) Regulation of cell death: thecalcium-apoptosislink,Nat.Rev.Mol.CellBiol.,4,552-565.

[63] J. E. Kim, J. S. Kang andW. J. Lee. (2012)VitaminC induces apoptosis in humancoloncancercell line,HCT-8via themodulationofcalciuminflux inendoplasmicreticulumandthedissociationofBadfrom14-3-3β,ImmuneNetw.,12,189-195.

[64] J. Gilster, K. Bacon, K. Marlink, B. Sheppard, C. Deveney and M. Rutten. (2004)Bismuth subsalicylate increases intracellular Ca2+, MAP-kinase activity, and cellproliferation in normal human gastric mucous epithelial cells, Dig. Dis. Sci., 49,370-378.

[65] R. J. DeBerardinis, J. J. Lum, G. Hatzivassiliou and C. B. Thompson. (2008) Thebiology of cancer:metabolic reprogramming fuels cell growth and proliferation,CellMetab.,7,11-20.

[66] D.-H. Kim, E. A. Rozhkova, I. V. Ulasov, S. D. Bader, T. Rajh,M. S. Lesniak and V.Novosad. (2010) Biofunctionalized magnetic-vortex microdiscs for targetedcancer-celldestruction,Nat.Mater.,9,165-171.

[67] S. Spremo. (2014) Novel bismuth complexes as potential anti-cancer drugs,BachelorofMedicinalChemistry(Advanced),UniversityofWollongong,Australia.

145

Chapter4. Synchrotronradiationinfraredmicrospectroscopy

studiesofliveHCT-8cellstreatedwithBiNSAIDs

andNSAIDs

In the last decade there has been an increase in the use of synchrotron radiation

infrared microspectroscopy (SR-IRMS) for the study of single mammalian cells to

discernbiomolecularchangesthatareassociatedwithdiseaseanddiseasetreatment.

This Chapter describes live cell SR-IRMS experiments performed at the Australian

Synchrotron(Clayton,Australia)fortheanalysisofHCT-8cellsfollowingexposureto

BiNSAIDs or NSAIDs in order to: (i) study global biomolecular variations,

(ii)determinewhetheranychangesarereflectiveofthemechanismsofcelldeath,and

(iii) identify whether the coordination of NSAIDs to Bi results in alterations in the

biomolecular effects of NSAIDs. A multivariate approach, Principal Component

Analysis (PCA),was applied to discriminate the spectral differences associatedwith

thedifferentcelltreatments.

Chapter4.Synchrotronradiationinfraredmicrospectroscopy

146

4.1 IntroductionIdentifying cellular events that occur as a result of exposure to a

chemotherapeuticagentisnecessarytounderstandhowthedrugaffectsacancercell.

Specific biological assays can be used to probemechanisms of action, a selection of

which are discussed in Chapter 2. Alternatively, Fourier transform infrared(FTIR)

spectroscopy,whichprobeschemicalchanges,isbeingincreasinglyutilisedasarapid,

non-destructive, label-free technique for studying the biochemistry of biological

materials [1]. The coupling of FTIR spectroscopywithmicroscopy and synchrotron

light has allowed researchers to simultaneously probe a large number of

macromolecules present within a single cell [1]. Importantly, detection of

simultaneous changes in themolecular composition of single cells can assist in the

identification of biological processes that can then be probed further using

complementarytechniques.

4.1.1 FTIRspectroscopyofbiologicalmolecules

FTIR spectroscopy is a technique traditionally employed by chemists and

materialsscientistsfortheidentificationandmolecularanalysisofchemicals.Selected

frequencies(energies)ofIRradiationareabsorbedbymoleculesatenergiesrequired

to cause vibrational excitation of bonds within a molecule [2]. The characteristic

vibrationsofeachchemicalgroupgiverise toaspectrumthatenableselucidationof

the identity of a molecule [2]. As demonstrated in Figure 4.1, IR active molecules

display two fundamentalmodes of vibrationwhich include: (i) stretching vibrations

(symmetric or asymmetricmotion of twoormore atomsbonded to a central atom)

resultingfromachangeinbondlengths,and(ii)bendingvibrations(includingoutof

planemotions, twisting andwagging, and in planemotions, rocking and scissoring)

resultingfromachangeinbondangles[2].


147

Figure4.1:FundamentalmodesofvibrationofIRactivemolecules.

The IR spectrum can be subdivided into three regions, the near-IR,mid-IR and

far-IR regions, encompassing the red end of the visible spectrum at 769 nm

(13000cm−1)tothebeginningofthemicrowaveregionat0.1mm(100cm−1)[2].The

mid-IRrange(4000–400cm−1)encompassesthemolecular‘fingerprint’regionthatis

frequentlyusedtostudystructuralandchemicalinformationoforganicmoleculesand

biologicalsamplesalike[2].

Eukaryotic cells are comprisedof protein,DNA,RNA, carbohydrates, and lipids.

Each of these biomolecules produces a unique, yet complex IR spectrum due to the

distinctivemolecularbondspresentintheirstructures,asdemonstratedinFigure4.2.

As such, the IR spectrum of amammalian cell exhibits a superposition of all of the

biomolecules contained in the cell (Fig. 4.2). Generalised assignments for

characteristicIRbandsassociatedwithbiologicalmaterials,includingeukaryoticcells,

arepresentedinTable4.1.


148

Figure4.2:TypicalIRsignaturesofbiologicalmaterials.AdaptedfromMilleretal.[3].


149

Table4.1:GeneralisedassignmentsofcharacteristicIRbandsofbiologicalandeukaryoticcellspectra[2,4].Wavenumber

(cm−1)Assignment Associated

biomolecule2955 νas(C−H)ofCH3 Lipid/phospholipid2930 νas(C−H)ofCH2 Lipid/phospholipid2898 ν(C−H)ofmethine Lipid/phospholipid2870 νs(C−H)ofCH3 Lipid/phospholipid2850 νs(C−H)ofCH2 Lipid/phospholipid1740 ν(C=O)ester Lipid/phospholipid1720 ν(C=O)ester DNA

1680–1715 ν(C=O) Nucleicacids1653 ν(C=O)/ν(C−N)/δ(N−H),amideI Protein

1567–1520 ν(C−N)/δ(N−H),amideII Protein1515 Tyrosineband Protein1470 δas(C−H)ofCH2 Lipid/phospholipid1420 νs(COO−) Protein

1310–1240 ν(C−N)/δ(N−H)/ν(C=O)/ν(O=C−N),amideIII Protein1232 νas(PO2−) Nucleicacids1170 νas(CO−O−C) Lipid1080 ν(C−O−P),νs(PO2−) Nucleicacids1050 ν(C−O) Carbohydrate

Key:ν=stretchingvibrations,δ=bendingvibrations,s=symmetric,as=asymmetric.

4.1.2 FTIRmicroscopyandsynchrotronsources

ThecouplingofmodernFTIRspectrometerswithpurposebuilt IRmicroscopes,

has allowed researchers to study samples on amicroscopic scale [1, 5]. One of the

limitations of conventional lab-based FTIR microspectroscopy is the low intrinsic

brightness which limits the smallest practical spot size (defined by the microscope

masking aperture) to approximately 20–40 μm at the expense of signal-to-

noise[1, 5]. This represents a major limitation when examining cells and cell

organelles [1, 5]. Conversely, a synchrotron light source produces lightwith higher

brilliance (increased photon flux) which when focused to narrow range emission

angles [5]withmicroscopeapertures(5–20μm)results in the inherentlybrightand

highly focused synchrotron light that facilitates the analysis ofmicrometre samples

with high signal-to-noise [1]. The main restrictions, however, stem from the

diffraction limit [1]. For instance, the diffraction limit in the mid-IR region (3500–

650cm−1)isapproximately1.4–7.7μmforaconventionalconfocalmicroscopewhich

approachesλ/2[1].


150

4.1.3 Infraredmicrospectroscopystudiesofdrugeffectsoncancercells

Morerecentdevelopmentsofinfraredmicrospectroscopy(IRMS)inconventional

lab-based settings, and the increasing utilisation of synchrotron radiation (SR) light

sourceshasallowedanumberofresearcherstostudytheeffectsofclinically-used(or

potential) anti-cancer drugs on cancer tissue or cells, at single-cell (and even

subcellular)resolution[6-14].ThestudyofdrugeffectsoncancercellsusingtheIRMS

approachhasanumberofperceivedadvantagesinaclinicalsetting. Forinstance,in

cancer treatments where multiple drug combinations are recommended for the

treatmentofspecifictypesofcancer(suchasnon-smallcell lungcancer), ithasbeen

proposed that IRMS could potentially be employed to determine themost effective

combinationofdrugstoadministertoeachindividualpatient[6].Additionally,IRMS

could provide information on early cellular events to varied doses of

chemotherapeuticswhichcouldallowclinicians todetermine theminimumeffective

dosetoadminister,reducingtheseverityofchemotherapeuticsideeffects[9].Inthis

context,IRMSconceivablyprovidesarapid,sensitiveandeffectivemeanstoimprove

thetreatmentofcancerwithlong-termbeneficialoutcomesforpatients[6].

AsIRMScanprovideaglobalviewoftheeffectsofdrugtreatmentonallcellular

biomolecules (i.e. proteins, lipids, nucleic acids, carbohydrates), it is considered that

the technique could potentially be used to identify possible modes of action of the

drug, as changes to IRMS spectra may be correlated to alterations in biological

functionspecifictothemodeofaction[8].Arecentreviewdiscussesthepotentialuse

of IRMS as a fast and easy technique to determine the mechanisms of action of

prospective candidate compounds in drug discovery programs [15], whereby the

authors propose that IRMS this could help widen the screening process beyond

conventionaltoxicityscreeningmethods.InordertodevelopIRMSforclinicaluse,the

characterisationandinterpretationofthespectralchangesinducedbydrugtreatment

oncancercellsneedstobefullydeveloped.Tothisend,anumberofresearchgroups

haveusedIRMSorSR-IRMStostudytheeffectsofvariouschemotherapeuticagentson

cancercellsinvitro[6-14].AselectionofthesestudiesissummarisedinTable4.2.On

the cellular level, IRMS has been used to study events which are important for the

fundamentalunderstandingof cellularprocesses suchas apoptosis/necrosis [16-21]

andstagesof cell cycle [16,22], thedifferencesbetweenproliferatingandquiescent

cells[23],differentiationofstemcells[24,25]andbloodcells[17],andtheeffectsof

ultravioletBradiationoncells[26,27].


151

Table 4.2: Summary of some IRMS or SR-IRMS studies that examined the effects ofchemotherapeuticagents(eitherpotentiallyorclinically-used)oncancercells.Author Drug(mechanismof

action)andCellline(cancertype)

Keyobservations

Bellisolaetal.[14]

• Imatinibmesylate(inhibitsABLtyrosinekinaseactivity)

• K562(humanerythroleukemic)

• HierarchicalClusterAnalysisshowedclusteringofthespectraofsingleK562cellstreatedwiththedrugintotwodistinctgroupscorrespondingtolivingandtoapoptoticcells

• TheamideIpeakwasshiftedtoalowerwavenumberandtherewasadifferenceintheabsorbanceassociatedwiththeν(C=O)modeofphospholipidsat1736 cm−1intheapoptoticcellpopulation

Chio-Srichanetal.[7]

• HypocrellinA(HA,lipid-solubleperyloquinonederivative,isolatedfromHypocrellabambusae)

• HeLa(humancervicalcancer)

• Principalcomponentanalysis(PCA)showedspectracollectedinthenucleusdifferedtothoseofthecytoplasm(nucleus=Óβ-typeproteinstructure)

• Irradiated,HA-treated(6h)nuclearspectraseparatedfromnon-irradiatedcontrol,andHA-treated(30min,6hand24h)spectra.Theloadingsplotindicatedirradiationcausedachangeinproteinsecondarystructure(Óβ-typestructure)

Derenneetal.[10]

• Methotrexate(MTX,antimetabolite–inhibitsfolicacidmetabolism)

• PC-3(humanprostatecancer)

• Studentt-testsindicatedthatthemostsignificantdifferencescausedbyMTXover24hwerelocatedbetween1300–1000cm−1

• MTXtreatmentfor48hcausedsignificantchangeintheproteinregionwithlimitedchangesinthenucleicacidregioncomparedto24hMTX-treatedspectra

• TheresultssuggestedthatMTXmetabolicperturbationsoccurintwostepsover48h

Drauxetal.2009[9]

• Gemcitabine(GMB,pyrimidineantimetabolite–inhibitsDNAreplicationandrepair)

• Calu-1(non-smallcelllungcancer)

• SpectraleffectsofGMBweredetectable,andclassifiable(usingclusteranalysis)atsub-lethaldosesthatinducedweak-moderategrowthinhibition(0.1–1nM,48–72h,butnot24h)andhighgrowthinhibition(10–100nM,24–72h)

• Effectsobservedoncellpopulationsweresimilartothoseobservedonsinglecells

Floweretal.[12]

• KF0101(cytotoxicgoldcomplex),cisplatin,MTX,paclitaxel(inhibitstubulindestabilisation)and5-fluorouracil(irreversibleinhibitorofthymidylatesynthase)

• A2780(humanovariancancer)

• Severalhundredsinglecontrolcellspectrawererequiredduetocellcycleheterogeneityamongcells

• PCAshowedthatdrug-treatedcellspectraclusteredinwell-definedareasindicatingacellcyclespecificresponse.Thisoverlappedwithasmallregionofcontrolcellswithsomespectraappearingoutsidethecontrolcellcluster,deemedtobe“cellcycleplus”behaviour,i.e.duetodrugresponsenotcellcycle

Gasperetal.[8]

• Ouabain(inhibitssodiumpumpactivity)

• PC-3(humanprostatecancer)

• Earlyouabaintreatment(6h)inducedlimitedchanges,mainlytoamideIandfingerprintregion(1500–1200cm−1)comparedto0hspectra

• Extendedouabaintreatment(24h)causedobviouschangestolipid-relatedbands,andashiftintheamideIbandcomparedto6-houabain-treatedspectra.However,fewfurthereffectswereobservedbetween36hvs.24houabain-treatedspectra

Sulé-Susoetal.[6]

• GMB(pyrimidineanti-metabolite)

• A549andCalu-1(non-smallcelllungcancers)

• Increasedintensityof1080cm−1followingGMBtreatmentinbothcelllines

• Increased1080/1050cm−1bandratiouptoLD50doses,plateauedathigherdosesinbothcelllines


152

Notably,thestudiespresentedinTable4.2wereperformedonfixedordriedcell

samples. Traditionally, IRMS sample preparation of mammalian cells involves

chemicalfixation(withchemicalssuchasformalinorethanol)ontoanIR-transparent

substrate (such asMirrR Low-E coated slides, zinc selenide (ZnSe), barium fluoride

(BaF2), or calcium fluoride (CaF2)windows) [28]. While a general consensus in the

literaturesupportsthevalidityofresultsobtainedfromfixedsamples,interestinthe

use of FTIR microspectroscopy studies employing live cells has grown in the last

decade [29-40]. Theadvantagesand limitationsofSR-IRMSstudiesofhydrated live

cellsampleswillbediscussedinthefollowingsection.

4.1.4 SR-IRMSoflivecells:advantagesanddisadvantages

TheIRMSanalysisoflivingcellsversusfixedcellshasgainedattentioninthelast

decadeforseveralreasons.Theseinclude:(i)theeliminationofartefactsarisingfrom

chemical fixation, (ii) the minimisation of physical artefacts such as resonant Mie

scatter (RMieS) in the spectra, and (iii) theanalysisof cells in real-time,providinga

means to monitor the biomolecular effects of cell events or drug treatment in

situ[41-43].ItisgenerallyrecognisedthatabrightlightsourceofIRphotons,suchas

asynchrotron, isrequiredtoobtaingoodsignal-to-noise fromsingle livecellspectra

duetoattenuationoftheIRsignalcausedbytheIRbeamasitpassesthroughtwoIR

transparentwindows,cellmediumandthehydratedcell[28]. However,itshouldbe

notedthatthecollectionoflivecellspectrahasalsobeensuccessfullyperformedusing

conventionalbench-topIRlightsources[33,35].

In the past decade, live cell experiments have been made possible by the

development and fabrication of purpose-built chambers for themaintenance of cell

hydration during data collection. The simplest designs involve placing the cells

between two IR transparent windows (e.g. CaF2, ZnSe, or diamond) separated by a

spacer (<12 μm) in a demountable holder. These setups have been utilised in a

number of studies [16, 32, 44-49] wherein these devices have been employed

successfully for short-term analyses of live cells. The holders are not generally

suitablefor long-term(>2h)measurementsduetothe lossofcellmedium[50]. To

this end, more elaborately designed micro-machined thin-reservoirs have been

employedforlivecellstudies[23,33],whileamulti-faceted,fullysealedfluidicdevice

was designed and employed in studies reported byVaccari et al. [34]. In the latter

system, the cells are introduced into the device via a syringe pump. Additionally,


153

lithographedbiocompatiblematerialswereusedtocreateseparatechambersforthe

cells, freshcellmediumandairbackgroundmeasurements [34]. Othergroupshave

developedmorecomplexflowchamberswithanopen-channelgeometry[51,52].

Oneofthereasonsthatsignificantefforthasfocusedonmaintaininghydrationis

tosimulatethebiochemistryofmammaliancells,andtheirbiomolecules,inparticular

nucleicacidbands.Forinstance,Whelanetal.[42]reportedthatdehydrationcausesa

B-liketoA-liketransition in theDNAofseveraleukaryoticcellsandthat thereverse

A-liketoB-liketransitionoccursuponrehydration.Thisworkalsoshowedthatupon

rehydration of the cells, sharp DNA-specific bands were observed; however, these

bands were previously broader and less intense when dehydrated. Furthermore,

Vaccari and co-workers [34] performed a study to investigate the effects of cell

fixationmethods(airdrying,ethanol,and formalin)on theSR-IRMSspectraofU937

monocytescomparedtotheSR-IRMSspectraofthelivingU937cells.Importantly,this

study showed that while formalin fixation produced spectra that were the most

similar to those from live cells (particularly in the lipidandprotein regions), effects

wereobservedintheDNAabsorptions. Forinstance,thebandcentredat1717cm−1

(reflectiveofbasepairingofnucleicacids in theB-form)waspronounced in the live

cellspectra,butabsentinthechemical-fixedanddriedcellspectra.

ThesecondadvantageofperformingSR-IRMSstudiesoflive,hydratedcellsisthe

minimisation of dispersion effects, which is an issue that affects fixed cell studies.

Dispersionartefacts suchasRMieScanarise in the spectraof single cells,whichare

typically8–30μmindiameter(andspherical)andoccurswhenthereisadifferencein

the refractive index between the cells and the surrounding environment [53]. This

occurswhentheIRwavelength(typically3–10μm,or3333–1000cm−1) issimilar in

sizetothediameterofthecellsuchthattheincidentIRbeamisscatteredasitcontacts

theouteredgesofthecellresultinginthedistortionofthecollectedIRspectrum[53].

Consequently,RMieS canbeobserved in a single cell spectrumas ahighlydistorted

baseline (particularly, between 2800–1700 cm−1) and derivative-like peak shapes

arisingfrompredominantbands[54,55].Suchdistortioncanleadtoalterationsinthe

positionandintensityofspectralfeatures,inparticulartheamideIband,whichisan

important biomarker for diagnostics and an indicator of therapeutic response [53].

TherecentdevelopmentofanalgorithmtocorrectfortheeffectsofRMieShasallowed

researcherstoremovetheeffectsofRMieSfromsinglecellspectra[53,56].However,

dispersioneffectssuchasRMieSaregenerallyminimisedinlivecellstudiessincethe


154

refractive index of the aqueous background closely matches that of the cell [28].

Therefore,thereisnoneedtoapplyanyRMieScompensationmethods,savinganalysis

time.

TraditionalIRMSstudiesoffixedcellsonlyallowforstaticmeasurementsanddo

notallowthesamecells tobestudiedat continuous timepoints. Incontrast, in situ

measurements associated with live cell studies are ideal for tracking drug-induced

changesinthesamecelloverthecourseofatreatmentperiod.Forinstance,theeffect

of arsenite treatment (100 μM) on K562 leukaemia cells has been analysed in

real-time in a purpose-built micro-fabricated chamber [32] at the Australian

Synchrotron[31].Theauthorswereabletotrackspectralchangesconsistentwiththe

biochemistry of apoptosis induced by arsenite treatment. Specifically, following

40minofexposure,adecreaseintheν(C=O)bandat1742cm−1inthearsenite-treated

cell spectrawas consistentwith lipidmembrane changes,while the decrease in the

intensitiesoftheνs(PO2−)andνas(PO2−)wasattributedtotheIRopaquenessofDNAin

an increasingly condensed state. Following 100 min incubation with arsenite, the

intensityoftheνs(PO2−)andνas(PO2−)wasattributedtoDNAfragmentation,anevent

inlaterstagesofapoptosis.AsignificantshiftintheamideIbandfrom1642cm−1to

1650 cm−1 in the arsenite-treated cell spectra was observed suggesting arsenate

induceschanges inthecompositionofproteins inthecell (fromprimarilyβ-sheetto

α-helix/randomcoil).

The IRMSanalysisof livecells isnotwithout itsdisadvantages;oneparticularly

difficultchallengetoovercomeis thestrongIRabsorptivityof thewatermolecule in

aqueousmedia.Thenormalmodesofliquidwaterincludetheasymmetricstretching

mode, the symmetric stretching mode, and the deformation (bending) mode that

appearat~3600cm−1,~3450cm−1,and~1643cm−1,respectively[41].Abroadband

centredat~2125cm−1,which is comprisedof a combinationof abendingbandand

the librationalmodesofwater, canalsobevisualised in cell spectrabetween2500–

1800cm−1[57]. ThewaterbandofparticularconcernforIRMSspectraisthestrong

deformationbandat1640cm−1whichcompletelyoverlapswiththeamideIbandand

partiallywiththeamideIIbandofproteins(Fig.3.3a),resultinginthedistortionofthe

intensity ratio of the amide I/amide II bands (Fig. 4.3b) [33]. There is currentlyno

general consensuson theapproach to take to correct theeffectsofwater saturation

from the spectrum of a cell, although a number of groups have developedMATLAB

algorithmstocorrectfortheabsorptionofwater[33,34];othergroupshaveoptedto


155

omit the amide region from chemometric analysis altogether due to the uncertainty

surrounding the influence of water on this region of the spectrum and a lack of a

reliablecompensationmethod[43]. Studiesassessingtheuseofnon-aqueousmedia

havealsobeenperformedasanalternative toaddress theproblemsassociatedwith

watercontribution.Forinstance,Sohetal.[58]havetestedtheeffectsofincubationof

MGH-U1 urothelial cancer cells and Caco-2 cells with paraffin mineral oil or

Fluorolube forup to120min. The results showed that cell viabilitywasunaffected

following 2-h incubation in Fluorolube, and differential uptake of drugswas similar

between cells incubated with or without Fluorolube [58]. FTIR mapping of cells

incubated in Fluorolube was performed in reflectance and transmission modes,

whereby the authors claimed that the latter proved best for spectroscopy [58].

However,acomparisonbetweenthespectraldifferencesbetweenFluorolubeandcell

mediumwasnotperformedorcommentedon[58]. Whileoil-basedmaterialscould

potentiallybeemployedfortheanalysisofsomeclinicalsamples,ideallycellmedium

wouldbebestemployedforcellstudiesmeasuringtime-dependenteffectsofdrugsin

order tomaintain an environment that closelymatches the conditions employed in

othercellbasedassays,particularlyinthecaseoflongertreatmentperiods(>4h).

Figure 4.3: (a) A single fixed HCT-8 cell spectrum (—) superimposed with a spectrumcollected fromcellmedium(—). Theoverlapof thebendingmodeofH2Owith theamide Iband of the cell spectrum can be visualised around 1620 cm−1. (b) The distortion of theamideI to amide II band intensity ratio of an uncorrected, single live HCT-8 cell spectrumcollected in cell medium (cell medium was used for background reference during spectralacquisition).


156

4.1.5 Chapteraims

The work in the previous Chapter suggests that the BiNSAIDs, Bi(tolf)3 and

Bi(mef)3, cause toxicity towardsHCT-8 cells that is approximately equivalent to the

respective free NSAIDs, tolfH and mefH, while Bi(difl)3 induced a greater toxic

responsecomparedtodiflH.However,theexactmechanismsbywhichtheBiNSAIDs

causetoxicitytotheHCT-8cellsarecurrentlyunknown.Itis,therefore,importantto

determine whether BiNSAIDs elicit toxicity through similar or new mechanisms of

actioncomparedtotheuncomplexedNSAIDs.

While several groups have recently reported the results from SR-IRMS studies

that employ live cells, and specifically the effects of drug compounds on cancer cell

lines, therearecurrentlynostudiesthathaveexaminedBi(III)complexesorNSAIDs

on livecancercellsusingthis technique. Additionally, there isnocurrentconsensus

on the pre-processing approach required to reduce the influence of strong

absorptivity of water bands arising from the aqueous environment. As the use of

SR-IRMS for studying the global biomolecular effects of drug-treated cells is still in

relative infancy, the current study was performed to determine the usefulness of

SR-IRMSfortheinvestigationofdominantbiomolecularchangesthatoccurfollowing

drug administration, with the focus on the BiNSAIDs and NSAIDs relevant to this

Thesis.

ThusthespecificaimsofthisChapterare:

1. To discern any differences in the results of principal component analysis

(PCA) performed on SR-IRMS spectra of HCT-8 cells using two different

approaches aimed to minimise the effects of background water: (i) pre-

processing using a water band subtractionmethod, or (ii) omission of the

amideregionfromtheanalysis.

2. Toestablish spectroscopically if thebiomolecular responsesofBiNSAIDs in

HCT-8 cells can be differentiated from those of the corresponding free

NSAIDsfollowingshort(4h)andprolonged(24h)periodsofdrugexposure.

3. TodeterminewhetheranyobservedchangesinducedbyBiNSAID/NSAIDare

consistentwithknownmechanismsofcelldeath.


157

4.2 MaterialsandMethods4.2.1 Materials

The materials employed for these studies are described in Sections 2.2.1 and

3.2.1. Additional materials (specific to the experiments in this Chapter) include

CELLSTAR® 24-well tissue culture plates sourced from Greiner Bio-One

(Kremsmünster, Austria) and polished CaF2 windows (12mm diameter, 0.5mm

thickness)manufacturedbyCrystranLimited(Poole,England).

4.2.2 Generalcellculture

Live HCT-8 cells were transported to the Australian Synchrotron (Clayton,

Australia) and cultured onsite using the cell culture procedures described in

Section2.2.3.

4.2.3 SR-IRMSstudies

4.2.3.1 Preparationofcellsamples

HCT-8cells(2.5×105cells,1mLgrowthmedium)weregrowndirectlyontoCaF2

windows (12mm diameter, 0.5mm thickness) that were located in 24-well cell

cultureplates.Sincethecellsdidnotadheretothewindowsrapidly,astainlesssteel

spacerwitha5mmdiameterholewasplacedontopofeachCaF2windowtocreatea

confinedarea for the cells to attach. Once theHCT-8 cellshadadhered, the spacers

were removed and the cells were incubated for a further period of time. No

differences in cell morphology were noted between the method employed in this

Chapter and the traditional cell culture method. Prior to treatment, the growth

mediumwasremovedandthecompounds(initiallydissolvedinDMSO,andaddedto

1mLRPMI-1640mediumtoproduceafinalDMSOconcentrationof2%v/v)orvehicle

(1mLRPMI-1640,2%v/vDMSO)wereaddedto thewell. HCT-8cellswere treated

witheachBiNSAIDataconcentrationequaltotheapproximateIC50(asdeterminedby

MTTassays, Section2.3.2), or anequimolar concentrationof the corresponding free

NSAID. At the end of the 4- or 24-h period, the cells were washed once with PBS

solution.

SR-IRMSspectraof liveHCT-8cellswerecollectedusingthedemountableliquid

cell shown in Figure 4.4, details of which (design and specification) have been

describedelsewhere[32].Firstly,the0.5-mmthickCaF2windowontowhichtheHCT-

8cellsweregrownwasplacedatthebottomofthesampleholderwiththecellsurface


158

facingupwards.Adropofcellmedium(20μL)wasdepositedinthecentreoftheCaF2

windowdirectlyabovethecells. Ontopofthiswasplacedthe0.5-mmCaF2window

with the spacer that had been patterned by polymer spin coating and UV

photolithography. This second CaF2 windowwas orientated such that the polymer

spacer was in contact with the underlying CaF2 window (Fig 4.4a). Next, a plastic

spacerwasplacedontopoftheCaF2windows(Fig.4.4b),afterwhichwasplacedthe

top of the Thermomicro-compression cell (Fig. 4.4c). Excess liquidwas expressed

throughagapinthemicro-fabricatedspacer[32]asthesampleholderwastightened

andasthecellscameintocontactwiththetopwindow.ItshouldbenotedthatHCT-8

cellsaregenerallyobservedasamixtureofindividualcellsorasmonolayersofcells,

whichistypicalforthegrowthpatternofthecellline.Cellsthatgrewinmonolayers

appearedtoflattenwhenthesampleholderwastightened.Itwasalsoobservedthat

themonolayersshowedsignsofdetachmentaftertreatment,whichwouldaccountfor

thisobservation. Nodifferencestothemorphologyofthesinglecellswerenoted. In

order to obtain results representative of the entire population, effort was taken to

collectcellspectrafrombothsinglecellsandcellsthatcouldbedefinedasindividual

cellsinthecellmonolayer.

Figure4.4: Construction of the compression cell holder components employed for SR-IRMSmeasurements:(a)TheCaF2windowontowhichaspacerhadbeenpatternedbypolymerspincoatingandUVphotolithographywasplacedontopofthewindowontowhichtheHCT-8cellsweregrown.ThecellswerehousedinthecentreoftheCaF2windowandthechannelsetchedintothespacerenabledafullyaqueousenvironmenttobemaintained;(b)Aplasticspacerwasplacedontopof theetchedCaF2window;(c)Thetopof theThermomicro-compressioncellwasplacedontopof theCaF2windowsandplasticspacerandtighteneduntil thecellswereobservedtocomeintocontactwiththetopetchedCaF2window.


159

The live cell SR-IRMSanalyseswere limited to approximately 100min for each

sample due to the loss of cellmedium from the demountable liquid cell used in the

present study. As such, the cellswere treated in cell culture plates for the desired

period(4hor24h),andtheCaF2substratewiththedropletoffreshcellmediumwas

placedinthedemountablecellforanalysis.

4.2.3.2 SR-IRMSdatacollection

AllFTIRspectroscopymeasurementswererecordedintransmissionmodeatthe

IRMbeamlineattheAustralianSynchrotron(Clayton,Australia).Thesetupemployed

a Bruker Vertex V80v spectrometer coupled to aHyperion 2000microscopewith a

liquidnitrogencoolednarrow-bandmercurycadmiumtelluride(MCT)detector. The

microscope XYZ stage was housed in a nitrogen-purged atmosphere to minimise

interference from atmospheric water vapor and carbon dioxide. Spectra were

acquired at 4cm−1 resolution over the range 3800–700cm−1 (Blackman–Harris

3-Termapodization,Mertzphasecorrectionandazerofillfactorof2)usingOPUS6.5

software(BrukerOpticsGmbH,Ettlingen,Germany).Theknife-edgeaperturewasset

to5×5µm2andcentredovereachcelltoensureonlycellmaterialwasanalysed.Each

cellspectrumwasrecordedusing128co-addedscans. Abackgroundspectrum(128

co-addedscans)ofcellmediumwasrecordedaftercollectingevery twocell spectra.

Cellspectraweregenerallycollectedwithin20–100minafterassemblyofthesample

holderasthecellmediumbegantoevaporatefromtheedgesofthesandwichedCaF2

windowsafterthisperiod.

4.2.3.3 SR-IRMSspectralprocessingandchemometricanalysis

ThefirststageofFTIRspectralprocessingwasperformedusingOPUSVersion7.0

(Bruker Optics GmbH, Ettlingen, Germany) software. All spectrawere restricted to

3020–1000cm−1. Although the sampleswereanalysed in aN2purgedenvironment,

weakCO2bandswereobservedinanumberofthespectra. Assuchtheatmospheric

compensationwasappliedtominimisetheappearanceofCO2.Inaddition,areference

spectrum collected from the cell medium background for each sample (using the

conditions described in Section 4.2.3.2) was subtracted from each individual cell

spectrum.Thiswasperformedintheinteractivemodeusingiterativestepstoensure

that the effects of the water spectrum were minimised without undue

overcompensationofthecellspectrum(Fig.4.5);thecorrectionvaluegenerallyvaried


160

from 0 to −0.09. Following water subtraction, each spectrum was rubberband

baseline corrected using 64 baseline points. The Unscrambler X software,

Version10.2 (Camo, Oslo, Norway) was employed for chemometric analysis. The

spectrawere analysed by PCA (full cross validation, 7 PCs) following derivatisation

(second derivative, Savitzky-Golay algorithm, 13 smoothing points) and unit vector

normalisation.

Figure4.5: Individual cell spectrum (—)before (toppanel) andafter (bottompanel)watersubtractiontoacorrectionvalueof−0.04usingOPUS7.0.Areferencespectrumcollectedfromthecellmediumbackgroundisrepresentedinthetoppanel(—).


161

4.3 Results

4.3.1 Comparisonofspectralcorrectionfactorsusedfordataanalysis

In this Section the SR-IRMS spectra of HCT-8 cells treated with RPMI-1640

containingDMSO (2% v/v, referred to herein as vehicle) andHCT-8 cells incubated

withRPMI-1640(withoutfetalbovineserum,penicillin/streptomycinorGlutaMAXTM

solutions,referredtohereinascellmedium)havebeenusedto:(i)describeandjustify

thecorrectionprocessesusedforthespectraldataanalysespresentedinthisChapter

(Sections4.3.1.1and4.3.1.2), and (ii)producecontroldata toestablishwhetherany

differencescouldbedetectedinthebiomolecularsignaturesofthetreatedHCT-8cells

over the course of the drug treatment periods (4 h, Section 4.3.2, or 24 h,

Section4.3.3).

In general, theSR-IRMSspectra collected from theHCT-8 cells (typical example

showninFig.4.6a)exhibitedwatersaturation.Thiswasobservedastheνs(OH)band

typically centred around 3400 cm−1 (not shown in Fig. 4.6a), a decrease in the

intensity ratio of the amide I/amide II bands, and the presence of the water

combination band between 2500–1800 cm−1, which creates a characteristic broad

curveinthespectrum(indicatedbytheleftmostarrowinFig.4.6a). Asexplainedin

Section4.1.4,thepresenceofthewaterbandoverlappingtheamideIband(andtoa

lesser extent the amide II band) presents one of the biggest challenges when

interpretingSR-IRMSspectraoflivecells.Twoapproachesweretakeninaneffortto

extractandinterpretthebiomoleculardifferencesbetweentheSR-IRMSspectraofthe

cellsamples:(i)waterbandcorrectionusingspectralsubtraction,and(ii)omissionof

theamideIandamideIIbandregion(1709–1486cm−1)fromPCA.Inadditiontothe

presence of strong water bands in the SR-IRMS spectra, high noise was observed

below1150cm−1asaresultofthepoortransmissionofIRlight.Thiswasduetothe

combination of the two CaF2windows and the small aperturewindow (5 × 5 μm2)

used in theseexperiments.As such, spectraldataatwavenumbersbelow1150cm−1

wereomittedbeforeperformingthePCApresentedinthisChapter.

4.3.1.1 Waterbandcorrectionusingaspectralsubtractionmethod

The averaged, baseline corrected and vector normalised SR-IRMS spectra of

HCT-8cellsincubatedwithcellmedium(4h),beforeandafterwaterbandsubtraction

are shown inFigure4.6a. As indicatedby the arrows inFig. 4.6a, themost striking

differenceobservedbetweenthenon-subtractedandwaterbandsubtractedSR-IRMS


162

spectra was the decrease in the absorbance of the amide II band (1548 cm−1) in

comparison to the amide I band (1653cm−1), and the flattening of the water

combination band between 2500–1800cm−1 (this result was expected as the water

combination band was used to determine the amount of correction to apply)

producingaSR-IRMSspectralprofileintheamideregionmorerepresentativeofthat

observedfromfixedcellspectra(seeFig.4.2).

Figure 4.6: (a) Averaged (n = 59), baseline corrected (rubberband baseline corrected,64baseline points), vector normalised HCT-8 cell spectra following 4-h incubation in cellmedium before (—), and after (—), water band subtraction. Black arrows indicate thespectral feature changes substantially affected by water band subtraction. (b) Unit vectornormalised, second derivative of the spectra shown in (a). The wavenumbers in blackrepresent thenon-correctedbandpositions,whilewavenumbers in red represent thewaterband subtracted band positions. The black arrows denote bands affected by water bandsubtraction.


163

Figure 4.6b shows the averaged, normalised, second derivative spectra of the

same SR-IRMS spectra presented in Figure 4.6a. The second derivative

representations of the averaged SR-IRMS spectra were typically examined as they

resolve theoverlappedabsorptioncomponents,whichprovideclearerelucidationof

band positions and band shifts [59, 60]. Compared to the non-subtracted second

derivative spectrum, a very slight reduction in the intensity of the amide II band

(1551cm−1), the regionbetween the amide I and amide II bands (1601–1591cm−1),

andtheν(C=O)band(1714cm−1)wereobservedfollowingwaterbandsubtraction(as

indicated by the arrows in Fig. 4.6b). Importantly, however, no differences were

observed in the band positions between the non-subtracted and water band

subtracted spectra (Fig. 4.6b). Following the application ofwater band subtraction,

these observations were similarly noted when the averaged, normalised, second

derivative SR-IRMS spectrum of the 4-h vehicle-treated cells was examined

(Appendix3.1),suggestingthattheroutineapplicationofthewaterbandsubtraction

methodproducesaconsistentresponseacrossdifferentsamples.

WhentheaverageSR-IRMSspectrumofHCT-8cellstreatedwithcellmediumwas

comparedtothatofthevehicle,beforewaterbandsubtraction(Appendix3.2), there

wasonlyaslightdifference in the intensitiesof themajorbands. Additionally,band

shifts of only 1–2 cm−1 were observed in the bands in the CH2/CH3 stretching

region(3000–2800 cm−1), the lipid ν(C=O) band (1742 cm−1) and the amide I band

(1653cm−1),asindicatedinAppendix3.2a. GiventhenoiseobservedintheSR-IRMS

spectrabelow1150cm−1,anyshiftsintheνs(PO2−)bandat1088cm−1areunlikelyto

be reliable and these observations have not been discussed further. Following the

application ofwater band subtraction, the intensities of the SR-IRMS spectra of cell

medium- and vehicle-treated cells showed closer agreement (Appendix 3.2b),

particularly in the fingerprint region (1750–1000 cm−1); however, there was an

increaseintheabsorbanceofthelipidbands,νas(CH3)andνas(CH2)at2958cm−1and

2926cm−1,respectively,intheaveragevehicle-treatedspectrum.Itisnoteworthythat

theν(C=O)bandsof lipidsandDNAat1742cm−1and1714cm−1, respectively,were

lessdefinedfollowingwaterbandsubtraction(Appendix3.2b);although,examination

of thesecondderivativespectrashowednoshift inwavenumber inrelationto these

bands (Appendix 3.3a-b). In fact, very little difference was observed between the

averaged second derivative spectra of cell medium- and vehicle-treated cells (4 h)

beforeandfollowingwaterbandsubtraction(Appendix3.3a-b).


164

4.3.1.2 Statisticalanalysisofdata:Principalcomponentanalysis(PCA)

Themultivariate data analysis technique, PCA,was employed in this study as a

means to distinguish any biomolecular differences between the cell populations

treatedwithBiNSAIDsorNSAIDs,orwithinthecellpopulationsthemselves.AstheIR

spectracontainthousandsofwavenumbers(orvariables)astatisticalmethodsuchas

PCAisemployedtoreducethecomplexityofthedataset,bycreatingasmallnumber

ofuncorrelatedartificialvariables,orprincipalcomponents(PCs),thatwillaccountfor

mostof thevariance in theobservedvariables [60]. Thesamplevarianceof thePCs

are representedbyeigenvaluesandare rankedaccording to theirmagnitude,where

the first eigenvalues account for a large portion of the variance of the data [60].

Assuch, the first PC (PC-1) is the linear combination with maximal variance (the

largesteigenvalue),andthesecondPC(PC-2)representsthenexthighestvariancein

anorthogonaldirectiontoPC-1[60,61]. PCAdisplaysthemainpatternsasascores

matrix(orplot),whereeachpoint(score)ontheplotrepresentsonecellspectrum;as

such, similar spectra cluster together, while dissimilar spectra appear further away

from each other [61, 62]. Additionally, PCA produces loadings plots that provide

information on the variables (spectral bands arising from differences in the

biomoleculargroupswithin thesamples in thecaseofSR-IRMSdata),whichare the

determinantsfortheclusteringvisualisedinthescoresplots[62]. Inthiswork,PCA

wasperformedonsecondderivativeSR-IRMSspectra. Assuch,positiveloadingsare

associatedwithscoresinthenegativespaceofthescoresplot,andnegativeloadings

are associated with positive scores [31]. Using PCA, the most significant spectral

differencesandtrendscanbeidentified[62].

Thestartingpointwas tousePCAasa tool toevaluate theeffectofwaterband

subtraction. PCA scores plotswere generated from spectrawithout andwithwater

bandsubtractionandcompared.Whileitwouldbeidealtocheckeachnon-subtracted

cell spectrum against the water band subtracted spectrum of each individually

correctedcell, thisprocesswouldbeextremelytime-consuminggiventhenumberof

spectra collected. The average spectra of the cell medium and vehicle-treated cell

spectra (Appendix 3.2 and Appendix 3.3) represent the average of over 50

unsynchronised living HCT-8 cells; as such, differences between the spectra of cells

within the population need to be examined in order to determine whether cells

subjectedtodifferentconditionsrepresentbiochemicallydistinctpopulations.


165

Figure 4.7 shows the PCA scores plot and corresponding loadings plots (3000–

2800cm−1,1790–1150cm−1)comparingcellmedium-andvehicle-treatedcellspectra,

without(Fig.4.7a)andwith(Fig.4.7b)waterbandsubtractionapplied. Whilesome

slightchangestothecoordinatesofseveralofthescorescanbeobserved(Fig.4.7b),

theoveralldistributionofthescoresappearsverysimilartothescoresplotobtained

from the spectrawithoutwaterband subtraction (Fig. 4.7a). Additionally, the same

bands in the loadings plots of both the non-subtracted (Fig.4.7a), and water band

subtracted cells (Fig. 4.7b)were responsible for the distribution of the scores along

thePC-1andPC-2axes. Consequently, these results indicate that there isnegligible

difference to theoutcomeof thePCAwhenwaterbandsubtraction is applied to the

SR-IRMSspectraoftheHCT-8cellscomparedtonon-subtractedspectra.

Figure4.7:PCAscoresplot(left)andcorrespondingPC-1(—)andPC-2(—)X-loadingsplots(right) of second derivative (Savitzky Golay, 13 smoothing points), unit vector normalisedSR-IRMSspectraobtained fromHCT-8 cells treated for4hwith cellmedium (n), or vehicle(RPMI-1640,DMSO2%v/v)(l):(a)beforewaterbandsubtraction,and(b)afterwaterbandsubtraction.PCAwasperformedonthespectra,encompassingallbiologicalregionsofinterest(3000–2800cm−1,1790–1150cm−1).


166

Whentheamidebandswereomitted(1709–1486cm−1)fromthePCA(Fig.4.8)of

second derivative SR-IRMS 4-h cell medium- or vehicle-treated spectra, without

(Fig.4.8a) andwith (Fig. 4.8b)water band subtraction applied, therewas negligible

difference observed between the scores plots of the non-subtracted or water band

subtracted spectra. Additionally, negligible differencewas observed in the loadings

plotofthewaterbandsubtractedspectra(Fig.4.8b)comparedtothenon-subtracted

spectra (Fig. 4.8a). Consequently, the analysis presented in the remainder of this

Chapterwasperformedonnon-subtractedspectra.

Figure4.8:PCAscoresplot(left)andcorrespondingPC-1andPC-2X-loadingsplots(right)ofsecond derivative (Savitzky Golay, 13 smoothing points), unit vector normalised SR-IRMSspectra obtained from HCT-8 cells treated for 4 h with cell medium (n), or vehicle(RPMI-1640,DMSO2%v/v)(l).PCAwasperformedonthespectra(3000–2800cm−1,1790–1710cm−1, and1485–1150 cm−1)with the omissionof the amide region (1709–1486 cm−1):(a)beforewaterbandsubtraction,and(b)afterwaterbandsubtraction.

4.3.1.3 EffectsofvehicleonliveHCT-8cells(4-hstudy)

Havingshown the rationale for thedataanalysis, thenext stepwas to interpret

theresultswithrespecttothecelltreatment.Figure4.7a(Section4.3.1.1)showsthe


167

PCAscoresplotandcorrespondingloadingsplots(3000–2800cm−1,1790–1150cm−1)

that compare4-h cellmedium- and vehicle-treated cell spectra,withoutwater band

subtractionapplied. Thescoresplotshowscompleteoverlapofthecellmediumand

vehicle scores and no distinct clustering of either group. This result suggests that

there were no substantial biomolecular differences between the two treatment

groups, and that heterogeneity among the SR-IRMS spectra of the cells within each

sampleisresponsibleforthedistributionofthescoresalongboththePC-1andPC-2

axis. ThePC-1 loadings plot indicated that the lipid-relatedbands, νas(CH2), νs(CH2)

andν(C=O)(2924cm−1,2852cm−1and1743cm−1,respectively),andloadingsrelated

totheamideIband(1662cm−1,1649cm−1and1637cm−1)arethepredominantbands

causing the distribution of the spectra along the PC-1 axis. The PC-2 loadings plot

(Fig.4.7a) indicated that the lipid-related bands, νas(CH2), νs(CH2) and ν(C=O)

(2926cm−1,2854cm−1and1745cm−1,respectively),andloadingsrelatedtotheamide

Iband(1662cm−1and1647cm−1)andamideIIband(1562cm−1and1539cm−1)are

the predominant bands causing the distribution of the spectra along the PC-2 axis.

These results suggest that variations in cellular proteins and lipidswere present in

each population. This result is not unexpected as the cells were not cell cycle

synchronised; thus the distribution of cells throughout different phases of the cell

cyclecouldbethelikelycauseofheterogeneityamongstthetwopopulations.Thisis

also consistent with other live cell studies that have shown that cells have distinct

biomolecularsignaturesatdifferentphasesofthecellcycle[43].

Figure4.8ashowsthePCAscoresandloadingsplotsofsecondderivativeSR-IRMS

4-h cellmedium- or vehicle-treated spectra following omission of the amide region

(3000–2800 cm−1, 1790–1710 cm−1, 1485–1150cm−1), without water band

subtraction.Boththecellmediumandthevehiclescoresweredistributedacrossboth

the positive and negative PC-1 space. Once again, this suggests that heterogeneity

exists within both cell populations, and that the biomolecular differences that exist

within each sample aremore dominant than the biomolecular differences that exist

between the two samples. In contrast to Figure 4.7a (amide region included), the

loadings plot in Figure 4.8a shows that there are strong lipid-related loadings at

2924cm−1(νas(CH2)),2839cm−1(νs(CH2)),and1745cm−1(ν(C=O))andaweakloading

at 1464cm−1 (δ(CH2)) that are responsible for the distribution of the scores along

PC-1. This observation suggests that variation in cellular lipids exists between the

SR-IRMSspectrainboththecellmedium-,andvehicle-treatedcellpopulations.Along


168

thePC-2space,therewasnocleardelineationofthetwosamples,although,thescores

obtained from the cellmedium-treated cells appearedpredominantly in thepositive

PC-2space,whilethescoresfromvehicle-treatedcellsappearedpredominantlyinthe

negative PC-2 space. Figure 4.8a shows that strong loadings in the region between

1250-1150cm−1were observed, whichwere not present in the loadings plot of the

PCAperformedwiththeamideregionincluded(Fig.4.7a);theloadingsat1252cm−1,

1238 cm−1, and 1227 cm−1 are associated with the νas(PO2−) band of nucleic acids.

Additionally, thederivative-likebands at2931 cm−1 and2847 cm−1 are indicativeof

band shifts of lipids. These observationsmay indicate that the 4-hDMSO (2% v/v)

exposuremayinducechangestolipidandnucleicacidconcentrationsinHCT-8cells.

4.3.1.4 EffectofvehicleonliveHCT-8cells(24-hstudy)

HCT-8 cells were incubated with cell medium or vehicle for 24-h in order to

correlatewiththetreatmentperiodemployedinthetoxicityanduptakeexperiments

performed with the HCT-8 cells (Chapter 2 and 3, respectively). The averaged,

baselinecorrectedandvectornormalisedSR-IRMSspectraofHCT-8cellstreatedwith

cellmediumorvehicle(24h),beforeandfollowingwaterbandsubtractionareshown

in Appendix 3.4. Similar to the 4-h samples (Appendix 3.2), only small differences

couldbeobservedwhencomparingthe24-haveragecellmediumandvehicle-treated

cellspectra,beforeorfollowingwaterbandsubtraction(Appendix3.4).Additionally,

theeffectsofwaterbandsubtractionwerealsoconsistentwiththeeffectsobservedon

the4-hsamples(Appendix3.2);specifically,theintensityratiooftheamideI/amideII

bands were affected, and the intensity of the ν(C=O) bands at 1743cm−1 and

1715cm−1wereslightlyreduced(Appendix3.4). Aslightshiftinthebandcentresof

the ν(C=O) bands was observed following water band subtraction (Appendix 3.4a

andb);however,nobandshiftwasnotedintherespectivesecondderivativespectra

(Appendix3.5aandb).Littledifferencewasobservedinthesecondderivativespectra

whencomparingthecellmedium-andvehicle-treatedcellswiththeexceptionofthe

line shape of the νas(PO2−) centred at 1240 cm−1 (and shoulder at 1227 cm−1)

(Appendix3.5aandb).Thismightindicatethatthepresenceofvehicleinducessmall

changestothenucleicacidcompositionoftheHCT-8cellsfollowing24-hincubation.

ThePCAscoresplotinFigure4.9a,whichincludestheamideregion(3000–2800cm−1,

and1790–1150cm−1),showedtherewasnodelineationofthe24-hcellmediumand

vehiclescoresalongthePC-1orPC-2axesbetweenthesamplesetssuggestingthatthe


169

biomolecular differences that existwithin each sample aremore dominant than the

biomoleculardifferencesthatexistbetweenthetwosamples.ThePC-1loadings(Fig.

4.9a) were attributable to CH2 stretches at 2924cm−1 and 2852cm−1 (νas(CH2) and

νs(CH2), respectively), and the ester stretch at 1743 cm−1 (ν(C=O)),which are bands

associatedwith lipids, amide-relatedbandsat1657cm−1,1549cm−1, and1398cm−1

(amide I,amide IIandνs(COO−)),and theν(C=O)bandat1716cm−1, related toDNA.

The derivative-like band shapes in the PC-2 loadings plot

(Fig.4.9a)indicatedthatshiftsinthepositionsoftheνas(CH2),ν(C=O),amideI,

amideIIandνas(PO2−)bandswerecontributingtothedistributionofthescoresalong

the PC-2 axis. These observations suggest that intra-sample variations in the

compositionoflipid,protein,andtheDNAexistwithinbothofthecellpopulations.

Figure4.9:PCAscoresplot(left)andcorrespondingPC-1(—)andPC-2(—)X-loadingsplots(right) of second derivative (Savitzky Golay, 13 smoothing points), unit vector normalisedSR-IRMS spectra obtained fromHCT-8 cells treated for 24 hwith cellmedium alone(n), orvehicle (RPMI-1640, DMSO 2% v/v) (l): (a) including the amide region (3000–2800 cm−1,1790–1150cm−1), and (b) excluding the amide region (3000–2800 cm−1, 1790–1710 cm−1,1485–1150cm−1).


170

Once again, this observation could potentially be attributed to the lack of cell cycle

synchronisation.

Figure 4.9b shows the PCA scores and loadings plots obtained using the cell

medium-,orvehicle-treated(24h)SR-IRMSspectrapresentedinFigure4.9a,butwith

theomissionof theamideregion(1709–1486cm−1). AswithFigure4.9a, thescores

plot showed no sign of delineation of the cell medium or vehicle scores from one

another (Fig.4.9b); again intra-sample variation appeared to be the predominant

sourceofthebiomoleculardifferencesinthesesamples. Intheabsenceoftheamide

region,lipid-relatedbands(2924cm−1(νas(CH2)),2854cm−1(νs(CH2)),and1743cm−1

(ν(C=O))dominatedthePC-1 loadingsplot(Fig.4.9b). Interestingly,omissionof the

amide region resulted in tighter distribution of the scores along the PC-2 axis as

showninthescoresplotinFigure4.9b.Theresults(Fig.4.9aandb)indicatethatthe

strongamideIloadingsmayberesponsibleforthewiderdistributionofscoresalong

the PC-2 axis in the scores plot when the amide region is included (Fig. 4.9a).

Asshown in Figure 4.9b, themost pronounced PC-2 loadingswere the νas(CH2) and

νs(CH2) bands,which appeared as a derivative-like shape, indicative of a band shift.

Someweak PC-2 loadings contributions from the νas(PO2−) band (1259–1217 cm−1)

werealsoevident. Overall, thePCAperformedwithout theamidespresent suggests

that intra-sample lipid variations exist in the cell spectra of the cell medium- and

vehicle-treated samples.Asno separationof the cellmedium-orvehicle-treatedcell

spectra scores is evident, the PCA results suggest that the intra-sample spectral

variationwithin each of these treatment populations is dominant over any spectral

variationsoccurringasaresultofthetwotreatments.

ThePCAresultsappear to suggest thatDMSOhas littleobservableeffecton the

SR-IRMSHCT-8 cell spectra following 24-h incubation, as compared to the SR-IRMS

spectraofHCT-8cellsincubatedwithcellmediumalone.

4.3.2 ComparisonoftheeffectsofBiNSAIDs,andtherespectiveNSAIDs,onSR-

IRMSspectraofliveHCT-8cells

4.3.2.1 Bi(tolf)3-andtolfH-treatedcells(4-htreatment)

The averaged SR-IRMS spectra (29–54 spectra recorded for each treatment) of

vehicle-,Bi(tolf)3-,andtolfH-treatedcellsshowedafewnotabledistinctionsfromeach

other following4-h treatmentas shown inAppendix3.6. Notably, theBi(tolf)3- and

tolfH-treatedaveragecellspectrashowedanincreaseinintensityoftheνas(CH2)and


171

νs(CH2) bands (2924 cm−1 and 2853 cm−1, respectively) compared to the vehicle-

treated average spectrum (Appendix 3.6a). In addition, the average tolfH-treated

spectrum showed a distinctive shift in the νas(PO2−) band (from 1240 cm−1 to

1232cm−1)whencomparedtothevehicle-treatedaveragespectrum(Appendix3.6a),

whiletheBi(tolf)3-treatedaveragespectrumshowedonlyaslightshift to1238cm−1.

The substantial shift in the νas(PO2−) was also identifiable in the average second

derivativespectrumoftolfH-treatedcells(Appendix3.6b),withalesssubstantialshift

in the Bi(tolf)3-treated average spectrum compared to the vehicle (Appendix 3.6b).

Achangeassociatedwiththenucleicacidsisobservedasathefrequencyshiftofthe

ν(C=O) band from 1716cm−1 (vehicle) to 1718 cm−1 and 1720 cm−1, in the

Bi(tolf)3-andtolfH-treatedaveragespectra,respectively(Appendix3.6b).

A comparison of the SR-IRMS spectra following 4-h treatment with vehicle,

Bi(tolf)3,ortolfH,isshowninthePCA(3000–2800cm−1,1790–1150cm−1)scoresand

loadingsplotsinFigure4.10a.AtfirstglanceitwasobservedthatboththeBi(tolf)3or

tolfHscoresoverlapwith thevehicle scoresand that there isnocleardelineationof

theclusteringbetweenthethreepopulations.However,oncloserinspectionthetolfH

scores appear to showan increasedproportionofpositivePC-1 scores compared to

thevehicle(whichappeartobepredominantlynegativePC-1scores).Thereappears

tobeevendistributionofBi(tolf)3inboththenegativeandpositivePC-1scores.The

PC-1loadingsplotinFigure4.10ashowsthatthepositivePC-1scoreswereassociated

withstrongnegativeloadingsfrommethylenevibrationsat2924cm−1and2854cm−1

(νas(CH2)andνs(CH2),respectively),andmoderatetoweaknegativeloadingsobserved

at1466cm−1(δ(CH2))and1745cm−1(ν(C=O)).Theseobservationsmayindicatethat

therewasadifferenceinlipidcompositioninsubpopulationsofthetolfH-treatedcells

andBi(tolf)3-treatedcells,comparedtothemajorityofthevehicle-treatedcells. Two

weakpositivePC-1loadingsrelatedtotheamideIband,at1693cm−1and1620cm−1,

wereassociatedwith theaforementionedscores,whichmaysuggestachange in the

compositionofcellularproteinsfollowingNSAIDtreatment.ThenegativePC-1scores

(consistingofthemajorityofthevehiclescores,andsomeoftheBi(tolf)3scores)were

correlatedwithmoderatetoweakpositiveloadingsassociatedwithproteinsincluding

1662cm−1and1639cm−1(amideI),1549cm−1(amideII),and1398cm−1(νs(COO−))

(Fig.4.10a).Aweakpositiveloadingat1248cm−1(νas(PO2−)),associatedwithnucleic

acids, was also correlatedwith the negative PC-1 scores. Along PC-2, therewas no

cleardistinctionofthethreesamplegroups;infact,therelativelyevendistributionof


172

the samples along PC-2 suggests intra-sample variationwas greater for each of the

threesamplesthantheinter-samplevariation.ThePC-2loadingsshowedderivative-

like loadings(indicativeofbandshifts) in theamide Iregionsuggesting thatcellular

proteincompositionwas thegreatestdeterminantof intra-samplevariability. Weak

contributionsfromlipidloadingsalsocontributedtointra-samplevariability,withno

contributions from the lower wavenumber region of the spectrum (<1500cm−1)

evidentintheloadingsplot(Fig.4.10a).

Figure4.10: PCA scoresplot and correspondingPC-1 andPC-2X-loadingsplots generatedfrom the second derivative (Savitzky Golay, 13 smoothing points), unit vector normalisedSR-IRMSspectraobtainedfromHCT-8cellsfollowing4htreatmentwithvehicle(RPMI-1640,DMSO2%v/v)(n),Bi(tolf)3(l),ortolfH(p)(45μMNSAID);(a)includingtheamideregion(3000–2800cm−1, 1790–1150 cm−1), and (b) excluding the amide region (3000–2800 cm−1,1790–1710cm−1,1485–1150cm−1).

Figure 4.10b shows the results of PCA comparing second derivative spectra

obtained from the4-h vehicle-,Bi(tolf)3-, or tolfH-treated cells,with theomissionof

theamideregion(1709–1486cm−1).AsthePCAscoresplotinFigure4.10bshows,the

tolfHscoresexhibitsomedelineationfromboththevehicleandBi(tolf)3scoresalong


173

PC-2 following removal of the amide region. However, closer inspection of the PCA

scores plot in Figure 4.10b reveals that a similar proportion of tolfH and Bi(tolf)3

scoresappearontheoppositesideofthePC-1axiswhencomparedtothePCAscores

plotinFigure4.10a(albeitonthereversesideofthePC-1axis).Additionallythetwo

PC-1loadingsplotsinFigure4.10showcomparablecontributionsfromsimilarbands

(although the PC-1 loadings plot appears to be inverted in Fig.4.10b). Thus, the

subpopulations of Bi(tolf)3 and tolfH scores were associated with strong positive

loadings from theνas(CH2), νs(CH2), andmoderate toweak loadings from theν(C=O)

and δ(CH2) bands of lipids, indicating a change in the lipid composition of

subpopulationsoftheNSAID-treatedcellscomparedtothevehicle-treatedcells. The

obviouschangeinthedistributionofthetolfHscoresalongPC-2wasaccompaniedby

an increase in the PC-2 loadings contributions from bands in the 1465–1150 cm−1

region following the removal of the amide region from the PCA (Fig.4.10b). The

positivePC-2scores(vehicleandBi(tolf)3scoresonly)wereassociatedwiththestrong

ν(C=O)loadingat1745cm−1,theνas(PO2−)bandofnucleicacidsat1246cm−1;andthe

νs(COO−) band at 1398cm−1. These changes demonstrate that lipids, proteins, and

nucleicacidsaredifferentlyaffectedbytolfHcomparedtoBi(tolf)3intheHCT-8cells

following4-htreatment.

4.3.2.2 Bi(tolf)3-andtolfH-treatedcells(24-htreatment)

The averaged SR-IRMS spectra of vehicle-, Bi(tolf)3, and tolfH-treated cells

(between 33–100 spectra recorded for each treatment) did not show very many

differencesfromeachotherfollowing24-htreatment,asdisplayedinAppendix3.7.It

wasevidentthattheaveragetolfH-treatedcellspectrumshowedreducedintensityin

the lipid region (3000–2800 cm−1), the ν(C=O) band (1740 cm−1) and the νas(PO2−)

band, indicating that tolfH treatment may reduce cellular lipid and nucleic acid

concentration. Incontrasttothe4-hsamples(Appendix3.6),therewereonlysubtle

differences in the line shape of the νas(PO2−) band in the 24-h second derivative

spectra of the average tolfH- and Bi(tolf)3-treated cells compared to the vehicle-

treatedcells(Appendix3.7).

Figure4.11showsthePCAcomparisonoftheSR-IRMSspectraofthesingleHCT-8

cells following 24-h treatment with vehicle, Bi(tolf)3, or tolfH. The scores plot

(Fig.4.11a) shows that the vehicle andBi(tolf)3 scoresweredistributed acrossPC-1

withnodelineationfromeachother;however,asmallBi(tolf)3subpopulationcluster


174

maybedistinguishedinthePC-1/PC-2negativespaces(Fig.4.11a). Whiletherewas

noobviousdelineationofthetolfHscoresfromtheothertwosamples,itwasobserved

that the tolfH scores appeared as predominantly positive PC-1 scores in a

comparatively tight cluster (Fig. 4.11a); additionally, the tolfH scores appeared

primarily in the negative PC-2 space. The negative PC-1 scores (comprised of

subpopulationsofvehicleandBi(tolf)3scores,Fig.4.11a)wereassociatedwithstrong

loadingsfromlipidbandsincludingνas(CH2),νs(CH2),ν(C=O)andδ(CH2).Conversely,

negativePC-1loadingsindicativeofproteinsat1655cm−1,1549cm−1and1398cm−1

(amideI,amideII,andνs(COO−),respectively)werecorrelatedwiththepositivePC-1

Figure 4.11: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots of secondderivative (Savitzky Golay, 13 smoothing points), unit vector normalised SR-IRMS spectraobtainedfromHCT-8cellsfollowing24htreatmentwithvehicle(RPMI-1640,DMSO2%v/v)(n), Bi(tolf)3 (l), or tolfH (p) (45 μMNSAID); (a) including the amide region (3000–2800cm−1, 1790–1150cm−1), and (b) excluding the amide region (3000–2800 cm−1, 1790–1710cm−1,1485–1150cm−1).


175

scores(comprisedofa largetolfHpopulation,Fig.4.11a). Theseresultssuggestthat

treatmentwithtolfHresultsinlipidandproteinchanges.ThePC-2loadingsplot

showedthatthemajorityofthetolfHscoresandtheBi(tolf)3subpopulationofscores

inthenegativePC-2spacewereassociatedwiththeνas(CH2),νs(CH2)andν(C=O)lipid-

relatedbands(Fig.4.11a).Thederivative-likeshiftintheamideIandamideIIbands

ofthePC-2loadingsplot(Fig.4.11a)suggeststhereisabandshiftbetweenthetreated

populationsinthePC-2negativespaceandthevehicle(andsomeBi(tolf)3scores)in

thePC-2positivespace.

Asshowninthescoresplot inFigure4.11b, therelativelytightclusteringof the

tolfH scores was still apparent following the removal of the amide region (1709–

1486cm−1)fromPCA.TheBi(tolf)3subpopulationcouldalsobeidentifiedinthePC-1

negative/PC-2 positive space (Fig. 4.11b). Again, strong lipid loadings arising from

νas(CH2), νs(CH2), ν(C=O), and δ(CH2) bands (2924 cm−1, 2852 cm−1, 1743 cm−1, and

1466cm−1,respectively)wereobservedandassociatedwiththePC-1negativescores

(comprised of subpopulations of vehicle and Bi(tolf)3 scores) in Figure 4.11b. The

removalof theamideregionfromthePCAincreasedthe intensityof thenucleicacid

(νas(PO2−)) PC-2 loadings at 1239 cm−1 and 1227 cm−1 and the lipid-related

νas(CO−O−C)loadingat1160cm−1(Fig.4.11b).Derivative-likeloadingswereobserved

inthePC-2loadingsfortheνas(CH2)andνs(CH2)lipid-relatedbandswhentheamides

were omitted (Fig. 4.11b), in contrast to the non-derivative bands in Figure 4.11a.

These observations suggest the lipid and nucleic acid SR-IRMS signatures of these

BiNSAID-andNSAID-treatedsubpopulationsdifferedfromthemajorityofthevehicle-

treatedcellspectrafollowing24-htreatment.

4.3.2.3 Bi(mef)3-andmefH-treatedcells(4-htreatment)

The averaged SR-IRMS spectra of vehicle-, Bi(mef)3, and mefH-treated cells

(between 42–83 spectra recorded for each treatment) showed some differences

following 4-h treatment as shown inAppendix 3.8. When compared to the average

vehicle-treated cell spectrum, the average Bi(mef)3-treated spectrum showed a

slightly increased intensity in the lipid region (3000–2800 cm−1) and the νas(PO2−),

whereas the average mefH-treated spectrum showed the opposite trend

(Appendix3.8a).However,theaveragesecondderivativespectraoftheBi(mef)3-and

mefH-treatedcellsshowedvery littlevariation in lineshapewhencomparedtoeach

other (Appendix3.8b); additionally, both the mefH and Bi(mef)3-treated average


176

spectrashowedthesamevariationinthelineshapeoftheδ(CH2)andνas(PO2−)bands

whencomparedtotheaveragevehicle-treatedspectrum(Appendix3.8b).

Figure 4.12a shows the PCA scores plot comparison of the SR-IRMS spectra

(3000–2800cm−1,1790–1150cm−1)ofsingleHCT-8cellsfollowing4-htreatmentwith

vehicle, Bi(mef)3, or mefH. There was no clear delineation observed between the

vehicle,Bi(mef)3 andmefH scorepopulations;however, themajorityof theBi(mef)3

and mefH PC-1 scores were positive, whereas the vehicle PC-1 scores were

predominantlynegative(Fig.4.12a).ThePC-1loadingsplot(Fig.4.12a)indicatedthat

the majority of the Bi(mef)3 and mefH scores are associated with strong loadings

arising from the νas(CH2), νs(CH2), and νas(C=O) bands (2924 cm−1, 2852 cm−1, and

1743cm−1, respectively), which may indicate that differences between the

mefH/Bi(mef)3-treatedcellsandthevehicle-treatedcellswereattributedto the lipid

bands. Conversely, the PC-1 scores primarily associated with vehicle-treated cells

were correlatedwith loadings attributed to amide I, amide II andνs(COO−)bandsof

proteins at 1637cm−1, 1533cm−1, and 1396 cm−1, respectively (Fig. 4.12a). The

relativelyevendistributionofthescoresofallthreesamplesacrossthenegativeand

positive PC-2 space suggests that intra-sample variation was dominant over inter-

sample variation. Moderate PC-2 loadings attributable to lipids (2926 cm−1, 2854

cm−1, and 1745 cm−1) and strong derivative-like amide I loadings (1660 cm−1, 1647

cm−1,and1635cm−1)suggestvariationinthelipidcompositionandproteinstructure

inthepopulationsofeachofthethreecellgroups.

AsshowninFigure4.12b,thePCAwasperformedwiththeomissionoftheamide

region (1709–1486cm−1) on the same SR-IRMS spectra analysed in Figure 4.12a.

Notably, inFigure4.12bthevehiclescoresclusterwasalmostseparatedfromthatof

the Bi(mef)3 and mefH scores, which were completely overlapped. A higher

proportion of vehicle PC-1 scoreswere positive, while the higher proportion of the

Bi(mef)3andmefHPC-1scoreswerenegative;however,someintra-samplevariation

wasevidentinPC-1,asindicatedbythespreadofthethreeseparatepopulationsover

thenegativeandpositivesidesofthePC-1space.AsshowninthePC-1loadingsplot

in Figure 4.12b, themajority of theBi(mef)3 andmefHPC-1 scoreswere associated

withstrongloadingsarisingfromtheνas(CH2),νs(CH2),andν(C=O)lipid-relatedbands,

similar to the results of the PCA inclusive of the amides (Fig. 4.12a). In the PC-2

loadingsplot(Fig.4.12b) itwasobservedthat theremovalof theamideregion from

thePCAincreasedtheintensityofthenucleicacidνas(PO2−)loadingsat1244cm−1and


177

1223cm−1 and the lipid νas(CO−O−C) loading at 1176 cm−1. Derivative-like loadings

wereobserved in thePC-2 loadings for the lipidbandsνas(CH2)andνs(CH2),and the

νas(PO2−) nucleic acid band, indicating a change in the band position between the

vehiclescoresandtheNSAIDscores.

Figure 4.12: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots of secondderivative (Savitzky Golay, 13 smoothing points), unit vector normalised SR-IRMS spectraobtained fromHCT-8cells following4h treatmentwithvehicle (RPMI-1640,DMSO2%v/v)(n), Bi(mef)3 (l), or mefH (p) (120 μM NSAID); (a) including the amide region (3000–2800cm−1, 1790–1150 cm−1), and (b) excluding the amide region (3000–2800 cm−1, 1790–1710cm−1,1485–1150cm−1).

4.3.2.4 Bi(mef)3-andmefH-treatedcells(24-htreatment)

The averaged SR-IRMS spectra of vehicle-, Bi(mef)3, and mefH-treated cells

(between54–95spectrarecorded foreach treatment)alsoshowedsomedifferences

fromeachother following24-h treatmentasshown inAppendix3.9. Itwasevident

that the averageBi(mef)3- andmef-treated cell spectra showed reduced intensity in

the νas(CH2), νs(CH2), and ν(C=O) bands, indicating that Bi(mef)3-treatment may


178

reduce cellular lipid concentrations (Appendix 3.9a). Additionally, therewere some

variations in the line shape of the νas(PO2−) band in the average second derivative

spectra of the mefH- and Bi(mef)3-treated cells (Appendix 3.9b) compared to the

vehicle-treatedcells.

A comparison of the SR-IRMS spectra of the single HCT-8 cells following 24-h

treatment with vehicle, Bi(mef)3, or mefH, is shown in the PCA scores plot in

Figure4.13a (examining the 3000–2800 cm−1 and 1790–1150 cm−1 regions). The

majority of the Bi(mef)3 scores appear in the negative PC-1 space and show some

delineation from the vehicle scores; however, the mefH scores appear to be

distributed across the positive and negative PC-1 space, overlapping with both the

Bi(mef)3scores,andsomeofthevehiclescores.

Figure 4.13: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots of secondderivative (Savitzky Golay, 13 smoothing points), unit vector normalised SR-IRMS spectraobtainedfromHCT-8cellsfollowing24htreatmentwithvehicle(RPMI-1640,DMSO2%v/v)(n), Bi(mef)3 (l), or mefH (p) (120 μM NSAID); (a) including the amide region (3000–2800cm−1, 1790–1150 cm−1), and (b) excluding the amide region (3000–2800 cm−1, 1790–1710cm−1,1485–1150cm−1).


179

The PC-1 loadings plot shows that νas(CH2), νs(CH2), and νas(C=O) lipid-related

loadingsat2924cm−1,2854cm−1and1743cm−1,respectively,areassociatedwiththe

vehicleandaproportionof themefHscoresonthepositivePC-1scores. Thestrong

loadingoriginated from the amide I band at 1637cm−1 andwas associatedwith the

Bi(mef)3, negative PC-1 scores. Intra-sample variation was observed along PC-2,

resulting from moderate lipid loadings, and strong amide I and amide II protein

loadings.

AsshowninFigure4.13b,thePCAwasperformedwiththeomissionoftheamide

region(1709–1486cm−1)onthesameSR-IRMSspectraanalysedinFigure4.13a.The

majorityoftheBi(mef)3scoresappearinthepositivePC-1spacewithasmallnumber

extending into thenegativePC-1space thatoverlapswith thevehicle scores. Again,

themefHscoresappeartobedistributedacrossthepositiveandnegativePC-1space,

overlappingwithboth thepositiveBi(mef)3scores,andsomeof thenegativevehicle

scores.SimilarlytotheamideinclusivePCA(Fig.4.13b),thePC-1loadingsplotshows

that lipid loadings at 2924 cm−1, 2854 cm−1 and 1743 cm−1 are associatedwith the

vehicleandthesmallproportionoftheBi(mef)3andmefHscoresonthepositivePC-1

scores. With the omission of the amide region, the PC-2 loadings plot has some

obviousdifferencestotheamideinclusivePC-2loadingsplot.Specifically,theνas(CH2)

andνs(CH2)bandsarenowderivative-likeinshape,indicatingabandshiftwithineach

sample. Additionally, the intensity of the νas(PO2−) loadings at 1240cm−1 and

1228cm−1,andtheνas(CO−O−C)loadingsat1184cm−1and1158cm−1wereincreased.

4.3.2.5 Bi(difl)3-anddiflH-treatedcells(4-htreatment)

AsshowninAppendix3.10,theaveragedBi(difl)3-anddiflH-treatedcellspectra

showedadecreaseintheintensityoftheνas(CH2)andνs(CH2)bands(2624cm−1and

2952cm−1)andlackofdefinitionoftheν(C=O)bandat1745cm−1(particularlyinthe

case of diflH). The averaged Bi(difl)3-treated second derivative spectrum showed

substantialdifferencesintheδ(CH2)andνas(PO2−)bandsat1466cm−1and1248cm−1,

respectively, compared to both the averaged vehicle- and diflH-treated second

derivative spectra (Appendix 3.10b). Also evident was the appearance of a slight

shoulder on the amide I band of the diflH-treated average second derivative

spectrum(Appendix3.10b).

Individual secondderivative spectraofvehicle-,Bi(difl)3- anddiflH-treatedcells

wereexaminedusingPCA,andtheresultingscoresandloadingsplots(examiningthe


180

3000–2800 cm−1 and 1790–1150 cm−1 regions) are presented in Figure 4.14a.

ThediflH scores were spread across both sides of the PC-1 axis, whereas both the

vehicleandBi(difl)3scoresshowedahigherdegreeofclusteringrelativetothediflH

scores.DisparityamongtheamideIbandsappearstobethemajorcontributortothe

distributionofthescoresalongPC-1,particularlyforthediflHscores.OnthePC-2axis

therewasnocleardelineationobservedbetweenthevehicle,Bi(difl)3anddiflHscore

populations;however,forthemajorityoftheBi(difl)3anddiflH(thelatterevenmore

so), the PC-2 scores were primarily positive values, whereas the vehicle scores

appearedaspredominantlynegativePC-2values.

Figure 4.14: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots of secondderivative (Savitzky Golay, 13 smoothing points), unit vector normalised SR-IRMS spectraobtained fromHCT-8cells following4h treatmentwithvehicle (RPMI-1640,DMSO2%v/v)(n), Bi(difl)3 (l), or diflH (p) (225 μM NSAID); (a) including the amide region (3000–2800cm−1, 1790–1150 cm−1), and (b) excluding the amide region (3000–2800 cm−1, 1790–1710cm−1,1485–1150cm−1).


181

It was observed in the PC-2 loadings plot that the amide I and amide II bands

were themain source of the variation,with thederivative-like shape of the amide I

band, indicative of a band shift. The PC-2 loadings arising from lipid bands were

associated with the positive PC-2 scores, primarily composed of Bi(difl)3 and diflH

scores.

The PCA scores plot following the omission of the amide region (1709–

1486cm−1) for the comparison of the vehicle-, Bi(difl)3, and diflH-treated cell

spectra(Fig. 4.14b) shows a substantial difference in the distribution of the three

sample populations compared to the amide inclusive PCA scores plot (Fig. 4.14a).

AsobservedinFigure4.14b,thevehicleanddiflHscoresappeartocompletelyoverlap,

whiletheBi(difl)3scoresareprimarilydistributedovernegativePC-1andPC-2values

and only partially overlapwith the other treatments. The distribution of the three

samplesacrossthePC-1axissuggeststhatintra-samplevariationispredominantover

anyinter-samplevariations;thePC-1loadingsplotindicatesthatthelipid-associated

bands, νas(CH2), νs(CH2), and ν(C=O) bands (2926cm−1, 2854cm−1, and 1745cm−1,

respectively) are the source of this variation (Fig. 4.14b). Inter-sample differences

couldbeobservedby theminimaloverlapof theBi(difl)3negativePC-2 scores from

the predominantly positive PC-2 vehicle and diflH scores (Fig. 4.14b). The PC-2

loadings plot indicates that the positive loadings attributable to lipids (δ(CH)2 at

1466cm−1), proteins (amideIII at 1321 cm−1), and nucleic acids (νas(C=O) at

1718cm−1) are associated with the PC-2 negative Bi(difl)3 scores (Fig. 4.14b).

Negative PC-2 loadings designated to lipids (νas(CH2) at 2920 cm−1, and νs(CH2) at

2850cm−1), proteins (νs(COO−) at 1396cm−1) and nucleic acids (νas(PO2−) at

1246cm−1)areassociatedwiththepositivescores,consistingprimarilyofvehicleand

diflHscores(Fig.4.14b).

4.3.2.6 Bi(difl)3-anddiflH-treatedcells(24-htreatment)

AsshowninAppendix3.11a,acomparisonoftheaverageSR-IRMspectraof24-h

vehicle-,Bi(difl)3-ordiflH-treatedcellsrevealedthatthediflH-treatedsampleshowed

astrongincreaseinthebandsassociatedwithlipids(3000–2800cm−1and1740cm−1).

Additionally, the averaged diflH-treated second derivative spectrum showed

substantial differences in the δ(CH2) band at 1466cm−1 compared to both the

averagedvehicle-andBi(difl)3-treatedsecondderivativespectra(Appendix3.11b).

Figure4.15ashowsthescoresandloadingsplots(examiningthe3000–2800cm−1


182

and 1790–1150 cm−1 regions) obtained following PCA of the individual second

derivative spectra obtained from the vehicle-, Bi(difl)3- and diflH-treated cells.

Althoughtherewassubstantialoverlapoftheclustersbetweenthethreesampleson

thescoresplot, thediflHscoresappearedalmostexclusivelyasnegativePC-1scores

(Fig. 4.15a). These scoreswere associatedwith strong νas(CH2), νs(CH2) and ν(C=O)

loadingsof lipids. ThepositivePC-1scorescomposedofvehicleandBi(difl)3 scores

were associated with moderate loadings from the amide I and amide II bands of

proteins.AllthreesamplesshowedscoresdistributionalongthePC-2axissuggesting

intra-samplevariationappearedtobethepredominantfactorinthePC-2distribution

of thesamples.The loadingsplotsuggests thatvariationswithin thepositionsof the

νas(CH2),νs(CH2),andamideIbandsexistedwithinthesepopulations.

Figure 4.15: PCA scores plot and corresponding PC-1 and PC-2 X-loadings plots of secondderivative (Savitzky Golay, 13 smoothing points), unit vector normalised SR-IRMS spectraobtained from HCT-8 cells following 24 h treatment with vehicle (RPMI-1640, DMSO2%v/v) (n), Bi(difl)3 (l), or diflH (p) (120 μM NSAID); (a) including the amide region(3000–2800cm−1, 1790–1150 cm−1), and (b) excluding the amide region (3000–2800 cm−1,1790–1710cm−1,1485–1150cm−1).


183

The PCA results obtained following omission of the amide region (1709–

1486cm−1) from the vehicle, Bi(difl)3 and diflH SR-IRMS spectra are presented in

Figure4.15b.SimilartoFigure4.15a,thediflHscoresappearedalmostexclusivelyas

PC-1 negative scores and remained associatedwith strong positive loadings arising

from lipids (νas(CH2), νs(CH2) and ν(C=O)). As was also observed in Figure 4.15a,

intra-samplevariationappearedtobethepredominantfactorinthedistributionofthe

samples across PC-2; the loadings plot (Fig. 4.15b) indicated that this resulted from

variations within the positions of νas(CH2), νs(CH2) and ν(C=O) bands within the

samples.Relativelylittlechangewasobservedfromtheloadingscontributionsinthe

1485–1150cm−1regionwhencomparing theamide inclusiveoramideomittedPC-1

and PC-2 loadings plots (Fig. 4.15a and Fig. 4.15b, respectively). The scores plots

indicatethat24-hincubationwithdiflHproducesbiomoleculareffects inHCT-8cells

which are relatively distinct from Bi(difl)3- and vehicle-treated cells and appear to

result from lipid-related differences. However, no significant differences were

observed for the Bi(difl)3-treated SR-IRM spectra when compared to the vehicle-

treatedSR-IRMspectra.

4.3.2.7 Summary of the IR spectral and PCA differences following BiNSAID and

NSAIDtreatment

The results described in Sections 4.3.2.1–4.3.2.6 have been summarised in

Table4.3 for ease of reference in the Discussion (Section 4.4). Briefly, the key

observationsemanatingfromtheanalysisoftheSR-IRMSstudiesinclude:

(i) TolfH treatment (4 h and 24 h) of HCT-8 cells predominantly resulted in

changestothelipidbandsandpossiblytheamidebands.

(ii)Bi(tolf)3treatment(4h)causedsomechangeinlipids;however,theeffectwas

less pronounced compared to tolfH. Additionally, the tolfH spectra were

distinguishable fromBi(tolf)3 and vehicle scores due to a change in lipids. The

24-htreatmentresultedinBi(tolf)3scoresthatwerewidelyscattered(compared

tothevehicleandtolfHscores)andpopulatedareasofvehicleandtolfHscores.

(iii) MefH and Bi(mef)3 treatment (4 h and 24 h) induced variation in the

spectrumofHCT-8 cells compared to vehicle-treated cells,whichwasprimarily

attributedtoaneffectonlipids.TheresponsebetweenmefHandBi(mef)3could

notbedistinguishedasthescoreswereoverlapped.


184

(iv) DiflH treatment (4 h) induced a difference in the spectrum of HCT-8 cells

comparedtovehicle,whichwasprimarilyattributedtoaneffectonamidesinthe

amide inclusive PCA. When the amide bands were omitted from the PCA, no

distinctionwasobservedbetweenthediflH-andvehicle-treatedcells. Following

24-h treatment, diflH induced a difference in the spectrum of HCT-8 cells

comparedtovehiclethatwasduetoanincreaseintheintensityofthelipidbands.

(v)PCAoftheamideinclusivespectrasuggestedtherewasnodifferencebetween

vehicleandBi(difl)3scores(4h).However,whentheamideswereomittedfrom

the PCA, 4-h Bi(difl)3 treatment induced a difference in the spectrum of HCT-8

cells compared to vehicle/diflH, whichwas primarily attributed to an effect on

lipids. PCA suggests there was no difference between the 24-h vehicle and

Bi(difl)3scores,andindicatedBi(difl)3hadadifferenteffecttodiflHatthechosen

treatmentconcentrations.


185

Table4.3:Sum

maryofaveragespectraltrendsandkeyPCAdifferencesbetweenvehicle-treatedcellsamplesversusNSAID/BiNSAID-treated

cellsamples.

Treatmenttime24h(Appendix4.7andFigure4.11)

Amideom

itted

• •TolfHscoresrelativelyclustered

+vePC-1,vehicledistributedacross

+veand−vePC-1

• •PC-1loadings:lipid,ν

as(CH2)(s),

ν s(CH2)(s),ν(C=O)

(m),andδ(CH

2)

(w);protein,νs(COO

− ) (w)

• •TolfHscoresrelativelyclustered

+vePC-2,vehiclescoresmainly−ve

PC-2;inter-sam

plevariationevident


as(CH2)(s),

ν s(CH2) (m

),ν(C=O)

(m),and

ν as(CO-O-C)(m

);protein,

ν s(COO

− ) (w);nucleicacids,ν as(PO 2

− )

(s) Resultsum

mary:TolfH

treatmentinducedaresponsethat

primarilyaffectedthelipids

• •VehicleandBi(tolf)3distributed

relativelyevenlyacross+veand−ve

PC-1

• •Greaternum

berofBi(tolf)3scores

+vePC-2,vehiclescoresmainly−ve

PC-2

• •PC-2loadings:AstolfH,seeabove

Resultsum

mary:Bi(tolf)3scores

werewidespread(com

paredto

thevehicleandtolfHscores)and

populatedareasofvehicleand

tolfHscores

IntensityofPCAloadingsbands:(s)=strong,(m)=moderate,(w

)=weak

Amideincluded

• •Avg.spectrum

tolfH:�intensityofν

as(CH2),

ν s(CH2),ν(C=O),andν as(PO 2

− )bands

• •TolfHscoresrelativelyclustered+vePC-1,vehicle

scoresdistributedacross+veand−vePC-1

• •PC-1loadings:lipidνas(CH2)(s),ν s(CH2)(s),and

ν(C=O)

(m);proteinam

ideI (m)andamideII(m)

• •TolfHscoresrelativelyclustered−vePC-2,vehicle

scoresmainly+vePC-2;

inter-samplevariationevident


as(CH2) (m

)andνs(CH2) (m

);protein,amideI(s)andamideII(m);nucleicacids,

ν(C=O)

(m)andνas(PO 2

− ) (w)

Resultsum

mary:TolfHtreatmentinduceda

responseinthespectrum

ofHCT-8cells

comparedtovehicle,whichwasattributedtoan

effectonlipidsandproteins

• •Littledifferencenotedintheavg.spectrum

ofvehicle

vs.Bi(tolf)3,exceptslight�intheintensityofν(C=O)

andν as(PO 2

− )bands

• •VehicleandBi(tolf)3scoresdistributedrelatively

evenlyacross+veand−vePC-1;nocleardelineation

orclustering

• •Greaternum

berofBi(tolf)3scores−vePC-2,vehicle

scoresmainly+vePC-2


Resultsum

mary:Bi(tolf)3scoresw

ere

widespread(com

paredtothevehicleandtolfH

scores)andpopulatedareasofvehicleandtolfH

scores


Amideom

itted

• •VehiclescoreslargelyPC-1andPC-2+ve,

tolfHscoreslargelyPC-1–ve;

inter-samplevariationevidentalongboth

PC-1andPC-2


as(CH2)(s),ν s(CH2)

(s),ν(C=O)

(m),δ(CH

2) (w);protein,

ν s(COO

− ) (w);nucleicacids,ν as(PO 2

− ) (w)

• •PC-2loadings:lipid,ν(C=O)(s),andδ(CH2)

(m);protein,νs(COO

− ) (m

);nucleicacids,

ν as(PO 2

− ) (m

)Resultsum

mary:TolfHtreatment

inducedaresponseinthespectrum

of

HCT-8cellscom

paredtovehicle,which

wasprimarilyattributedtoaneffecton

lipids

• •VehiclescoreslargelyPC-1+ve,Bi(tolf)3

scoreslargelyPC-1−ve(lessdelineationof

Bi(tolf)3scoresfromvehiclescores

comparedtotolfHvs.vehicle);som

einter-

samplevariationevidentonPC-1


• •Noseparationfrom

vehicleonPC-2

Resultsum

mary:Bi(tolf)3treatment

causedsomechangeinlipids;how

ever,

theeffectwaslesspronouncedcompared

totolfH.Additionally,thetolfHspectra

weredistinguishablefrom

Bi(tolf)3and

vehiclescoresduetoachangeinlipids

Amideincluded

• •Avg.spectrum

tolfH:�intensityof

ν as(CH2)andνs(CH2),shiftinνas(PO 2

− )

band(1240cm

−1�1232cm

−1)vs.vehicle

• •VehiclescoreslargelyPC-1–ve,tolfH

scoreslargelyPC-1+ve;




(s),ν(C=O)

(m),andδ(CH

2) (w);protein,

amideI (m),am

ideII(w),andν s(COO

− )

(w);nucleicacids,ν as(PO 2

− ) (w)

Resultsum

mary:TolfHtreatment


of

HCT-8cellscom



lipidsw

ithsomecontributionfrom

proteins

• •Avg.spectrum

Bi(tolf)3:�intensityof

ν as(CH2)andνs(CH2),shiftinνas(PO 2

− )

band(1240cm

−1�1238cm

−1)vs.vehicle

• •VehiclescoreslargelyPC-1−ve,Bi(tolf)3

scoreslargelyPC-1+ve(although,greater

overlapwithvehiclethantolfH);some



Resultsum

mary:Bi(tolf)3treatment

causedsomechangeinlipids/proteins;

however,theeffectwasless

pronouncedcom

paredtotolfH

vehicle

vs.tolfH

(45μM

)

vehicle

vs.

Bi(tolf)3

(15μM

)


186

Table4.3cont.:SummaryofaveragespectraltrendsandkeyPCAdifferencesbetweenvehicle-treatedcellsamplesversusNSAID/BiNSAID-

treatedcellsamples.


Amideom

itted

• •VehiclescoreslargelyPC-1–

ve,mefHscoreslargelyPC-1

+ve;someinter-sample

variationevident


as(CH2)

(s),ν s(CH2)(s),ν(C=O)

(m),and

δ(CH

2) (w);protein,νs(COO−)

(w)

Resultsum

mary:MefH

treatmentinduceda


of

HCT-8cellsthatdifferedto

vehicle;thiswasprimarily

attributedtoaneffecton

lipids

• •VehiclescoreslargelyPC-1–

ve,mefHscoreslargelyPC-1

+ve;inter-samplevariation

evident

• •PC-1loadings:Asm

efH,see

above

Resultsum

mary:Sam

eas

mefHresult(seeabove)


)=weak

Amideincluded

• •Avg.spectrum

:mefH�intensity

ν as(CH

2),ν

s(CH

2),andν(C=O)bands

vs.vehicle

• •VehiclescoreslargelyPC-1+ve,mefH

scoreslargelyPC-1–ve;

someinter-samplevariationevident


as(CH2)(s),


(m),andδ(CH

2)

(w);protein,amideI (m),

amideII(w),ν s(COO

− ) (w)

Resultsum

mary:MefHtreatment


ofHCT-8cellsthatdifferedto

vehicle;thiswasprimarilyattributed

toaneffectonlipids

• •Avg.spectrum

Bi(mef) 3�intensity

ν as(CH

2),ν

s(CH

2),andν(C=O)bandsvs.

vehicle

• •VehiclescoreslargelyPC-1+ve,

Bi(mef) 3scoreslargelyPC-1–ve;inter-

samplevariationevident

• •PC-1loadings:Asm

efH,seeabove

Resultsum

mary:Sam

easmefH

result(seeabove)


Amideom

itted

• •VehiclescoreslargelyPC-1–ve,mefHscores

largelyPC-1+ve


as(CH2)(s),ν

s(CH

2)(s),

ν(C=O)

(m),andδ(CH

2) (w);protein,νs(COO−)(w)

• VehiclescoreslargelyPC-2–ve,mefHscores

largelyPC-2+ve;

inter-samplevariationevidentalongbothPC-1and

PC-2


as(CH2)(s),ν s(CH2) (m

),ν(C=O)

(m),andν as(CO-O-C) (m);protein,νs(COO−)

(w);nucleicacidsν(C=O) (m),andν as(PO

2−) (m)

Resultsum

mary:MefHtreatmentinduceda


ofHCT-8cellsthat

differedtovehicle;thisw

asprimarily

attributedtoaneffectonlipids

• •PC-1scoresandloadingsobservations:As

mefH,seeabove

Resultsum

mary:Sam

easmefHresult(see

above)

Amideincluded

• •Avg.spectrum

:mefH�intensity

ν as(CH

2),ν

s(CH

2),andνas(PO 2

− )bands

vs.vehicle

• •VehiclescoreslargelyPC-1–ve,

mefHscoreslargelyPC-1+ve;inter-

samplevariationevident


as(CH2)(s),


(w),and

δ(CH

2) (w);protein,amideI (s),

amideII(w),andν s(COO

− ) (w)

Resultsum

mary:MefHtreatment


ofHCT-8cellsthatdifferedto

vehicle;thiswasprimarily


• •Avg.spectrum

:Bi(mef) 3�intensity

ν as(CH

2)νs(CH

2),andνas(PO 2

− )bands

vs.vehicle

• •PC-1scoresandloadings

observations:Asm

efH,seeabove

Resultsum

mary:Sam

easmefH

result(seeabove)

vehiclevs.

mefH

(120μM)

vehiclevs.

Bi(mef) 3

(40μM

)


187

Table4.3cont.:SummaryofaveragespectraltrendsandkeyPCAdifferencesbetweenvehicle-treatedcellsamplesversusNSAID/BiNSAID-

treatedcellsamples.


Amideom

itted

• •VehiclescoreslargelyPC-1+ve,

diflHscoresPC-1–ve;inter-sample

variationevident


as(CH2)(s),


(m),and

δ(CH

2) (w);protein,νs(COO−) (w)

Resultsum

mary:DiflHtreatment

inducedadifferenceinthe

spectrum

ofHCT-8cellscom

pared

tovehicle,whichwasprimarily


• •Bi(difl) 3scoreswereoverlapped

withthevehiclescoresalongPC-1

• •Vehiclescorespreferentially

appearedon+vePC-2;thissuggests

thatasubpopulationofBi(difl)

3

scores(–vePC-2)differsfromthe

vehiclescores


as(CH2)(s),

ν s(CH2) (m

),andν(C=O)

(m),and

δ(CH

2) (w);protein,νs(COO−) (w);

nucleicacidsν

as(PO 2

− ) (w)

Resultsum

mary:PCAsuggests

therewasnodifferencebetween

vehicleandBi(difl) 3scores,and

indicatedBi(difl) 3hadadifferent

effecttodiflHatthechosen

treatmentconcentrations


)=weak

Amideincluded

• •Avg.spectrum

diflH�intensityofthelipid

bands,ν as(CH

2),ν

s(CH

2),andν(C=O)vs.

vehicle

• •VehiclescoreslargelyPC-1+ve,diflH

scoresPC-1–ve;inter-samplevariation

evident



(s),andν(C=O)

(m);protein,amideI (m),

amideII(m),andν s(COO

− ) (w)

Resultsum

mary:DiflHtreatment

inducedadifferenceinthespectrum

of

HCT-8cellscom



lipids

• •Littlevariationintheavg.spectrum

of

Bi(difl) 3vs.vehicle

• •Bi(difl) 3scoreswereoverlappedwiththe

vehiclescoresalongPC-1

• •Vehiclescorespreferentiallyappearedon

–vePC-2;thissuggeststhata

subpopulationofBi(difl)

3scores(+vePC-2)

differsfrom

thevehiclescores



(m),andν(C=O)

(m);protein,amideI(s),

andν s(COO

− ) (w);nucleicacids

ν as(PO

2−)(w)

Resultsum

mary:PCAsuggeststhere

wasnodifferencebetweenvehicleand

Bi(difl) 3scores,andindicatesBi(difl)

3hadadifferenteffecttodiflHatthe

chosentreatmentconcentrations


Amideom

itted

• •DiflHscoreswereoverlappedwiththe

vehiclescores;bothweredistributed

equallyalongPC-1.Intra-sam

ple

variationwasevidentalongPC-1.No

separationofdiflHandvehiclescores

alongPC-2


as(CH2)(s),


(m),andδ(CH

2) (w);

protein,νs(COO−) (w);nucleicacids,

ν as(PO

2−)(w). Lipidsw

eretheprimary

sourceofintra-sam

plevariationwhen

theam

idebandsw

ereexcluded

Resultsum

mary:PCAsuggests

therewasnodifferencebetween

vehicleanddiflHscoreswhenthe

amideswereom

itted

• •PC-1scoresandloadings

observationsasdiflH(seeabove)

• •SomedelineationofBi(difl)

3scores

(–ve)fromthevehicle(anddiflH)

scores(+ve)w

asevidentalongPC-2;

inter-samplevariationwasevident


as(CH2) (m

),ν s(CH2) (m

),δ(CH

2) (w),and

ν as(CO-O-C) (m);protein,νs(COO−) (m);

nucleicacidsν(C=O) (m),and

ν as(PO

2−) (m)

Resultsum

mary:Bi(difl)

3treatment

inducedadifferenceinthespectrum

ofHCT-8cellscom

paredto

vehicle/diflH;whichwasprimarily


Amideincluded

• •Avg.spectrum

:diflH:�intensityofν

as(CH2),

ν s(CH2)andν(C=O)bandsvs.vehicle

• •VehiclescoresclearlyclusteredanddiflH

scoresdistributedacross–veand+vePC-1;

intra-samplevariationevident.Variationinthe

amideIband(s)appearedtobethecauseofthe

variation(diflHscores)

• •VehiclescoreswerelargelyPC-2–ve,anddiflH

scoreswerePC-2+ve;

inter-samplevariationevidentinPC-2


as(CH2)(m

),andν s(CH2)

(m);protein,amideI (s)andamideII(m)

Resultsum

mary:DiflHtreatmentinduceda

differenceinthespectrum

ofHCT-8cells

comparedtovehicle,whichwasprimarily

attributedtoaneffectonam

ides

• •Avg.spectrum

Bi(difl)

3:�intensityofν

as(CH2)

ν s(CH2),andν(C=O)bandsvs.vehicle

• •Bi(difl) 3scoreswereoverlappedwiththe

vehiclescoresalongPC-1;well-clusteredinto

comparisontothediflHscores

• •VehiclescoreswerelargelyPC-2–ve,and

Bi(difl) 3scoreswerelargelyPC-2+ve;som

einter-samplevariationevidentinPC-2

• •PC-2loadings:AsdiflH,seeabove

Resultsum

mary:PCAsuggeststherewas

nodifferencebetweenvehicleandBi(difl) 3

scores

Vehicle

vs.diflH

(225μM)

Vehicle

vs.

Bi(difl) 3

(75μM

)


188

4.4 DiscussionPreviousinvitrostudieshaveshownthatNSAIDscanexertanti-neoplasticeffects

throughanumberofbiochemicalpathways(discussedinSection1.2.4).Inthiswork

SR-IRMS has been employed to determine whether gross biomolecular changes

inducedbyNSAIDs/BiNSAIDscanbedistinguished.Thisapproachhasbeenexplored

based on previous studies by other researchers who have correlated biomolecular

effectsobservedbySR-IRMSwithspecificdrugtreatments(Table4.2). SR-IRMSwas

specificallyemployedinthisstudytodeterminewhetherthetechniquecouldbeused

to:(i)observeglobalmolecularchangesintheliveHCT-8cellsthathavebeenexposed

to theNSAIDsor the correspondingBiNSAIDs, (ii) determinewhether anyobserved

changes induced by the BiNSAID/NSAID are consistent with knownmechanisms of

celldeath, and (iii)determinewhetherBiNSAIDs/NSAIDs inducesimilarordifferent

biomolecularresponses.

Livecellstudieswereperformedinordertoexaminethecellsinastatethatmost

closelymimickedthatusedfortheotherbiochemicalassaysperformedaspartofthis

Thesis,without any potential confounding IR effects from a fixation/drying process.

ItwasclearfromtheSR-IRMSstudyoflivecellsthatanumberoffactorsneededtobe

considered when determining the usefulness of SR-IRMS for the abovementioned

purposes. These include:(i) thedataanalysis,specificallythewaterbandcorrection

strategies (discussed in Section 4.4.1) and (ii) the significance of the biomolecular

changes induced by vehicle, and BiNSAID/NSAID treatments, with respect to

intra-sample variation (discussed in Section 4.4.2). Finally, the experimental

approach/designandimprovementsforfutureexperimentswithrespecttotheresults

arediscussedinSection4.4.3.

4.4.1 Dataanalysis:Waterabsorptioncorrectionstrategy

AspreviouslyhighlightedinSection4.1.4,theeffectsoftheaqueousbackground

cancreateproblemswhenanalysingtheSR-IRMSspectraoflivecellsincellmedium.

Inparticular,thewaterdeformationbandcentredat1620cm−1,whichoverlapswith

the amide I absorbancebandneeds tobe addressed. Many researchers [33, 43, 57]

have highlighted the issue of water absorption in the collection and processing of

SR-IRMS spectra obtained from living single cells. Conflicting approaches in the

literatureledtotheuseofthetwoanalyticalstrategiespresentedinthisThesis.


189

In the present study it was observed that when comparing the average cell

spectra of a sample corrected for water band absorption to the same sample set

withoutwaterband correction, little tonodifferences in thepositionof the amide I

and amide II bands were observed in the second derivative spectra. This result is

similar tothatreportedbyMarcsisinetal.[33],wherebytheauthorsconcludedthat

the spectral integrity of the data was maintained; consequently, they employed

spectra without any water band correction in PCA. In the present study (Section

4.3.1), PCA was performed on spectra uncorrected, and corrected for the water

absorption, inorder to furtherdeterminewhether therewereanydifferences in the

PCA scores and loadings plots following correction for thewater absorption. Some

subtledifferenceswereobservedinthecoordinatesofsomeoftheindividualscoresin

the PCA scores plot following correction forwater absorption; however, the overall

trendsremainedthesameregardlessofwhetherwaterbandcorrectionwasappliedin

all of the sample sets tested. The results from the present study suggest that the

biomolecularinformationintheamideregioncontributedbycellSR-IRMSmayremain

extractable using PCA, despite the contributions from the water band centred at

1620cm−1.

Water correction strategies have been debated in a number of studies. For

instance,Whelanet al. [43] argue that anywater band ratioing is unreliable as it is

unconventional. Given that the water band influences the amide region, the

researchers excluded the amide region from PCA obtained from their live cell cycle

study[43]. Consequently, in thepresentstudyPCAwasalsoperformedonSR-IRMS

thatexcludedtheamideregion,andcomparedtothePCAperformedonspectrathat

includedtheamideregion.Inmostcases,thescoresplotwasalteredmarkedlywhen

the amide region was excluded, which can be expected as the amide I band is the

strongestbandinthespectrumandthestrongestmeasureofproteinsinthespectrum.

However, in all cases the explained variance of PC-1 (observed on the x-axis of the

scoresplotsinFig.4.7−4.15)wassubstantiallyincreasedfollowingtheomissionofthe

amideregion,whichmayindicatethattheamideIregionwascontributingmorenoise

than information. Additionally,with the omission of the amide region, the loadings

plotsshowedpredominantcontributions fromlipid-relatedbands,and inmostcases

increasedthestrengthofthecontributionsfrombandsbetween1485–1150cm−1.

Therearetwooutcomesthatrequirefurtherexamination.Firstly,theomissionof

theamidebandsappearstohaveasubstantialeffectontheoutcomeofPCA;thusthe


190

omissionoftheamidebandsmayinfactresultintheoversightofusefuldatafromPCA

producing results less representative of the true biomolecular differences between

samples. Secondly, the effectiveness of the water band correction is difficult to

evaluateandmaystillnotprovideatruebiochemicalrepresentation.

In conclusion, the water band correction using spectral subtraction strategy

employed in thepresent studyappears tohave subtle effectson theoutcomeof the

PCA results when compared to the respective non-corrected spectra. If a suitable

water compensation strategy can be determined in the future, and employed

universally by researchers who perform SR-IRMS studies of live cells, the results

obtainedinthepresentstudyshouldbereanalysedwiththisstrategy.Forthepresent

study, however, it was opted to perform PCA on SR-IRMS spectra that were not

correctedforthewaterbandastheresultsinSection4.3.1revealedlittledifferencein

theresultantPCA.

4.4.2 ThebiomoleculareffectsoftreatmentonliveHCT-8cells

4.4.2.1 Vehicle-treatedcells

One of the notable results of the PCA of the SR-IRMS cell spectra, evident even

within control cell populations was intra-sample variation. There are a number of

reasonsthatcancontributetointrinsicvariationswithinacellpopulationsubjectedto

thesametreatment.Twowell-studiedcausesare:(i)cellcycle,and(ii)biomolecular

changesduetocelldeath.

At any given time in an unsynchronised population of mammalian cells,

subpopulationsof cellswill be found inoneof thedifferentphasesof the cell cycle:

gap1 (G1), synthesis (S), gap 2 (G2), mitosis (M) or quiescence (G0) [43].

FTIRdemonstratedthatmammaliancellsundergoingcelldivisionshowdifferencesin

eachdifferentphaseof thecell cycle [16,22]. Furthermore,Whelanetal. [43]have

shown through live cell studies that biomolecular changes could be distinguished

within each specific phase by examining the SR-IRMS spectra of the cells every two

hourspost-mitosis.Inthepresentstudynoattemptwasmadetosynchronisethecells

inanyparticularphaseof the cell cycle,but instead the cellswerepreparedusinga

similarproceduretothatemployedforallotherassaysperformedinthisThesis.Itis

reasonable to suggest that lipid, protein and nucleic acid content of theHCT-8 cells

willbeinfluencedbythecellcyclephaseandwillthereforeresultinvariationsinthe

SR-IRMSspectraofeachindividualcellwithinthepopulation.


191

CelldeathisanotherfactorthatmayinfluencetheSR-IRMSspectraofindividual

cells. For instance, studies have shown that apoptosis and necrosis can be

distinguished in the IRMS spectra of single cells [18, 20, 21]. Importantly, in the

presentstudy,itwasestablishedfromflowcytometryassays(presentedinChapter2)

thatlateapoptotic/deadHCT-8cellsdetachedfromthecellsubstrate.IntheSR-IRMS

studyreported in thisChapteranydeadcellswouldhavebeenremovedbywashing

stepsperformedprior to insertion into the cell chamber for analysis. Thus, the cell

populationthatremainsadheredtothesubstrateprimarilycontainscellsofrelatively

good health or in the early stages of apoptotic cell death as confirmed by flow

cytometryanalysisoftheadherentcellpopulation(Appendix4.12).

Whiletherearevariationsinthebiochemistryofthecellsduetocellcycleandcell

death(whichcorrelatesinvariationsintheSRIRMSspectra),thecontrolcellsamples

thatwere incubatedwith cellmedium, in the absence of vehicle or drugs,would be

consideredtobethe‘normal’distributionofacultureofHCT-8cells.Floweretal.[12]

alsoreportedasimilarheterogeneityinthescoresplotsofspectraassociatedwiththe

controlcells.Encouragingly,theresultsfromthePCAloadingsplotsgeneratedinthis

study (Fig.4.9) indicate that the main source of variation of the cell medium and

vehicle cells is the lipid bands which causes the dominant variance across the

PC-1axis. This isconsistentwiththePCAresultsofWhelanetal. [43]whoreported

thesameobservationwhentheyperformedaPCAofG1,S,andG2cells.Itis,therefore,

likely that the distribution of the vehicle scores is indicative of cell cycle variation

withinthepopulation.

It was essential to prepare non-treated (cell medium only) cells and vehicle-

treated(DMSO2%v/vinRPMI-1640)cellsinordertodetermineifanybiomolecular

changes occurred in response to DMSO exposure. It should be noted that all cell

experimentsperformed in thisThesis includedavehicle-treated ‘control’ inorder to

directlycomparewiththeBiNSAID/NSAIDtreatments(preparedasDMSOsolutionsin

cellmedium toachievedissolutionof thedrug). A shorter4-h treatment timeframe

(aswellasthetypical24-htreatment)wasevaluatedinordertodeterminewhether

any spectral changes indicative of biomolecular changes could be detected in the

absence of cytotoxic effects (as indicated by MTT assays, Appendix 1.2). Vehicle

treatment appeared to effect the SR-IRMS spectra of the HCT-8 cells after 4-h as

evidenced by the PCA scores plots (Fig. 4.8); however, the delineation identified

between cellmedium- and vehicle-treated cellswas only observedwhen the strong


192

loadings from the amide region (Fig. 4.7) were omitted from the PCA (Fig.4.8).

Followingomissionoftheamideregion,thePC-2loadingsplots(Fig.4.8)showedthat

the νas(CH2), νs(CH2), νas(C=O), andνas(CO−O−C)bands, and the νas(PO2−)band,were

theprimarycontributorstothevariancebetweenthecellmediumandvehiclescores.

DMSOisawater-miscible,polarsolvent,whichhasoftenbeenusedincellbiology

for the dissolution of hydrophobic compounds, cryoprotection, inducing cell

differentiation, free radical scavenging, membrane penetration, and as a cell

fusogen[63]. The effect of DMSO on lipids, and in particular, membrane

phospholipids, has been investigated in a number of studies [64-68]. Molecular

dynamicsimulationsperformedusingdioleoylphosphatidylcholinemembranesmixed

with20mol%ofcholesterolandlowDMSOfractions(≤10mol%)resultedinsporadic

and very transient hydrophobic pores which were arranged in the form of single-

molecule water columns that cross the membrane and disappear [67].

Incomplimentarymicroscopyexperiments,smallundulationsofthecellmembraneof

DC-3F cells treated with 10 vol% (2.74 mol%) DMSO were identified following

1-h[67]. Gurtovenko et al. [69] report that atomic-scale molecular dynamic

simulationsperformedonlipidbilayerscomprisingdipalmitoylphosphatidylcholinein

aqueous solution with DMSO at low concentrations showed decreased membrane

thickness and increased membrane fluidity (2.5−7.5 mol%), while moderate

concentrations (10−20mol%) induced formation of transientwater pores, and high

concentrations (25−100 mol%) destroyed the bilayer structure of membranes.

Morerecently,neutronreflectometrystudiesassessingtheeffectofDMSO(2%v/v)on

lipid membrane mimetic bilayers comprised of POPC have been performed [68].

Interestingly, scattering length density profiles indicated that DMSO (2%v/v) in

D2O/PBScausednostructuralchangestothebilayer.However,asubstantialdecrease

in the scattering length density profile of the hydrophobic tail region indicated that

DMSOenters the tail regionof thePOPCbilayerandsubsequentlydisplaces theD2O

molecules [68]. Given that the formationof transienthydrophobicpores across cell

membranesreportedlyoccursathigherconcentrationsofDMSO(>10vol%)[69]than

that utilised in the present study, it is conceivable that variation in the SR-IRMS

spectra was potentially due to the occupation of DMSO below the head groups of

membrane phospholipids and the displacement of water molecules resulting in a

difference in theSR-IRMSsignatureobserved in the lipid-relatedbandsbetween the

4-hcellmedium-andvehicle-treatedPCAscores.


193

Interestingly,nodelineationbetweenthecellmedium-treatedorvehicle-treated

scores was observed following 24-h treatment with or without the amide bands

includedinthePCA(Fig.4.9).Plasmamembranedisruptionisoftenaccompaniedby

rapid resealing of the cell membrane in order to ensure cell survival [70]. Work

performed by Togo et al. [71, 72] suggests that cells are able to reseal wounded

membranesmore rapidly following repeated disruption to the cellmembranewhen

compared to the rate of repair following the initial wound due to an increase in

exocytosis. It isplausiblethat followinganextendedperiodoftreatment(>4h)the

HCT-8 cells could become tolerant to DMSO exposure and the cell membrane may

‘recover’toanextent. Perhapsthelackofbiomoleculardistinctionasdeterminedby

PCAbetween thespectraof cells treatedwithcellmediumorvehicle following24-h

may be due to gradual displacement of DMSO from the cell membrane. A recent

report[73]suggeststhatexperimentalquantificationofthepropertiesofDMSO–water

bondsatthelipidmembranesurfaceremainspoorlyunderstood. Thusfurtherwork

isneededtoexplicitlyconfirmthepresentfindings.

Thereareanumberofreports[65,74,75]thatdiscussthevaryingtoleranceand

biochemical effects exhibited by different cell lines when exposed to DMSO. For

instance, the effect of DMSO (2%) on the membrane dynamics and phospholipid

composition of the two different cell lines, FLC (Friend leukemia cells) and Raji

(lymphoblastoid) cells has been investigated [65]. Following 5-day incubationwith

DMSO only the phospholipid composition of FLC cells, but not the Raji cells was

affected.Thecoloncancercells,Caco-2/TC7,havebeenshowntobetoleranttoDMSO

for concentrations up to 10%v/vwithout the induction ofmembrane leakage [74].

Inanotherstudy,prolongedDMSO(2%v/v)treatmenthasbeenreportedtoinducea

number of biochemical changes in SW480 and SW620 cell lines (primary colonic

tumorandlymphnodemetastasis,respectively,isolatedfromthesamepatient)[75];

however, cell viability was not affected for up to four days following incubation in

DMSO [75]. Clearlyprevious research shows that the cell response toDMSOcanbe

variable, notonlydependentonDMSOconcentrationand timeof exposure, but also

dependent on the specific cell line. The results of the present study show that the

viabilityofDMSO-treatedHCT-8cellsisslightlyaffectedfollowing24-htreatment(as

shown by MTT assay, Chapter 2); however, no significant biomolecular differences

could be detected by PCA of the SR-IRMS spectra. For the purpose of the current

SR-IRMS study, the lack of distinction between cellmedium andDMSO scores is an


194

importantresultasitsuggeststhatDMSOexposureof24-hdoesnotsignificantlyalter

theSR-IRMSspectraandthebiomolecularintegrityoftheHCT-8cellsusingthistool.

This indicates that any biomolecular changes observed when comparing the

DMSO-treated cells to the BiNSAID or NSAID-treated cells are not dominated by

presenceofDMSO.

4.4.2.2 BiNSAID/NSAID-treatedcells

A few key observations resulted from the PCA analysis of vehicle-treated cell

spectra compared to BiNSAID/NSAID-treated spectra. These included the findings

that: (i) inmostcases theremovalof theamideregionresulted insubtlechanges to

thedistributionofthescoresplots,(ii)inmostcases4-hdrugtreatmentcausedsubtle

biomolecularchangesdespitenoobviousphysicalchangestothecellsorviabilityloss

(described in Chapter 2), (iii) the changes observed following 24-h drug treatment

differed from the changes observed at 4-h, and (iv) differences could be observed

whencomparingtheBiNSAIDscorestotherespectiveNSAIDscores.

It should be noted that the cell medium containing the drugs was removed

followingthedefinedincubationperiods(4hor24h)andthecellswerewashedprior

toplacementinthecellholderforSR-IRMSanalysis. Thiswasperformedinorderto

prevent any SR-IRMS detection of drug in the cellmedium. Other researchers [35]

havereportedthatthedetectionlimitofFTIRinstrumentationisnotsensitiveenough

todetecttheextremelylowconcentrationsofdrugthatwouldresideintheindividual

cell.Thisisanimportantpointasitindicatesthatanysignificantchangesdetectedin

thecellspectraweretheresultoftheactionsofthedrugswithinthecellsandnotdue

tothedetectionofthedrugitself.

When the spectral analyses were performed for the full spectral range

(3000−2800 cm−1, 1790−1150 cm−1) differences were commonly observed in the

amide bands following treatment with BiNSAID/NSAID versus vehicle. These

differencesweremost obvious for cells treatedwith Bi(tolf)3 (subtle for 4 h), tolfH

(subtlefor4handmoderatefor24h),Bi(mef)3(moderatefor4h),mefH(moderate

for 4 h), and diflH (strong for 4 h), and less so for Bi(difl)3,while no changeswere

observed in Bi(tolf)3 (24h). The most significant changes were observed as band

shiftsintheamideIbandfrom1655to1653cm−1(tolfHandmefH(4h),andBi(mef)3

anddiflH(24h))orwereassociatedwiththedevelopmentofshouldersat~1635cm−1

insamplestreatedfor24-hwithtolfH,Bi(mef)3andmefH(24h).Thefrequencyofthe


195

amideIbandhasbeenreportedtorepresentthedominantproteinstructureswithin

inacell.Forinstance,commonproteinsecondarystructuresandtheircorresponding

vibrationalfrequencyrangesinclude:α-helix(1660−1649cm−1),β-sheet(1637−1615

cm−1), random coil (1648−1638 cm−1), turn (1680−1660 cm−1), and β-antiparallel

(1692−1680 cm−1) [76]. Thus the development of a shoulder at~1635 cm−1 in the

24-h tolfH-, Bi(mef)3-, and mefH-treated cell spectra is potentially indicative of an

increase in β-secondary structures. An increase in the proportion of β-sheet

structureshasbeenidentifiedintheearlystagesofapoptosisinA549[77],HeLa[17]

and U-87 MG cells [78]. This is in agreement with the present study where a

substantialproportionofBiNSAID/NSAID-treated(atIC50dosesfor24h)HCT-8cells

were identified as early apoptotic in flow cytometry experiments (discussed in

Chapter2).Unfortunately,however,ashasbeenestablishedbyVaccarietal.[34]and

Whelanetal.[43]inSection4.1.4thewaterbandalsoeffectstheamideIandamideII

bands;assuch,whilstitappearedthatcorrectionforthewaterbandshowednoeffect

on these bands, it cannot be definitively proven that these changes are solely

attributedtomolecularchangesassociatedwithamidebands.

In most cases the other biomolecular changes associated with BiNSAID/NSAID

treatmentofcellswereconsistentregardlessofwhethertheanalysiswasperformed

forspectrawhichincludedoromittedtheamideregions.Themostsignificantchanges

observed in both of these analyses were associated with CH2/CH3 region (3000−

2800cm−1), ν(C=O) (1745 cm−1) and νas(PO2−) (1245 cm−1) that canbe attributed to

lipids.Therewasonlyoneinstancewherethelipidchangeswerenotobservedinboth

datasetanalysesandthiswastheamideinclusiveanalysesofthe4-hBi(difl)3/diflH-

treatedcellspectra(Fig.4.14a).Instanceswherelipidchangeswereobservedinclude:

tolfH(4hand24h),Bi(tolf)3(4h),mefH(4hand24h),Bi(mef)3(4hand24h),and

diflH(24h).

Interestingly,the4-hBi(tolf)3/tolfH(Fig.4.10)and4-hBi(mef)3/mefH(Fig.4.12)

PCA results showed that the loadings bands associated with the lipids were also

stronglyassociatedwiththeBiNSAID/NSAID-treatedscores(asopposedtothevehicle

scores). Furthermore, it was observed in the average Bi(tolf)3-, tolfH and Bi(mef)3-

treated (Appendix 3.6 and Appendix 3.7) spectra that there was an increase in the

intensity of the νas(CH2) and νs(CH2) bands, a possible indication that lipid

concentration is increased in the BiNSAID/NSAID-treated cell populations. The 4-h

Bi(difl)3 (Fig. 4.14), showed the opposite trend whereby the lipid bands were


196

associated with the vehicle spectra. This is not surprising given that the diflH

structure differs significantly from the structures of mefH/tolfH (Fig. 1.2).

Importantly, asno changes to the viability of theBiNSAID/NSAID-treated cellswere

observed after 4-h treatment (as determined byMTT assays, Chapter 2), it is likely

that the biomolecular changes observed at this time-point are associated with

moleculareventsthatprecedeapoptosis.Factorsthateffectlipidsinclude:(i)changes

to the lipid membrane due to interaction/uptake of the drug molecules, or (ii) a

downstreammetaboliceffectoncethedrugisinsidethecell(e.g.lipidsynthesis).

TheSR-IRMSspectralchangesobservedinthelipidbandsdetectedinresponseto

NSAIDtreatmentareconsistentwithreports[79-81]thatNSAIDsareabletointeract

withlipids,andinparticular,phospholipids,whichconstitutealargeproportionofthe

lipids comprising the cellmembrane. Experimental evidence [79] suggests that the

deleteriouseffectsofNSAIDsontheGItissuecouldbeduetomechanismsthatinvolve

direct disruption to the surface-active properties of phospholipids of the gastric

mucosa, in addition to COX-1 inhibition. Studies by Lichtenberger et al. [80] have

shown that 30-min exposure of AGS human gastric carcinoma cells to aspirin or

ibuprofen (both 0.5 mM), or naproxen (1.5mM) altered the fluidity and packing

densityofthecellmembraneinaheterogeneousmannerresultinginportionsofthe

membrane becoming hydrophilic.While the exposure time and treatment

concentrationsvariedfromthepresentstudythereisclearevidencethatNSAIDscan

directly disrupt the mammalian cell membrane. Suwalsky et al. [81] used X-ray

diffraction to show that mefH perturbed lipid membrane mimetics consisting of

dimyristoylphosphatidylcholine or dimyristoylphosphatidylethanolamine

bilayers. Additionally, the researchers used fluorescence resonance energy transfer

measurementstoshowthatmefHwasabletoincreasefluidityofthehydrophobiccore

ofdimyristoylphosphatidylcholineliposomes.Itwasdeterminedthatthisoccurredas

a result of intercalation of mefH into the hydrophobic acyl chains of the

dimyristoylphosphatidylcholineliposome[81].

Interestingly,theoppositetrend(fromthatofthe4-htreatments)wasobserved

when examining the 24-h Bi(tolf)3/tolfH (Fig. 4.11) and 24-h Bi(mef)3/mefH

(Fig.4.13) PCA results whereby the loadings bands associated with the lipids were

also strongly associated with the vehicle-treated scores (as opposed to the

BiNSAID/NSAID-treatedscores).Assuch,thetolfH-,Bi(mef)3-andmefH-treatedcells

appeartobeassociatedwithalowerlipidconcentrationthanthevehicle-treatedcells.


197

Thiswas supported by the reduction in the absorbance of the νas(CH2), νs(CH2) and

ν(C=O) bands in the SR-IRMS spectra (Appendix 3.7 and Appendix 3.9). Flow

cytometry assays (Appendix 3.12) showed that following 24 h treatment with the

same BiNSAID/NSAID concentrations, there was an increase in the proportion of

apoptotic cells (~16−40% within the adherent cell population). Phase contrast

microscopy also revealed morphological changes such as cell shrinkage and the

formation of membrane blebs (Fig. 2.7). Thus, in addition to changes to the lipid

membrane due to the interaction/uptake of the drug molecules, or effects on lipid

metabolismcausedbydruguptake,itisfeasiblethatthelipidbandswilladditionally

be altered as a result of apoptosis. The classic example of an alteration in lipid

distribution associated with cell apoptosis is the ‘flipping’ of phosphatidylserine

species from the interior cell membrane to the exterior leaflet of the cell

membrane[82]. Additionally,changesinmembranefluiditycommonlyoccurincells

undergoingapoptosis[83].Assuch,theobserveddecreaseinlipidabsorbancemaybe

aresultofapoptosis;however,thisobservationisincontrasttoanumberoffixedcell

studies that have demonstrated that the intensity of the lipid band increases in

apoptoticcells[17,20,84,85].LivecellstudiesperformedbyBirardaetal.[39]have

shownthatserumstarvationorcarbonyl cyanidem-chlorophenylhydrazoneexposure

induced apoptosis in U937 monocytes and caused an increase in cellular lipid

concentration.Theauthorsspeculatedthattheincreaseinthelipidcontentobserved

by spectroscopicmeasurementswas due to the accumulation of lipid droplets. The

decrease(ratherthanincrease)inlipidabsorbanceinthepresentstudymaysuggest

thatlipiddropletaccumulationisgenerallynotaneventassociatedwithcelldeathin

theHCT-8cellline.

Given the similarity in the structures of tolfH and mefH (Fig. 1.2), it is

unsurprising that tolfH and mefH would result in a similar biomolecular effect on

HCT-8 cells. Additionally, the overlap of the scores clusters in the Bi(mef)3/mefH

series is perhaps not surprising given the instability of the Bi(mef)3 complex

(concluded by the NMR data presented in Chapter 2) resulting in free mef in the

treatmentmedium. However, oneof themost surprising resultswere those for the

24-h diflH treatmentwhere itwas observed that therewas an increase in the lipid

intensity associated with diflH treatment, but no change associated with Bi(difl)3

treatment compared to the vehicle (Fig. 4.15 and Appendix 3.11). It was apparent

fromtheanalysesthatthediflH-treatedHCT-8cellsshoweddifferencesfromvehicle-


198

treated cells but this was not the case for Bi(difl)3-treated cells. The SR-IRMS

differencesobservedfordiflH-versustolfH/mefH-treatedcellscannotbeexplainedby

theproportionofdyingcellsbasedondecreasedtoxicitysinceinallcasestheIC50was

applied. However, these differences might be explained by variations in the

mechanismofactionofdiflHcomparedtomefHandtolfH.Thisisnotsurprisinggiven

thediflHstructuredifferssignificantlyfromthestructuresofmefHandtolfH(Fig.1.2).

It isnotclearwhy the24-hdiflH-treatedcellsshowedsuchastrong increase in

theintensityoftheCH2/CH3bandregion.AlivecellcyclestudyperformedbyWhelan

et al. showed that an increase in lipid band intensity in the SR-IRMS spectra of

synchronised cells appeared 17-h post-mitosis, i.e. in the late-S phase of the cell

cycle[43]. Assuch, it isplausible that the increase in the lipidband intensityof the

24-hdiflH-treatedcellspectramightbeduetotheaccumulationofHCT-8cellsinthe

S-phaseofthecellcycle.However,NSAIDswithasimilarchemicalstructure,including

aspirin and salicylate, have been reported to arrest the cell cycle in the G0/G1

phase[86-88]. The increased intensityof theCH2/CH3bandregion, indicativeof the

lipidband,wasconsistentwithBirardaetal.[39]wherebytheauthorsspeculatethat

theincreaseinthelipidcontentobservedbyspectroscopicmeasurementswasdueto

theaccumulationoflipiddroplets. Furtherexperimentsemployingfluorescentprobes

couldbeperformedinthefuturetostudycellcyclearrestandthepossibleformation

of lipid droplets in diflH-treated HCT-8 cells, in order to better understand this

SR-IRMSobservation.

4.4.3 Limitationsoftheexperimentalapproachandfutureimprovements

The experimental design is an important factor that must be considered when

undertaking any scientific investigation. Mignolet et al. have recently commented

upon the variation in experimental design (and analysis procedures), cell handling,

and spectroscopy, in relation to the evaluation of data published in the field of

biospectroscopy[15].ThepreparationoflivecellsforSR-IRMSanalysisinthepresent

experimentwasparticularlychallengingforanumberofreasons.

Eachtreatmentsamplewaspreparedwithacontrolthatwaspreparedfromthe

same subculture, thus the passage number was kept consistent. However, the

applicationoftreatmentshadtobestaggeredby2−3hoursgiventhetimerequiredto

assemble the cell holder (typically 15−20min), set up the holder in themicroscope

andcollect theSR-IRMSspectraof thecells (typically60−90min). Thecell samples


199

prepared following 2−3hours may, therefore, experience some changes in the cell

cycleprogressionandtheconcentrationofmetabolitesinthecellmedium.Whilethe

experiments in Chapter 4 were designed to mimic those of Chapters 2 and 3 in

hindsight, one important consideration for future experiments might also be the

preparationofsynchronisedcellsamplestotrytoreduceintra-samplevariation.This

wasnotpossibleforthecurrentstudy,however,duetobeamtimelimitations.

Vaccari and co-workers [89] have commented on the issues with demountable

liquid holders in recent publications. Essentially, the perfect closure of the

demountabledeviceisnotguaranteed,whichcanpotentiallyresultinvariationinthe

optical path length throughout the device. Importantly,water subtraction accuracy

canbe affectedby changes in path lengthbetween samples andbuffer spectra [34].

Additionally,withthedemountabledesign(suchastheoneusedinthepresentstudy)

it is difficult to standardise ‘tightening’ when the device was assembled with a cell

sample contained inside [89]. Whelan et al. [43] performed live cell studies at the

AustralianSynchrotronthatemployedthesamedeviceastheoneinthepresentstudy

attheandhavecommentedthat itsassemblycouldbeacauseforvariationbetween

replicate samples. Birarda et al. [89] have previously demonstrated that the

compressionofU937monocytecellsinadevicethatcontainedachamberheightless

thanhalfthatofthediameterofthecell(causingstrongdeformationofthecells)ledto

changesinAmI/AmII,AmII/lipidandνas(PO2−)/νs(PO2−)ratios;thusmechanicalstress

can induce cellular deformation and consequent biomolecular alterations that are

detectableusingSR-IRMS.Iffutureexperimentsweretobeperformed,theuseofaa

multi-faceted, fully sealed fluidic device, such as that reported byVaccari et al. [34]

maybemoresuitable.


200

4.5 ChaptersummarySR-IRMSwasemployedfortheanalysisofliveHCT-8cellsfollowingexposureto

BiNSAIDs or NSAIDs in order to determine if it could be used to study global

biomolecular changes and allude to potential mechanisms of action of these

compounds. A multivariate approach, PCA, was applied to discern the spectral

differences associated with the analysis of the data and different cell treatments.

Specificobservationsregardingtheanalyticalapproachincludedthefindingsthatthe:

(i) Subtraction of a water band spectrum from the cell spectrum of each

individualcellonlysubtlyaffectedthePCAresults.

(ii) Omission of the amide region of the HCT-8 cell spectra from PCA caused

substantial changes to thedistributionof the sample scoresplot. Inmany

cases, weak loadings bands became more prominent in the loadings plot

uponomissionoftheamidebands.

The subsequent PCA showed that there was significant intra-sample variation,

althoughsomedelineationcouldbediscernedinthefollowingtreatmentgroups:

(i) The comparison of cellmedium-only or vehicle-treatedHCT-8 cell spectra

indicated there were subtle effects, attributable to the vehicle (2% v/v

DMSOincellmedium),onthespectrafollowing4-htreatment,butnot24-h

treatment.

(ii) Following4-htreatments:

a. Bi(tolf)3 and tolfH induced similar biomolecular effects on the HCT-8

cells in comparison to the vehicle,whichwereprimarily attributed to

effects on the lipidswith some contribution fromproteins. However,

followingremovaloftheamideregion,discriminationbetweenthetwo

compounds was apparent and attributable to ν(C=O) and νas(PO2−),

suggestive of difference in the effect of the two compounds on

phospholipids.

b. Bi(mef)3ormefHinducedsimilareffectsontheHCT-8cells,whichwere

primarilyattributedtolipids.

c. DiflH induced biomolecular effects on the HCT-8 cells, which were

primarily attributed to proteins. However, following amide removal

from PCA analysis, diflH no longer showed any changes from the

vehicle.TreatmentwithBi(difl)3wasdistinguishablefromdiflHandthe


201

vehicle following amide removal, which was primarily attributed to

lipids.

(iii) Following24-htreatments:

a. TolfH induced biomolecular effects on the HCT-8 cells, which were

primarily attributed to lipids with some contribution from

proteins. Treatment of the cells with Bi(tolf)3 induced a varied

responsewithintheBi(tolf)3-treatedpopulation.

b. TreatmentwithBi(mef)3andmefHinducedsimilarbiomoleculareffects

ontheHCT-8cells,whichwereprimarilyattributedtolipids.

c. TreatmentwithBi(difl)3resultedinnodistinguishabledifferencefrom

the vehicle-treated cells. However, treatment with diflH was

distinguishable fromBi(difl)3 and vehicle scores,whichwas primarily

attributedtolipids.

The ambiguity surrounding the inclusion or omission of amide bands from

SR-IRMSspectraobtainedfromlivemammaliancellswasinvestigatedinthepresent

study using two methods reported in the literature. While this study may not

necessarilybeusedtoprovidefurtheradviceonthe inclusionofamidebands in live

cell studies, and thevalidityof thebiomolecular results theyprovide, this studyhas

providedadirectcomparisonaddingcredencetobothproceduresreportedbyother

investigators.

The SR-IRMS results obtained in the present study are preliminary in nature

showing intra- and inter-sample variations reflective of biomolecular differences in

the cell treatment groups. On the whole, the PCA results from the SR-IRMS study

suggest that treatmentwith BiNSAIDs or NSAIDs affect the lipids following 4-h and

24-h treatment. AsSR-IRMSprovidedaglobaloverviewofallbiomolecularchanges

that occurred in the HCT-8 cells following treatment with vehicle, BiNSAIDs or

NSAIDs(including cell cycle, cell death and drug-induced biomolecular changes),

additionalexperimental techniques shouldpotentiallybeemployed to confirm these

observations.Forinstance,astudytocomparetheeffectsoftheBiNSAIDs/NSAIDson

lipidsinisolatedsystems(e.g.cell-freelipidbilayers)isbeingconductedbyDillonand

Brown([68]andunpublishedresults),andanadditional invitro lipidomicstudyhas

beenperformedonisolatedlipidsfromdrug-treatedcells(seeChapter5).


202

References

[1] D.McNaughtonandB.R.Wood.(2012)Synchrotroninfraredspectroscopyofcellsandtissue,Aust.J.Chem.,65,218-228.

[2] B. Stuart. (2004) Biological applications, In Infrared spectroscopy:Fundamentalsandapplications,pp137-165,JohnWiley&Sons,Ltd.

[3] L.M.Miller,G.D.SmithandG.L.Carr.(2003)Synchrotron-basedbiologicalmicrospectroscopy: From the mid-infrared through the far-infraredregimes,J.Biol.Phys.,29,219-230.

[4] D.Naumann.(2001)FT-infraredandFT-ramanspectroscopyinbiomedicalresearch,Appl.Spectrosc.Rev.,36,239-298.

[5] P. Dumas and L. Miller. (2003) Biological and biomedical applications ofsynchrotroninfraredmicrospectroscopy,J.Biol.Phys.,29,201-218.

[6] J. Sulé-Suso, D. Skingsley, G. D. Sockalingum, A. Kohler, G. Kegelaer,M.Manfait and A. J. El Haj. (2005) FT-IR microspectroscopy as a tool toassess lung cancer cells response to chemotherapy, Vib. Spectrosc., 38,179-184.

[7] S. Chio-Srichan, M. Réfrégiers, F. Jamme, S. Kascakova, V. Rouam andP.Dumas. (2008) Photosensitizer effects on cancerous cells: A combinedstudy using synchrotron infrared and fluorescencemicroscopies,Biochim.Biophys.ActaGen.Subj.,1780,854-860.

[8] R.Gasper, J.Dewelle,R.Kiss,T.Mijatovic andE.Goormaghtigh. (2009) IRspectroscopy as a new tool for evidencing antitumor drug signatures,Biochim.Biophys.ActaBiomembr.,1788,1263-1270.

[9] F. Draux, P. Jeannesson, C. Gobinet, J. Sule-Suso, J. Pijanka, C. Sandt,P.Dumas,M.ManfaitandG.D.Sockalingum.(2009)IRspectroscopyrevealseffect of non-cytotoxic doses of anti-tumour drug on cancer cells, Anal.Bioanal.Chem.,395,2293-2301.

[10] A. Derenne, R. Gasper and E. Goormaghtigh. (2010) Monitoring ofmetabolism perturbation in prostate PC-3 cancer cells by sub-lethalconcentrationsofmethotrexate,Spectroscopy,24.

[11] A.Derenne,R.GasperandE.Goormaghtigh. (2011)TheFTIRspectrumofprostatecancercellsallowstheclassificationofanticancerdrugsaccordingtotheirmodeofaction,Analyst,136,1134-1141.

[12] K. R. Flower, I. Khalifa, P. Bassan, D. Demoulin, E. Jackson, N. P. Lockyer,A.T.McGown,P.Miles,L.VaccariandP.Gardner.(2011)SynchrotronFTIRanalysisofdrugtreatedovarianA2780cells:anabilitytodifferentiatecellresponsetodifferentdrugs?,Analyst,136,498-507.

[13] C.Hughes,M.D.Brown,F. J.Ball,G.Monjardez,N.W.Clarke,K.R.Flowerand P. Gardner. (2012) Highlighting a need to distinguish cell cyclesignatures from cellular responses to chemotherapeutics in SR-FTIRspectroscopy,Analyst,137,5736-5742.

[14] G.Bellisola,M.DellaPeruta,M.Vezzalini,E.Moratti,L.Vaccari,G.Birarda,M.Piccinini,G.CinqueandC.Sorio.(2010)TrackinginfraredsignaturesofdrugsincancercellsbyFouriertransformmicrospectroscopy,Analyst,135,3077-3086.

[15] A. Mignolet, A. Derenne, M. Smolina, B. R. Wood and E. Goormaghtigh.(2016)FTIRspectralsignatureofanticancerdrugs.Candrugmodeofactionbeidentified?,Biochim.Biophys.ActaProteinsProteom.,1864,85-101.


203

[16] H.-Y. N. Holman, M. C. Martin, E. A. Blakely, K. Bjornstad andW.R.McKinney.(2000)IRspectroscopiccharacteristicsofcellcycleandcelldeath probed by synchrotron radiation based Fourier transform IRspectromicroscopy,Biopolymers,57,329-335.

[17] J. Zhou, Z.Wang, S. Sun,M. Liu andH. Zhang. (2001) A rapidmethod fordetecting conformational changes during differentiation and apoptosis ofHL60 cells by Fourier-transform infrared spectroscopy, Biotechnol. Appl.Biochem.,33,127-132.

[18] N. Jamin, L.Miller, J.Moncuit,W.-H. Fridman, P. Dumas and J.-L. Teillaud.(2003)Chemicalheterogeneityincelldeath:CombinedsynchrotronIRandfluorescence microscopy studies of single apoptotic and necrotic cells,Biopolymers,72,366-373.

[19] F.GasparriandM.Muzio.(2003)MonitoringofapoptosisofHL60cellsbyFourier-transforminfraredspectroscopy,Biochem.J.,369,239-248.

[20] U. Zelig, J. Kapelushnik, R. Moreh, S. Mordechai and I. Nathan. (2009)Diagnosis of cell death bymeans of infrared spectroscopy,Biophys. J., 97,2107-2114.

[21] A. Lamberti, C. Sanges and P. Arcari. (2010) FT-IR spectromicroscopy ofmammalian cell culturesduringnecrosis andapoptosis inducedbydrugs,Spectroscopy,24.

[22] S. Boydston-White, M. Romeo, T. Chernenko, A. Regina, M. Miljković andM.Diem. (2006) Cell-cycle-dependent variations in FTIRmicro-spectra ofsingle proliferating HeLa cells: Principal component and artificial neuralnetworkanalysis,Biochim.Biophys.ActaBiomembr.,1758,908-914.

[23] J.R.Mourant,Y.R.Yamada,S.Carpenter,L.R.Dominiqueand J.P.Freyer.(2003) FTIR spectroscopy demonstrates biochemical differences inmammalian cell cultures at different growth stages, Biophys. J., 85,1938-1947.

[24] P. Heraud, E. S. Ng, S. Caine, Q. C. Yu, C. Hirst, R. Mayberry, A. Bruce,BR.Wood,D.McNaughton,E.G.StanleyandA.G.Elefanty.(2010)Fouriertransforminfraredmicrospectroscopyidentifiesearlylineagecommitmentindifferentiatinghumanembryonicstemcells,StemCellRes.,4,140-147.

[25] D.Ami,A.Natalello,P.Mereghetti,T.Neri,M.Zanoni,M.Monti,S.M.DogliaandC.A.Redi.(2010)FT-IRspectroscopysupportedbyPCA–LDAanalysisforthestudyofembryonicstemcelldifferentiation,Spectroscopy,24.

[26] D.Pozzi,P.Grimaldi,S.Gaudenzi,L.DiGiambattista,I.Silvestri,S.MorroneandA.CongiuCastellano.(2007)UVB-radiation-inducedapoptosisinJurkatcells: A coordinated fourier transform infrared spectroscopy-flowcytometrystudy,Radiat.Res.,168,698-705.

[27] L.DiGiambattista,P.Grimaldi,S.Gaudenzi,D.Pozzi,M.Grandi,S.Morrone,I. Silvestri andA.CongiuCastellano. (2010)UVBradiation inducedeffectsoncellsstudiedbyFTIRspectroscopy,Eur.Biophys.J.,39,929-934.

[28] M. J. Baker, J. Trevisan, P. Bassan, R. Bhargava,H. J. Butler, K.M. Dorling,P.R.Fielden,S.W.Fogarty,N. J.Fullwood,K.A.Heys,C.Hughes,P.Lasch,P.L.Martin-Hirsch,B.Obinaju,G.D.Sockalingum, J.Sulé-Suso,R. J.Strong,M. J.Walsh,B.R.Wood,P.GardnerandF.L.Martin. (2014)UsingFouriertransform IR spectroscopy to analyze biologicalmaterials,Nat. Protoc., 9,1771-1791.


204

[29] M. Miljkovic, M. Romeo, C. Matthaus and M. Diem. (2004) Infraredmicrospectroscopy of individual human cervical cancer (HeLa) cellssuspendedingrowthmedium,Biopolymers,74,172-175.

[30] D.A.Moss,M.KeeseandR.Pepperkok.(2005)IRmicrospectroscopyoflivecells,Vib.Spectrosc.,38,185-191.

[31] K.L.Munro,K.R.Bambery,E.A.Carter,L.Puskar,M. J.Tobin,B.R.WoodandC.T.Dillon. (2010)Synchrotronradiation infraredmicrospectroscopyof arsenic-induced changes to intracellular biomolecules in live leukemiacells,Vib.Spectrosc.,53,39-44.

[32] M. J. Tobin, L. Puskar, R. L. Barber, E. C. Harvey, P. Heraud, B. R. Wood,K.R.Bambery, C. T. Dillon and K. L. Munro. (2010) FTIR spectroscopy ofsingle live cells in aqueous media by synchrotron IR microscopy usingmicrofabricatedsampleholders,Vib.Spectrosc.,53,34-38.

[33] E. J. Marcsisin, C. M. Uttero, M. Miljkovic and M. Diem. (2010) Infraredmicrospectroscopyoflivecellsinaqueousmedia,Analyst,135,3227-3232.

[34] L.Vaccari,G.Birarda, L.Businaro, S. Pacor andG.Grenci. (2012) Infraredmicrospectroscopyoflivecellsinmicrofluidicdevices(MD-IRMS):towardapowerfullabel-freecell-basedassay,Anal.Chem.,84,4768-4775.

[35] E. J.Marcsisin,C.M.Uttero,A. I.Mazur,M.Miljkovic,B.BirdandM.Diem.(2012) Noise adjusted principal component reconstruction to optimizeinfraredmicrospectroscopyofindividuallivecells,Analyst,137,2958-2964.

[36] D. E. Bedolla, S. Kenig, E. Mitri, P. Ferraris, A. Marcello, G. Grenci andL.Vaccari.(2013)DeterminationofcellcyclephasesinliveB16melanomacellsusingIRMS,Analyst,138,4015-4021.

[37] P. Zucchiatti, E. Mitri, S. Kenig, F. Billè, G. Kourousias, D. E. Bedolla andL.Vaccari. (2016) Contribution of ribonucleic acid (RNA) to the fouriertransform infrared (FTIR) spectrum of eukaryotic cells, Anal. Chem., 88,12090-12098.

[38] E. Mitri, S. Kenig, G. Coceano, D. E. Bedolla, M. Tormen, G. Grenci andL.Vaccari. (2015) Time-resolved FT-IR microspectroscopy of proteinaggregationinducedbyheat-shockinlivecells,Anal.Chem.,87,3670-3677.

[39] G.Birarda,D.E.Bedolla,E.Mitri,S.Pacor,G.GrenciandL.Vaccari.(2014)ApoptoticpathwaysofU937 leukemicmonocytes investigatedby infraredmicrospectroscopyandflowcytometry,Analyst,139,3097-3106.

[40] E.Mitri, G. Birarda, L. Vaccari, S. Kenig,M. Tormen andG. Grenci. (2014)SU-8bondingprotocolforthefabricationofmicrofluidicdevicesdedicatedtoFTIRmicrospectroscopyoflivecells,LabChip,14,210-218.

[41] G. Birada,G. Grenci andL. Vaccari. (2010) Synchrotron radiation infraredmicrospectroscopy (SR-IRMS) of living cells in physiological environment,In Microscopy: Science, Technology, Applications and Education (A.Méndez-Vilas and J. Díaz, Eds.), pp 422-432, Formatex Research Centre,Padajoz.

[42] D.R.Whelan,K.R.Bambery,P.Heraud,M.J.Tobin,M.Diem,D.McNaughtonandB.R.Wood. (2011)Monitoring thereversibleB toA-like transitionofDNA in eukaryotic cells using Fourier transform infrared spectroscopy,NucleicAcidsRes.,39,5439-5448.

[43] D. R. Whelan, K. R. Bambery, L. Puskar, D. McNaughton and B. R. Wood.(2013) Synchrotron Fourier transform infrared (FTIR) analysis of singlelivingcellsprogressingthroughthecellcycle,Analyst,138,3891-3899.


205

[44] P.Heraud,B.R.Wood,M. J.Tobin, J.Beardall andD.McNaughton. (2005)Mappingofnutrient-inducedbiochemicalchangesinlivingalgalcellsusingsynchrotron infrared microspectroscopy, FEMS Microbiol. Lett., 249,219-225.

[45] D.A.Moss,M.KeeseandR.Pepperkok.(2005)IRmicrospectroscopyoflivecells,VibrationalSpectroscopy,38,185-191.

[46] N.Gierlinger,L.Goswami,M.Schmidt, I.Burgert,C.Coutand,T.RoggeandM. Schwanninger. (2008) In situ FT-IR microscopic study on enzymatictreatmentofpoplarwoodcross-sections,Biomacromolecules,9,2194-2201.

[47] M.J.Nasse,S.Ratti,M.GiordanoandC.J.Hirschmugl.(2009)Demountableliquid/flow cell for in vivo infrared microspectroscopy of biologicalspecimens,Appl.Spectrosc.,63,1181-1186.

[48] L.Quaroni,T.ZlatevaandE.Normand.(2011)Detectionofweakabsorptionchanges frommolecular events in time-resolvedFT-IR spectromicroscopymeasurementsofsinglefunctionalcells,Anal.Chem.,83,7371-7380.

[49] R. Zhao, L. Quaroni and A. G. Casson. (2010) Fourier transform infrared(FTIR)spectromicroscopiccharacterizationofstem-likecellpopulationsinhuman esophageal normal and adenocarcinoma cell lines, Analyst, 135,53-61.

[50] B.R.Wood.(2016)TheimportanceofhydrationandDNAconformationininterpreting infrared spectra of cells and tissues, Chem Soc Rev, 45,1980-1998.

[51] H.Y.Holman,R.Miles,Z.Hao,E.Wozei,L.M.AndersonandH.Yang.(2009)Real-time chemical imaging of bacterial activity in biofilms usingopen-channel microfluidics and synchrotron FTIR spectromicroscopy,Anal.Chem.,81,8564-8570.

[52] H.-Y.N.Holman,H.A.Bechtel,Z.HaoandM.C.Martin.(2010)SynchrotronIRspectromicroscopy:chemistryoflivingcells,Anal.Chem.,82,8757-8765.

[53] P. Bassan, A. Kohler, H. Martens, J. Lee, H. J. Byrne, P. Dumas, E. Gazi,M.Brown, N. Clarke and P. Gardner. (2010) Resonant Mie scattering(RMieS) correction of infrared spectra from highly scattering biologicalsamples,Analyst,135,268-277.

[54] B. Mohlenhoff, M. Romeo, M. Diem and B. R. Wood. (2005) Mie-typescattering and non-Beer-Lambert absorption behavior of human cells ininfraredmicrospectroscopy,Biophys.J.,88,3635-3640.

[55] M.Romeo,B.MohlenhoffandM.Diem.(2006)Infraredmicro-spectroscopyof human cells: Causes for the spectral variance of oral mucosa (buccal)cells,Vib.Spectrosc.,42,9-14.

[56] P. Bassan, A. Kohler, H.Martens, J. Lee, E. Jackson, N. Lockyer, P. Dumas,M.Brown, N. Clarke and P. Gardner. (2010) RMieS-EMSC correction forinfraredspectraofbiologicalcells:extensionusingfullMietheoryandGPUcomputing,J.Biophotonics,3,609-620.

[57] L.Vaccari,G.Birada,G.Grenci,S.PacorandL.Businaro.(2012)Synchrotronradiation infrared microspectroscopy of single living cells in microfluidicdevices: advantages, disadvantages and future perspectives, J. Phys. Conf.Ser.,359,012007.

[58] J.Soh,A.Chueng,A.Adio,A.J.Cooper,B.R.BirchandB.A.Lwaleed.(2013)Fourier transform infrared spectroscopy imaging of live epithelial cancercellsundernon-aqueousmedia,J.Clin.Pathol.,66,312-318.


206

[59] K.Malek, B. R.Wood andK. R. Bambery. (2014) FTIR imaging of tissues:techniques and methods of analysis, In Optical Spectroscopy andComputational Methods in Biology and Medicine (M. Baranska, Ed.),pp419–473,Springer,Netherlands.

[60] D. Ami, P. Mereghetti and S. M. Doglia. (2012) Multivariate analysis forfourier transform infrared spectra of complex biological systems andprocesses, In Multivariate Analysis in Management, Engineering and theSciences,INTECH,OpenAccessArticle.

[61] M.Diem,A.Mazur,K.Lenau,J.Schubert,B.Bird,M.Miljkovic,C.KrafftandJ.Popp. (2013) Molecular pathology via IR and Raman spectral imaging,J.Biophotonics,6,855-886.

[62] R.BroandA.K.Smilde.(2014)Principalcomponentanalysis,Anal.Methods,6,2812-2831.

[63] Z. W. Yu and P. J. Quinn. (1994) Dimethyl sulphoxide: A review of itsapplicationsincellbiology,Biosci.Rep.,14,259-281.

[64] N. D. Weiner, M. Y. Lu and M. Rosoff. (1972) Interactions of dimethylsulfoxidewithlipidandproteinmonolayers,J.Pharm.Sci.,61,1098.

[65] H. Tapiero, G. Zwingelstein, A. Fourcade and J. Portoukalian. (1983) Theeffect of dimethyl sulfoxide on the membrane dynamics and thephospholipid composition of two different cell lines,Ann. N. Y. Acad. Sci.,411,383-388.

[66] T. J. Anchordoguy, J. F. Carpenter, J. H. Crowe and L. M. Crowe. (1992)Temperature-dependent perturbation of phospholipid bilayers bydimethylsulfoxide,Biochim.Biophys.ActaBiomembr.,1104,117-122.

[67] M.-A. l. de Ménorval, L. M. Mir, M. L. Fernández and R. Reigada. (2012)Effectsofdimethyl sulfoxide incholesterol-containing lipidmembranes:Acomparativestudyofexperimentsinsilicoandwithcells,PLoSOne,7,1-12.

[68] T.Brown. (2016)Bismuth-NSAIDsascolorectal cancerchemopreventives:Insights into their membrane interactions and uptake mechanisms,Bachelor of Medicinal Chemistry (Advanced) with Honours, University ofWollongong,Australia.

[69] A. A. Gurtovenko and J. Anwar. (2007) Modulating the structure andproperties of cell membranes: the molecular mechanism of action ofdimethylsulfoxide,J.Phys.Chem.B,111,10453-10460.

[70] T. Togo. (2006) Disruption of the plasma membrane stimulatesrearrangement of microtubules and lipid traffic toward the wound site,J.CellSci.,119,2780-2786.

[71] T.Togo,J.M.Alderton,G.Q.BiandR.A.Steinhardt.(1999)Themechanismoffacilitatedcellmembraneresealing,J.CellSci.,112,719-731.

[72] T.Togo,J.M.AldertonandR.A.Steinhardt.(2003)Long-termpotentiationof exocytosis and cell membrane repair in fibroblasts,Mol. Biol. Cell, 14,93-106.

[73] A.M.Schrader,S.H.Donaldson,Jr.,J.Song,C.Y.Cheng,D.W.Lee,S.HanandJ. N. Israelachvili. (2015) Correlating steric hydration forces with waterdynamics through surface force and diffusion NMR measurements in alipid-DMSO-H2Osystem,Proc.Natl.Acad.Sci.U.S.A.,112,10708-10713.

[74] G.DaViolante,N.Zerrouk,I.Richard,G.Provot,J.C.ChaumeilandP.Arnaud.(2002)Evaluationofthecytotoxicityeffectofdimethylsulfoxide(DMSO)onCaco2/TC7colontumorcellcultures,Biol.Pharm.Bull.,25,1600-1603.


207

[75] Y. S. Kim,D. Tsao, B. Siddiqui, J. S.Whitehead, P. Arnstein, J. Bennett andJ.Hicks. (1980) Effects of sodium butyrate and dimethylsulfoxide onbiochemicalpropertiesofhumancoloncancercells,Cancer,45,1185-1192.

[76] J. F. Neault and H. A. Tajmir-Riahi. (1998) Interaction of cisplatin withhuman serum albumin. Drug binding mode and protein secondarystructure,Biochim.Biophys.Acta,1384,153-159.

[77] G.Abdel-RaoufAhmed,F.A.R.KhorshidandT.A.Kumosani.(2009)FT-IRspectroscopy as a tool for identification of apoptosis-induced structuralchanges in A549 cells treated with PM 701, Int. J. Nano Biomater., 2,396-408.

[78] L.Buriankova,Z.Nadova,D.Jancura,M.Refregiers,I.Yousef,J.MikesandP.Miskovsky. (2010) Synchrotron based Fourier-transform infraredmicrospectroscopy as sensitive technique for the detection of earlyapoptosisinU-87MGcells,LaserPhys.Lett.,7,613.

[79] M.N.Giraud,C.Motta,J. J.Romero,G.BommelaerandL.M.Lichtenberger.(1999) Interaction of indomethacin and naproxen with gastric surface-active phospholipids: a possible mechanism for the gastric toxicity ofnonsteroidal anti-inflammatory drugs (NSAIDs), Biochem. Pharmacol., 57,247-254.

[80] L.M.Lichtenberger,Y.Zhou,V.Jayaraman,J.R.Doyen,R.G.O'Neil,E.J.Dial,D. E. Volk, D. G. Gorenstein, M. B. Boggara and R. Krishnamoorti. (2012)Insight into NSAID-induced membrane alterations, pathogenesis andtherapeutics: characterization of interaction of NSAIDs withphosphatidylcholine,Biochim.Biophys.Acta,1821,994-1002.

[81] M.Suwalsky,M.Manrique-Moreno,J.Howe,K.BrandenburgandF.Villena.(2011)Molecularinteractionsofmefenamicacidwithlipidbilayersandredbloodcells,J.Braz.Chem.Soc.,22,2243-2249.

[82] K. Segawa and S. Nagata. (2015) An apoptotic 'eat me' signal:phosphatidylserineexposure,TrendsCellBiol.,25,639-650.

[83] S. Baritaki, S. Apostolakis, P. Kanellou, M. T. Dimanche-Boitrel, D. A.Spandidos and B. Bonavida. (2007) Reversal of tumor resistance toapoptotic stimuli by alteration of membrane fluidity: therapeuticimplications,Adv.CancerRes.,98,149-190.

[84] K. Z. Liu, L. Jia, S. M. Kelsey, A. C. Newland and H. H. Mantsch. (2001)Quantitative determination of apoptosis on leukemia cells by infraredspectroscopy,Apoptosis,6,269-278.

[85] S. Gaudenzi, D. Pozzi, P. Toroa, I. Silvestri, S. Morrone and A. Castellano.(2004)CellapoptosisspecificmarkerfoundbyFouriertransforminfraredspectroscopy,Spectroscopy,18,415–422.

[86] W. Fan, J. Li, J. Chen, L. Zhu, Y.Wang, B. Sun, B. Hua, C. Guo and Z. Yan.(2017)Aspirininhibitstheproliferationofsynovium-derivedmesenchymalstemcellsbyarrestingthecellcycleintheG0/G1phase,Am.J.Transl.Res.,9,5056-5062.

[87] M.Du,W.Pan,X.Duan,P.YangandS.Ge.(2016)Lowerdosageofaspirinpromotes cell growth and osteogenic differentiation in murine bonemarrowstromalcells,J.DentalSci.,11,315-322.

[88] R.A.Perugini,T.P.McDade,F.J.Vittimberga,Jr.,A.J.DuffyandM.P.Callery.(2000) Sodium salicylate inhibits proliferation and induces G1 cell cycle


208

arrestinhumanpancreaticcancercell lines,J.Gastrointest.Surg.,4,24-32,discussion32-23.

[89] G. Birarda, G. Grenci, L. Businaro, B. Marmiroli, S. Pacor, F. Piccirilli andL.Vaccari. (2010) Infraredmicrospectroscopy of biochemical response oflivingcellsinmicrofabricateddevices,Vib.Spectrosc.,53,6-11.

209

Chapter5. Lipidomic profiling of glycerophospholipids using

ESI-MS/MS of HCT-8 cells treated with BiNSAIDs

andNSAIDs

This Chapter describes shotgun lipidomic profiling of the organic extracts of HCT-8

cells using electrospray ionisation tandem mass spectrometry (ESI-MS/MS). This

study was undertaken to compliment SR-IRMS studies reported in Chapter 4 that

showed that the most prominent biomolecular changes in BiNSAID/NSAID-treated

cells was that associated with lipids. The purpose of the lipidomic study was to

determine how treatmentwith BiNSAIDs or NSAIDs affects the spectral profile and

cellular concentration of the common phospholipids, phosphatidylcholine (PC),

phosphatidylethanolamine(PE)andphosphatidylserine(PS).

Chapter5.LipidomicprofilingofglycerophospholipidsusingESI-MS/MS

210

5.1 IntroductionPhospholipids are one of the major structural lipids found in eukaryotic cell

membranes and are essential for maintaining the biochemical and biophysical

propertiesofthecellmembrane[1].Theeukaryoticcellmembraneiscomprisedofa

phospholipid bilayer, embedded with other lipid molecules (namely sterols and

glycolipids) and membrane-embedded proteins [1]. The lipid bilayer provides

individual cells with a selective barrier from the extracellular environment and a

communication point between the intracellular and extracellular environments [1].

Additionally, lipid bilayers form discrete organelles within the intracellular space.

Theyareessentialforcompartmentalisationofspecificchemicalreactions,increasing

the biochemical efficiencywithin the cell [1]. Aswell as their structural role,many

phospholipidsaretheprecursorsforsecondarymessengers,whichplayanimportant

role in cellular functions including cell proliferation, apoptosis, signal transduction

pathwaysandtheregulationofmembranetrafficking[2].

Giventhe importanceofphospholipids inmaintaining thephysical integrityand

homeostasis of cells, changes to the composition of phospholipids associated with

apoptosis, in particular drug-induced apoptosis, have been increasingly investigated

[3-5]. The apoptoticprocess causes changes to thephospholipid compositionof the

outersurfaceofthecellmembraneandthephospholipiddistributionwithinthelipid

bilayer (most notably the externalisation of negatively-charged PS) [6]. Recently,

attention has focused on the importance of phospholipid composition in diseases,

includingcancerandtheirassociatedtreatment[7-9]. For instance,cancercellsthat

have a higher abundance of saturated phospholipids[10], exhibit changes in lipid

membrane dynamics and protein signal transduction, which can protect the cancer

cell fromoxidative stressby lowering the riskof lipidperoxidation. Additionally, as

lipidsmediateanumberofcellularprocessesthroughsignalingpathways,alterations

in lipid composition and signaling can result in changes to theway the cancer cells

respondtocertainchemotherapies[11].

Theresults fromtheSR-IRMSstudy(Chapter4) implied that therewasaglobal

effect on cellular lipids followingHCT-8 cell treatmentwithBiNSAIDs/NSAIDs. The

anti-inflammatoryeffectsofNSAIDsarisethroughtheinteractionwithlipidmediators

resulting in downstream consequences for arachidonic acid, which is derived from

phospholipidsintheplasmamembraneviaphospholipaseA2[12].Additionally,there

are several studies [13-16] suggesting that NSAIDs are capable of interaction with


211

phospholipids.Theseincludetheireffectonhydrophobicity,fluidityandpermeability

properties of membrane phospholipids; as such these effects could ultimately

contribute to the cell injury/toxicity observed following NSAID exposure.

Consequently, investigating changes to the cellular composition of common

phospholipids is a logical place to begin examining lipid-related effects of BiNSAIDs

andtherespectivefreeNSAIDsonthelipidomeofHCT-8cells.

Recently, lipidomics (or lipid profiling) has been increasingly utilised to study

various tissue and cellular lipidomes [17]. This has been aided by recent

advancementsinmassspectrometry(MS)techniques,whichhaveenabledresearchers

to study lipids with high sensitivity, allowing for the accurate characterisation,

identification, and quantification of thousands of lipid species in biological

samples[17].

5.1.1 Shotgun lipidomics:Electrospray ionisationmassspectrometry(ESI-MS)

andtandemMSfortheexaminationofphospholipids

While several MS approaches exist for the analysis of lipids, electrospray

ionisation tandem mass spectrometry (ESI-MS/MS) represents one of the most

commonly utilised techniques for achieving detailed structural analysis of

phospholipids in biological samples [18]. Two common approaches applied for

ESI-MS/MSanalysisoflipidsare:(i)thepartialseparationofthecrudelipidextractvia

high-performance liquid chromatography followed by infusion of the mobile phase

into theESI source [18];or (ii) thedirect infusionof the sample into theESI source

[19].Thelatterapproach,knownasshotgunlipidomics[19],wasthechosenapproach

inthestudydescribedinthisChapter.

The general protocol for the preparation of the lipid extracts from biological

materialsforshotgunlipidomicanalysisusingESI-MS/MSissummarisedinFigure5.1

where the subsequent phospholipid analysis is performedusing a triple quadrupole

(QqQ)instrumentwithanESIsource.AnESImassspectrumfromacrudelipidextract

providesaprofileofthelipidsthationiseinthechosenmode(i.e.positiveornegative).

However,sincethemassaccuracyofsomeMSinstrumentusedfortheseexperiments

may be low and the crude lipid extract (with potentially thousands of lipids) is

complicated, thecomplementary techniqueESI-MS/MS isoftenemployed to confirm

molecular identification[20]. ESI-MS/MS can be achieved through the use of aQqQ

massspectrometer,showninFigure5.1. InatraditionalESI-MS/MSexperimentina


212

QqQ, the ion of interest (parent ion) is isolated inQ1, followedby collision-induced

dissociation (CID) inq2 (collisioncell) (Fig.5.1). This involves fragmentationof the

[M+H]+ or [M−H]− ions by collisionwith an inert gas (such as argon) in q2,with

massanalysis of the resulting ionic fragments occurring in Q3. Specifically, CID of

phospholipids in positive ionmode causes cleavage of the phosphate-glycerol bond

resulting ineliminationofthepolarheadgroupasachargedorneutralspecies[21].

The polar head group fragments (or product ions) are characteristic of each

phospholipid;therefore,targetedscans(knownasprecursorionorneutrallossscans)

enablethedetectionandstructuralelucidationofeachoftheprecursorphospholipid

molecules contained in a samplewhile excluding other (sometimesmore abundant)

ions[17].

Figure 5.1: Schematic representation of the shotgun lipidomic profilingmethod by tandemmassspectrometryusingatriplequadrupole(QqQ)MSinstrument.


213

5.1.2 Thestructureofglycerophospholipids

The general chemical structures of several common glycerophospholipids are

showninFigure5.2a.Glycerophospholipidsareamphipathicmoleculesthatcontaina

polarheadgrouplinkedtoa3-carbonglycerolbackboneinthesn-3position,andtwo

fattyacylmoleculesattachedtothesn-1andsn-2positionsoftheglycerolbackbone.

The fatty acyl molecules are bonded at the sn-1 position via an ester (diacyl

glycerophospholipids), alkyl ether (1-O-alkyl) or alkenyl ether (1-O-alkenyl)

linkage(Fig. 5.2a). The chemical structures of the PC, PE and PS head groups are

shown in Figure 5.2b. The head groups provide the glycerophospholipid with a

hydrophilic region. The final components of phospholipids are the fatty acyl

molecules, which comprise the hydrophobic region of the glycerophospholipid

molecule(Fig.5.2c).Fattyacylmoleculesareclassifiedbytheirchainlength(number

of carbons, typically 4−30). The number and position of double bonds, including

saturated (no double bonds), monounsaturated (one double bond) and

polyunsaturated(more thanonedoublebond),are furtherused toclassify fattyacyl

chains. The number of glycerophospholipid head groups, and the varying carbon

chain lengthsandnumber/positionofdoublebondsof fattyacids, essentiallymeans

thatover10,000uniquemolecularspeciesofglycerophospholipidscouldexist.


214

Figure5.2:Chemicalstructureof:(a)Glycerophospholipids,wheretheglycerolbackboneisinblack, thehydrophilicregion ishighlighted inblueandthehydrophobic fattyacylchainsarehighlighted in red. (b) Common glycerophospholipid head groups. (c) A fatty acid (18carbons in length) containing no double bonds (saturated), a single double bond(monounsaturated),andtwodoublebonds(polyunsaturated).


215

5.1.3 Nomenclatureofglycerophospholipids

Theglycerophospholipidnomenclatureused in this thesis is inaccordancewith

Fahyetal.[22].Glycerophospholipidsarenamedfirstlybytheirheadgroupandthen

bytheirfattyacylchains;forinstance,amoleculewithaphosphocholineheadgroup

andtwofattyacylchainscomprisedof16carbons(sn-1position)and18carbonswith

one double bond (sn-2 position) would be referred to as PC (16:0/18:1). The

experiments reported in this Chapterwereused to identify the total composition of

theglycerophospholipidmolecule,butdonotidentifytheindividualfattyacylchains.

In this instance, the sum composition (i.e. total quantity of carbons and number of

doublebonds)ispresentedinthename.Forexample,aPCmoleculecontainingatotal

of 34carbons in the fatty acyl chains and one double bond will be presented as

PC34:1.

It must be noted that one of the limitations of the commonly employed MS

techniques(asinthisstudy)istheinabilitytodistinguishisobaricspecieswitheither

odd-chainorether-linkedfattyacylgroups.Inaddition,theMSmethoduseddoesnot

allow for definitively distinguishing between glycerophospholipids containing an

1-O-alkenyletherversusglycerophospholipidswitha1-O-alkylether-linkedfattyacyl

groupcontainingadoublebond,asit isnotpossibletodeterminethepositionofthe

double bond along the fatty acyl chain. Therefore, it can only be determined that a

doublebondispresentinanether-linkedfattyacylchain. ThismeansaPCmolecule

withm/z746 canbe assigned to threedifferentmolecules; PC33:1 (odd-chain), PC

P-34:0 (1-O-alkenyl ether), or PCO-34:1 (1-O-alkyl ether). Odd-chain fatty acids

constitute veryminor components inmammals [23] andhavebeen reported at low

concentrationsinhumancancercelllines[3].Assuch,moleculeswithisobaricspecies

will bedesignatedas ether-linked, by the “O-”prefix, for the alkyl ether linked fatty

acid, and the “P-” prefix for the 1-O-alkenyl ether, or plasmalogen (e.g. PC O-34:1/

PCP-34:0).

5.1.4 Biologicalimportanceofglycerophospholipids

Ineukaryoticcellstheglycerophospholipidsaccountforapproximately60mol%

oflipidmass[19].Themostcommonphospholipidclassesfoundinmammaliancells

arePCandPE,whilePS,phosphatidylglycerol,diphosphatidylglycerol(orcardiolipin),

phosphatidylinositol and phosphatidic acid comprise less abundant phospholipid

speciesinmammaliancells[1].Inmosteukaryoticmembranes,PCandPEcollectively


216

comprise approximately 75mol%of total phospholipidmass, and are present in an

approximately3:2(PC:PE)molarratio[19].GiventheabundanceofPCandPEwithin

thecell,thesetwoglycerophospholipidswerechosenforlipidomicexaminationinthe

presentstudy.

As phospholipids are the major structural lipids found in eukaryotic cell

membranes, thevariation in themolecular structuresofphospholipidmoleculeshas

implications for cellularmembrane properties includingmembrane fluidity and the

modulationofmembrane-boundproteinactivity[1].Phospholipidsusuallydistribute

asymmetricallyinthecellmembraneswherePCcanbedominantlyfoundintheouter

monolayer, whilst PE and PS are mainly present in the inner monolayer of plasma

membranes [19]. The composition of the phospholipids contained in the lipid

membranes surrounding individual types of organelles can also differ [1]. For

instance,PEisparticularlyenrichedininnermembranesofmitochondria(∼35–40%

of totalphospholipids)comparedwithotherorganelles(∼17–25%)[24]. Duetothe

relativesizeofthePCheadgroup,PCmoleculesaretypicallycylindricalinshape[25].

Alternatively,PEmoleculescontainasmallerheadgroup,resultinginatypicallycone

shaped molecule [26]. PE molecules alone cannot form membranes, rather their

insertion into the membrane is essential for important cellular processes such as

membranefusionandfissionandtheembeddingofmembraneproteins[26].PEisthe

mostabundant lipidonthecytoplasmic layerofcellularmembranes,withsignificant

roles in cellular processes such as membrane fusion, cell cycle, autophagy, and

apoptosis [27]. PE molecules, similarly to PS, are translocated to the outer cell

membraneduringapoptosis [27-29]. Phospholipidsalsoplaya role incell signaling

pathways. For instance, the formation of the lipid mediators, prostaglandins and

leukotrienes, are derived from arachidonic acid sourced from the lipid

membrane[30].

As shown in Figure 5.2, there are two types of ether-containing

glycerophospholipids, 1-O-alkyl and 1-O-alkenyl. The ether-linked phospholipids

predominantly contain either a phosphocholine or a phosphoethanolamine head

group. Upto50%ofglycerophospholipidscontainether-linkedfattyacylchains,but

their functionsremain largelyunknown[31]. Ether-containingPCmolecules tendto

containthe1-O-alkylbondandaretypifiedbyplateletactivatingfactor[32,33]. The

1-O-alkenylglycerophospholipidscontainafattyacylgroupthathasavinyletherbond

(i.e. a double bond on the first carbon from the ether end, Fig. 5.2a) and are often


217

termed plasmalogens. Plasmalogens are particularly susceptible to oxidative attack

due to the presence of the vinyl-ether bond; as such, plasmalogens have a reported

antioxidant effect towards reactive species including ROS and iron-induced

peroxidation[33].Incertaincelltypes(e.g.,inflammatorycells,neurons,andtumour

cells), up to 70% of PE molecules contain an ether-linkage, rather than an acyl-

linkage[34].

5.1.5 ChapterAims

Themechanism(s)ofNSAID-inducedapoptosisincancercellsinvitroandinvivo

hasyet tobe fullydetermined,particularly in thecaseofCOX-2deficientcancercell

lines(discussedinSection1.2.4).IthasbeenshownthatanumberofNSAIDsinteract

directlywithglycerophospholipids,resultinginchangestothehydrophobicity,fluidity

andpermeabilitypropertiesofmembranephospholipids[13-16].GiventhatBiNSAID-

andNSAID-inducedapoptosisintheHCT-8cellswasidentifiedinChapter2,andthat

lipidvariationwasprominentintheSR-IRMSresults(Chapter4)itwasofinterestto

determine whether BiNSAID- and NSAID-induced apoptosis caused changes to the

glycerophospholipidcompositionofthemostabundantcellularglycerophospholipids,

PC,PEandPS,inHCT-8cells.

In this study, theanalysisof cellularglycerophospholipidswasperformedusing

shotgun lipidomics. The effect of two BiNSAIDs and their respective NSAIDs were

chosenforexaminationusinglipidextraction.Bi(tolf)3wasassessedduetothefactit

exhibitedhighesttoxicityagainsttheHCT-8cells,andBi(difl)3waschosenasthediflH

representedasalicylateacidstructureversusthefenamatestructuralgroupof tolfH.

Additionally,Bi(difl)3 exhibitedunusual toxicitywith respect to themolar IC50 ratio,

BiNSAID:NSAID(aspertheMTTassayresultspresentedinChapter2).

ThespecificaimsofthisChapterwereto:

1. DeterminewhetherBiNSAIDtreatmentaffectedthecompositionandrelative

quantity of abundant glycerophospholipid (PC, PE and PS)molecules in the

HCT-8 colon cancer cell line, and compare the effects to the corresponding

NSAIDtreatment.

2. ComparewhethertheeffectsonPC,PEandPScompositionandtheirrelative

quantitiesdifferedbetweenBiNSAIDtreatment(Bi(tolf)3andBi(difl)3).


218


Allof thechemicalsandmaterialsused for thecell cultureand treatmentof the

cell samples in this Chapter have been previously described in Section 2.2.1.

Additionalchemicals(specifictotheexperimentsinthisChapter)includemagnesium

chloride hexahydrate (MgCl2.6H2O, ≥99%) and 2-mercaptoethanol (>98%), which

were purchased from Sigma Aldrich (St Louis, USA). QuickStartTM Bradford Dye

Reagent,1×andQuickStartTMbovineserumalbumin(BSA)standardset(2mg,1.5mg,

1mg,0.75mg,0.5mg,0.25mgand0.125mg)wereacquiredfromBio-Rad(Hercules

CA, USA). Analytical grade ammonium acetate(≥97%), chloroform (≥99.5%), and

methanol(≥99.7%)andweresourcedfromAjaxFinechem(SevenHills,Australia). A

methanolic internal standard mix (comprised of PC (19:0/19:0), 20µM; PE

(17:0/17:0), 20 µM; PS (17:0/17:0), 20 µM) was generously provided by

AssociateProfessorToddMitchell(SchoolofMedicine,UniversityofWollongong)and

prepared by Dr Jennifer Saville (School of Chemistry, University of Wollongong).

TheglycerophospholipidstandardswereoriginallyprocuredfromAvantiPolarLipids

(Alabaster,USA).

5.2.2 Celllines


5.2.3 Preparationofcelllipidextracts

HCT-8 cells (1×106 cells, 5mL)were seeded in60-mmcell culturedishesand

incubated for24h,afterwhichthegrowthmediumwasremovedandthecellswere

washed twice with PBS solution. Treatment solutions of the BiNSAID (Bi(tolf)3 or

Bi(difl)3) or NSAID (tolfH or diflH) were added to each dish (3 mL RPMI-1640,

2%DMSOv/v),whilethecontrolcellswereincubatedwithvehicle(3mLRPMI-1640,

2%DMSOv/v)for24h.HCT-8cellsweretreatedwiththeapproximateIC50dosesof

each BiNSAID (as determined by MTT assays, Section 2.3.2), or an equimolar

concentration of the corresponding free NSAID. Following treatment, any detached

cellswerecollected. Theadherent cellswerewashedwithPBSsolution (1mL)and

anydisplacedcellswerecollected. Trypsin(0.25%inPBS,2mL)wasthenaddedto

eachdishand the remainderof the cellswas combinedwith thepreviouswashings.

The resultant suspension was centrifuged (300 g, 5 min) and the cell pellet was


219

washed with PBS solution to remove any trypsin. The cells were resuspended in

treatment medium(1mL) and the cell suspension was divided into two 500-µL

aliquots;onealiquotwasimmediatelyassessedforproteincontent,whilethesecond

aliquotwasstored(37°C,5%CO2)forlipidextractionandESI-MS/MSanalysis.

ThecellsinthefirstaliquotwerecentrifugedandwashedoncewithPBSsolution.

The cells were then resuspended in ice-cold osmotic lysis buffer (200 µL, 10 mM

HEPES, 5 mM MgCl2.6H20 and 20 mM 2-mercaptoethanol) and incubated on

ice(30min). Following this incubation the swollen cells were sheared 10 times

througha29-gaugesyringetocausedisruptionofthecellmembrane.Lysisefficiency

wasconfirmedbylightmicroscopy.Thedisruptedcellswerecentrifuged(20,000g,5

min) to collect lysed cellular debris and the supernatant (20 µL) was assayed for

proteinconcentrationwithBradfordreagentusingBSAasastandard.

Thesecondaliquotofcells,tobeusedforlipidextraction,wasremovedfromthe

incubatorandthecellswereresuspended.Acalculatedvolumeofthecellsuspension

(normalised to protein concentration) was removed and the cells were centrifuged

andwashedwith PBS solution. Theywere then resuspended inmethanol (500 µL)

andstoredovernightat−80°C.Thecellsamples(inmethanol)weretransferredtoa

glass test tube, chloroform(1mL)wasaddedand the solutionwasvortexedbriefly.

Thecellsampleswereplacedonanorbitalshakerat lowspeedfor1htoallowlipid

extraction to occur. Samples were spikedwith amethanolic internal standardmix

(50µL) such that he final concentration of each internal standard in all the cell

samples was 50nmol/mg protein. Following brief vortexing, ammonium acetate

solution (1mL, 0.15 M) was added and the solution was vortexed and centrifuged

(5min,2,000g). Theorganicphasewas removedand1mLof chloroform:methanol

(2:1 v/v) was added to the aqueous phase and vortexed and centrifuged (5 min,

2,000g).Theorganicphasewascombinedwiththepreviouslycollectedorganicphase

andafreshaliquotofammoniumacetate(500μL)wasadded.Theresultantsolution

wasvortexedandcentrifuged(5min,2,000g)followingwhich,theaqueousphasewas

removedusingaPasteurpipette.Carewastakentoensurethattheorganicphasewas

notdisturbed.Theorganicphasewasthendriedundernitrogenbeforereconstitution

in chloroform:methanol (1:2v/v, 500 μL). Samples were stored at −80°C until

ESI-MS/MSanalysiswasperformed.


220

5.2.4 Massspectrometry

ESI-MS/MS analysis was performed on a Waters QuattroMicro™ QqQ mass

spectrometerwithMasslynx 4.0 software (Waters,Milford, USA). Prior to infusion,

aqueousammoniumacetate(1M,25μL)wasaddedtoallsamplestoproduceafinal

concentrationof50mM. ForESI-MSanalysis,sampleswereinfusedataflowrateof

10μL/min. Table5.1 contains the parameters used for ESI-MS and targeted ion

scan(MS/MS) analysis in positive ionmode. The collision energies for the targeted

ionscansusedinthisprojectareshowninTable5.2.Spectraweretypicallyobtained

overamassrangeofm/z600-900inpositiveionmode.ForESI-MS/MS,ascanrateof

200Th/swasused,and100scanswerecombined. Thespectrawerethenprocessed

usingbackgroundsubtractionandsmoothingwithaSavitsky-Golayalgorithm.

Table 5.1: Instrument parameters used for ESI-MS and targeted ion scan (MS/MS)analysisontheWatersQuattroMicroTMtriplequadrupolemassspectrometer.

Parameter MS MS/MSCapillary(kV) 3.0 3.0Cone(V) 50.0 50.0Extractor(V) 2.00 2.00RFlens(V) 0.0 0.0Sourcetemperature(°C) 80 80Desolvationgasflowrate(L/h) 320 320LM1resolution 15.0 15.0HM1resolution 15.0 15.0Ionenergy1 0.5 1.0Entrance 50 −2Collisionenergy 2 seeTable4.2Exit 50 1LM2resolution 15.0 15.0HM2resolution 15.0 15.0Ionenergy2 0.6 1Multiplier(V) 650 650Gascellpiranipressure <1.0e−4 3.0e−3

Table5.2:Targetedion(MS/MS)scans.

ScanType Targetedion Collisionenergy(eV) Featureidentified*Precursor m/z184.1 35 PCNeutralloss 141.0Da 25 PENeutralloss 185.0Da 22 PS*PC,phosphatidylcholine;PE,phosphatidylethanolamine;PS,phosphatidylserine.


221

5.2.5 Dataanalysis

In order to determine the limit of detection (LOD) and limit of quantification

(LOQ)oftheinstrumentforeachsetofsamplesthatwererunoverdifferentdays,the

LODandLOQwerecalculatedforallprecursorscansandneutrallossscansfollowing

the conclusion of each run. Following background subtraction and smoothing, the

averagenoisewascalculatedinaregionofthespectrumwherenoionizedlipidswere

expected by averaging the total noise intensity over a 3 Th region. Themean and

standarddeviation(s.d.)werecalculatedfromtheaverageofthetotalnoiseintensity

andtheLODandLOQweredeterminedusingthefollowingequations:

LOD=mean+(3×s.d.)(1)

LOQ=mean+(10×s.d.)(2)

Thecentroid functionwasapplied to thespectrausingpeakareas todetermine

theareaunderthecurve.AnypeakswithintensitybelowtheLODwerenotincluded

infurtheranalysis.PeakintensitiesabovetheLOD,butbelowtheLOQ,werereported

intheidentification,butwereexcludedfromquantitativeanalysis.

AnExcel template (providedbyAssociateProfessorToddMitchell)wasused to

simulate isotopicdistributionsand thenecessarycorrectionswereapplied toenable

quantification of individual and total lipids within the PC, PE and PS classes as

described in previous studies [35]. The relative ion abundance of the three

isotopologues was determined using the MassLynx isotope modeling function by

calculating the 13C isotopicdistribution foreachmoleculeand inputasa ratioof the

molecular ion. The analysis beganwith the smallest ion in order to correct for an

isotopologueofa lowerm/z ioncontributingat themono-isotopicpeakofaheavier

ion.Theabundancecouldthenbesummedforallisotopes,eliminatingdiscrimination

based on size and isotopic ratio differences. In order to quantify each individual

glycerophospholipid present within the class, the isotope sum for each ion was

normalisedtotheisotopesumoftheinternalstandard.


Statisticalanalysiswasperformedusingone-wayANOVA,followedbytheTukey-

Kramermultiple comparison test using GraphPad Prism, Version 5.04 forWindows

(GraphPadSoftware,La Jolla,USA). Theresultswereanalysedas thevehicleversus

BiNSAIDversusthecorrespondingNSAID,inordertodrawastatisticalcomparisonat

thesameNSAIDtreatmentconcentration.Inthecaseswheretheionabundancewas


222

not detected above the LOD or LOQ in at least 75% of the samples, the ion was

excludedfromthestatisticalanalysis[36].Resultsarepresentedasmean±standard

deviation(s.d.).AvalueofP≤0.05wasconsideredstatisticallysignificant.


223

5.3 Results5.3.1 Phosphatidylcholine(PC)

The effect of BiNSAID and NSAID treatment on the individual and total PC

moleculesinHCT-8cellswasexaminedintwoways:(i)inspectingthechangesinthe

overallprofileof thePCmolecules(i.e. changes in therelativeabundanceofeachPC

moleculeasaproportionofthetotalofPC,Fig.5.3−5.5)and(ii)calculatingtheoverall

changestotherelativeconcentrationofPCmolecules(Fig.5.6).

TheprofilesofthePCmolecules(calculatedasapercentageoftotalPC)detected

above the LOD using ESI-MS/MS in HCT-8 cell lipid extracts treated with vehicle,

Bi(tolf)3ortolfHareshowninFigure5.3,whiletheprofilesfollowingtreatmentwith

vehicle, Bi(difl)3 or diflH are shown in Figure 5.4. For clarity, only molecules that

accounted for >0.5% relative abundance of the total PC molecules have been

presentedinFigures5.3and5.4. AsummaryofalltheidentifiedPCmoleculesabove

theLODcanbefoundinAppendix4.1.Asshownintheleft-mostgraphsofFigures5.3

and 5.4, the three most abundant PC molecules present in the vehicle- and

BiNSAID/NSAID-treated HCT-8 cell lipid extracts were PC 34:1 (m/z 760),

PC32:1(m/z 732) and PCP-32:0/O-32:1 (m/z 718). Notably, Bi(tolf)3 and tolfH

treatmentexhibitedverysimilarprofilechanges,whichdifferedsignificantlyfromthe

profile obtained from vehicle-treated cells (Fig. 5.3). Additionally, the Bi(difl)3 and

diflH treatment exhibited very similar profile changes, which differed significantly

from the profile obtained from vehicle-treated cells(Fig 5.4). However, the profile

changes induced by Bi(tolf)3/tolfH (Fig. 5.3) compared to Bi(difl)3/diflH (Fig. 5.4)

differedsomewhat.Forinstance,therelativeabundanceofthemostpredominantPC

molecule,PC34:1,decreasedintheBi(tolf)3/tolfH-treatedcells(Fig5.3),butincreased

intheBi(difl)3/diflH-treatedcells(Fig5.4)comparedtothevehicle.


224

Figure5.3:Profileofphosphatidylcholine(PC)moleculesinHCT-8cellsfollowing24-htreatmentw

ithvehicle(RPMI-1640,DMSO

2%v/v),Bi(tolf)3(15µM),ortolfH(45µM

),calculatedasapercentageoftotalPC.Dataarepresentedasmean±s.d.(n=3(vehicle)

orn=4(Bi(tolf)3andtolfH)replicatesfrom

oneexperiment.Significancewithrespecttothecontrolisindicatedbylp≤0.05

(BiNSAIDandNSAID),up≤0.05(BiNSAIDonly),ornp≤0.05(NSAIDonly).Forclarityonlymoleculesthataccountedfor>0.5%

relativeabundanceofthetotalPCmoleculeshavebeenpresented.N

ote:Oddorether-linkedfattyacylchain(e.g.PCO-34:0/PC33:1)

cannotbedifferentiatedinthisexperiment.


225

Figure5.4:Profileofphosphatidylcholine(PC)moleculesinHCT-8cellsfollowing24-htreatmentw

ithvehicle(RPMI-1640,DMSO

2%v/v),Bi(difl) 3(75µM),ordiflH(225µM

),calculatedasapercentageoftotalPC.Dataarepresentedasmean±s.d.(n=4

replicatesfrom

oneexperiment).Significancewithrespecttothecontrolisindicatedbylp≤0.05(BiNSAIDandNSAID),up≤0.05

(BiNSAIDonly),ornp≤0.05(NSAIDonly).Forclarityonlymoleculesthataccountedfor>0.5%

relativeabundanceofthetotalPC

moleculeshavebeenpresented.Note:Oddorether-linkedfattyacylchain(e.g.PCO-34:0/PC33:1)cannotbedifferentiatedinthis

experiment.


226

In order to identify if fatty acid composition accounts for the changes seen in

Figures 5.3 and 5.4, the PCmolecules were classified into the following categories:

those that contained fatty acidswithnodouble bonds (saturated), onedouble bond

(monounsaturated), ≥two double bonds (polyunsaturated), or an ether-linked fatty

acid. Theseweresummedwithin theaforementionedcategoriesand theresultsare

presented in Figure 5.5. It is important to note that in this context, the sum

composition of the two fatty acids have been reported and therefore saturated,

monounsaturated and polyunsaturated refers to the total number of double bonds

present. Although the identityofbothacylchainscouldnotbedeterminedwith the

methodologyusedinthepresentstudy,ithasbeenreportedthatthetypicalPCcarries

onesaturatedandoneunsaturatedchain[1].Following24-htreatmentwithBi(tolf)3

or tolfH, there was a slight, but significant decrease (Bi(tolf)3, P ≤ 0.01;

tolfH,P≤0.001) in the proportion of PC molecules containing monounsaturated or

polyunsaturated fatty acids compared to the cells treated with vehicle

alone(Fig.5.5a).Thesechangeswereconcurrentwithasignificantincrease(Bi(tolf)3,

P ≤ 0.01; tolfH, P≤ 0.001) in the proportion of PC molecules that contained

ether-linked fattyacids following24-h treatmentwithBi(tolf)3or tolfHcompared to

treatmentwith vehicle alone (Fig. 5.5a). No significant change to the proportion of

saturated PC molecules was observed (Fig. 5.5a). Additionally, no significant

differences were observed between the changes in the PC composition in

Bi(tolf)3-treatedcellswhencomparedtotolfH-treatedcells(Fig.5.5a).

Interestingly,Figure5.5bshowsthattheoppositetrendscouldbeidentifiedwhen

theBi(difl)3-anddiflH-treatedsampleswerecomparedtothevehicle-treatedsamples

wherebytherewasasmall(butsignificant)increaseinthepercentageofPCmolecules

containingmonounsaturated(Bi(difl)3,P≤0.01;diflH,P≤0.001),orpolyunsaturated

fattyacids (bothP≤0.001)compared to thecells treatedwithvehiclealone. These

observations (Fig. 5.5b) were concurrent with a significant decrease in saturated

(P≤0.001) and ether-linked PC molecules (both P≤0.001). Furthermore, while

Bi(difl)3 and diflH induced the same overall trends (i.e. concurrent increase or

decrease in relative proportion of PC in each fatty acid category, Fig. 5.5b), the

composition of PC between the Bi(difl)3 and diflH differed significantly in all cases

(unsaturated, P≤0.001; monounsaturated, P ≤ 0.01; polyunsaturated P ≤ 0.01; and

ether-linked,P≤0.001).


227

Figure 5.5: The proportion of phosphatidylcholine (PC) molecules with fatty acid chainscontainingsaturated,monounsaturated,polyunsaturatedorether-linkedfattyacidscalculatedasapercentageofthetotalPCmoleculesidentifiedinHCT-8celllipidextractstreatedfor24hwith: (a) vehicle (RPMI-1640, 2%DMSO v/v), Bi(tolf)3 (15 µM) or tolfH (45 µM); and(b)vehicle (RPMI-1640, 2%DMSOv/v), Bi(difl)3 (75 µM) or diflH (225 µM). Data arepresentedasmean±s.d.((a)n=3(vehicle)orn=4(Bi(tolf)3andtolfH)replicatesfromoneexperiment; and (b) n = 4 (vehicle, Bi(difl)3 and diflH) replicates from one experiment.Significance with respect to the control (lower marking) or corresponding NSAID (uppermarking)isindicatedby**P≤0.01,***P≤0.001;ns=notsignificant.

Figure5.6showsthecomparisonoftherelativeconcentrationofPC(normalised

totherespectivetotalPCconcentrationofthevehicle-treatedsamples)extractedfrom

the HCT-8 cells following 24-h treatment with vehicle, BiNSAIDs or NSAIDs. A

significantdecreaseinthetotalPCconcentrationwasobservedbetweentheBi(tolf)3-,

tolfH-anddiflH-treatedcellscomparedtothevehicle-treatedcells(Bi(tolf)3andtolfH,

P≤0.001;diflH,P≤0.05)(Fig.5.6).However,nosignificantdecreasewasobservedin

thetotalPCconcentrationoftheBi(difl)3-treatedcellscomparedtothevehicle-treated

cells(P>0.05,Fig.5.6b).Additionally,thetotalPCconcentrationinthetolfH-treated

cells was found to be significantly lower than that of the Bi(tolf)3- treated cells

(P≤0.05)whenthetwotreatmentswerecompared(Fig.5.6a).


228

Figure5.6:Thenormalised concentrationof phosphatidylcholine (PC)molecules containingsaturated, monounsaturated, polyunsaturated or ether-linked fatty acids in HCT-8 cell lipidextracts treated for 24hwith: (a) vehicle (RPMI-1640,DMSO2%v/v), Bi(tolf)3 (15µM) ortolfH (45 µM); and (b) vehicle (RPMI-1640, DMSO 2% v/v), Bi(difl)3 (75 µM) or diflH(225µM).ThevaluesarenormalisedtotherespectivevehicletotalPCconcentration.Dataarepresentedasmean±s.d.((a)n=3(vehicle)orn=4(Bi(tolf)3andtolfH)replicatesfromoneexperiment; and (b) n = 4 (vehicle, Bi(difl)3 and diflH) replicates from one experiment).Significance with respect to the control (lower marking) or corresponding NSAID (uppermarking)isindicatedby*P≤0.05,**P≤0.01,***P≤0.001;ns=notsignificant.

When the PC molecules were analysed by each fatty acid category (right-hand

side, Fig. 5.6) an overall decreases in the concentration of PCmolecules containing

fatty acyl chains with saturated, monounsaturated, polyunsaturated and an ether-

linked fatty acyl chains were observed in the Bi(tolf)3- and tolfH-treated cells

compared to the vehicle-treated cells(Fig. 5.6a). Notably, the diflH-treated samples

showed the same trends (Fig. 5.6b); however, the decrease was not significant

(P>0.05) in the monounsaturated and polyunsaturated fatty acyl chain categories.

TheBi(difl)3-treatedsamplesshowedadecreaseintheconcentrationofPCmolecules

containing fatty acids with saturated (P ≤ 0.01, Fig.5.5b). However, there was no


229

significantdecrease intheconcentrationofPCmoleculeswithmonounsaturatedand

ether-linkedfattyacids(P>0.05,Fig.5.5b).TheslightincreaseinconcentrationofPC

molecules with polyunsaturated fatty acids in the Bi(difl)3-treated samples was not

foundtobesignificantcomparedtothevehicle-treatedsamples(P>0.05,Fig.5.5b).

5.3.2 Phosphatidylethanolamine(PE)

Figure5.7shows theprofilesof thePEmolecules (calculatedasapercentageof

totalPE)detectedabovetheLODusingESI-MS/MSinHCT-8celllipidextractstreated

withvehicle,BiNSAIDs,orNSAIDs.Asindicatedintheleft-mostgraphsofFigure5.7,

the most abundant PE molecules identified in HCT-8 cells treated with vehicle,

BiNSAIDsorNSAIDswerePE34:1,PE36:2,PE36:1,andPE38:4. Asummaryofall

the identified PE molecules above the LOD is shown in the Appendix 4.1. The

calculated LOD and LOQ values were slightly higher for the Bi(difl)3/diflH-treated

samples compared to theBi(tolf)3/tolfH-treatedsamples. Asa result ionabundance

didnot reach theLOD thresholdandseveralmoleculesofvery lowabundancewere

omittedfromtheresultspresentedinFigure5.7b.

AnalysisofPEprofilesrevealschangestotherelativeabundanceofPEmolecules

invehicle-treatedcellscomparedtotheBi(tolf)3-andtolfH-treatedcells(Fig.5.7).The

mostprominentdifferenceisseenintheether-linkedPEmolecules,PEP-38:5/O-38:6

(m/z750) and PE P-38:4/O-38:5 (m/z 752), which were approximately twice the

abundance in the Bi(tolf)3- and tolfH-treated cell samples compared to the vehicle-

treated cell sample (P≤ 0.05). Additionally, other lower abundance ether-linked

molecules showed an increase in relative abundance of PE molecules following

treatmentwithBi(tolf)3andtolfH, includingPEP-32:0/O-32:1,PEP-36:4/O-36:5,PE

P-36:3/O-36:4, and PEP-40:5/O-40:6 (Fig. 5.7a). This was accompanied by a

significantdecrease in theproportionofdiacylglycerolPEmolecules,specificallyPE

34:2,PE36:2,PE38:5,andPE38:4.

Analysis of PE profiles also revealed changes to the relative abundance of PE

molecules between vehicle-treated cells, and the Bi(difl)3- and diflH-treated cell

samples (Fig. 5.7b). Similarly to the Bi(tolf)3/tolfH-treated series (Fig. 5.7a), a

significantincreasewasobservedintheabundanceoftheether-linkedPEmolecules,

PEP-38:5/O-38:6 (m/z 750) and PE P-38:4/O-38:5 (m/z 752) in the Bi(difl)3- and

diflH-treatedcellsamplescomparedtothevehicle-treatedcellsample(P≤0.05).


230

Figure5.7:Profileofphosphatidylethanolamine(PE)moleculesinHCT-8cellsfollowing24htreatmentwith:(a)vehicle(DMSO2%v/v),Bi(tolf)3(15µM),ortolfH(45µM);and(b)vehicle(RPMI-1640,DMSO2%v/v),Bi(difl)3(75µM),ordiflH(225µM),calculatedasapercentageoftotalPE.Dataarepresentedasthemean±s.d.((a)n=3(vehicle)orn=4(Bi(tolf)3andtolfH)replicates from one experiment; (b) n = 4 (vehicle, Bi(difl)3 and diflH) replicates from oneexperiment).Significancewithrespecttothecontrolisindicatedbylp<0.05(bothBiNSAIDandNSAID),up<0.05(BiNSAIDonly),ornp<0.05(NSAIDonly).Note:Oddorether-linkedfattyacylchains(e.g.PEO-34:0/PE33:1)cannotbedifferentiatedinthisexperiment.


231

Theproportionof fattyacyl chainunsaturationorether-linked fattyacyl chains

calculatedasapercentageofthetotalPEmoleculesidentifiedinthevehicle-,Bi(tolf)3-

andtolfH-treatedHCT-8cell lipidextracts ispresentedinFigure5.8a. Followingthe

24-htreatmentwithBi(tolf)3ortolfH,therewasaslight(butinsignificant)decreasein

the percentage of monounsaturated PE molecules, but a significant decrease (P ≤

0.001) in the percentage of polyunsaturated PE molecules compared to the cells

treated with vehicle alone (Fig. 5.8a). These changes were concurrent with a

significant increase (P ≤ 0.001) in the percentage of PEmolecules containing ether-

linkedfattyacylchainsfollowingthe24-htreatmentwithBi(tolf)3ortolfH(Fig.5.8a).

NosignificantvariationwasobservedbetweentheBi(tolf)3-ortolfH-treatedsamples.

The changes in the PE profile of the Bi(difl)3/diflH series (Fig. 5.8b) differed

slightlyfromtheBi(tolf)3/tolfHseries(Fig.5.8a)wherebyBi(difl)3anddiflHtreatment

resultedinasignificantdecrease(P≤0.001)inthepercentageofmonounsaturatedPE

molecules,butasmallerdecrease(Bi(difl)3,P≤0.001;diflH,P>0.05,notsignificant)

inthepercentageofpolyunsaturatedPEmoleculescomparedtothecellstreatedwith

vehicle alone. However, similar to the result in Fig. 5.8a, a significant increase

(P≤0.001)inthepercentageofPEmoleculescontainingether-linkedfattyacylchains

wasobservedfollowing24-htreatmentwithBi(difl)3ordiflH(Fig.5.8b).Asobserved

in Figure 5.8b, Bi(difl)3 treatment caused more pronounced changes to the overall

proportion of PE containingmonounsaturated (P ≤ 0.01) and polyunsaturated (P ≤

0.05)fattyacylchainswhencomparedtotreatmentwithdiflH.


232

Figure 5.8: The proportion of phosphatidylethanolamine (PE) molecules containingmonounsaturated,polyunsaturatedorether-linkedfattyacidscalculatedasapercentageofthetotal PE molecules identified in HCT-8 cell lipid extracts treated for 24 h with: (a) vehicle(RPMI-1640,DMSO2%v/v),Bi(tolf)3 (15µM)or tolfH (45µM); or (b) vehicle (RPMI-1640,DMSO 2% v/v), Bi(difl)3 (75µM) or diflH (225 µM). Data are presented as mean ± s.d.((a)n=3(vehicle)orn=4(Bi(tolf)3andtolfH)replicatesfromoneexperiment;and(b)n=4(vehicle,Bi(difl)3anddiflH)replicatesfromoneexperiment).Significancewithrespecttothecontrol (lowermarking)or correspondingNSAID (uppermarking) is indicatedby *P≤0.05,**P≤0.01,***P≤0.001;ns=notsignificant.


233

A comparison of the relative concentration of PE (normalised to the respective

total concentration of the vehicle-treated samples) from the HCT-8 cells following

24-htreatmentwithvehicle,BiNSAIDsorNSAIDsispresentedinFigure5.9.Following

24-h treatment, an overall decrease in the total PE concentration of the Bi(tolf)3-

(P>0.05),tolfH-(P≤0.01)anddiflH-treated(P>0.05,notsignificant)cellscompared

to thevehicle-treated cells (Fig. 5.9)wasobserved. However, anoverall increase in

the concentration of PE was observed in Bi(difl)3-treated cells compared to the

vehicle-treated cells, although this was not statistically significant (Fig. 5.9b). As

showninFigure5.8a,thetotalPEwassignificantlydecreasedbytolfHtreatmentwhen

comparedtoBi(tolf)3treatment.

WhenthePEmoleculesweregroupedaccordingtofattyacidcategory(Fig.5.9),it

was apparent that there was an overall decrease in the concentration of

monounsaturatedPEmoleculesintheBi(tolf)3-(P≤0.05),tolfH-(P≤0.01)anddiflH-

treated(P≤0.05)cellscomparedtothecellstreatedwithvehiclealone.Nodecrease

was observed following Bi(difl)3 treatment (Fig. 5.9b). As shown in Figure 5.9a,

Bi(tolf)3andtolfHtreatmentalsoresultedinasignificantdecrease(bothP≤0.001)in

the relative concentration of polyunsaturated PE molecules relative to the vehicle

treatment. TreatmentwithdiflHalsoresultedinadecrease(P>0.05)intherelative

concentration of polyunsaturated PE molecules relative to the vehicle treatment;

however, no change was observed in the Bi(difl)3-treated samples (Fig. 5.9b). As

Figure 5.9 shows, all BiNSAID and NSAID treatments resulted in an increase in the

concentrationofether-linkedfattyacidPEmoleculescomparedtothevehicle-treated

cells, however, thiswas only statistically significant in the BiNSAID-treated samples

compared to the controls (P ≤ 0.01). In both treatment sets the BiNSAID-treated

samplesshowedasignificant(P≤0.05)increaseintheconcentrationofether-linked

moleculeswhencomparedtotheNSAIDtreatmentalone(Fig.5.9aand5.9b).


234

Figure 5.9: The normalised concentration of phosphatidylethanolamine (PE) moleculescontainingmonounsaturated, polyunsaturated or ether-linked fatty acids in HCT-8 cell lipidextracts treated for 24hwith: (a) vehicle (RPMI-1640,DMSO2%v/v), Bi(tolf)3 (15µM) ortolfH(45 µM); and (b) vehicle (RPMI-1640, DMSO 2% v/v), Bi(difl)3 (75 µM) or diflH(225µM).ThevaluesarenormalisedtotherespectivetotalPEvehicleconcentration.Dataarepresentedasmean±s.d.((a)n=3(vehicle)orn=4(Bi(tolf)3andtolfH)replicatesfromoneexperiment; and (b) n=4 (vehicle, Bi(difl)3 and diflH) replicates from one experiment).Significance with respect to the control (lower marking) or corresponding NSAID (uppermarking)isindicatedby*P≤0.05,**P≤0.01,***P≤0.001;ns=notsignificant.


235

5.3.3 Phosphatidylserine(PS)

TherelativeabundancesofPSmolecules(calculatedasapercentageoftotalPS)

detectedabovetheLODarepresentedinFigure5.10.Asummaryofalltheidentified

PS molecules above the LOD can be found in Appendix 4.1. Due to the low

concentrationofPSpresentinthecellsamples,onlyasmallnumberofPSmolecules

couldbereliablydetectedabovetheLOD(Fig.5.10).ThemostabundantPSmolecules

identifiedwerePS34:1(m/z762)andPS36:1(m/z790)(Fig.5.10).Theproportion

ofthemostabundantPSmolecule,PS36:1,wasincreasedinallBiNSAID-andNSAID-

treatedsamplescomparedtothevehicle-treatedsamples(Fig.5.10),althoughthiswas

onlysignificantintheBi(tolf)3-treatedsamples(P≤0.05).Therelativeabundanceof

PS36:2 was significantly decreased in the Bi(tolf)3- and NSAID-treated

samples(P≤0.05)comparedto thevehicle. AsFigure5.10shows, thechange in the

proportionofPS34:1differedbetweenthetwotreatmentsetswherebyitsabundance

significantly increased in the Bi(tolf)3- and tolfH-treated samples compared to the

vehicle, butwas slightly (but not significantly) decreased in the Bi(difl)3- and diflH-

treated samples compared to the vehicle. Additionally, therewas somevariation in

the detection of the lower abundance molecules between the two treatment

sets(Fig.5.10).

Figure 5.10: Profile of phosphatidylserine (PS) molecules in HCT-8 cells following 24 htreatmentwith:(a) vehicle (RPMI-1640,DMSO2%v/v),Bi(tolf)3 (15µM),or tolfH(45µM);and(b)vehicle(RPMI-1640,DMSO2%v/v),Bi(difl)3(75µM),ordiflH(225µM),calculatedasapercentageof totalPS. Dataarepresentedas themean±s.d, ((a)n=3 (vehicle)orn=4(Bi(tolf)3andtolfH)replicatesfromoneexperiment;and(b)n=4(vehicle,Bi(difl)3anddiflH)replicatesfromoneexperiment).Significancewithrespecttothecontrolisindicatedbylp≤0.05(bothBiNSAIDandNSAID),up≤0.05(BiNSAIDonly),ornp≤0.05(NSAIDonly).


236

Fig5.11showsthecomparisonoftherelativeconcentrationofPS(normalisedto

therespectivetotalconcentrationofthevehicle-treatedsamples)fromtheHCT-8cells

following the 24-h treatment with vehicle, BiNSAIDs or NSAIDs. A decrease was

apparentinthetotalPSconcentrationoftheBi(tolf)3-(P≤0.05),tolfH-(P≤0.01)and

diflH-treated(P≤0.05)cellscomparedtothevehicle-treatedcells(Fig.5.11).Similar

tothePEresults(Fig.5.9),anincreaseinthetotalconcentrationofPSwasobservedin

Bi(difl)3-treatedcellscomparedtothevehicle-treatedcells(Fig.5.11b);however,this

resultwasnotstatisticallysignificant(P>0.05).

Figure5.11:Thenormalisedconcentrationofphosphatidylserine(PS)moleculesinHCT-8celllipidextractstreatedfor24hwith:(a)vehicle(RPMI-1640,DMSO2%v/v),Bi(tolf)3(15µM)or tolfH (45 µM); and (b) vehicle (RPMI-1640, DMSO 2% v/v), Bi(difl)3 (75 µM) or diflH(225µM).ThevaluesarenormalisedtotherespectivetotalPEvehicleconcentration.Dataarepresentedasmean±s.d.((a)n=3(vehicle)orn=4(Bi(tolf)3andtolfH)replicatesfromoneexperiment; and (b) n = 4 (vehicle, Bi(difl)3 and diflH) replicates from one experiments).Significance with respect to the control (lower marking) or corresponding NSAID (uppermarking)isindicatedby*P≤0.05,**P≤0.01,ns=notsignificant.

5.3.4 Summaryofresults

The overall trends observed in Sections 5.3.1-5.3.3 are presented in Tables 5.3

and5.4.InTable5.3and5.4,aredcolourindicatesadecreaseinprevalence,whilea

green colour represents an increase, andgrey indicatesno change. The intensity of

thecolourreflectsthesignificancelevel,wherebythedarkestshadeindicatesthemost

significance(P≤0.001). Table5.3showsthePCandPEprofilesof theBi(tolf)3-and


237

tolfH-treated samples exhibited the same general overall changes compared to the

vehicle-treated cells. The Bi(difl)3- and diflH-treated cells also showed the similar

trendswhen compared to eachother (Table5.3). ThePC results forBi(difl)3/diflH-

treated cells exhibited changes that were opposite to those observed for the

Bi(tolf)3/tolfH-treatedcells(Table5.3).Incontrast,thesameoverallchangeswere

Table 5.3: Summary of the overall trend in glycerophospholipid (PC and PE) profile (% oftotal) following treatment of HCT-8 cells with BiNSAIDs of NSAIDs (24 h) compared tovehicle-treatedcells(DMSO2%v/v,24h).

Bi(tolf)3,15µM

tolfH,45µM

Bi(difl)3,75µM

diflH,225µM

PC Saturated − á â â Monounsaturated â â á á Polyunsaturated â â á á Ether-linked á á â â PE Monounsaturated â â â â Polyunsaturated â â â â Ether-linked á á á á

- Nochange â Decreasedabundance,P>0.05 â DecreasedabundanceP≤0.01

â Decreasedabundance,P≤0.001 á Increasedabundance,P≤0.05

á Increasedabundance,P≤0.01 á Increasedabundance,P≤0.001

Table 5.4: Summary of the overall changes in glycerophospholipid (PC, PE or PS)concentrationfollowingtreatmentofHCT-8cellswithBiNSAIDsofNSAIDs(24h)comparedtovehicle-treatedcells(DMSO2%v/v,24h).

Bi(tolf)3,15µM

tolfH,45µM

Bi(difl)3,75µM

diflH,225µM

PC Saturated â â â â Monounsaturated â â − â Polyunsaturated â â á â Ether-linked â â â â Total â â â â PE Monounsaturated â â − â Polyunsaturated â â − â Ether-linked á á á á TotalPE â â á â PS TotalPS â â á â

- Nochange â Decreasedconcentration,P>0.05 â Decreasedconcentration,P≤0.05

â Decreasedconcentration,P≤0.01 â Decreasedconcentration,P≤0.001

á Increasedconcentration,P>0.05 á Increasedconcentration,P≤0.01


238

observed for PE composition following Bi(tolf)3/tolfH and Bi(difl)3/diflH

treatment(Table5.3).

Whenexamining theoverallchanges to therelativeconcentrationofPC,PEand

PSinducedbyBiNSAIDorNSAIDtreatmentinTable5.4,itwasobservedthatBi(tolf)3,

tolfHanddiflHexhibitedasimilaroveralltrend,wherebyanoveralldecreasein

concentration is observed in all PC and PE categories (with the exception of an

increase in PE ether-linked molecules), and total PS concentration following

treatment. As observed in Table 5.4, the Bi(difl)3-treated cell samples displayed

several differences from the three other treatments examined, whereby some

categoriesshowedanoverallincreaseornochangeinPC,PEandPSconcentration.


239

5.4 DiscussionThe overall aim of the experiments described in this Chapterwas to study the

most abundant glycerophospholipid classes in BiNSAID- and NSAID-treated HCT-8

cellsusingmethodsestablishedwithinthelaboratory[35-38].Thepurposeofthiswas

tofurtherinvestigatetheresultsobtainedinChapter4whichshowedlipidchangesin

BiNSAID/NSAID-treated cells. Specifically lipidomic analysis of the

glycerophospholipids, PC, PE and PS, using ESI-MS/MS was applied to two sets of

BiNSAID/NSAID-treatedHCT-8cell samples todetermine: (i)whetherchangescould

bedetectedintheglycerophospholipidprofilesofthedrug-treatedcellscomparedto

the vehicle-treated cells; (ii) if the relative concentrations of glycerophospholipids

were affected by BiNSAID/NSAID treatment; (iii) whether each BiNSAID and the

parent NSAID induced similar or different changes in the glycerophospholipid

composition or concentration; and (iv) whether any glycerophospholipid changes

induced by drug treatmentwere similar or differed between the twoBiNSAIDs and

correspondingNSAIDs examined. Thediscussionwill focus on these outcomeswith

respect to several key observations that stemmed from changes to the relative

concentrations of glycerophospholipids (Section5.4.1), the changes within the

glycerophospholipid saturation categories (Section 5.4.2), and the changes in the

abundance and relative concentrations of ether-linked glycerophospholipids

(Section5.4.3).

5.4.1 EffectofBiNSAIDandNSAIDtreatmentonphospholipidconcentration

The results in Section 5.3.1−5.3.3 indicated that the ESI-MS/MS technique was

sensitive enough to detect changes in the overall composition of the

glycerophospholipids selected for analysis. Additionally, the use of lipid internal

standardsallowedfortherelativequantificationofindividualPC,PEandPSmolecules.

ThetechniquewasparticularlysuccessfulforanalysingPCmolecules,whichgenerally

exhibit the highest abundance in mammalian cells [1]. The analysis of the lower

abundance molecules in the PE and PS samples approached the LOD for the

experimental conditions used. As such, a more sensitive instrument such as a

nano-ESI [39] would improve the detection of low abundance molecules if future

investigations were required. Overall, however, the ESI-MS/MS experiments

performedinthisChapterwereasuitablestartingpointtoinvestigatechangesinthe

dominantlipidsandpromisingresultshaverevealedsomeinterestingchangesinthe


240

relativeconcentrationsofthePC,PEandPSmoleculesinHCT-8cellsamplesfollowing

treatmentwithBiNSAIDsandNSAIDs.

The overall decrease in the relative concentration of glycerophospholipids

followingBi(tolf)3, tolfHanddiflH treatment (summarised inTable5.4) is consistent

with other studies wherein a decrease in the concentration of PC and PE has been

reported in a number of cell lines undergoing apoptosis [3, 40, 41]. For instance,

Anthony et al.[40] showed that drug-induced apoptosis results in inhibition of PC

synthesis in HL-60 leukaemia cells. Niebergall and Vance [41] determined that the

percentageofapoptoticcellsincreasedwhenbothmembranePCandPEdecreasedin

mutantChinesehamsterovarycellline,MT58.Finally,LiandYuan[3]reportthatthe

concentrationofPCandPEdecreased inHeLahumancervicalcancercells following

treatmentwith theanti-cancerdrugpaclitaxel.However,nochangewasobserved in

the concentration of PS in the study[3], which contrasts the results in Section

5.3.3. Theresults fromthe tolfH-treatedcells isconsistentwith theresults fromthe

SR-IRMSstudies,whichsuggestedthatfollowing24-htreatmentwithtolfHtheHCT-8

cells have lower lipid concentration than the vehicle-treated cells. The results from

the lipidomicanalysisofHCT-8cells indicate that theSR-IRMSobservationcouldbe

contributed toby loweredglycerophospholipid (PC,PEandPS) concentration in the

tolfH-treatedcells.

Interestingly,Bi(difl)3treatmentresultedinnosignificantchangestoPC,PEorPS

concentration inHCT-8cells (Table5.4). However, thediflH-treatedcellsamples(at

anequimolar treatmentconcentration),showedasignificantdecrease in therelative

concentration of PC and PS, similar to the Bi(tolf)3/tolfH-treated samples (albeit a

slightlylesssignificantdecrease,Table5.4).TheresultsobservedfromtheESI-MS/MS

analysis for the Bi(difl)3-treated cells were consistent with some of the other

experimental observations reported in this Thesis (PGE2 assays and SR-IRMS

experiments, Chapter 2 and Chapter 4, respectively), whereby Bi(difl)3 displayed a

different effect on the HCT-8 cells compared to the Bi(III) fenamate complexes,

Bi(tolf)3andBi(mef)3.ThedifferenceinthebiomolecularresponseoftheHCT-8cells

toBi(difl)3versusdiflHwasevidentintheSR-IRMSresults(Chapter4),althoughthere

issomedeviationfromtheresultsobservedfromthelipidomicanalysis(Section5.3).

Specifically,theSR-IRMSresultssuggestedthat24-htreatmentoftheHCT-8cellswith

diflHcausedanincreaseinlipids,whichwashypothesisedtooccurasaresultofthe

accumulationof lipiddroplets. Lipiddropletsarecomposedofacorecontainingthe


241

neutral lipids, triacylglycerols and cholesterol esters, surrounded by a phospholipid

monolayer (generally comprised of PC, PE, phosphatidylinositol and lysoPC, and

lysoPE)[42]. The results obtained from the lipidomic analysis (Section5.3) suggest

thatthisincreaseisnotduetoanincreaseinPC,PEorPSconcentration.Thisimplies

that instead an increase in the relative concentration of neutral lipids (such as

triacylglycerols and cholesterol esters)may be responsible for the large increase in

lipidSR-IRMSbandsobservedinthespectraof the24-hdiflH-treatedcells. Assuch,

futureworkcouldpotentiallyexpandthelipidomicanalysistoincludeneutralspecies

includingtriacylglycerols,cholesterolestersanddiacylglycerols.

5.4.2 TheeffectsofBiNSAIDandNSAIDtreatmentonfattyacylchainsaturation

WhileDNAisconsideredanimportanttargetforchemotherapeutics,researchers

have also considered the importance of the plasma membrane as a target for

anticancer therapies for anumberofdecades [43,44]. As there is little evidence to

suggest thatNSAIDs (orBiNSAIDs, Section3.3.2) interactwithDNA, othermeans of

cell death have been previously discussed, as described in Section 1.2.4. It is

importanttoconsidertheplasmamembraneasatargetforNSAIDs,sinceanumberof

NSAIDshavebeenshowntointeractwithlipids,andinparticular,phospholipids[13,

15,45]aspreviouslydiscussedinSection4.4.4.2.Oneparticularlyimportantproperty

of theplasmamembraneis fluidity.Thiscomesfromrecognisingthatmoleculesthat

modulatemembrane fluiditymay sensitise resistant tumour cells to chemotherapy-

induced cell death [46]. The level of unsaturation of fatty acids is an important

influence on the fluidity of membranes. For instance, the greater the number of

double bonds a fatty acyl molecule contains, the more physical space it occupies,

therefore influencing the fluidity of the bilayer. As such, changes in the relative

abundanceandconcentrationsofthesaturatedandunsaturatedPCandPEmolecules

werecomparedbetweenvehicle-,BiNSAID-,andNSAID-treatedcells.

Inthecurrentstudy,theBi(tolf)3-andtolfH-treatedcellsdisplayedadecreasein

the relative abundance and concentration of PC and PE molecules containing

monounsaturated and polyunsaturated fatty acids compared to the vehicle-treated

cells(Table5.3).Thisisconsistentwithotherstudies[3]thatreportedthatadecrease

inpolyunsaturatedfattyacidsprimarilycorrelatedwiththeoverallreductioninthePE

concentration in HeLa cells following paclitaxel-induced apoptosis. Li and Yuan [3]

also observed that in the PC class both polyunsaturated fatty acids and


242

monounsaturated fattyacidscontributedalmostequally to theobserveddecrease in

concentration in the HeLa cells, with a substantial decrease in saturated PCs also

observed. Themarkedreductioninthequantityofunsaturatedfattyacidssuggested

therewasapotentialdecreaseofmembranefluidityinHeLacellsfollowingpaclitaxel-

inducedapoptosis[3].

In contrast, no significant decreases were observed in the concentration of PC

moleculescontainingpolyunsaturatedfattyacidsfollowingtreatmentwithBi(difl)3or

diflH in theHCT-8 cell lipid extracts. Instead, the abundance of the unsaturatedPC

moleculeswassignificantly increased in theBi(difl)3/diflH-treatedcellscompared to

the vehicle-treated cells, which was opposite to the trend observed in the

Bi(tolf)3/tolfH-treated samples. These resultsmight suggest thatBi(difl)3 anddiflH-

treatedcells,experienceanincreaseinmembranefluidity.Ithasbeensuggestedthat

polyunsaturatedfattyacylchainsaremoresusceptibletoperoxidation,renderingcells

moresusceptibletooxidativestress-inducedcelldeath[10].Therefore,anincreasein

theabundanceofpolyunsaturatedPCandPEmoleculesstimulatedbyBi(difl)3ordiflH

treatmentmaybeanadvantageforinducingcelldeathintheHCT-8cellsastheremay

be increased potential for lipid peroxidation to occur. Future experiments could be

performed to examine whether the generation of reactive oxygen species (ROS) is

implicated in theapoptoticpathways inducedby theBiNSAIDs/NSAIDsexamined in

thisThesisandwhetherthereisanypotentialcorrelationwithlipidperoxidation.

The differences in the abundance and the saturation levels of the PC and PE

moleculesmay,inturn,resultinvariationsinfluiditybetweenthemembranesofthe

Bi(tolf)3-andBi(difl)3-treatedcells.ItisnotyetcleariftheinsertionofBi(tolf)3versus

Bi(difl)3 into thephospholipid bilayer surrounding the cellmay affect the fluidity of

thecellmembrane,orifcellulareventsrelatedtocelldeath(oracombinationofboth

factors) cause theobservedchanges. Asmentionedpreviously (Section3.4.1),work

hascommencedtoexaminetheinteractionsofBiNSAIDswithbilayermimeticsusing

neutronreflectometry(BrownandDillon[47]andunpublishedwork).Theresultsto

date show that Bi(tolf)3 interacts with the head groups of lipid bilayer membrane

mimics and partitions into the tail region of the membranes (comprised of POPC

molecules). While theseresultsweredetermined inacell-freesystem, itshowsthat

theBi(tolf)3complex isabletodisplace lipidmoleculesandpotentiallypartition into

lipid bilayers. Further studies are planned to introduce other integral membrane

components (includingcholesterolandmembrane-boundproteins) inorder tomore


243

closely mimic the cell membrane (personal communication). Additionally, it is

tempting tospeculate thatsubsequentcellmembranevariations(suchasmembrane

fluidityandtheimplicationsforactiveorpassivedruguptake)mayberesponsible(in

part)forthelargeproportionaldifferenceintheuptakeinBiobservedintheBi(tolf)3

andBi(difl)3-treatedcells(Chapter3),althoughfurtherworkisneededtodetermineif

thisisthecase.

5.4.3 TheeffectsofBiNSAIDandNSAIDtreatmentonether-linkedmolecules

One of themost interesting observations from the combined PC and PE profile

comparisonswas the effect that theBiNSAID andNSAID treatment imparted on the

relativeabundance(Fig.5.5andFig.5.8)andrelativeconcentration(Fig.5.6andFig.

5.9)ofether-linkedmolecules.Somecautionmustbetakenwheninterpretingether-

linkedmolecules due to the inability to distinguish thesemolecules from odd-chain

molecules as explained in Section 5.1.4. However, it is reported that plasmalogens

constitute ~15−20% of total phospholipids in cell membranes [32]. As such, it is

reasonable to assume that a substantial number of plasmalogens, and other ether-

linkedspecies,arepresentintheHCT-8cells.

TheresultsfromtheBiNSAID-andNSAID-treatedcellsindicateageneralincrease

intheabundanceofether-linkedPEmolecules(Table5.3),accompaniedbyanoverall

increase in the relative concentration of the ether-linked PE molecules (Table 5.4).

Additionally, it was observed that Bi(tolf)3 and tolfH caused an increase in the

abundanceofPCmoleculescontainingether-linkedfattyacidswhencomparedtothe

percent of the total PC (Fig. 5.5), although the relative concentration of all PCs

decreased (Fig. 5.6). An increase inPC andPE ether-linkedphospholipidshasbeen

observed in hematopoietic progenitor cells undergoing apoptosis following growth

factorwithdrawal, whichwas accompanied by a decrease in the relative content of

glycerophospholipids [48]. Fuchs et al. [48] speculated that an increase in ether-

linkedphospholipidsmayserveasareservoir forarachidonoyl residuesreleasedby

thebreakdownofdiacyl-glycerophospholipids,inordertoreducefurthermetabolism

ofarachidonicacidtopro-inflammatoryprostaglandins.TheBiNSAID/NSAID-treated

HCT-8cellsmaycauseaccumulationofarachidonicacidsincetheyactattheactivesite

of the COX-1/COX-2 enzymes to prevent arachidonic acid conversion to

prostaglandins. Additionally, the HCT-8 cells may upregulate arachidonic acid to

compete with the NSAIDs as tolfH and diflH are competitive/reversible COX-1/-2


244

inhibitors.Assuch,theaforementionedhypothesisproposedbyFuchsetal.[48]may

potentiallyprovideoneexplanation for the increase inether-linkedmolecules in the

BiNSAID/NSAID-treated HCT-8 cells since the cells would require a mechanism to

removeexcessarachidonicacid.

Schonfeldetal.[49]observedtheenhancedsynthesisofPEplasmalogensinPC12

cellsstimulatedwithnervegrowthfactoranddocosahexaenoicacid.Theauthors[49]

speculated that an increase in plasmalogen content was a possible reason for

enhanced resistance of nerve growth factor-differentiated cells to oxidative stress

allowing the cells to accumulate excessive intracellular Fe without undergoing cell

death. Anumberofstudies[50-54]havereportedthatNSAIDscancauseuncoupling

of oxidative phosphorylation mitochondria, an event that is synonymous with the

productionofROSproductionandoxidativestressduringapoptosis. Assuch, future

experimentsshouldbeperformedtodetermineifBiNSAID-inducedapoptosisleadsto

ROS production in HCT-8 cells in order to shed light onto the importance of ether-

linkedphospholipidproduction.

Oneofthedisadvantagesofusingconstantneutrallossof141Da(asperformed

inthepresentstudy),inordertodetermineether-linkedPEcontentinlipidmixtures,

isthatplasmalogenPEdonotundergoneutrallossofthePEheadgrouptothesame

extent as diacyl-PEmolecules [21]. This ismost likely attributed to the unique gas

phaseionchemistryimpartedbythevinylethersubstituentwhichaltersthefavorable

CID mechanism of the neutral loss of 141 Da from PE species [21]. As such, the

experimentalconditionsemployedinthepresentstudymaypotentiallyunderestimate

the presence of ether-linked PE species. Overall, the results from the ESI-MS/MS

analysisoftheHCT-8cellssuggestedthatether-linkedPCandPEmoleculesmayplay

an important role in response to BiNSAID/NSAID treatment and the implications of

thisresponseshouldbeinvestigatedfurther.

5.4.4 Overallsummaryoftheresultsandfuturedirections

The lipidomic analysis of vehicle-, BiNSAID- or NSAID-treated HCT-8 cells

suggested that, in addition to the glycerophospholipid changeswhich occurred as a

result of apoptosis, there were also additional effects resulting from each drug

treatment.Acombinationoftheseeffectshasresultedinsubstantialdifferencesofthe

PCandPEprofiles(andtoalesserextentthePSprofile)oftheBiNSAID-andNSAID-

treated cells compared to the vehicle-treated cells. The lipidomic analysis of HCT-8


245

cell lipid extracts suggests that Bi(tolf)3/tolfH treatment results in the activation of

different apoptotic responses compared to Bi(difl)3/diflH that influence the cellular

composition and relative concentrations of PC, PE andPS in theHCT-8 cells. These

results further demonstrate the difference in the biological response induced by

Bi(tolf)3 and Bi(difl)3 on the HCT-8 cells, whichwas observed in other experiments

presentedinthisThesis(reportedinChapters2-4).

Whilethereweresomeconsistenciesbetweentheobservationsobtainedfromthe

two techniques, the connection is speculative at this stage since it remains unclear

whether the glycerophospholipid changes detected using ESI-MS/MS analysis are

directlyresponsibleforthechangesdetectedintheSR-IRMSexperiments(Chapter4).

Gasper et al. were not able to correlatemass spectra and infrared bands related to

νas(CH2),νs(CH2),νas(CH3),νs(CH3)(i.e.between3025and2800cm−1)[5].Assuch,the

authors concluded that the overall global changes in lipid chains could not be

correlatedwithanyparticularlipidspeciesresolvedbymassspectrometryasalllipid

contributionsareoverlappedinthisspectralrange[5].Consequently,theresultsfrom

the ESI-MS/MS experiments performed in this study should be considered as a

complimenttothedataobtainedusingSR-IRMS(Chapter4).

While the results from the current study show that there are changes to the

profile and relative concentrations of PC, PE and PS in HCT-8 cells treated with

BiNSAIDs or NSAIDs when compared to vehicle-treated cells, it could not be

determinedwhetherthesechangesareadirectorindirecteffectofthedrugtreatment.

That is to say, it remains unknownwhetherBiNSAID orNSAID interactionwith the

identified glycerophospholipids leads to apoptosis, or whether the BiNSAID- or

NSAID-induced apoptosis results in the observed changes to the profiled

glycerophospholipids.Furtherexperiments(lipidomicandapoptosisassays)couldbe

performedatearliertimepoints(<24h)todetermineifglycerophospholipidprofile

changes are observedprior to the appearance of early apoptotic biomarkers (i.e.PS

exposure). Any futureESI-MS/MSexperimentsshouldalsoexaminetheeffectsofBi

alone and other anti-cancer agents, such as cisplatin, to draw a direct comparison

between these drugs and the BiNSAIDs/NSAIDs in HCT-8 cells. Unfortunately, the

preparation of the aforementioned samples was limited by time and resources (in

particular, the expense of phospholipid internal standards required to prepare an

increasednumberofsamples) in thepresentstudy. If thisbarriercanbeovercome,

further ESI-MS/MS experiments could also be expanded to other lipid species with


246

relativelyhighabundance inmammaliancells including: sphingomyelins, cholesterol

esters,triacylglycerols,diacylglycerols,andcardiolipins.


247

5.5 ChapterSummaryLipidomicanalysisoftreatedHCT-8cellsusingESI-MS/MSwasusedtodetermine

whetherBiNSAID/NSAIDtreatmentaffectsthecompositionofcommonphospholipids,

namelyPC,PEandPS.Specificobservationsincluded:

(i) Treatment of the HCT-8 cells with BiNSAIDs or NSAIDs significantly

altered the abundance of several individual PC, PE and PS molecules

comparedtotreatmentwithvehiclealone.

(ii) ThechangestotheabundanceoftheindividualPC,PEandPSmoleculesin

theoverallprofilesoftheBiNSAID-,andtherespectiveNSAID-treatedcells

alsodisplayedmanysimilarchangeswhencomparedtothevehicle.There

were some differences between the changes induced by Bi(tolf)3/tolfH

treatmentscomparedtotheBi(difl)3/diflHtreatments,suggestingthatthe

HCT-8 cells respondeddifferently to the fenamate compounds compared

tothesalicylatecompounds.

(iii) Changes were observed in the total concentration of PC, PE and PS

between the BiNSAID and NSAID-treated cells compared to the vehicle-

treated cells. However, when the PC and PE molecules were grouped

according to fatty acid categories, itwas observed that therewere some

differences to the overall trend (i.e. increase or decrease) between the

Bi(difl)3 and diflH-treated cells compared to the vehicle-treated cells,

whereastheBi(tolf)-andtolfH-treatedcellsinducedsimilarchanges.

(iv) The overall abundance of the ether-linked PC and PE molecules was

increasedbyallBiNSAIDandNSAIDtreatments,whichwasaccompanied

byan increase in the relative concentrationof ether-linkedPEmolecules

(butnotether-linkedPCmolecules).

Insummary,BiNSAIDandNSAIDtreatmentresultedinchangestothePC,PEand

PS composition of HCT-8 cells that was detected using ESI-MS/MS analysis. The

generalobservation fromthe lipidomicanalysiswas thatBiNSAID-andNSAIDsalter

glycerophospholipid abundance and concentration inHCT-8 cancer cells. Similar to

the results obtained from the MTT assays, flow cytometry, and PGE2 assays

(Chapter2),thepresenceofBiintheBi(tolf)3complexdidnotappeartosubstantially

alter theeffectsof tolfH,whereby theglycerophospholipid changes inducedby tolfH

wereminimally affected by the presence of Bi in theBi(tolf)3 complex. In contrast,

therewassomevariationbetweentheresultsobtainedfromBi(difl)3anddiflH-treated


248

HCT-8 cells. As such, furtherwork into the implications of phospholipid changes in

drug-inducedapoptosisshouldbeinvestigatedasthesemayhaveimplicationsforthe

mechanismofactionfortheBiNSAIDsandNSAIDs(inadditiontoCOXinhibition).


249

5.6 References[1] G. vanMeer,D. R. Voelker andG.W. Feigenson. (2008)Membrane lipids:where

theyareandhowtheybehave,Nat.Rev.Mol.CellBiol.,9,112-124.[2] K. M. Eyster. (2007) Themembrane and lipids as integral participants in signal

transduction: lipid signal transduction for the non-lipid biochemist,Adv. Physiol.Educ.,31,5-16.

[3] X.LiandY. J.Yuan.(2011)LipidomicanalysisofapoptoticHeLacells inducedbyPaclitaxel,Omics,15,655-664.

[4] M. Almada, M. R. Domingues, M. L. Dória, B. M. Fonseca, N. A. Teixeira andG.Correia-da-Silva. (2015)Lipidomicapproach towardsdecipheringanandamideeffectsinratdecidualcell,J.CellPhysiol.,230,1549-1557.

[5] R. Gasper, G. Vandenbussche and E. Goormaghtigh. (2011) Ouabain-inducedmodifications of prostate cancer cell lipidome investigated with massspectrometryandFTIRspectroscopy,Biochim.Biophys.Acta,1808,597-605.

[6] G.Kroemer,L.Galluzzi,P.Vandenabeele,J.Abrams,E.S.Alnemri,E.H.Baehrecke,M. V. Blagosklonny, W. S. El-Deiry, P. Golstein, D. R. Green, M. Hengartner,R.A.Knight, S.Kumar, S.A. Lipton,W.Malorni,G.Nunez,M.E. Peter, J. Tschopp,J.Yuan,M.Piacentini,B.Zhivotovsky,G.MelinoandNomenclatureCommitteeonCell Death. (2009) Classification of cell death: recommendations of theNomenclatureCommitteeonCellDeath2009,CellDeathDiffer.,16,3-11.

[7] M.L.Dória,C.Z.Cotrim,C.Simões,B.Macedo,P.Domingues,M.R.DominguesandL.A.Helguero.(2013)Lipidomicanalysisofphospholipidsfromhumanmammaryepithelialandbreastcancercelllines,J.CellPhysiol.,228,457-468.

[8] E.Marien,M.Meister,T.Muley,S.Fieuws,S.Bordel,R.Derua,J.Spraggins,R.VandePlas,J.Dehairs,J.Wouters,M.Bagadi,H.Dienemann,M.Thomas,P.A.Schnabel,R.M.Caprioli,E.WaelkensandJ.V.Swinnen.(2015)Non-smallcelllungcancerischaracterized by dramatic changes in phospholipid profiles, Int. J. Cancer, 137,1539-1548.

[9] K.A.Veselkov,R.Mirnezami,N.Strittmatter,R.D.Goldin,J.Kinross,A.V.M.Speller,T.Abramov,E.A. Jones,A.Darzi,E.Holmes, J.K.NicholsonandZ.Takats. (2014)Chemo-informatic strategy for imaging mass spectrometry-based hyperspectralprofiling of lipid signatures in colorectal cancer,Proc.Natl. Acad. Sci. U.S.A., 111,1216-1221.

[10] E.Rysman,K.Brusselmans,K.Scheys,L.Timmermans,R.Derua,S.Munck,P.P.VanVeldhoven,D.Waltregny,V.W.Daniels,J.Machiels,F.Vanderhoydonc,K.Smans,E.Waelkens, G. Verhoeven and J. V. Swinnen. (2010)De novo lipogenesis protectscancer cells from free radicals and chemotherapeutics by promotingmembranelipidsaturation,CancerRes.,70,8117-8126.

[11] S. Beloribi-Djefaflia, S. Vasseur and F. Guillaumond. (2016) Lipid metabolicreprogrammingincancercells,Oncogenesis,5,e189.

[12] J.Balsinde,M.V.WinsteadandE.A.Dennis.(2002)PhospholipaseA2regulationofarachidonicacidmobilization,FEBSLett.,531,2-6.

[13] L. M. Lichtenberger, Y. Zhou, V. Jayaraman, J. R. Doyen, R. G. O'Neil, E. J. Dial,D.E.Volk,D.G.Gorenstein,M.B.BoggaraandR.Krishnamoorti.(2012)InsightintoNSAID-induced membrane alterations, pathogenesis and therapeutics:characterization of interaction of NSAIDs with phosphatidylcholine, Biochim.Biophys.Acta,1821,994-1002.

[14] A.Sade,S.Tunçay,İ.Çimen,F.SevercanandS.Banerjee.(2012)CelecoxibreducesfluidityanddecreasesmetastaticpotentialofcoloncancercelllinesirrespectiveofCOX-2expression,Biosci.Rep.,32,35-44.

[15] M.Suwalsky,M.Manrique-Moreno,J.Howe,K.BrandenburgandF.Villena.(2011)Molecular interactionsofmefenamic acidwith lipidbilayers and redblood cells,J.Braz.Chem.Soc.,22,2243-2249.


250

[16] M. A. Barrett, S. Zheng, G. Roshankar, R. J. Alsop, R. K. R. Belanger, C. Huynh,N.Kučerka and M. C. Rheinstädter. (2012) Interaction of aspirin (acetylsalicylicacid)withlipidmembranes,PLoSOne,7,e34357.

[17] B. Brügger. (2014) Lipidomics: analysis of the lipid composition of cells andsubcellular organelles by electrospray ionization mass spectrometry, Ann. Rev.Biochem.,83,79-98.

[18] M. Pulfer and R. C. Murphy. (2003) Electrospray mass spectrometry ofphospholipids,MassSpectrom.Rev.,22,332-364.

[19] X.HanandR.W.Gross. (2005)Shotgun lipidomics: electrospray ionizationmassspectrometricanalysisandquantitationof cellular lipidomesdirectly fromcrudeextractsofbiologicalsamples,MassSpectrom.Rev.,24,367-412.

[20] G. Hart-Smith and S. J. Blanksby. (2011)Mass analysis, InMass Spectrometry inPolymerChemistry,pp5-32,Wiley-VCHVerlagGmbH&Co.KGaA.

[21] K.A.ZemskiBerryandR.C.Murphy.(2004)Electrosprayionizationtandemmassspectrometry of glycerophosphoethanolamine plasmalogen phospholipids, J. Am.Soc.MassSpectrom.,15,1499-1508.

[22] E. Fahy, S. Subramaniam, R. C. Murphy, M. Nishijima, C. R. Raetz, T. Shimizu,F.Spener,G.vanMeer,M.J.WakelamandE.A.Dennis.(2009)UpdateoftheLIPIDMAPScomprehensiveclassificationsystemforlipids,J.LipidRes.,50,S9-14.

[23] B. Jenkins, J. A. West and A. Koulman. (2015) A review of odd-chain fatty acidmetabolism and the role of pentadecanoic acid (C15:0) and heptadecanoic acid(C17:0)inhealthanddisease,Molecules,20,2425-2444.

[24] J. E. Vance. (2015) Phospholipid synthesis and transport in mammalian cells,Traffic,16,1-18.

[25] E. van den Brink-van der Laan, J. A. Killian and B. de Kruijff. (2004) Nonbilayerlipidsaffectperipheralandintegralmembraneproteinsviachangesinthelateralpressureprofile,Biochim.Biophys.Acta,1666,275-288.

[26] G.vanMeerandA.I.P.M.deKroon.(2011)Lipidmapofthemammaliancell,J.CellSci.,124,5-8.

[27] K. Emoto,N. Toyama-Sorimachi,H. Karasuyama,K. Inoue andM.Umeda. (1997)Exposureofphosphatidylethanolamineonthesurfaceofapoptoticcells,Exp.CellRes.,232,430-434.

[28] A. Marconescu and P. E. Thorpe. (2008) Coincident exposure ofphosphatidylethanolamineandanionicphospholipidsonthesurfaceofirradiatedcells,Biochim.Biophys.Acta1778,2217-2224.

[29] N. Maulik, V. E. Kagan, V. A. Tyurin and D. K. Das. (1998) Redistribution ofphosphatidylethanolamine andphosphatidylserine precedes reperfusion-inducedapoptosis,Am.J.Physiol.,274,H242-248.

[30] C. D. Funk. (2001) Prostaglandins and leukotrienes: advances in eicosanoidbiology,Science,294,1871-1875.

[31] J. Leßig and B. Fuchs. (2009) Plasmalogens in biological systems: their role inoxidative processes in biological membranes, their contribution to pathologicalprocessesandagingandplasmalogenanalysis,Curr.Med.Chem.,16,2021-2041.

[32] N.E.BravermanandA.B.Moser.(2012)Functionsofplasmalogenlipidsinhealthanddisease,Biochim.Biophys.ActaMol.BasisDis.,1822,1442-1452.

[33] P.Brites,H.R.WaterhamandR.J.A.Wanders.(2004)Functionsandbiosynthesisofplasmalogensinhealthanddisease,Biochim.Biophys.ActaMol.CellBiol.Lipids,1636,219-231.

[34] J. E. Vance. (2008) Phosphatidylserine and phosphatidylethanolamine inmammaliancells: twometabolically relatedaminophospholipids, J. LipidRes., 49,1377-1387.

[35] J. M. Deeley, T. W. Mitchell, X. Wei, J. Korth, J. R. Nealon, S. J. Blanksby andR.J.Truscott. (2008)Human lens lipidsdiffermarkedly from thoseof commonlyusedexperimentalanimals,Biochim.Biophys.Acta,1781,288-298.


251

[36] S.E.Norris,M.G.Friedrich,T.W.Mitchell,R. J.W.TruscottandP.L.Else.(2015)Humanprefrontalcortexphospholipidscontainingdocosahexaenoicacidincreaseduring normal adult aging, whereas those containing arachidonic acid decrease,Neurobiol.Aging,36,1659-1669.

[37] J. T. Saville, Z. Zhao, M. D. Willcox, S. J. Blanksby and T. W. Mitchell. (2010)Detectionandquantificationoftearphospholipidsandcholesterolincontactlensdeposits: the effect of contact lens material and lens care solution, Invest.Ophthalmol.Vis.Sci.,51,2843-2851.

[38] J.R.Hughes, J.M.Deeley,S. J.Blanksby,F.Leisch,S.R.Ellis,R. J.W.TruscottandT.W. Mitchell. (2012) Instability of the cellular lipidome with age, Age, 34,935-947.

[39] B. Brügger, G. Erben, R. Sandhoff, F. T. Wieland and W. D. Lehmann. (1997)Quantitative analysis of biologicalmembrane lipids at the lowpicomole level bynano-electrospray ionization tandem mass spectrometry, Proc. Natl. Acad. Sci.U.S.A.,94,2339-2344.

[40] M.L.Anthony,M.ZhaoandK.M.Brindle.(1999)Inhibitionofphosphatidylcholinebiosynthesis following induction of apoptosis in HL-60 cells, J. Biol. Chem., 274,19686-19692.

[41] L. J. Niebergall and D. E. Vance. (2012) The ratio of phosphatidylcholine tophosphatidylethanolamine does not predict integrity of growing MT58 Chinesehamsterovarycells,Biochim.Biophys.Acta,1821,324-334.

[42] T.FujimotoandR.G.Parton.(2011)Notjustfat:Thestructureandfunctionofthelipiddroplet,ColdSpringHarb.Perspect.Biol.,3,a004838.

[43] G.AranciaandG.Donelli. (1991)Cellmembranesastarget foranticanceragents,Pharmacol.Res.,24,205-217.

[44] T.R.TrittonandJ.A.Hickman.(1990)Howtokillcancercells:membranesandcellsignalingastargetsincancerchemotherapy,CancerCells,2,95-105.

[45] M.N.Giraud,C.Motta,J.J.Romero,G.BommelaerandL.M.Lichtenberger.(1999)Interaction of indomethacin and naproxen with gastric surface-activephospholipids: a possible mechanism for the gastric toxicity of nonsteroidalanti-inflammatorydrugs(NSAIDs),Biochem.Pharmacol.,57,247-254.

[46] S.Baritaki,S.Apostolakis,P.Kanellou,M.T.Dimanche-Boitrel,D.A.SpandidosandB.Bonavida.(2007)Reversaloftumorresistancetoapoptoticstimulibyalterationofmembranefluidity:therapeuticimplications,Adv.CancerRes.,98,149-190.

[47] T.Brown.(2016)Bismuth-NSAIDsascolorectalcancerchemopreventives:Insightsinto theirmembrane interactionsanduptakemechanisms,BachelorofMedicinalChemistry(Advanced)withHonours,UniversityofWollongong,Australia.

[48] B. Fuchs, J. Schiller andM. A. Cross. (2007) Apoptosis-associated changes in theglycerophospholipidcompositionofhematopoieticprogenitorcellsmonitoredby31P NMR spectroscopy and MALDI-TOF mass spectrometry, Chem. Phys. Lipids,150,229-238.

[49] E. Schonfeld, I. Yasharel, E. Yavin and A. Brand. (2007) Docosahexaenoic acidenhances ironuptakebymodulating iron transporters and accelerates apoptoticdeathinPC12cells,Neurochem.Res.,32,1673-1684.

[50] Y. Masubuchi, S. Yamada and T. Horie. (1999) Diphenylamine as an importantstructure of nonsteroidal anti-inflammatory drugs to uncouple mitochondrialoxidativephosphorylation†,Biochem.Pharmacol.,58,861-865.

[51] Y.Masubuchi, S.YamadaandT.Horie. (2000)Possiblemechanismofhepatocyteinjury induced by diphenylamine and its structurally related nonsteroidal anti-inflammatorydrugs,J.Pharmacol.Exp.Ther.,292,982-987.

[52] S.Somasundaram,G.Sigthorsson,R.J.Simpson,J.Watts,M.Jacob,I.A.Tavares,S.Rafi,A.Roseth,R.Foster,A.B.Price,J.M.WrigglesworthandI.Bjarnason.(2000)Uncouplingofintestinalmitochondrialoxidativephosphorylationandinhibitionof


252

cyclooxygenasearerequiredforthedevelopmentofNSAID-enteropathyintherat,Aliment.Pharmacol.Ther.,14,639-650.

[53] G.Galati,S.Tafazoli,O.Sabzevari,T.S.ChanandP.J.O'Brien.(2002)IdiosyncraticNSAIDdruginducedoxidativestress,Chem.Biol.Interact.,142,25-41.

[54] Y. Li, X.M.Qi, X. Xue, X. F.Wu, Y. F.Wu,M. Chen, G. Z. Xing, Y. Luan and J. Ren.(2009) The relationship between diphenylamine structure and NSAIDs-inducedhepatocytesinjury,Toxicol.Lett.,186,111-114.

253

Chapter6. ExplorationoftheinvitrotoxicityofnewBi(III)

aminoarenesulfonate,indole-carboxylate,and

hydroxamatecomplexes

Following promising results (Chapter 2) that showed that a selection of BiNSAIDs

exhibitedtoxicitytowardstheHCT-8coloncancercells,otherBicomplexeswerealso

investigated for prospective anti-cancer activity against this cell line. Several non-

NSAIDBi(III)complexes,whichincludedhomoleptictris-Bi(III)aminoarenesulfonates,

indole-carboxylates,andhydroxamates,wereselectedforthisinvitroscreeningbased

on the fact that the ligandswere derivatives of biologically active compounds. The

purpose of the work described in this Chapter was to identify new lead Bi(III)

complexes to provide direction for future studies (such as those described in

Chapters2−5).

Chapter6.Anti-canceractivityofBi(III)complexes

254

6.1 IntroductionAs previously discussed in Section 1.4.1.4, several classes of Bi(III) complexes

haveshownbothinvitroandinvivoanti-cancerpotential.ThefocusinChapters2−5

has been on the anti-cancer potential of Bi complexes of NSAIDs based on the

hypothesisthatCOX-2over-expressionisassociatedwithcancerandthatNSAIDsare

strong inhibitors of COX. It is evident from the literature (Chapter 1) that Bi(III)

complexesofother ligandsalsodisplayanti-cancerpotential invitro. Todate, there

have been no investigations of homoleptic Bi(III) complexes of aminoarenesulfonic

acids (Section 6.1.1) [1], indole-carboxylic acids (Section 6.1.2) [2], and hydroxamic

acids(Section6.1.3)[3,4].WhilemostoftheseligandsarenotknowntoinhibitCOX,

they have been shown to be biologically active, as discussed in Sections 6.1.1−6.1.3.

Assuch,theBi(III)complexesoftheseligandswarrantinvestigation.

6.1.1 SulfonicacidsandBi(III)aminoarenesulfonates

A number of simple sulfonic acid (general formula RSO3H) molecules play

important roles in biological systems. For instance, the glycine analogue,

aminomethanesulfonic acid (Fig. 6.1), has demonstrated similar hepatoprotective

effects as glycine in isolated Kuppfer (liver) cells [5]. The aspartate analogue,

L-cysteicacid (Fig. 6.1), is a potent and effective agonist towards several rat

metabotropicglutamatereceptors[6]. Taurine(Fig.6.1), isnaturallypresent inhigh

levels in the brain and is believed to play a role in neurological processes including

osmoregulation, antioxidant activity, neuromodulation, and the control of calcium

influx[7]. Inaddition,experimentalevidencesuggeststhattaurineandthesynthetic

analogue, homotaurine (Fig. 6.1), reduce protein aggregation of amyloid-β proteins

involved in the formation of plaques in Alzheimer’s disease[8, 9]. Recently, it has

been shown that a synthetic sulfonic acid, anthraquinone-2-sulfonic acid (Fig.6.1)

confers protection in primary rat cortical neurons exposed to hydrogenperoxide or staurosporine;asaresultthissulfonicacidmaypresentanoveltherapeuticagentfor

neurologicalprotection[10].

Ofrelevancetothisworkarereportsoftheanti-cancerpotentialofanumberof

sulfonicacidderivatives,whereby itwas shown that, in some instances, the sulfonic

acid derivative showed improved activity compared to the original drug [11, 12].

Forinstance,indirubin-5-sulfonicacid(Fig.6.1),aderivativeoftheanti-leukemiadrug

indirubin,wasfoundtobeamorepotentinhibitorofcyclin-dependentkinaseswhen


255

compared to indirubin [11].However, indirubin-5-sulfonic acid showed limited

activitywhentestedagainsttheT-lymphoblasticJurkatcellline,whichwasbelievedto

bearesultofpoorcelluptakeofthecompound[11].Inanotherstudy,thevitaminE

analogue,RRR-α-tocopheryloxybutylsulfonicacid(Fig.6.1),wasdesignedtoprevent

hydrolysisbygastricesterases,aproblemwhichisbelievedtocompromisetheinvivo

anti-tumouractivityofα-vitaminEsuccinate[12].Theanalogueshowedsignificantly

reduced tumour volumes compared to α-vitaminE succinate following the 6-week

treatmentofPC3prostatecancercellxenograftednudemice[12].Theserumlevelsof

RRR-α-tocopheryloxybutylsulfonicacid-andα-vitaminEsuccinate-treatedmicewere

determined tobe15μMand2μMafteronemonthofdosing, respectively,which is

consistent with the improved bioavailability and anti-tumour activity of the

analogue[12].

Figure6.1:Thechemicalstructuresofbiologicallyactivesulfonicacids.


256

Importantly, a wide range of sulfonic acids have been shown to be relatively

benign inmammals; for example, LD50 doses in excess of 1000mg/kg bodyweight

(often exceeding 5000mg/kg bodyweight) are required to induce acute toxicity in

rats dosed orally, regardless of the sulfonic acid tested [13]. In addition, a large

number of sulfonic acids possess little to no mutagenic and carcinogenic activity

towardsmammaliancells[13].

TheAndrewsgrouphaveinvestigatedsulfonatesasligandsforBi(III)complexes

in thedevelopmentof anti-bacterialdrugswith improvedpotencyover traditionally

used Bi compounds (BSS, CBS and RBC) for the treatment of H. pylori

infections[1,14-16].Tothisend,severalhomoleptictris-Bi(III)aminoarenesulfonate

complexes of the form [Bi(O3S-RN)3] (where O3S-RN represents the deprotonated

ligands of: ortho-aminobenzenesulfonic acid (oAB-SO3H, Fig. 6.2); meta-

aminobenzenesulfonic acid; para-aminobenzenesulfonic acid (pAB-SO3H, Fig.6.2);

6-amino-3-methoxybenzenesulfonicacid;2-pyridinesulfonicacid;4-amino-3-hydroxy-

1-naphthalenesulfonic acid (4A3HN-SO3H, Fig.6.2); 2-amino-1-naphthalensulfonic

acid(2AN-SO3H,Fig.6.2);5-amino-1-naphthalensulfonicacid(5AN-SO3H,Fig.6.2);or

5-isoquinolinesulfonicacid)weresynthesisedandtestedforinvitroanti-bacterial

Figure6.2:ThechemicalstructuresoftheaminoarenesulfonicacidsusedtosynthesisthefiveBi(III) aminoarenesulfonate complexes, (Bi(oAB-SO3)3, Bi(pAB-SO3)3, Bi(4A3HN-SO3)3,Bi(2AN-SO3)3,andBi(5AN-SO3)3),investigatedinthepresentstudy.


257

activityagainstH.pylori [1]. Importantly, theassociatedpooraqueoussolubilitydid

not appear to affect the bactericidal properties of the Bi(III) aminoarenesulfonate

complexes against three standard laboratory strains of H. pylori [1].The Bi(III)

aminoarenesulfonatecomplexesshowedincreasedanti-bacterialactivitycomparedto

BSS,CBS,RBC,andtheuncomplexedaminoarenesulfonicacids[1].

Inthepresentstudy,fiveBi(III)aminoarenesulfonatecomplexeswerechosenfor

invitroanti-cancerscreeningbasedontheireaseofsolubilityinthevehicleusedinthe

MTT assays (Chapter 2). The complexes, (Bi(oAB-SO3)3, Bi(pAB-SO3)3,

Bi(4A3HN-SO3)3, Bi(2AN-SO3)3, and Bi(5AN-SO3)3), and the corresponding

uncomplexedaminoarenesulfonicacids,wereselectedforscreening.

6.1.2 IndolesandBi(III)indole-carboxylates

Indoles are heterocyclic compounds that consist of a fused six-membered

andfive-membered nitrogen-containing pyrrole ring. Naturally occurring indole

molecules that confer biological activity include tryptophan (Fig. 6.3), which is the

precursorof theneurotransmitter serotonin. Indole-3-lactic acid (ILA-H, Fig. 6.3) is

oneexampleof anumberof tryptophanmetabolites found inhumanplasma, serum

andurine[17,18].Adiverserangeofdrugsexistthatincorporateanindolemoietyin

theirstructures(reviewedin[19]). Indole-containingmoleculeshaveawidevariety

of medicinal uses, including treatment of asthma, cancer, depression, HIV infection,

hypertension,microbialinfection,andinflammation(reviewedin[19]).Ofparticular

interest to the work performed in this Thesis are the anti-cancer and

anti-inflammatoryeffectsof indole-basedcompounds. For instance,somecommonly

used NSAIDs (e.g. indoH, Fig. 1.2) contain an indole moiety. While several high

molecularweight indole-containingmoleculesare currentlymarketedasanti-cancer

drugs (e.g. Vincristine, Fig. 6.3) [19], a relatively simple, low molecular weight

compound, indole-3-carbinol (Fig.6.3), has been extensively studied as a potential

anti-canceragentagainstseveralcancercelltypes,namelybreast,colon,endometrium

andprostate(reviewedin[20]).

Pathaketal.[2]haverecentlysynthesisedandevaluatedtheinvitroanti-bacterial

and anti-parasitic activity (against H. pylori and Leishmania promastigotes,

respectively) of several homolepticBi(III) complexesderived from indole-carboxylic

acids, namely Bi(2-(1H-indol-3-yl)acetate)3, Bi(3-(1H-indol-3-yl)propanoate)3,

Bi(4-(1H-indol-3-yl)butanoate)3, Bi(1-methyl-1H-indole-3-carboxylate)3, or Bi(2-(1H-


258

indol-3-yl)-2-oxoacetate)3 (referred to herein as Bi(IGA)3, where IGA-H = indole-3-

glyoxylicacid(Fig.6.3)).

Inthepresentstudy,twoBi(III)indole-carboxylates,Bi(IGA)3andBi(ILA)3,willbe

assessedforanti-canceractivityagainsttheHCT-8cell linesincetheirpotentialanti-

cancerpropertieshadnotbeenassessedpreviously.

Figure 6.3: The chemical structures of biologically active indoles, and the indole-carboxylicacidsused to synthesise the twoBi(III) indole-carboxylate complexes,Bi(ILA)3andBi(IGA)3,investigatedinthisstudy.

6.1.3 HydroxamicacidsandBi(III)hydroxamates

Hydroxamicacids(generalformulaRCONR´OH)areweakacidsthatrepresentan

importantfamilyoforganicbioligands[21].Hydroxamicacidsplayaroleinavariety


259

ofbiologicalprocesses,whichislargelyduetotheabilityofhydroxamicacidstoform

stable chelates with metals such as Fe(III) [21, 22]. One of the first recognised

biologicalrolesofhydroxamicacidswastheiruseassiderophoresinmicrobialFe(III)

sequestration and transport [21]. Hydroxamic acids are used in the treatment of

bacterial infections [23], thalassemia [24], and inflammation [25]. Additionally,

hydroxamic acid molecules have been investigated as potential treatments for a

varietyofailmentsincludingcancer[26],malaria[27],andAlzheimer’sdisease[28].

The structurally simple hydroxamic acid, acetohydroxamic acid (Fig.6.4), is a

potent inhibitor of nickel-dependent ureases [29, 30]. In 1983, the FDA approved

acetohydroxamic acid for the treatment of urinary tract infections [23].

Benzohydroxamic acid (BHA-H,Fig. 6.4)has alsobeen shown to inhibit urease [31].

The similarly structured salicylhydroxamic acid (SHA-H, Fig.6.4) is employed in the

treatment of urinary tract infections and urolithiasis [32]. SHA-H has also shown

potentialinthetreatmentofCandidaalbicansinfection[32,33].

Importantly hydroxamic acids also inhibit enzymes including Zn-dependent

MMPs[34,35],histonedeacetylases,andCOX[36-38],whichhassparkedinterest in

their use as potential anti-cancer drugs. The hydroxamic acid-based drugs,

Batimastat(Fig. 6.4) and Marimastat (Fig.6.4), are MMP inhibitors that have been

investigated in Phase III clinical trials for the treatment of cancer [39, 40].

Unfortunately,neitherdrughasprogressedbeyondPhaseIIItrialsduetopoorefficacy

and adverse side effects [40]. Vorinostat (Merck & Co., Inc., Fig. 6.4) is an orally

delivered, potent inhibitor of histonedeacetylase thatwas approvedby the FDA for

the treatment of cutaneous T-cell lymphoma in 2006 [41, 42]. More recently, the

histone deacetylase inhibitor, Belinostat (Spectrum Pharmaceuticals, Fig. 6.4) was

approved by the FDA for the treatment of peripheral T-cell lymphoma [43].

Thedevelopment of diverse new ranges of hydroxamic acid-based molecules that

inhibithistonedeacetylasescanbefoundinarecentreview[44].

A number of hydroxamic acids inhibit the activity of COX [36-38], an enzyme

implicated in the formation and progression of a number of types of cancer (as

discussed in Section 1.2.3). Notably, some hydroxamic acid molecules are able to

interactwithboththePOXandCOXactivesites;forexample,tepoxalin(Fig.6.4)


260

Figure6.4:Thechemicalstructuresofaselectionofbiologicallyactivehydroxamicacids,andthehydroxamicacidsused to synthesise the threeBi(III)hydroxamatecomplexes,Bi(BHA)3,Bi(SHA)3,andBi(O-AcSHA)3,investigatedinthisstudy.

isadual-inhibitoroftheCOXandPOXactivesites[36],andhasalsobeenreportedto

inhibitlipoxygenase[25].Thereisaninterestindevelopinghydroxamicacidsasnew

generationCOXinhibitorssincehydroxamicacidsmaypotentiallyprovidepotentCOX

inhibition through dual COX and POX active site inhibitionwhile also reducing side


261

effectssuchastopicalirritation,partiallyduetothefacthydroxamicacidsareweaker

acids than the analogous carboxylic acids [21]. Twohydroxamic acids developed

specifically to target COX, O-acetylsalicylhydroxamic acid (O-AcSHA-H, Fig.6.4) and

triacetylsalicylhydroxamicacid(Fig.6.4),exhibitedirreversible,non-selectivebinding

tobothCOX-1andCOX-2withcomparableor improvedpotencytoaspH[37,38]. It

has been shown that the mechanism of COX inhibition by O-AcSHA-H and

triacetylsalicylhydroxamic acid is the same as that exhibited by aspH whereby

Ser530(COX-1) is irreversibly acetylated [37, 38]. Given this evidence, it may be

possible that hydroxamic acid inhibitors of COX may confer potential anti-cancer

activity,inamannersimilartotraditionalNSAIDs,particularlyagainstcancercelllines

thatexpressupregulatedlevelsofCOX-2.

Due to the ability of several hydroxamic acids (e.g. acetohydroxamic acid,

BHA-Hand SHA-H) to inhibit bacterial urease and their use as anti-bacterial agents,

Pathaketal.[3,4]havesynthesisedseveralBi(III)hydroxamatecomplexesunderthe

premise that theymayexhibit improvedanti-bacterial activity againstH.pylori than

currentlyusedBi(III)compounds,BSS,CBSandRBC. Additionally, thetoxicityofthe

Bi(III) hydroxamates has been assessed against human fibroblasts [4]; however,

currently no studies have investigated the anti-cancer activity of these complexes

against human cancer cell lines. Three Bi(III) hydroxamate complexes, Bi(BHA)3,

Bi(SHA)3, andBi(O-AcSHA)3,werechosen forassessment in thepresent studybased

ontheireaseofsolubilityinthevehicleusedintheMTTassays(Chapter2).

6.1.4 Scopeofthischapter

FollowingresultsthatshowedthatBiNSAIDsexhibitedinvitrotoxicitytowardthe

HCT-8coloncancercellline(Chapter2)itwasofinteresttoexpandtherangeofBi(III)

compoundsand identifyotherpotential leadcompounds. Assuch,Bi(III)complexes

derivedfrombiologicallyrelevantligandswerestudiedwiththespecificaimsof:

1. Assessing the in vitro toxicity of Bi(III) aminoarenesulfonate complexes,

Bi(III) indole-carboxylatecomplexes,andBi(III)hydroxamatecomplexes,

and

2. Comparing the toxicity of the above-mentioned Bi complexes to the

uncomplexed aminoarenesulfonic acids, indole-carboxylic acids and

hydroxamicacidsusingthehumancoloncancercellline,HCT-8.


262


The Bi(III) aminoarenesulfonate complexes, Bi(oAB-SO3)3, Bi(pAB-SO3)3,

Bi(4A3HN-SO3)3,Bi(2AN-SO3)3,andBi(5AN-SO3)3,weresynthesisedandcharacterised

by Dr Madleen Busse (Monash University, Australia) as previously described[1].

TheBi(III) indole-carboxylate complexes, Bi(IGA)3, and Bi(ILA)3, and the Bi(III)

hydroxamatecomplexes,Bi(BHA)3,Bi(SHA)3,andBi(O-AcSHA)3,weresynthesisedand

characterised by Dr Amita Pathak (Monash University, Australia) as previously

described [2-4]. All aminoarenesulfonic acids (oAB-SO3H, pAB-SO3H, 4A3HN-SO3H,

2AN-SO3H, and 5AN-SO3H), indole-carboxylic acids (IGA-H and ILA-H), and

hydroxamic acids (BHA-H, SHA-H, and O-AcSHA) were supplied by ProfessorPhilip

Andrews (MonashUniversity, Australia) andwere originally purchased fromSigma-

Aldrich(StLouis, USA) or TCI America (Portland, USA) and used without further

purification as previously described [1-4]. The materials used for the cell culture

experimentsdescribedinthisChapterhavebeenpreviouslydescribedinSection2.2.1.

6.2.2 CellLines


6.2.3 MTTAssay

TheMTTassaywasperformedaspreviouslydescribedinSection2.2.4.


263

6.3 ResultsThe MTT assay was used to evaluate the effects of the Bi(III) complexes and

ligandson cell viability inHCT-8 cells following24h exposure. Figure6.5a-d shows

the representative concentration-response curves, graphed for the concentration of

each ligand (acknowledging that three mol of ligand are present per mol of Bi(III)

complex). Forclarity,theMTTresultsforthelessactivecompoundsareprovidedin

Appendix5.1.TheIC50valuesdeterminedafter24-hexposuretotheBi(III)complexes

andligandsaresummarisedinTable6.1.

The Bi(III) aminoarenesulfonates and corresponding aminoarenesulfonic acids

demonstrated little to no activity towards HCT-8 cells (Appendix 5.1), with the

exception of Bi(4A3HN-SO3)3 and 4A3HN-SO3H (Fig. 6.5a). The complexation of

4A3HN-SO3 to Bi(III) increased the potency of the free 4A3HN-SO3H towards the

HCT-8cellsbyapproximately11times(Table6.1).

TheBi(III) indolates, Bi(IGA)3 andBi(ILA)3, exhibited low IC50 values (Fig. 6.5b,

Table6.1)of10µMand17µM,respectively. AsevidentinFigure6.5b,however,the

ligands exhibited low toxicity. For instance, IGA-H demonstrated no significant

reduction in MTT conversion up to 100µM (Fig. 6.5b), and while this decreased

between 100-1000 µMwith a small increase between 1000 µM to 2000 µM, IGA-H

failed to achieve greater than 50% inhibition of MTT conversion (Fig. 6.5b).

Comparatively, ILA-H demonstrated consistently low toxicity up to the highest

concentrationassessed,2000µM(Fig.6.5b).Inbothinstances,itisapparentthatthe

Bi(III)indole-carboxylatesexhibitedgreatertoxicitythanpredictedfortheligands.

AssummarisedinTable6.1,theIC50valuesobtainedfortheBi(III)hydroxamate

complexes, Bi(BHA)3 and Bi(SHA)3, were approximately 10 and 15 times higher,

respectively,thanBi(IGA)3.However,Bi(BHA)3demonstratedthelowestIC50valueof

the threeBi(III)hydroxamates tested,andnearly5.5 timesgreateractivity than free

BHA-H (Table 6.1). Additionally, Bi(SHA)3 demonstrated greater toxicity than free

SHA-H, whereby the latter did not inhibit MTT conversion below 50% over the

concentration range tested (up to 2000 µM, Fig. 6.5d). As observed in Figure6.5d,

Bi(O-AcSHA)3inducedsomereductionincellviabilitycomparedtothevehicle-treated

cells,but failedtoproduce50%inhibition(maximuminhibition~52%at1000µM).

UncomplexedO-AcSHA-H (IC50 = 94 ± 13 µM) showed substantially greater toxicity

towardstheHCT-8cellsincomparisontoBi(O-AcSHA)3(Fig.6.5d).


264

Finally,thecompounds,Bi(4A3HN-SO3)3,Bi(IGA)3,andBi(ILA)3,weremoretoxic

towardsHCT-8cellsthantheanti-cancerdrug,cisplatin(IC50=50µM,Section2.3.2),

andtheirrespectivefreeligands,asindicatedbytheirlowerIC50values(Table6.1and

Section2.3.2).

Figure 6.5: Concentration-response curves obtained from the MTT assays of HCT-8 cellstreated (24 h) with the specified compounds and corresponding Bi(III) complexes:(a)4A3HN-SO3H or Bi(4A3HN-SO3)3; (b) IGA-H, Bi(IGA)3, ILA-H, or Bi(ILA)3; (c) BHA-H orBi(BHA)3; and (d) SHA-H, Bi(SHA)3,O-AcSHA-H or Bi(O-AcSHA)3. Error bars represent thestandard deviation from the mean (n = 6). One representative of three independentexperimentsisshown.


265

Table6.1: IC50 valuesdeterminedbyMTT toxicity assays following treatment (24h) ofHCT-8coloncancercellswiththespecifiedcompounds.

IC50(µM)a IC50ratioLH:Bi(L)3Ligand(LH) Complex[Bi(L)3]

AminoarenesulfonicacidsoAB-SO3

>2000

>2000

-

pAB-SO3 >2000 >2000 -4A3HN-SO3 131±12 12±0.2 10.92AN-SO3 >2000 >2000 -5AN-SO3 >2000 >2000 -Indole-carboxylicacidsIGA

>2000

10±2

>200

ILA >2000 17±1 >118HydroxamicacidsBHA

582±32

106±8

5.5

SHA >2000 159±13 >12.6O-AcSHA 94±13 >1000 <0.1

aIC50valuewasdeterminedusingatleastthreeindependentMTTassays.


266

6.4 Discussion6.4.1 Bi(III)aminoarenesulfonates

The MTT assay results (Section 6.3.1, Table 6.1) showed that the Bi(III)

aminoarenesulfonates and the corresponding free ligands showed little activity

towardsHCT-8cells,with theexceptionofBi(4A3HN-SO3)3and4A3HN-SO3H. Since

mostof theseBi-aminosulfonatesexhibitedno toxicityagainsthumancancercells in

the present study, and have exhibited anti-bacterial activity in other studies [1], it

would be worthwhile to determine if the complexes also show a lack of activity

towardsnon-canceroushumancellsasthiswouldbeabeneficialpropertyfortheuse

oftheBi(III)aminoarenesulfonatesforthetreatmentofH.pyloriinfection.

There are no recent publications that report the in vitro or in vivo anti-cancer

activityof theaminoarenesulfonicacidsused in thepresent study [45-49]. As such,

the reasons why Bi(4A3HN-SO3)3 and 4A3HN-SO3H showed substantial toxicity

comparedtotheotherBi(aminoarenesulfonates)3(Table6.1)canonlybespeculated.

FactorssuchaspKaandlogPmaybelesslikelytocontributetothevariationinactivity

asthefivetestedsulfonicacidsdisplaysimilarpKavalues(rangingfrom1−3,[1])and

logPvalues[45-49]. Additionally, the influenceof chemical substituentsneeds tobe

considered.Forinstance,thepresenceofaphenylring,asulfonylgroupandanamino

grouparecommonfeaturesamongthechemicalstructuresoftheaminoarenesulfonic

acids tested. While4A3HN-SO3Hsharesanaphthaleneskeletonwith2AN-SO3Hand

5AN-SO3H, the presence of the hydroxyl group is unique to the structure of

4A3HN-SO3H. It isnotclearwhetherthehydroxylgroup(andtheextraH-bondingit

provides),oritsadjacentpositiontotheaminogroupbearsinfluenceontheactivityof

4A3HN-SO3H.However,itismorelikelythatthepooraqueoussolubilitycontributes

tothe lackof invitroactivityof theoAB-SO3,pAB-SO3,2AN-SO3H,and5AN-SO3Hand

therespectiveBi(III)complexesagainsttheHCT-8cells.

Interestingly,complexationof4A3HN-SO3HwithBi(III)resultedinanincreasein

toxicityofapproximately11-fold(Table6.1). Thiswasfarhigherthanthatobserved

for Bi(difl)3 and diflH (Section 2.3.2) where it was speculated that the increase in

toxicity of Bi(difl)3might be due to an additive effect between the Bi and the diflH

basedonthecomparableIC50valuesofBiCl3(383µM)anddiflH(401µM).Insteadit

couldbespeculatedthatBi(4A3HN-SO3)3 is likelytoremain intactuponenteringthe

cell. While it is beyond the scope of the aims of this Chapter, the stability of the

complexes in cellmedium and cellular uptake of Bi and potential targets should be


267

examined in future work. For instance, it has previously been shown that the

compound,sodium4-amino-3-hydroxy-1-naphthalene-sulfonate,whichisstructurally

similar to 4A3HN-SO3H, caused a substantial decrease inmitogenesis in Balb/c3T3

fibroblasts that were stimulatedbyacidic fibroblast growth factor (an important

angiogenesis-promoting polypeptide) [50]. These results give credence to a

mechanismoftoxicityof4A3HN-SO3Hthat involvestargetinganenzymeessential to

thecancercellgrowth.

The fact that the toxicity of Bi(4A3HN-SO3)3 was comparable to the

BiNSAIDs(Section2.3.2) provides precedence to explore the non-NSAID derived

Bi(4A3HN-SO3)3 in future experiments. Consequently, studies of the toxicity of

Bi(4A3HN-SO3)3towardsnormalhumancelllines(e.g.healthycoloncellsandhuman

peripheralbloodmononuclearcells(hPBMCs))wouldbeworthwhiletodeterminethe

potential effect on healthy colon cells if administered as a chemotherapeutic or

chemopreventive agent. Thebenefit of screeningBi(4A3HN-SO3)3 onnon-cancerous

cellswillbe two-fold,whereby the resultswillnotonlyestablish selectivity towards

cancercells,butwillalsoprovideinsight intowhetheranypotentialsideeffectsmay

resultfromtheadministrationofthecomplexasananti-bacterialagent,asoriginally

proposedbyBusseetal.[1].

6.4.2 Bi(III)indole-carboxylates

The MTT results (Section 6.3.1, Table 6.1) showed that the indole-carboxylic

acids, IGA-H and ILA-H, only displayed high toxicity towards HCT-8 cells when

coordinated to Bi as the Bi(III)complexes. Furthermore, the resultant complexes

displayed some of the lowest IC50values (highest toxicity) of all the compounds

assessed. Since ILA-H is a product of tryptophan metabolism [17], it was not

surprisingthatILA-HdidnotshowanytoxicitytowardstheHCT-8cells(upto2mM).

The comparatively lower IC50 values of theBi(IGA)3 andBi(ILA)3 (IC50 = 10 µMand

17µM, respectively) cannot be explained by Bi(III) toxicity alone since it was

previously established (Chapter 2) thatwhenBi(III)was bound to the non-toxic Cl−

anion the IC50value (383µM)wasconsiderablyhigher. If theBi(IGA)3andBi(ILA)3

complexesremainintact,alternativetoxicitypathways,mechanismsortargetswould

beresponsibleforthetoxicity.

Importantly, the homoleptic Bi(III) indole-carboxylate complexes, including

Bi(IGA)3, showed little toxicity towardshuman fibroblasts (up to 100µM, 48h) [2],


268

making them excellent candidates for further testing as chemopreventive and

chemotherapeuticproperties. Therefore, further investigation into theirmechanism

oftoxicityincludingdeterminingmodeofcelldeathandpotentialintracellulartargets

shouldbeperformed. Thereisalsoscopetoexpandthecompoundlibraryforfuture

testing as Pathaketal. [2] have synthesised several other Bi(III) indole-carboxylate

complexes.

6.4.3 Bi(III)hydroxamates

The results of theMTT assays (Section6.3.1, Table 6.1) showed that theBi(III)

hydroxamates, Bi(BHA)3 and Bi(SHA)3, exhibited moderate toxicity towards HCT-8

cellswithincreasedactivitycomparedtothecorrespondinghydroxamicacids,BHA-H

and SHA-H, at equimolar concentrations. Interestingly, Bi(O-AcSHA)3 demonstrated

substantiallyreducedtoxicitycomparedtothefreeO-AcSHA-Hligand,whichwasthe

only Bi(III) complex observed to exhibit such a reduced effect (Sections 2.3.2

and6.3.1).

ThereareseveralpossiblereasonsthatBi(O-AcSHA)3displayedreducedtoxicity

comparedtoO-AcSHA-H.Firstly,itwasobservedthat,followingthetreatmentperiod,

Bi(O-AcSHA)3-treated wells showed signs of precipitation in cell medium at

approximately300µM.ThereforethepoorsolubilityofBi(O-AcSHA)3maycontribute

toitslimitedtoxicity.Secondly,aspreviouslydiscussedinSection6.1.3,O-AcSHA-His

anirreversibleCOXinhibitorsimilartoaspH[37]. If themechanismoftoxicityofO-

AcSHA-Hproceedsvia inhibitionofCOXthen it isconceivablethat the loweractivity

couldbe the resultof the inabilityof theacetyl group toacetylate theSer530of the

active site of COX if the Bi(O-AcSHA)3 remains intact inside the cell. In order to

explicitly prove that this is the case, further work to determine the exact chemical

nature of the complex in the cell would be required; however, this would be quite

challenging. In addition, COX inhibition assays (with and without cells) should be

performed. Additionally, ELISA assays may be utilised to assess the ability of

Bi(O-AcSHA)3 andO-AcSHA-H to inhibitPGE2production inorder todeterminehow

effectively the compounds inhibit intracellular COX in vitro (as performed in

Chapter2).

The improved toxicity of BHA-H (IC50 = 582 μM) over SHA-H (IC50 > 2000 μM)

may be attributed to the difference in the chemical structure and the resultant

physicochemical properties. The same trend was observed in another study [51],


269

wherebyBHA-HwasfoundtobemorecytotoxicthanSHA-H,towardsneuroblastoma

IMR-32cells (IC50=218μMand249μM, respectively) following48-h treatment. In

the same study the Ru(III) complexes, [RuCl(H2O)BHA2] and [RuCl(H2O)SHA2], also

showed toxicity towards the IMR-32cells (IC50=166μMand221μM, respectively);

however,theRu(III)complexeswouldbeconsideredlesstoxicwhenaccountingthat

therearetwomolesofligandpermoleofRu(III). Itisnotablethat[RuCl(H2O)BHA2]

wasdeterminedtobemoretoxicthan[RuCl(H2O)SHA2],whichisconsistentwiththe

trendobservedfortheBi(BHA)3andBi(SHA)3complexesassessedinSection6.3.1.

The chemical structures of BHA-H and SHA-H differ only by the presence of a

phenolic –OH group in the SHA-H structure. Pathak et al. [4] determined that the

bidentatecomplexationofBi(III)totheseligandsoccursthroughthecarbonylandthe

hydroxamicacid–OHgroup. ThepKavaluesof thehydroxamicacidgroupofBHA-H

and SHA-H have been reported as 8.71 and 7.40, respectively [52]. The SHA-H

hydroxamicacidprotonismoreacidicduetointramolecularbondingoftheconjugate

basewiththephenolicOH(pKa9.81)[52].DuetothehigherpKa,theBHAligandmay

elicit a stronger affinity for Bi at physiological pH (which is maintained in the cell

medium) compared to SHA; hence the increased propensity for coordination may

improvethestabilityandtoxicityoftheBi(BHA)3complex.

Alternatively, it cannotbe ruledout that theBHA-Hstructureprovidesahigher

affinitytowardsanenzymeimportanttocancercellfunction.Forinstance,BHA-Hand

SHA-H both inhibit the enzyme ribonucleotide reductase [53], which catalyses the

formationofdeoxyribonucleotidesfromribonucleotides.Inonestudy,SHA-Hactivity

was higher than BHA-H (IC50= 150μM and 400 μM, respectively) when assessed

againstpurifiedenzymes [53];however, theseexperimentsdonot take intoaccount

thecellmediumorintracellularenvironmentofthecompounds.BHA-Hhasalsobeen

shown tobe apotent inhibitorof histonedeacetylase6 [54].Whiledetermining the

aqueousstability,uptakeandmechanismsofactionwasbeyondthescopeoftheaims

ofthisChapter,thesecriteriacouldbepotentiallyexaminedinfuturestudies.


270

6.5 ChaptersummaryThe overall aim of this Chapter was to determine the cytotoxicity of Bi(III)

derivativesofaminoarenesulfonicacids,indole-carboxylicacids,andhydroxamicacids

against the human colorectal cancer cell line, HCT-8, in order to identify new lead

compounds for future studies. The results obtained from theMTT assays indicated

thatthetoxicityofthecomplexesishighlyvariablewithrespecttotheligandsinthe

Bi(III)complex.Specificobservationsincludedthefollowing:

i. Bi(III) aminoarenesulfonates displayed no toxicity towards HCT-8 cells

with the exception of Bi(4A3HN-SO3)3, which demonstrated reasonably

high toxicity, which was superior to the anti-cancer drug, cisplatin.

Additionally, the IC50 value obtained for Bi(4A3HN-SO3)3 was

approximately11timeslowerthan4A3HN-SO3H.

ii. Bi(III) indole-carboxylates, Bi(IGA)3 and Bi(ILA)3, displayed the highest

toxicities towardstheHCT-8cells,whichweresuperior to theanti-cancer

drug, cisplatin. Bi(III) coordination enhanced the toxicity of the free

ligands,whichshowedlittletoxicityagainsttheHCT-8cells(upto2mM).

iii. Bi(III) hydroxamates, Bi(BHA)3 and Bi(SHA)3, showed higher toxicity

against the HCT-8 cells compared to the free ligands, BHA-H and SHA-H.

Conversely, the toxicity of Bi(O-AcSHA)3was lower thanO-AcSHA-H. The

IC50valuesofBi(BHA)3andBi(SHA)3were2-3timeslowerthancisplatin.

This Chapter confirms for the first time that Bi(III) derivatives of selected

aminoarenesulfonic acids, indole-carboxylic acids, and hydroxamic acids display

cytotoxicity towards a human colon cancer cell line in vitro. TheBi(IGA)3, Bi(ILA)3,

andBi(4A3HN-SO3)3complexesshowedpromisingtoxicity,whichwascomparableto

Bi(tolf)3, assessed inSection2.3.2. This is an important result andwarrants further

investigation into the solution stability, drug uptake,mechanisms of action, and the

toxicity of theseBi(III) complexes against normal cells (e.g. fibroblast, hPBMCs) and

othercancercell lines. Althoughmanyofthecomplexestestedinthisstudyshowed

positiveactivityagainst theHCT-8coloncancercells, theactivityofBi(4A3HN-SO3)3

againstanon-cancerouscelllineremainstobetestedinordertoassessthedegreeof

selectivityandsubsequentadversesideeffects.Furtherstudiesinvestigatingwhether

theseBi(III)complexespossessanyanti-tumouractivityinvivocouldalsopotentially

bepursued.


271

6.6 References[1] M.Busse,I.Trinh,P.C.Junk,R.L.FerreroandP.C.Andrews.(2013)Synthesisand

characterisation of bismuth(III) aminoarenesulfonate complexes and theirpowerful bactericidal activity against Helicobacter pylori, Chemistry, 19,5264-5275.

[2] A. Pathak, V. L. Blair, R. L. Ferrero, L. Kedzierski and P. C. Andrews. (2017)Structural influencesontheactivityofbismuth(III) indole-carboxylatocomplexestowardsHelicobacterpyloriandLeishmania,J.Inorg.Biochem.,177,266-275.

[3] A. Pathak, V. L. Blair, R. L. Ferrero, M. Mehring and P. C. Andrews. (2014)Bismuth(III) benzohydroxamates: powerful anti-bacterial activity againstHelicobacter pylori and hydrolysis to a unique Bi34 oxido-cluster[Bi34O22(BHA)22(H-BHA)14(DMSO)6],Chem.Commun.,50,15232-15234.

[4] A.Pathak,V.L.Blair,R.L.Ferrero,P.C.Junk,R.F.TaborandP.C.Andrews.(2015)Synthesisandstructural characterisationofbismuth(III)hydroxamatesand theiractivityagainstHelicobacterpylori,DaltonTrans.,44,16903-16913.

[5] S.Ishizaki-Koizumi,I.Sonaka,Y.Takei,K.IkejimaandN.Sato.(2004)Theglycineanalogue,aminomethanesulfonicacid,inhibitsLPS-inducedproductionofTNF-αinisolated rat Kupffer cells and exerts hepatoprotective effects in mice, Biochem.Biophys.Res.Commun.,322,514-519.

[6] Q. Shi, J. E. Savage, S. J. Hufeisen, L. Rauser, E. Grajkowska, P. Ernsberger, J. T.Wroblewski, J.H.NadeauandB.L.Roth.(2003)L-homocysteinesulfinicacidandother acidic homocysteine derivatives are potent and selective metabotropicglutamatereceptoragonists,J.Pharmacol.Exp.Ther.,305,131-142.

[7] R. J.Huxtable. (1989)Taurine in thecentralnervoussystemand themammalianactionsoftaurine,Prog.Neurobiol.,32,471-533.

[8] I.Santa-María,F.Hernández,F.J.MorenoandJ.Avila.(2007)Taurine,aninducerfor tau polymerization and a weak inhibitor for amyloid-β-peptide aggregation,Neurosci.Lett.,429,91-94.

[9] F.Gervais,R.Chalifour,D.Garceau,X.Kong,J.Laurin,R.McLaughlin,C.Morissetteand J. Paquette. (2001) Glycosaminoglycan mimetics: a therapeutic approach tocerebralamyloidangiopathy,Amyloid,8Suppl1,28-35.

[10] T. C. Jackson, J. D. Verrier and P.M. Kochanek. (2013) Anthraquinone-2-sulfonicacid(AQ2S)isanovelneurotherapeuticagent,CellDeathDis.,4,e451.

[11] R.Hoessel,S.Leclerc,J.A.Endicott,M.E.Nobel,A.Lawrie,P.Tunnah,M.Leost,E.Damiens, D. Marie, D. Marko, E. Niederberger, W. Tang, G. Eisenbrand and L.Meijer. (1999) Indirubin, the active constituent of a Chinese antileukaemiamedicine,inhibitscyclin-dependentkinases,Nat.CellBiol.,1,60-67.

[12] J.Ni,T.Mai, S.-T. Pang, I.Haque,K.Huang,M.A.DiMaggio, S. Xie,N. S. James,D.Kasi,S.R.ChemlerandS.Yeh.(2009)InvitroandinvivoanticancereffectsofthenovelvitaminEetheranalogueRRR-α-tocopheryloxybutylsulfonicacidinprostatecancer,Clin.CancerRes.,15,898-906.

[13] H.Griem,J.Ahlers,R.Bias,B.Broecker,H.Hollander,H.-P.Gelbke,H.-J.Klimisch,I.Mangelsdorf, A. Paetz, N. Schong, G. Stropp, R. Vogel, C. Weber, K. Ziegler-Skylakakis and E. Bayer. (1994) Toxicity and ecotoxicity of sulfonic acids:structure-activityrelationship,Chemosphere,28,2203-2236.

[14] P.C.Andrews,M.Busse,G.B.Deacon,R.L.Ferrero,P.C.Junk,K.K.Huynh,I.Kumarand J. G.Maclellan. (2010) Structural and solution studies of phenylbismuth(III)sulfonate complexes and their activity againstHelicobacter pylori,Dalton Trans.,39,9633-9641.

[15] P.C.Andrews,M.Busse,G.B.Deacon,R.L.Ferrero,P.C.Junk,J.G.MacLellanandA.Vom. (2012) Remarkable in vitro bactericidal activity of bismuth(III) sulfonatesagainstHelicobacterpylori,DaltonTrans.,41,11798-11806.


272

[16] P.C.Andrews,M.Busse,P.C. Junk,C.M.ForsythandR.Peiris. (2012)Sulfonato-encapsulated bismuth(III) oxido-clusters from Bi2O3 in water under mildconditions,Chem.Commun.,48,7583-7585.

[17] I. Morita, M. Kawamoto, M. Hattori, K. Eguchi, K. Sekiba and H. Yoshida. (1990)Determinationof tryptophanand itsmetabolites inhumanplasmaandserumbyhigh-performance liquid chromatography with automated sample clean-upsystem,J.Chromatogr.B,526,367-374.

[18] National Center for Biotechnology Information. PubChem Compound Database;CID=92904,Available from: https://pubchem.ncbi.nlm.nih.gov/compound/92904[accessedJan.4,2018].

[19] S.Biswal,U. Sahoo, S. Sethy,H.K. S.Kumar andM.Banerjee. (2012) Indole:Themoleculeofdiversebiologicalactivities,AsianJ.Pharm.Clin.Res.,5,1-6.

[20] J.-R. Weng, C.-H. Tsai, S. K. Kulp and C.-S. Chen. (2008) Indole-3-carbinol as achemopreventiveandanti-canceragent,CancerLett.,262,153.

[21] CelineJ. Marmion, D. Griffith and KevinB. Nolan. (2004) Hydroxamic acids − anintriguingfamilyofenzymeinhibitorsandbiomedicalligands,Eur.J.Inorg.Chem.,2004,3003-3016.

[22] E. C. O'Brien. (1997) Metal complexes of salicylhydroxamic acid andO-acetylsalicylhydroxamicacid,Inorg.Chim.Acta,266,117-120.

[23] E.M.F.MuriandT.G.Barros.(2013)Hydroxamicacidsasinhibitorsofureaseinthe treatment of Helicobacter pylori infections, In Hydroxamic Acids: A UniqueFamily of Chemicals with Multiple Biological Activities (S. P. Gupta, Ed.),pp-241-254,SpringerBerlinHeidelberg.

[24] P.S.DobbinandR.C.Hider.(1990)Ironchelationtherapy,Chem.Br.,26,565-568.[25] D. C. Argentieri, D. M. Ritchie, M. P. Ferro, T. Kirchner, M. P. Wachter, D. W.

Anderson, M. E. Rosenthale and R. J. Capetola. (1994) Tepoxalin: a dualcyclooxygenase/5-lipoxygenase inhibitor of arachidonic acid metabolism withpotent anti-inflammatory activity and a favorable gastrointestinal profile,J.Pharmacol.Exp.Ther.,271,1399-1408.

[26] W. P. Steward. (1999) Marimastat (BB2516): current status of development,CancerChemother.Pharmacol.,43Suppl,S56-60.

[27] Z. I. Cabantchik. (1995) Iron chelators as antimalarials: the biochemical basis ofselectivecytotoxicity,Parasitol.Today,11,74-78.

[28] S.H.Chen,H.M.Wu,B.Ossola,N.Schendzielorz,B.C.Wilson,C.H.Chu,S.L.Chen,Q.Wang, D. Zhang, L. Qian, X. Li, J. S. Hong andR. B. Lu. (2012) Suberoylanilidehydroxamic acid, a histonedeacetylase inhibitor, protectsdopaminergicneuronsfromneurotoxin-induceddamage,Br.J.Pharmacol.,165,494-505.

[29] W. N. Fishbein and P. P. Carbone. (1965) Urease catalysis. II. Inhibition of theenzymebyhydroxyurea,hydroxylamine,andacetohydroxamicacid,J.Biol.Chem.,240,2407-2414.

[30] H. P. Makkar, O. P. Sharma, R. K. Dawra and S. S. Negi. (1981) Effect ofacetohydroxamicacidonrumenureaseactivityinvitro,J.DairySci.,64,643-648.

[31] D.M.Keogan, B. Twamley,D. Fitzgerald-Hughes andD.M.Griffith. (2016)Novelclass of Bi(III) hydroxamato complexes: synthesis, urease inhibitory activity andactivityagainstH.pylori,DaltonTrans.,45,11008-11014.

[32] S. Pang, S. Tristram and S. Brown. (2011) Salicylhydroxamic acid inhibits thegrowthofCandidaalbicans,Int.J.Biol.LifeSci.,7,40-46.

[33] S.-Y. M. Pang, S. Tristram and S. Brown. (2011) Inhibition of the growth ofpathogenicCandidaspp.bysalicylhydroxamicacid,Int.J.Biol.LifeSci.,7,1-7.

[34] I. Botos, L. Scapozza,D. Zhang, L. A. Liotta andE. F.Meyer. (1996)Batimastat, apotentmatrixmealloproteinaseinhibitor,exhibitsanunexpectedmodeofbinding,Proc.Natl.Acad.Sci.U.S.A.,93,2749-2754.

[35] S. M. Wojtowicz-Praga, R. B. Dickson and M. J. Hawkins. (1997) Matrixmetalloproteinaseinhibitors,Invest.NewDrugs,15,61-75.


273

[36] S. S. Tam,D.H. Lee, E. Y.Wang,D. G.Munroe andC. Y. Lau. (1995)Tepoxalin, anovel dual inhibitor of the prostaglandin-H synthase cyclooxygenase andperoxidaseactivities,J.Biol.Chem.,270,13948-13955.

[37] P. J.Loll,C.T.Sharkey,S. J.O'Connor,C.M.Dooley,E.O'Brien,M.Devocelle,K.B.Nolan,B.S.SelinskyandD. J.Fitzgerald.(2001)O-acetylsalicylhydroxamicacid,anovelacetylatinginhibitorofprostaglandinH2synthase:structuralandfunctionalcharacterizationofenzyme-inhibitorinteractions,Mol.Pharmacol.,60,1407-1413.

[38] C. M. Dooley, M. Devocelle, B. McLoughlin, K. B. Nolan, D. J. Fitzgerald and C. T.Sharkey. (2003) A novel family of hydroxamate-based acylating inhibitors ofcyclooxygenase,Mol.Pharmacol.,63,450-455.

[39] J.A.Sparano,P.Bernardo,P.Stephenson,W.J.Gradishar,J.N.Ingle,S.ZuckerandN.E.Davidson.(2004)RandomizedPhaseIIItrialofmarimastatversusplaceboinpatientswithmetastaticbreastcancerwhohaverespondingorstablediseaseafterfirst-line chemotherapy: Eastern cooperative oncology group trial E2196, J. Clin.Oncol.,22,4683-4690.

[40] M.L.Rothenberg,A.R.NelsonandK.R.Hande.(1998)Newdrugsonthehorizon:matrixmetalloproteinaseinhibitors,Oncologist,3,271-274.

[41] P.A.MarksandR.Breslow.(2007)Dimethylsulfoxidetovorinostat:developmentofthishistonedeacetylaseinhibitorasananticancerdrug,Nat.Biotech.,25,84-90.

[42] H.-J. Kim and S.-C. Bae. (2011) Histone deacetylase inhibitors: molecularmechanismsofactionandclinicaltrialsasanti-cancerdrugs,Am.J.Transl.Res.,3,166-179.

[43] O.A.O'Connor,S.Horwitz,T.Masszi,A.VanHoof,P.Brown,J.Doorduijn,G.Hess,W. Jurczak,P.Knoblauch,S.Chawla,G.Bhat,M.R.Choi, J.Walewski,K.Savage,F.Foss, L. F. Allen and A. Shustov. (2015) Belinostat in patients with relapsed orrefractory peripheral T-cell lymphoma: results of the Pivotal Phase II BELIEF(CLN-19)study,J.Clin.Oncol.,33,2492-2499.

[44] D. Pal and S. Saha. (2012) Hydroxamic acid – a novel molecule for anticancertherapy,J.Adv.Pharm.Technol.Res.,3,92-99.

[45] National Center for Biotechnology Information. PubChem Compound Database;CID=6926, Available from: https://pubchem.ncbi.nlm.nih.gov/compound/6926[accessedJan.4,2018].



[48] National Center for Biotechnology Information. PubChem Compound Database;CID=6670, Available from: https://pubchem.ncbi.nlm.nih.gov/compound/6670[accessedJan.42018].


[50] C.Fernandez-Tornero,R.M.Lozano,M.Redondo-Horcajo,A.M.Gomez,J.C.Lopez,E. Quesada, C. Uriel, S. Valverde, P. Cuevas, A. Romero and G. Gimenez-Gallego.(2003) Leads for development of new naphthalenesulfonate derivatives withenhanced antiangiogenic activity: crystal structure of acidic fibroblast growthfactor in complex with 5-amino-2-naphthalene sulfonate, J. Biol. Chem., 278,21774-21781.

[51] R. Kaushal and Sheetal. (2016) In vitro anticancer and antibacterial activities ofoctahedral ruthenium(III) complexes with hydroxamic acids. Synthesis andspectroscopiccharacterization,Russ.J.Gen.Chem.,86,360-367.


274

[52] E. C. O'Brien, E. Farkas, M. Jose Gil, D. Fitzgerald, A. Castineras and K. B. Nolan.(2000) Metal complexes of salicylhydroxamic acid (H2SHA), anthranilichydroxamic acid and benzohydroxamic acid. Crystal and molecular structure of[Cu(phen)2(Cl)]ClPh2Sha, a model for a peroxidase-inhibitor complex, J. Inorg.Biochem.,79,47-51.

[53] B. Van't Riet, G. L.Wampler and H. L. Elford. (1979) Synthesis of hydroxy- andamino-substitutedbenzohydroxamicacids: inhibitionofribonucleotidereductaseandantitumoractivity,J.Med.Chem.,22,589-592.

[54] F. F.Wagner, D. E. Olson, J. P. Gale, T. Kaya, M.Weïwer, N. Aidoud, M. Thomas,E.L.Davoine, B. C. Lemercier, Y.-L. Zhang and E. B. Holson. (2013) Potent andselective inhibition of histone deacetylase 6 (HDAC6) does not require asurface-bindingmotif,J.Med.Chem.,56,1772-1776.

Chapter7.

ConclusionsandFutureDirections

Chapter7.ConclusionsandFutureDirections

276

7.1ConclusionsCRC is the secondmost common cancer in both men and women in Australia.

AstheincidenceofCRChasnotimprovedsincethe1980s,chemoprevention(theuse

ofapharmacologicalornutritionalagenttoprevent,reverse,ordelaytheprogression

ofcarcinogenesis)hasbeensuggestedasaprimarypreventionstrategytoreducethe

burden of this disease. The upregulation of COX-2, an enzyme expressed during

inflammation response, has been linked to a number of cancers, including CRC.

Assuch,inhibitionofCOX-2maypresentatargetforchemoprevention/chemotherapy.

To this end, over-the-counter COX inhibitors, NSAIDs, have been investigated as

potential chemopreventives. Several epidemiological and clinical studies provided

evidencethatNSAIDs(inparticular,aspirinandsulindac)reducetheriskofCRCwhen

taken regularly over periods of months to years. However, side effects, such as

GIbleeding, CV abnormalities and hepatotoxicity, limit the daily use of NSAIDs over

prolonged periods of time. As such, a method to make NSAIDs safer for long-term

administration is needed, without diminishing the chemotherapeutic potential of

theNSAIDs.

Severalstudieshaveshownthatanumberofmetal-NSAIDcomplexesreducethe

formationofGIulcerationcomparedtothenon-complexedNSAIDsinanimalmodels.

Importantly,[Cu2(indo)4(DMF)2]reducedaberrantcryptformationintherodentcolon

cancer model, while also exhibiting GI sparing properties.Bismuth, a group 15

post-transitionmetal,hasbeenusedtotreatanarrayofGIdiseasesforcenturies,with

anacceptabletoxicityprofileatlowdoseswhenadministeredoverextendedperiods.

Thus, it ishypothesised that thecombinationofBi andNSAID ina single compound

may potentially relieve the adverse GI side effects of NSAIDs if used daily as a

chemopreventive agent. To this end, the suitability of Bi complexes of

NSAIDs (BiNSAIDs) as potential chemotherapeutics or chemopreventives needs to

beexamined.

In order to determine the suitability of the BiNSAIDs, Bi(tolf)3, Bi(mef)3 and

Bi(difl)3, as a potential chemotherapeutic or chemopreventive agents against CRC, it

wasnecessarytoevaluatetheinvitropotentialoftheBiNSAIDsasapreludetofuture

work that would involve in vivo studies. The experiments described in this Thesis

were divided into two goals: (i) to establish the chemotherapeutic potential of

BiNSAIDsbydeterminingwhetherBiNSAIDsareabletokilltransformedcells(toxicity

and cell death assays), and (ii) to establish whether BiNSAIDs demonstrate


277

chemopreventivepotentialbydeterminingwhetherBiNSAIDsareable to inhibit the

production of PGE2 (and thus inhibit COX) in transformed cells. A human epithelial

ileocecalcoloncancercell line,HCT-8,wasemployedasamodelCRCcell lineforthe

evaluation of the BiNSAIDs, and the respective uncomplexed NSAIDs, tolfH, mefH

anddiflH.

OnespecificaimofthisprojectwastocomparethetoxicityofBiNSAIDstothatof

the uncomplexed NSAIDs. The toxicity of the BiNSAIDs and NSAIDs towards the

HCT-8cellswasassessedusingtheMTTassay(Chapter2). Theresultsshowedthat

the toxicity ranking of the BiNSAIDs paralleled those of the respective free

NSAIDs:diflH<mefH<tolfH.WhiletheIC50valuesoftheBiNSAIDs(rangingbetween

16−81μM) were lower than the free NSAIDs (ranging between 37−403μM), it was

apparentthatthetoxicityoftheBiNSAIDswasreflectiveofthemolarratioofthethree

NSAIDmoleculescontainedintheBiNSAIDs,withtheexceptionofBi(difl)3.Oneofthe

most importantresults fromtheMTTassayswasthemicromolar toxicityofBi(tolf)3

and Bi(mef)3 towards HCT-8 cells which was comparable to the anti-cancer drug,

cisplatin(IC50=16μM,41μM,and50μM,respectively).

Themechanismof cell deathwas investigatedusing phase contrastmicroscopy

and AnnV/7AAD staining (Chapter 2). The results from these experiments showed

that treatment with all BiNSAIDs and NSAIDs induced morphological changes

indicativeofcelldeathviaapoptosis.WhenconsideredwiththeresultsfromtheMTT

assays,itappearedmostlikelythattheNSAIDisthemaincontributorofthetoxicityof

theBiNSAIDcomplexes.TheexperimentsdescribedinChapter2showedforthefirst

time that BiNSAIDs exhibit potential anti-cancer activity towards a human colon

cancer cell line in vitro and that cell death occurred via apoptotic pathway.

Furthermore, the toxicityof theBiNSAIDswas similaror improvedcompared to the

respectiveNSAIDs,indicatingthatthecomplexationwithBididnotnegativelyimpact

thebioactivityoftheNSAIDs.

The second focus of the in vitro studieswas to determine if theBiNSAID could

inhibitCOXactivityinthecoloncancercellssincethiswastheunderlyingmechanism

that hadbeenhypothesised forNSAID chemopreventive behaviour. As such, assays

wereperformedtodeterminewhethercomplexationwithBiaffectstheabilityofthe

NSAIDs, tolfH, mefH and diflH, to interact with COX in vitro, by quantifying PGE2

productionbyHCT-8cells(Chapter2).Theinhibitionoftheprostaglandin,PGE2,was

studiedsincePGE2hasanestablishedroleinthepromotionofcancercellgrowthand


278

fortification (Fig. 1.5). The results showed that the inhibitionofPGE2productionof

HCT-8cellstreatedwiththeBiNSAIDsparalleledthoseoftherespectivefreeNSAIDs:

diflH < mefH < tolfH. There are two important conclusions from these

experiments.Firstly,theresultsindicatedthattheanti-cancereffectsoftheexamined

NSAIDswere not substantially diminished by the coordination of Bi to the NSAIDs.

Furthermore, the results confirmed that treatment of HCT-8 cells with BiNSAIDs

caused a reduction in PGE2 (via COX inhibition),which is one of themechanisms of

actionhypothesisedfortheNSAID-chemopreventionofCRC.

Once the biological potential of the BiNSAIDs had been established in vitro,

Biuptakeandlocalisationstudieswereperformedtofurtherprobetheirintracellular

behaviour. Bi uptake was quantified using GFAAS (Chapter 3) where the highest

cellular Bi was observed following treatment with Bi(tolf)3. Since the NMR studies

(Chapter 2) indicated that Bi(tolf)3 was themost stable BiNSAID in cell medium, it

appears that Bi uptake is most likely facilitated by the NSAID. Treatment with

Bi(mef)3andBi(difl)3resultedinsubstantial,butmuchloweruptakeofBicomparedto

Bi(tolf)3, which correlated with their lower toxicities and increased reactivity

(potentialdissociation)incellmediumasdeterminedbyNMRspectroscopy.

Microprobe SRXRF imaging (Chapter 3) showed that the intracellular fate of Bi

was similar for all BiNSAID treatments irrespective of the coordinated NSAID.

Forinstance, all BiNSAID-treated cells showed Bi accumulation in the cytoplasm

within24-hexposure,typicallyindiscretehotspots.Anumberofpossiblereasonsfor

thelocalisation(includingtargetingofthenucleusandmitochondria)wereconsidered

andtwomorefeasibleoptionswereconcludedbasedoncorrelationmapsandreports

fromtheliterature.ThefirstwasthathotspotscorrelatedwithlysosomessincetheBi

locationandthesizeofthehotspots(0.3–5.8μm2)wereconsistent.Itwaspostulated

that thispackagingofBimightbe thecellsmechanismofdetoxification. Asmanyof

thehotspotswerelocatednearthenucleus,andacknowledgingthatCOX-2islocated

the on nuclear membrane, the SRXRF results also indicate that Bi is capable of

intracellular movement. In the instance that the BiNSAID has remained intact this

would suggest that BiNSAIDs could locate in the vicinity of and target COX-2. It is

importanttonote,however,thattheexperimentalmethodsutilisedinthisstudyonly

provided informationabout theuptakeand localisationofBi. Consequently, it isnot

possible tomakeanyevidence-based conclusionsabout theuptakeand intracellular

fateoftheNSAIDs.WhilethiscouldbeaddressedbytaggingtheNSAID,thistoowould


279

beinconclusiveifthetaggingalteredthecellulardistributionoftheBiNSAID.

SR-IRMSwasutilisedtoprobecellularbiomolecularchangesassociatedwiththe

treatment of liveHCT-8 cellswith BiNSAIDs andNSAIDs (Chapter 4). A number of

factors were considered when analysing the spectra obtained from the live cell

populations. Firstly, while live cell analysis can reduce spectral artifacts, such as

RMieS,theIRspectraofwatercontainsavibrationalmode(visualisedasabroadband

at~1643cm−1)thatoverlapswiththeamideIband,andtoalesserextenttheamideII

band.Theconcernisthatthiscanresultinthedistortionoftheintensityratioofthe

amide I/amide IIbands, thusconfoundingbiomolecularchangesresulting fromdrug

treatment.Inordertodeterminewhetherthewaterbandsubstantiallyinfluencesthe

biological information contained in the amide region, a correction strategy was

appliedtoremovethewaterbandsfromthecellspectra.TheresultantPCAscoresand

loadings plots revealed very little difference before and after the application of the

waterbandcorrectionappliedinthisstudy.However,duetocontinueddebateinthe

literatureabouttheidealcorrectionstrategy,moreworkwillneedtobeperformedin

thisarea.

Subsequently,theSR-IRMScellspectrawereanalysedintwoways,wherebythe

spectrawereanalysedwithandwithouttheinclusionoftheamideregion.Itwasclear

fromthePCAscoresplotsthatintra-sampleandinter-samplevarianceexistedwithin

the vehicle-, BiNSAID-, and NSAID-treated populations. Examination of the PCA

loadings plots indicated that in the majority of BiNSAID- and NSAID-treated cell

populations,theloadingsassociatedwithlipidsdominatedthePCAloadingsplots.

Following4-htreatment,Bi(tolf)3andtolfHinducedsimilarbiomoleculareffects

on theHCT-8 cells in comparison to the vehicle,whichwereprimarily attributed to

effects on the lipids with some contribution from proteins. However, following

removal of the amide region, discrimination between the two compounds was

apparentandattributabletoν(C=O)andνas(PO2−),suggestiveofdifferenceintheeffect

ofthetwocompoundsonphospholipids.Bi(mef)3ormefHinducedsimilareffectson

theHCT-8cells,whichwereprimarilyattributedtolipids.DiflHinducedbiomolecular

effects on the HCT-8 cells, which were primarily attributed to proteins. However,

followingamideremovalfromPCAanalysis,diflHnolongershowedanychangesfrom

thevehicle. TreatmentwithBi(difl)3wasdistinguishable fromdiflHand thevehicle

followingamideremoval,whichwasprimarilyattributedtolipids.

Following24-htreatment,tolfHinducedbiomoleculareffectsontheHCT-8cells,


280

which were primarily attributed to a possible reduction in lipids with some

contribution from proteins. Treatment of the cells with Bi(tolf)3 induced a varied

responsewithintheBi(tolf)3-treatedpopulation. TreatmentwithBi(mef)3andmefH

induced similar biomolecular effects on the HCT-8 cells, which were primarily

attributedtolipids.ThisresultmightbeareflectionoftheBi(mef)3dissociationsuch

that theeffectofBi(mef)3 treatment isprimarilyattributable tomefH. Interestingly,

treatment with Bi(difl)3 resulted in no distinguishable difference from the vehicle-

treatedcellSR-IRMSspectra.However,treatmentwithdiflHwasdistinguishablefrom

Bi(difl)3 and vehicle scores, which was primarily attributed to a strong association

with lipids. This was due to a large increase in the lipids contributions that were

visualisedintheaveragedSR-IRMSspectraofthe24-hdiflH-treatedcells.

Recognisingthatanoverallchange in lipidswas thedominantobservation from

theSR-IRMSexperiments,ESI-MS/MSwasusedasacomplimentarytechniquetoshed

lightonwhether lipidchangesweretrulyassociatedwithBiNSAID/NSAIDtreatment

rather than intra-sample variation. The most abundant glycerophospholipids in

mammaliancells,PC,PEandPS,werechosenasa startingpoint tobeginexamining

theeffectsofBiNSAID/NSAIDtreatment.

The changes to the abundanceof the individualPC, PE andPSmolecules in the

overall profiles of the Bi(tolf)3- and Bi(difl)3-treated cells, and the respective

NSAID-treated cells, displayedmany similar changeswhen compared to the vehicle-

treated cells. There were some differences between the changes induced by

Bi(tolf)3/tolfH treatments compared to the Bi(difl)3/diflH treatments. When the PC

and PE molecules were grouped according to fatty acid categories, the observed

changesintheBi(tolf)-andtolfH-treatedcellsweresimilartoeachotherandshowed

a significantdecrease in the contributionofunsaturatedPCandPEmolecules,buta

significantincreaseinether-linkedmolecules.TheresultsobtainedfromtheBi(difl)3-

and diflH-treated cells showed similar changes in the grouped PE profile, but

conversely, the reverse trendwas observed in the groupedPCprofileswhereby the

contribution of the unsaturated PC and PE molecules significantly increased, but a

significant decrease in ether-linkedmoleculeswas observed. Thismay suggest that

while Bi(tolf)3 and Bi(difl)3 both induced cell death via apoptosis, the biomolecular

events that precede or occur during apoptosis differ between the two complexes

wherebytheHCT-8cellsresponddifferentlytothefenamatecompoundscomparedto

thesalicylatecompounds.


281

Additionally, changeswereobserved in the relativeconcentrationofPC,PEand

PS in the BiNSAID and NSAID-treated cells compared to the vehicle-treated cells.

However, when the PC and PE molecules were grouped according to fatty acid

categories, it was observed that there were some differences to the overall trend

between the Bi(tolf)3 and tolfH-treated cells compared to the vehicle-treated cells.

Specifically, an overall decrease in the relative concentrations of PC, PE and PSwas

also observed. This result was consistent with literature reports that PC and PE

concentrations are reduced in apoptotic cells, and the SR-IRMS studies, which

suggested that tolfH (and potentially Bi(tolf)3) were associated with lower lipid

content.

In contrast, some variation was observed between the response elicited by

Bi(difl)3comparedtodiflH,wherebythediflH-treatedcellsexhibitedlowerPC,PEand

PS than the Bi(difl)3-treated cells.This was inconsistent with the SR-IRMS

experimentsthatshowedthatdiflH-treatedcellswereassociatedwithincreasedlipid

content in the PCA results (24-h results). As such, this result may indicate that

glycerophospholipids (specifically PC, PE and PS) were not associated with the

increase in the lipid signalsand thatother lipids, includingneutral lipidclasses (e.g.

cholesterolesters,triacylglycerides)shouldbeexaminedinfuturestudies.

Encouraged by the promising in vitro anti-cancer activity of the BiNSAIDs

againstthe HCT-8 cells, and recognising that Bi itself possesses anti-cancer

activity (Section1.4.1.4), several Bi(III) complexes of aminoarenesulfonates,

indole-carboxylates, hydroxamates were screened for potential anti-cancer activity

invitrousingtheMTTassay. Thecomplexeswerechosenfor investigationsincethe

aforementionedchemicalclassescompriseexamplesofbiologicallyactivemolecules,

includinganti-cancerdrugs.TheresultsfromtheMTTassaysshowedthatofthefive

Bi(III)-aminoarenesulfonate complexes tested, only Bi(4A3HN-SO3)3 and its ligand,

4A3HN-SO3H, displayed toxicity (IC50 = 12μM and 131 μM, respectively) in the

concentrationrangetested(upto2mM).TheBi(III)indole-carboxylates,Bi(IGA)3and

Bi(ILA)3,displayedtwoofthehighestIC50values(10μMand17μM,respectively)of

all the tested Bi(III) complexes assayed in this Thesis, and were superior to the

anti-cancerdrug,cisplatin(IC50=50μM)whentestedintheHCT-8cells.Incontrast,

the free ligands, IGA-H and ILA-H showed little to no toxicity against the HCT-8

cells(upto2mM)indicatingthatthecomplexationtoBiisessentialfortoxicity.The

Bi(III)hydroxamates,Bi(BHA)3andBi(SHA)3, showedhigheractivity (IC50=106μM


282

and 159μM, respectively) against the HCT-8 cells than that expected for themolar

equivalent of the free ligands BHA-H and SHA-H (IC50 = 582 μM and > 2,000 μM,

respectively). Conversely, the toxicity of Bi(O-AcSHA)3 was lower than O-AcSHA-H

(IC50=>1,000μMand94μM,respectively).

ThestudiesreportedinthisThesissuggestthatBi(tolf)3isastrongcandidatefor

further testing in an animalmodel. Specifically, the Bi(tolf)3 complexwas themost

potentBiNSAID towards cancer cells in vitro (displaying toxicity at aphysiologically

relevant concentration), and appeared to maintain stability in the aqueous

environment used to simulate physiological conditions in vitro. Additionally, the

Bi(tolf)3 showed similar ability to inhibit the production of PGE2 (Chapter 2) and

inducedsimilarchangestothephospholipidprofilecomparedtotolfHatanequimolar

treatmentconcentration.

7.2FutureDirectionsOverall, the results reported in this Thesis indicate that BiNSAIDs display

potential chemotherapeutic and chemopreventive properties against transformed

cells invitro. InadditiontothefuturedirectionssuggestedinthepreviousChapters,

thelongertermdirectionoftheprojectsshouldincludetwogeneralaims:(i) further

invitro work, including screening against other cancer and healthy cell lines, and

(ii)ultimately,invivostudies.

ThestudiesreportedinthisThesiswerepurposelylimitedtoonecelllinetogain

a fundamental understanding of the interactions of BiNSAIDswith a representative

CRCcellline.Infuturestudies,itwouldbeofinteresttoassesswhethertheBiNSAIDs

display greater potency towards cell lines that overexpress COX-2 (e.g.HT-29 colon

cancercells)comparedtoacelllinethatexhibitsnoexpressionofCOX-2(e.g.HCT-116

colon cancer cells), and thus whether there is any correlation between COX-2

expressionandBiNSAIDtoxicity.TheseexperimentshavebeencommencedbyDillon,

BrownandRanson(personalcommunication).Itisanticipatedthatthesestudieswill

provide insight into whether the inhibition of COX-2 or other off-target effects are

essential for the in vitro toxicity of BiNSAIDs. Furthermore, BiNSAIDs could also be

assessed in vitro against other types of GI cancer cell lines, including esophageal,

gastric and other cancers of the upper intestinal tract. It will also be essential to

determine whether the BiNSAIDs display any toxicity towards non-cancerous cells

including primary fibroblasts (as a non-cancerous colon cell line) and


283

hPBMCs(isolatedfromhumanblood).

InorderfortheBiNSAIDstohaveanypotentialforuseaschemotherapeuticsor

chemopreventives,severaloutcomesmustfirstbeestablishedinvivo. Theseinclude:

(i) comparing the GI toxicity of the BiNSAIDs to the free NSAIDs following dosing

(short- and long-term) in order to establish if the BiNSAIDs protect against GI

ulceration; (ii) comparing the BiNSAIDs and NSAIDs for the prevention of tumour

formation in the GI tract as ameasure of chemoprevention, and (iii)measuring the

accumulation of Bi in the GI tract and vital organs (such as the liver, kidneys, and

brain)inordertodeterminethesafetyoftheBiNSAIDsfollowinglong-termdosing.

Based on the work presented here, studies to examine the aforementioned

outcomeshavebeencommenced(initiallyforBi(indo)3,indoH,Bi(asp)3andaspH)by

Dillon and Brown, et al. using rodent models (personal communication).

ThepreliminarydatashowsthatBiNSAIDsareabletoreducetheformationoflesions

in the small intestine of rats following a single dose (unpublished results). A long-

term study is planned whereby the ability of the BiNSAIDs to prevent tumour

formation in an azoxymethane/dextran sulfate sodium colitis mouse model.

Asdiscussed in Section 1.4.2, the accumulation of Bi in animal models has been

established andBi toxicity has occurred in humans (albeit patientswho have taken

extremelylargedosesofBi). Assuch,it isimportantthatanimalstudiesaddressthe

potentialforaccumulationandtoxicityofBifollowingtreatmentwithBiNSAIDs.The

accumulationofBi in theorgansof themicedosedwithBiNSAIDswill beexamined

using GFAAS to provide insight into the potential for Bi toxicity, and therefore, the

safetyoftheBiNSAIDs(DillonandBrownetal.,personalcommunication).

Overall, the results reported in this Thesis provide an indication that BiNSAIDs

displaychemotherapeuticandchemopreventionpropertiesagainsttransformedcells.

TheSR-IRMSandlipidESI-MS/MSexperimentsindicatethatthemechanismofaction

of the BiNSAIDs may be multi-faceted and may provide a more robust

chemotherapeuticapproachthansometraditionalchemotherapeuticdrugs,whereby

cancer cells often develop chemoresistance. Additionally, the employment of

chemopreventivemedications(particularly inpopulationsatriskofdevelopingCRC)

could be beneficial to society and reduce the costs associated with treating cancer.

Additional in vivo work will need to be performed in order to further reveal the

therapeuticpotentialoftheBiNSAIDs.

Appendix1

284

Appendix1.AdditionalMTTassayandPGE2assayresults

Appendix1.1:Concentration-responsecurveobtainedfromtheMTTassayofDMSO(%v/vinRPMI-1640)inHCT-8cellsafter24-htreatment.Errorbarsrepresentthestandarddeviationfromthemean(n=6).Onerepresentativeoftwoindependentexperimentsisshown.

Appendix1

285

Appendix1.2:Concentration-responsecurvesobtainedfromtheMTTassaysof:(a)tolfHandBi(tolf)3;(b)mefHandBi(mef)3;and(c)diflHandBi(difl)3;inHCT-8cellsafter4-htreatment.Error bars represent the standard deviation from themean (n = 6). One representative ofthreeindependentexperimentsisshown.

Appendix1

286

Appendix1.3:Concentration-responsecurvesobtainedfromtheMTTassaysofcisplatin,BiCl3and BSS after 24-h treatment. Error bars represent the standard deviation from themean(n=6).Onerepresentativeofthreeindependentexperimentsisshown.

Appendix 1.4: PGE2 production in HCT-8 cells after 4-h treatment with BSS. Results areexpressedasmean±SD(n=3fromthreeindependentexperiments).

Appendix1

287

Appendix1.5:Concentration-responsecurveobtainedfromtheMTTassaysofaspHinHCT-8cellsafter24-htreatment.Errorbarsrepresentthestandarddeviationfromthemean(n=6).Onerepresentativeofthreeindependentexperimentsisshown.

Appendix2

288

Appendix2. Additional microprobe SRXRF thin-sectioned cellmaps

Appendix2.1:GraphitefurnaceoperatingconditionsusedfortheanalysisofBi.

Stage Temperature(°C)

Ramp(sec)

Hold(sec)

Gasflow(mL.min-1)

Dry1 110 1 30 250Dry2 130 15 30 250Ash 1100 10 20 250Atomisation 1700 0 5 0Cleanout 2450 1 3 250

Appendix2

289

Appendix 2.2:Microprobe SRXRF elementalmaps for toluidine-blue stained thin-section ofHCT-8cellsfollowing24-htreatmentwith:(a)vehicleonly;(b)Bi(difl)3(40μM);(c)Bi(mef)3(40μM);(d)Bi(tolf)3(40μM),and(e)BSS(40μM).Theelementsarespecifiedtotheleftofeachrow.Operatingconditionsinclude:beamenergy=13.45keV,beamsize=0.3×0.3μm2;step size = 0.3 μm; dwell time = 2.5 s/pt and scan dimensions (H × V) = (a) 13 × 13 μm2;(b)19×19μm2;(c)17×14μm2;(d)21×22μm2and(e)12×14μm2.

Appendix2

290

Appendix 2.3:Microprobe SRXRF elementalmaps for toluidine-blue stained thin-section ofHCT-8cellsfollowing4-htreatmentwithBi(mef)3(40μM).Theelementsarespecifiedtotheleft of each row. Operating conditions include: beam energy = 13.45 keV, beam size =0.3×0.3μm2; step size = 0.3μm; dwell time = 2.5 s/pt and scan dimensions (H × V) = (a),14×11μm2;(b)22×14μm2.

Appendix2

291

Appendix2.5:Correlative lightmicrographofatoluidine-bluestainedthin-sectionofHCT-8cells following 24-h treatment with: (a) Bi(difl)3 (40 μM); (b) Bi(mef)3 (40μM); and(c)Bi(tolf)3 (40 μM) and the corresponding microprobe SRXRF maps of P, Zn, Bi, and thecolocalisation (Col) of the three elements. Operating conditions include: beam energy =13.45keV, beam size = 0.3 × 0.3μm2; step size = 0.3μm; dwell time = 2.5 s/pt and scandimensions (H × V) = (a) 19 × 16μm2 (i) and 19 × 19μm2 (ii); (b) 22 × 15μm2 (i) and17×14μm2(ii);(c)31×21μm2(i)and21×22μm2(ii).

Appendix3

292

Appendix3.AdditionalSR-IRMSspectra

Appendix3.1: (a)Averaged(n=54),baselinecorrected(rubberbandbaselinecorrected,64baseline points), vector normalised HCT-8 cell spectra following 4-h incubation in vehicle(RPMI-1640, DMSO 2% v/v) before (—) and after (—) water band compensation. Blackarrows indicate the spectral features affected by water compensation. (b) Unit vectornormalised,secondderivativeofthespectrashownin(a).

Appendix3

293

Appendix 3.2: Averaged, baseline corrected (rubberband baseline corrected, 64 baselinepoints), vector normalised, HCT-8 cell spectra following 4-h incubation in cell medium(RPMI-1640;— ,n=54),orvehicle(RPMI-1640,2%DMSOv/v;— ,n=59):(a)withouttheapplication of water band subtraction, and (b) following water band subtraction. Thewavenumbers in black represent the medium-treated band positions, while band positionsthatwereshiftedinthevehicle-treatedspectraarerepresentedinred.

Appendix3

294

Appendix3.3:Averagedsecondderivative(SavitzkyGolay,13smoothingpoints),unitvectornormalised, HCT-8 cell spectra following 4-h incubation in cell medium (RPMI-1640; — ,n=54),orvehicle(RPMI-1640,2%DMSOv/v;— ,n=59):(a)withouttheapplicationofwaterband correction, and (b) following water band correction. The wavenumbers in blackrepresent themedium-treatedbandpositions,while bandpositions thatwere shifted in thevehicle-treatedspectraarerepresentedinred(sh=shoulder).

Appendix3

295

Appendix 3.4: Averaged, baseline corrected (rubberband baseline corrected, 64 baselinepoints), vector normalised HCT-8 cell spectra following 24-h incubation in cell medium(RPMI-1640alone;— ,n=39),orvehicle(RPMI-1640,2%DMSOv/v;— ,n=54):(a)withoutthe application of water band correction, and (b) following water band correction. Thewavenumbers in black represent the medium-treated band positions, while band positionsthatwereshiftedinthevehicle-treatedspectraarerepresentedinred.

Appendix3

296

Appendix3.5:Averagedsecondderivative(SavitzkyGolay,13smoothingpoints),unitvectornormalised,HCT-8cellspectrafollowing24-hincubationinRPMI-1640alone(— ,n=39),orvehicle (RPMI-1640, 2%DMSO v/v;— , n = 54): (a) without the application of water bandcorrection,and(b)followingwaterbandcorrection.Thewavenumbersinblackrepresentthemedium-treatedbandpositions,whilebandpositionsthatwereshiftedinthevehicle-treatedspectraarerepresentedinred(sh=shoulder).

Appendix3

297

Appendix3.6:(a)Averaged,baselinecorrected(rubberbandbaselinecorrected,64baselinepoints),vectornormalisedHCT-8cellspectrafollowing4-hincubationinvehicle(RPMI-1640,DMSO2%v/v,— ,n=59),Bi(tolf)3(15µM, — ,n=45)ortolfH(45µM, — ,n=29).(b)Unitvectornormalised,secondderivativeofthespectrashownin(a).

Appendix3

298

Appendix3.7:(a)Averaged,baselinecorrected(rubberbandbaselinecorrected,64baselinepoints), vector normalised HCT-8 cell spectra following 24-h incubation in vehicle(RPMI-1640,DMSO2%v/v,— , n = 54), Bi(tolf)3 (15 µM, — , n = 33) or tolfH (45 µM, — ,n=100).(b)Unitvectornormalised,secondderivativeofthespectrashownin(a).

Appendix3

299

Appendix3.8:(a)Averaged,baselinecorrected(rubberbandbaselinecorrected,64baselinepoints),vectornormalisedHCT-8cellspectrafollowing4-hincubationinvehicle(RPMI-1640,DMSO2%v/v;— ,n=83),Bi(mef)3(40µM, — ,n=42)ormefH(120µM, — ,n=54).(b)Unitvectornormalised,secondderivativeofthespectrashownin(a).

(a)

(b)

Appendix3

300

Appendix3.9:(a)Averaged,baselinecorrected(rubberbandbaselinecorrected,64baselinepoints), vector normalised HCT-8 cell spectra following 24-h incubation in vehicle(RPMI-1640,DMSO2%v/v;— ,n=54),Bi(mef)3(40µM, — ,n=95)ormefH(120µM, — ,n=90).(b)Unitvectornormalised,secondderivativeofthespectrashownin(a).

(a)

(b)

Appendix3

301

Appendix3.10.(a)Averaged,baselinecorrected(rubberbandbaselinecorrected,64baselinepoints),vectornormalisedHCT-8cellspectrafollowing4-hincubationinvehicle(RPMI-1640,DMSO2%v/v,— ,n=59),Bi(difl)3(75µM, — ,n=34)ordiflH(225µM, — ,n=41).(b)Unitvectornormalised,secondderivativeofthespectrashownin(a).

Appendix3

302

Appendix3.11:(a)Averaged,baselinecorrected(rubberbandbaselinecorrected,64baselinepoints), vector normalised HCT-8 cell spectra following 24-h incubation in vehicle(RPMI-1640,DMSO2%v/v,— , n=54),Bi(difl)3 (75µM, — , n=66)ordiflH (225µM, — ,n=38).(b)Unitvectornormalised,secondderivativeofthespectrashownin(a).

Appendix3

303

Appendix 3.12: Percentage of early apoptotic and late apoptotic cells based on flowcytometric analysis of cell death in the adherent only HCT-8 cells treated for 24 h withvehicle(RPMI-1640,2%v/vDMSO;Control),BiNSAIDs(Bi(tolf)3(15µM),Bi(mef)3(40µM)orBi(difl)3(75µM));ortherespectiveequimolarfreeNSAIDs.Eachsegmentrepresentsthemean±s.d.(n=3replicatesfromoneexperiment).Statisticalsignificanceofthepopulationofdyingcells(encompassingearlyapoptotic,lateapoptoticandnecroticcellpopulations)comparedtothe control is indicated by *** = P ≤ 0.001. Statistical significance of the BiNSAID-treatedsamplescomparedtotherespectiveNSAID-treatedsamplesisindicatedby‡=P≤0.001.Theviable(AnnV−/7AAD−)andearlynecrotic(AnnV−/7AAD+)cellpopulationshavebeenomittedfromthegraphforclarity.

Appendix4

304

Appendix4.AdditionalESI-MS/MSresultsAppendix4.1:AnalytesidentifiedabovetheLODinHCT-8celllipidextractstreatedwithvehicle(2%v/vDMSO,24-h)usingESI-MS/MS.

Analyte Analyte Analyte Precursorm/z184.1 m/z Precursorm/z184.1 m/z Neutralloss141.0Da m/zPC27:0 664 PCO-38:4 794 PEO-32:0 676PC28:0 678 PCO-38:3/37:4 796 PE32:1 690PC29:0 692 PCO-38:2p 798 PEO-34:1 702PC30:1 704 PCO-38:1p 800 PEO-34:0/33:1 704PC30:0 706 PCO-38:0p 802 PE34:2 716PCO-32:1 716 PC38:6 806 PE34:1 718PCO-32:0p 718 PC38:5 808 PEO-36:4 724PC31:0 720 PC38:4 810 PEO-36:3 726PC32:2 730 PC38:3 812 PEO-36:2/35:3 728PC32:1 732 PC38:2 814 PEO-36:1/35:2 730PC32:0 734 PC38:1 816 PEO-36:0/35:1 732PCO-34:2 742 PCO-40:5 820 PE36:4 740PCO-34:1 744 PCO-40:4 822 PE36:3 742PCO-34:0/33:1 746 PCO-40:3 824 PE36:2 744PC33:0 748 PCO-40:0 830 PE36:1 746PC34:3 756 PC40:7 832 PEO-38:6/36:0 748PC34:2 758 PC40:6 834 PEO-38:5 750PC34:1 760 PC40:5 836 PEO-38:4 752PC34:0 762 PC40:2 842 PEO-38:3/37:4 754PCO-36:4 766 PE38:6 764PCO-36:3 768 Neutralloss185.0Da m/z PE38:5 766PCO-36:2/35:3 770 PS32:1 734 PE38:4 768PCO-36:1/35:2 772 PS33:1 748 PE38:3 770PCO-36:0/35:1 774 PS34:1 762 PEO-40:6/38:0 776PC35:0 776 PS35:1 776 PEO-40:5 778PC36:6 778 PS36:2 788 PEO-40:4 780PC36:5 780 PS36:1 790 PE40:6 792PC36:4 782 PS38:4 812 PE40:5 794PC36:3 784 PS38:3 814 PC36:2 786 PS38:2 816 PC36:1 788 PS38:1 818 PCO-38:6/36:0 790 PS40:6 836 PCO-38:5 792 PS40:5 838 PS40:1 846

Appendix5

305

Appendix5.AdditionalMTTassayresults

Appendix 5.1: Concentration-response curves obtained from MTT assays of HCT-8 cellstreated (24 h) with the specified aminoarenesulfonic acids and corresponding Bi(III)complexes: (a) oAB-SO3H or Bi(oAB-SO3)3; (b) pAB-SO3H or Bi(pAB-SO3)3; (c)2AN-SO3H orBi(2AN-SO3)3; and (d) 5AN-SO3H or Bi(5AN-SO3)3. Error bars represent the standarddeviation from themean (n = 6). One representative of three independent experiments isshown.

In vitro investigations of the potential chemotherapeutic and ...

Documents

Transcript of In vitro investigations of the potential chemotherapeutic and ...