Serum insensitive, intranuclear protein delivery by the multipurpose cationic lipid S aint-2

Regulation and epigenetic modulation ofEpCAM gene expression in ovarian cancer

Ieneke van der Gun

Cover: Immuno�uorescence staining for EpCAM on ovarian cancer cells

Coverdesign: Peter van der Sijde, B.T.F. van der Gun

Lay-out: Peter van der Sijde, Groningen

Printed by: Drukkerij van Denderen, Groningen

ISBN: 978-90-367-4467-6 (printed version)ISBN: 978-90-367-4466-9 (digital version)

© 2010, B.T.F. van der GunNo parts of this thesis may be reproduced or transmitted in any forms or by any means, electronicor mechanical, including photocopying, recording or any information storage and retrieval system,without permission of the author and the publisher holding the copyrights of the articles.

The research presented in this thesis was �nancially supported by grants from the European Commis-

sion’s Fifth and Sixth Framework Program (Contract QRLT-2001-0448 and COOP-CT-2005-017984).

The author gratefully acknowledges the �nancial support for printing of this thesis by:

RIJKSUNIVERSITEIT GRONINGEN

Regulation and epigenetic modulation ofEpCAM gene expression in ovarian cancer

Proefschrift

ter verkrijging van het doctoraat in deMedische Wetenschappen

aan de Rijksuniversiteit Groningenop gezag van de

Rector Magni!cus, dr. F. Zwarts,in het openbaar te verdedigen op

maandag 13 september 2010om 13.15 uur

door

Bernardina Theresia Francisca van der Gun

geboren op 12 juli 1960te Schalkwijk

Promotores: Prof. dr. M.G. RotsProf. dr. L.F.M.H. de Leij

Copromotores: Dr. P.M.J. McLaughlinDr. M.H.J. Ruiters

Beoordelingscommissie: Prof. dr. A. KissProf. dr. H.J. HaismaProf. dr. H. Hollema

Paranimfen: Alice Arendzen

Margo Takke

CONTENTS

Chapter 1 General introduction to the thesis 9

Chapter 2 EpCAM in carcinogenesis: the Good, the Bad or the Ugly 17

Chapter 3 Transcription factors and molecular epigenetic marks associated with 39the EpCAM gene in ovarian cancer

Chapter 4 Serum insensitive, intranuclear protein delivery by the multipurpose 57cationic lipid Saint-2J Control Release 2007;123(3) 228-38

Chapter 5 Persistent down-regulation of the pancarcinoma-associated Epithelial 77Cell adhesion Molecule via active intranuclear methylationInt J Cancer 2008;123(2) 484-9

Chapter 6 Targeted DNA methylation by a DNA methyltransferase coupled to a 91Triple helix Forming Oligonucleotide to downregulate the EpithelialCell Adhesion MoleculeBioconjugate Chem. 2010; in press

Chapter 7 Sustained downregulation of EpCAM gene expression by siRNA 105targeting a coding region

Chapter 8 Summary & General discussion and perspectives 117

Nederlandse samenvatting 129

Dankwoord 137

Publications 141

9

General introduction to the thesis

Chapter 1

10

Figure 1. Principle of gene speci!c silencing via targeted DNA methylation. Gene directed methylation canbe induced by a DNA methyltransferase coupled to a sequence speci!c DNA binding domain. The gene speci!cDNA binding domain targets the DNA methyltransferase to a speci!c location in the gene. Subsequently site-directed DNA methylation can take place, inducing silencing of the gene.

The Epithelial Cell Adhesion Molecule and cancer

Initially, EpCAM was known to be a tumor associated antigen because of its frequent discovery in a

wide variety of carcinomas.4 However, it appears that EpCAM is also present on most normal epithelial

cells and that it plays a role in many ’normal’processes. During embryogenesis, EpCAM is involved in

the process of maturation and di"erentiation.5 In addition, the upregulation of EpCAM expression on

epithelial cells observed during in#ammatory processes5 suggests a role of EpCAM in regeneration

after tissue damage, i.e. cell proliferation and di"erentiation. Since it is clear that changes in EpCAM

expression are associated with development and regeneration of epithelia, one can envision that

Chapter 1

1

In the industrialized world, one in three people will develop cancer during their lifetimes. Despite

the progress that has been made in reducing incidence and mortality rates, cancer is still the second

leading cause of death in industrialized countries. Although a lot has been discovered about the

molecular mechanisms underlying the development and progression of cancer, a great deal is still

poorly understood. During the last decade it became evident that epigenetic dysregulation plays a

role as equally important as genetic changes like mutations. In this respect, the frequently observed

overexpression of the Epithelial Cell Adhesion Molecule (EpCAM) on most carcinomas1 seems to

be associated with epigenetic changes without underlying genetic defects. Moreover, the EpCAM

protein has recently regained interest as a marker for cancer stem cells.2 Since cancer stem cells

expressing EpCAM are more tumorigenic than EpCAM negative stem cells3, and because cancer

stem cells are radiation and drug resistant, downregulation of EpCAM by epigenetic interference

might be a promising approach to decrease oncogenic potential of tumor cells in a permanent way.

Aim of this thesis

The aim of this thesis was to de!ne and interfere in the (epi)genetic regulation of the gene that

code for the EpCAM protein. We designed a novel strategy based on epigenetic rewriting to silence

EpCAM expression in a permanent manner as indicated in Figure 1.

DNA Methyltransferase

Target DNA

sequence specific DNA-binding

domain

permanent genesilencing

spreading

11

1

Introduction

defects in the regulation of EpCAM expression may induce tissue remodeling. If this is chronically

activated this may eventually lead to cancer. Currently, the overexpression of EpCAM on most

carcinoma types is used as a diagnostic marker with some prognostic signi�cance6 and as a target

in antibody based clinical trials.4;7 The emerging function of EpCAM in cell proliferation, migration

and possibly cancer initiation8 broadens the interest to use EpCAM not only as an immunotarget

but also as a target for gene silencing. Especially, its observed expression on cancer stem cells from

pancreas9, breast10, liver11, colon12 and prostate3 tumors, could be taken as an indication that such

intervention could stop tumor initiation. Since transient downregulation of EpCAM expression has

been shown to decrease oncogenic potential of tumors cells11;13-16, permanent downregulation of

EpCAM expression in vivo might add a promising therapeutic approach to the current options for

treatment of cancer.

Epigenetics and cancer

Epigenetics is de�ned as a heritable change in gene expression that does not involve any changes

in the DNA sequence.17 It is well established nowadays, that epigenetic alterations play important

roles in the development of cancer.17;18 Epigenetic factors that control gene expression include

nucleosome positioning, microRNAs, DNA methylation and histone modi�cations. Positioning of

nucleosomes along the DNA regulates gene expression by altering the accessibility of regulatory

DNA sequences to transcription factors.19 MicroRNAs (miRNAs) are mostly short non-coding RNAs

that regulate gene expression by binding to a “seed sequence” in the 3’-untranslated region of the

target mRNA resulting in degradation of the mRNA or inhibition of its translation. However, binding

to the 5’-untranslated region and to coding regions has also been described.20 Alterations in the

level or type of miRNA expression can cause for example activation of oncogenes. Like normal

genes, the expression of miRNAs can be regulated by epigenetic mechanisms. In this thesis we focus

on epigenetic control by DNA methylation and histone modi�cations.

DNA methylation occurs in cytosines preceding guanines, the so called CpG dinucleotides.

CpGs are relatively rare in the genome but tend to cluster in islands which are usually located in the

5’-regulatory region of many genes. Methylation of CpG islands in promoters leads to transcriptional

silencing of genes. Both, hypomethylation as well as hypermethylation of DNA are associated with

carcinogenesis. Failure to repress tissue-restricted genes appropriately by loss of DNA methylation

could result in loss of tissue speci�city and promote cancer formation.21 For example, a number of

protein coding genes are overexpressed in ovarian cancer due to loss of DNA methylation including

synuclein-γ22, claudin-323 and claudin-4.24 At the other hand, hypermethylation of normally

unmethylated tumor suppressor genes correlates with a loss of their expression in cancer cell lines

and in primary tumors.17 DNA methylation together with histone modi�cations a"ect chromatin

structure and thereby in#uence accessibility of regulatory proteins. The DNA is wrapped around an

octamer of four histone proteins (H2A, H2B, H3, H4). The histones are subjected to a variety of post-

translational modi�cations including phosphorylation, acetylation, ubiquitylation and methylation.

12

Chapter 1

1

With regard to transcriptional regulation, increases in histone acetylation are generally associated

with increased gene expression, while the e�ect of histone methylation depends on the a�ected

amino acid and its position in the histone tail. For example trimethylation of lysine 4 of histone

3 is associated with transcriptional activation, whereas trimethylation of lysine 27 of histone 3 is

associated with transcriptional silencing. Many more histone modi�cations have been characterized

as repressive or active marks.25

In contrast to cancer-associated genetic mutations, epigenetic modi�cations are in principle

reversible. Inhibitors of DNA methylation and histone deacetylases have been investigated to

induce re-expression of tumor suppressor genes and reversal of malignant phenotypes. The drugs

decitabine and vorinostat, a DNA methylation and histone deacetylase inhibitor, respectively, are

both US Food and Drug Administration approved for the treatment of hematologic malignancies.26;27

Many clinical trials suggest that epigenetic drugs in combination with chemotherapy or other

biological agents could improve current anti-cancer therapies.27;28 However, because of the genome

wide e�ects of epigenetic drugs, the concomitant induction of side e�ects is dose limiting. This

dose limited toxicity might be overcome by targeting of the epigenetic modi�er to a speci�c gene

achieved by epigenetic editing.

Epigenetic editing and cancer

Because of the aberrant expression of EpCAM on many carcinomas, frequently associated with

increased oncogenic potential as possibly originating from the tumorigenic features of EpCAM

expressing cancer stem cells, downregulation of EpCAM by epigenetic interference might be a

promising approach to decrease oncogenic potential of tumor cells.

In contrast to short-interference (si)RNA-mediated silencing, epigenetic interference is expected

to be long-lasting. SiRNA-mediated silencing is based on induction of sequence speci�c cleavage

of perfectly complementary messenger RNA. Because mRNA molecules are constantly produced,

this way of silencing is transient. Upon activation by double-stranded (ds)RNA that is processed into

siRNA duplexes by Dicer29, the duplexes are incorporated into Argonaute 2 (Ago 2) and the RNA-

Induced Silencing Complex (RISC). One strand of the siRNA is eliminated by Ago 2, while the other

strand is used to recognize the target RNA, and the perfect base-pairing then allows cleavage of the

target mRNA in the cytoplasm by Ago 2.

Epigenetic editing is a strategy resulting in long-term gene speci�c silencing. Using this

approach, molecular epigenetic marks are overwritten by an e�ector domain targeted to speci�c

genes by a sequence speci�c DNA-binding motif. As e�ector domains, histone modi�ers like the

histone methyltransferases SUV39H1 and G9a or DNA methyltransferases DNMT1, DNMT3a and 3b

can be used to induce gene silencing. In this thesis we make use of the prokaryotic DNA (cytosine-5)

methyltransferase, which methylates cytosines in CpG dinucleotides.30 Since M.SssI has the same

base and sequence speci�city as mammalian DNA methyltransferases like the above mentioned

DNMTs, this enzyme is an excellent tool to study the e�ect of DNA methylation. Gene silencing

13

1

Introduction

via DNA methylation requires in principle only one treatment because the DNA methyltransferase

present in the cell will copy the new methylation mark with each replication cycle. To demonstrate

this permanent downregulation on EpCAM expression via DNA methylation, we compared the

e�ect of a single delivery of M.SssI with a single transfection with EpCAM speci�c siRNA.

To direct M.SssI to the EpCAM gene, we made use of a Triple helix-Forming Oligonucleotide (TFO)

which can form a triple helix with the target double strand DNA31. TFOs bind in the major groove to

the polypurine strand of double-stranded DNA via Hoogsteen hydrogen bonds either in a parallel or

in an anti-parallel fashion. Triple helix-forming oligonucleotides have been used as a tool to inhibit

transcription initiation or elongation by preventing binding of transcription factors to the DNA or by

blocking the RNA polymerase complex. TFOs have also been used to target cleaving, cross-linking

reagents or anticancer agents to unique target sequences.31

In order to achieve chemically coupling of the M.SssI to the TFO, a C-terminal 6xHis-cysteine

tag was introduced and the two internal cyteines were replaced resulting in C141S and C368A

substitutions. One of the replaced cysteines (C141S) is the active site cysteine, which is an additional

advantage since the C141S mutation decreases the enzyme activity to 2-5% of M.SssI.32 In this way,

the binding speci�city of the TFO-C141S conjugate to the DNA is expected to be dominated by the

TFO instead of by the recognition site of C141S.

Permanent gene speci�c silencing by siRNAs

An alternative approach to induce targeted DNA methylation is via siRNA designed to target

promoters.33 The mechanisms by which siRNA triggers DNA methylation or histone modi�cations

in human cells are still being de�ned.34 It has been shown that siRNA-mediated transcriptional gene

silencing (TGS) in human cells, induced an increase in the repressive histone modi�cations: H3K9

and H3K27 methylation.35 The antisense strand siRNA interacted with a complex that contained

DNA methyltransferase DNMT3a, H3K27me3 and the targeted promoter.36 Several putative

mechanisms for siRNA mediated TGS are described.33 SiRNAs can induce both methylation of DNA

and histones, although it is not known which occurs �rst, siRNA-mediated TGS is likely to be involved

in a chromatin remodeling complex as methylation and deacetyltransferase inhibitors reversed the

siRNA-mediated TGS.37

Outline of this thesis

In Chapter 2, we �rst present a comprehensive introduction to the biological role of EpCAM in

carcinogenesis, tumor progression and metastasis in a broad range of carcinoma types. Furthermore,

we summarize current literature regarding transcription factors and the (epi)genetic regulation of

the EpCAM gene to identify potent regulatory factors which play a role in endogenous EpCAM gene

expression. As the epigenetic regulation of EpCAM was unknown for ovarian cancer, we de�ned

the DNA methylation level and the histone modi�cations associated with the EpCAM promoter for

14

Chapter 1

1

a panel of ovarian cancer cell lines with di�erential EpCAM expression (Chapter 3). In addition, we

analyzed whether transcription factors described to play a role in ovarian cancer are associated with

the EpCAM gene in living cells.

Since we aimed to deliver a DNA methyltransferase coupled to a Triple helix Forming

Oligonucleotide into tumor cells, we tested in Chapter 4 whether the non-viral gene transfer vector

SAINT-2:DOPE is also able to deliver functional proteins. Importantly, at that time no other agent was

available to deliver functional active proteins in the presence of serum, which is a serious limitation

for in vivo application. Since DNA methyltransferases need to enter the nucleus to exert their

function, we examined whether the deliver agent SAINT-2:DOPE is also capable of nuclear functional

protein delivery. To this extend, we explored if delivery of the DNA methyltransferase M.SssI by

SAINT-2:DOPE results in active methylation of the EpCAM gene leading to downregulation of EpCAM

expression (Chapter 5). Alternatively, we investigated whether the EpCAM promoter is sensitive for

DNA methylation by examining if the DNA methylation inhibitor 5-aza-2’-deoxycytidine induces or

upregulates endogenous EpCAM expression.

Two di�erent approaches to achieve targeted gene speci!c silencing are presented in Chapter

6 and 7. First, we explored whether treatment with a TFO targeting domain, conjugated to the

mutated DNA methyltransferase C141S results in methylation of the target CpGs (Chapter 6).

Subsequently, the functional e�ect of such targeted CpG methylation on EpCAM promoter activity

was investigated by reporter gene expression. To provide insights in two alternative gene silencing

approaches, we compared genome-wide inheritable DNA methylation with gene speci!c transient

siRNA treatment in a subset of cells over time (Chapter 7).

Finally, the results described in this thesis are summarized and discussed, and future perspectives

are outlined in Chapter 8.

15

1

Introduction

References

1. Went PTH, Lugli A, Meier S, Bundi M, Mirlacher M, Sauter G et al. Frequent EpCAM protein expression inhuman carcinomas. Human Pathology 2004;35:122-8.

2. Visvader JE, Lindeman GJ. Cancer stem cells in solid tumors: accumulating evidence and unresolvedquestions. Nature Rev.Cancer 2008;8:755-68.

3. Gires O, Klein CA, Baeuerle PA. On the abundance of EpCAM on cancer stem cells. Nat Rev Cancer2009;9:143.

4. Baeuerle PA, Gires O. EpCAM (CD326) !nding its role in cancer. Br.J Cancer 2007;96:417-23.

5. Trzpis M, McLaughlin PMJ, de Leij LMFH, Harmsen MC. Epithelial Cell Adhesion Molecule: More than aCarcinoma Marker and Adhesion Molecule. American Journal of Pathology 2007;171:386-95.

6. Went P, Dirnhofer S, Schopf D, Moch H, Spizzo G. Expression and prognostic signi!cance of EpCAM.J.Cancer Mol 2008;3:169-74.

7. Seimetz D, Lindhofer H, Bokemeyer C. Development and approval of the trifunctional antibodycatumaxomab (anti-EpCAMá+áanti-CD3) as a targeted cancer immunotherapy. Cancer TreatmentReviewsIn Press, Corrected Proof.

8. Munz M, Baeuerle PA, Gires O. The emerging role of EpCAM in cancer and stem cell signaling. CancerResearch 2009;69:5627-9.

9. Li C, Heidt DG, Dalerba P, Burant CF, Zhang L, Adsay V et al. Identi!cation of Pancreatic Cancer Stem Cells.Cancer Research 2007;67:1030-7.

10. Al-Hajj M, Wicha MS, ito-Hernandez A, Morrison SJ, Clarke MF. Prospective identi!cation of tumorigenicbreast cancer cells. Proc.Natl Acad.Sci.USA 2003;100:3983-8.

11. Yamashita T, Ji J, Budhu A, Forgues M, Yang W, Wang HY et al. EpCAM-Positive Hepatocellular CarcinomaCells Are Tumor-Initiating Cells With Stem/Progenitor Cell Features. Gastroenterology 2009;136:1012-24.

12. Dalerba P, Dylla SJ, Park IK, Liu R, Wang X, Cho RW et al. Phenotypic characterization of human colorectalcancer stem cells. Proceedings of the National Academy of Sciences 2007;104:10158-63.

13. Du W, Ji H, Cao S, Wang L, Bai F, Liu J et al. EpCAM: A Potential Antimetastatic Target for Gastric Cancer.Digestive Diseases and Sciences 2009;x:1033-8.

14. Munz M, Kieu C, Mack B, Schmitt B, Zeidler R, Gires O. The carcinoma-associated antigen EpCAMupregulates c-myc and induces cell proliferation. Oncogene 2004;23:5748-58.

15. Osta WA, Chen Y, Mikhitarian K, Mitas M, Salem M, Hannun YA et al. EpCAM is overexpressed in breastcancer and is a potential target for breast cancer gene therapy. Cancer Res. 2004;64:5818-24.

16. Yanamoto S, Kawasaki G, Yoshitomi I, Iwamoto T, Hirata K, Mizuno A. Clinicopathologic signi!cance ofEpCAM expression in squamous cell carcinoma of the tongue and its possibility as a potential target fortongue cancer gene therapy. Oral Oncology 2007;43:869-77.

17. Esteller M. Epigenetics in cancer. N.Engl.J Med. 2008;358:1148-59.

18. Sharma S, Kelly TK, Jones PA. Epigenetics in cancer. Carcinogenesis 2010;31:27-36.

19. Jiang C, Pugh BF. Nucleosome positioning and gene regulation: advances through genomics3. Nat.Rev.Genet. 2009;10:161-72.

20. Valeri N, Vannini I, Fanini F, Calore F, Adair B, Fabbri M. Epigenetics, miRNAs, and human cancer: a newchapter in human gene regulation3. Mamm.Genome 2009;20:573-80.

21. De SC, Loriot A. DNA hypomethylation in cancer: Epigenetic scars of a neoplastic journey. Epigenetics.2010;5.

16

Chapter 1

1

22. Gupta A, Godwin AK, Vanderveer L, Lu A, Liu J. Hypomethylation of the Synuclein{gamma} Gene CpGIsland Promotes Its Aberrant Expression in Breast Carcinoma and Ovarian Carcinoma. Cancer Research2003;63:664-73.

23. Honda H, Pazin MJ, D’Souza T, Ji H, Morin PJ. Regulation of the CLDN3 gene in ovarian cancer cells. CancerBiol.Ther. 2007;6:1733-42.

24. Barton CA, Hacker NF, Clark SJ, O’Brien PM. DNA methylation changes in ovarian cancer: Implications forearly diagnosis, prognosis and treatment. Gynecologic Oncology 2008;109:129-39.

25. Kouzarides T. Chromatin Modi!cations and Their Function. Cell 2007;128:693-705.

26. Jabbour E, Issa JP, Garcia-Manero G, Kantarjian H. Evolution of decitabine development:accomplishments, ongoing investigations, and future strategies. Cancer 2008;112:2341-51.

27. Lane AA, Chabner BA. Histone deacetylase inhibitors in cancer therapy. J Clin.Oncol. 2009;27:5459-68.

28. Matei DE, Nephew KP. Epigenetic therapies for chemoresensitization of epithelial ovarian cancer.Gynecologic Oncology 2010;116:195-201.

29. de FA, Vornlocher HP, Maraganore J, Lieberman J. Interfering with disease: a progress report on siRNA-based therapeutics. Nat Rev.Drug Discov. 2007;6:443-53.

30. Darii MV, Cherepanova NA, Subach OM, Kirsanova OV, Rasko T, Slaska-Kiss K et al. Mutational analysis ofthe CG recognizing DNA methyltransferase SssI: Insight into enzyme-DNA interactions. Biochim.Biophys.Acta 2009;1794:1654-62.

31. Duca M, Vekho" P, Oussedik K, Halby L, Arimondo PB. The triple helix: 50 years later, the outcome. NucleicAcids Research 2008;36:5123-38.

32. Rathert P, Rasko T, Roth M, Slaska-Kiss K, Pingoud A, Kiss A et al. Reversible inactivation of the CG speci!cSssI DNA (cytosine-C5)-methyltransferase with a photocleavable protecting group. Chembiochem.2007;8:202-7.

33. Kawasaki H, Taira K, Morris KV. siRNA induced transcriptional gene silencing in mammalian cells17. Cell Cycle 2005;4:442-8.

34. Morris KV. siRNA-mediated transcriptional gene silencing: the potential mechanism and a possible role inthe histone code. Cell Mol.Life Sci 2005;62:3057-66.

35. Verdel A, Vavasseur A, Le GM, Touat-Todeschini L. Common themes in siRNA-mediated epigeneticsilencing pathways. Int.J Dev.Biol 2009;53:245-57.

36. Weinberg MS, Villeneuve LM, Ehsani A, Amarzguioui M, Aagaard L, Chen ZX et al. The antisense strandof small interfering RNAs directs histone methylation and transcriptional gene silencing in human cells.RNA. 2006;12:256-62.

37. Morris KV, Chan SW, Jacobsen SE, Looney DJ. Small interfering RNA-induced transcriptional genesilencing in human cells. Science 2004;305:1289-92.

17

Chapter 2

EpCAM in carcinogenesis: the Good, the Bad or the Ugly

Bernardina T.F. van der Gun1, Lieuwe J. Melchers2,3 , Marcel H.J. Ruiters4, Lou F.M.H. de Leij1,Pamela M.J. McLaughlin1, and Marianne G. Rots1

1 Epigenetic Editing, Dept. of Pathology and Medical Biology, University Medical Center Groningen, TheNetherlands

2 Dept. of Pathology and Medical Biology, University Medical Center Groningen, The Netherlands3 Dept. of Oral & Maxillofacial Surgery, University Medical Center Groningen, The Netherlands4 Synvolux Therapeutics Inc., Groningen, The Netherlands

provisionally accepted by Carcinogenesis

18

Chapter 2

2

ABSTRACT

The Epithelial Cell Adhesion Molecule (EpCAM) is a membrane glycoprotein that is highly expressed

on most carcinomas and frequently used as a diagnostic and prognostic marker for a variety of

carcinomas. Moreover, EpCAM has been identi�ed as a marker of cancer initiating cells. In this review,

we summarize the evidence for the double role of EpCAM in carcinogenesis: as an adhesion molecule

it mediates homophilic adhesion interactions thereby preventing metastasis. In addition, EpCAM

abrogates E-cadherin mediated cell-cell adhesion and thus might, in association with claudin-7,

promote metastasis. Upon cleavage of the extracellular domain of EpCAM, the intracellular domain

functions as a part of a transcriptional complex inducing c-myc and cyclin A and E. In line with this

seemingly controversial role, EpCAM overexpression has been associated with both decreased and

increased overall survival of patients. Modulation of EpCAM expression also re�ects its dual role as,

depending on the cell type, either induction or downregulation of EpCAM lowers the oncogenic

potential. As epigenetic dysregulation seems to underlie aberrant EpCAM expression, epigenetic

editing provides a novel approach to investigate the biological role of EpCAM and might expand the

options for EpCAM as a therapeutic target in cancer.

19

2


INTRODUCTION

The Epithelial Cell Adhesion Molecule (EpCAM) is a transmembrane glycoprotein originally

discovered on colon carcinomas.1 In healthy individuals, EpCAM expression is restricted to most

normal epithelia, but in most human carcinomas EpCAM is overexpressed to varying degrees.2 The

diagnostic and prognostic characteristics of EpCAM have been demonstrated by many independent

research groups2;3 and the EpCAM overexpression is exploited in several EpCAM directed antibody

or vaccine based clinical trials for a wide variety of carcinomas.4 Recently, EpCAM has been identi�ed

as a marker for cancer initiating stem cells5, which makes it an even more interesting target for

cancer therapy.

The biological function of EpCAM is largely unknown. EpCAM is able to abrogate E-cadherin-

mediated cell-cell adhesion by disrupting the link between α-catenin and F-actin thereby

loosening cell-cell adhesion.6 In addition, association of EpCAM with claudin-7 interferes with

EpCAM-mediated homotypic cell-cell adhesion, promoting cell motility, proliferation, survival,

carcinogenesis and metastasis formation.7 Furthermore, it has been shown that upon cleavage of

the extracellular domain of EpCAM, the intracellular domain functions as part of a transcriptional

complex inducing c-myc and cyclin A and E expression.8 These �ndings support a role for EpCAM

as an oncogene. Indeed, EpCAM overexpression is associated with decreased overall survival

of patients with di�erent types of cancer.9-12 In contrast to its promoting role regarding tumor

formation, EpCAM is also described as a tumor protecting protein. EpCAM was proposed to function

as a cell adhesion molecule since EpCAM is able to mediate homophilic adhesive interactions13,

thereby preventing cell-scattering. Due to these adhesive properties, EpCAM is likely to play a role

in inhibition of invasion.13;14 Indeed, loss of EpCAM contributed to increased migratory potential15

and EpCAM expression on metastases was lower compared to primary tumors.16 Moreover, EpCAM

overexpression in some carcinoma types is associated with improved patient survival.17-21

Mechanistic studies to investigate the role of EpCAM by enforced modulation of EpCAM

expression also re�ect the dual role of EpCAM. Murine colorectal carcinoma cells transfected with

murine EpCAM cDNA increased cell-cell adhesion, attenuated tumor cell invasion in matrigel and

decreased tumor incidence and metastasis when inoculated in the spleen of the mice.14 These data

suggest that EpCAM expression antagonizes tumor growth and metastasis. In contrast, induction of

EpCAM expression into human epithelial kidney cells as well as into murine �broblast showed an

enhanced metabolism and colony formation capacity compared with the empty vector transfected

cells.22 Furthermore, in four di�erent carcinoma types, downregulation of EpCAM expression

utilizing antisense or siRNA, decreased cell proliferation, migration and invasiveness.22-26

Based on the above observations, EpCAM plays a paradoxical role in carcinogenesis, acting as

a tumor suppressive gene or as an oncogene, possibly depending on the microenvironment. Since

epigenetic regulation is associated with aberrant EpCAM expression, recent advances in epigenetic

interference27;28 might be a promising novel approach to either up- or downregulate EpCAM

20

Chapter 2

2

expression, depending on the tumor type. This review comprehensively summarize the current

knowledge, showing a gradual change in scienti�c opinion from studies proposing a protective role

of EpCAM to studies proposing a promoting role in carcinogenesis. We describe the (epi)genetic

events involved in EpCAM regulation and discuss the carcinoma types which might bene�t from

future (epi)genetic therapy, either inducing or repressing EpCAM expression.

1. Biological role of EpCAM in carcinogenesis

The highly overexpressed tumor-associated antigen on carcinomas, currently referred to as EpCAM,

has been discovered multiple times.4With each discovery, EpCAM received the name of the respective

monoclonal antibody or cDNA clone, leading to many synonyms: Epithelial GlycoProtein-2 (EGP-

2), Epithelial Speci�c Antigen (ESA), GA733-2 and Tumor-Associated Calcium Signal Transducer 1

(TACSTD1) among others.4 EpCAM localizes to the basolateral membrane in normal epithelial tissue,

whereas EpCAM displays di!erent distributions in carcinoma.29 EpCAM is a transmembrane protein

consisting of an extracellular domain (EpEX), a single transmembrane domain and a short 26-amino

acid intracellular domain (EpICD).4 The extracellular domain comprises an epidermal growth factor-

like (EGF) domain, a thyroglobulin repeat domain followed by a cysteine-poor domain. The EGF-

like and thyroglobulin domains form a globular structure and are required for the homophilic

cell-cell adhesion of EpCAM.30 EpCAM mediated homotypic cell-cell adhesion is also in"uenced by

association with the tight-junction protein claudin-7, the variant isoform of the cell-matrix adhesion

protein CD44v6 and the tetraspanin CD9, which complex facilitates metastasis formation.31 EpCAM is

capable of abrogating the E-cadherin mediated adhesions and to rearrange the cytoskeleton of the

cell.6 EpCAM has found to be hyperglycosylated in carcinoma tissue as compared with autologous

epithelia.32;33

Recently, EpCAM was identi�ed as a signal transducer8: upon cleavage of EpEX, EpICD associates

with the adaptor protein FHL2 (four and a half LIM domain protein 2), β-catenin and the transcription

factor Lef-1, the complex binds to the DNA at the Lef-1 consensus site inducing c-myc and cyclin A

and E expression.8 The oncogenic potential of EpICD was demonstrated in a mouse xenograft model,

in which Hek293 cells stably expressing EpCAM or EpICD produced nearly equivalent tumors.8 In

addition, EpCAM has been identi�ed as a marker for cancer initiating stem cells. EpCAM expressing

pancreatic cancer stem cells showed a 100-fold enhanced tumorigenic potential compared with

EpCAM negative pancreatic cancer stem cells.34 Also, EpCAM positive hepatocellular carcinoma

stem cells, but not EpCAM negative hepatocellular carcinoma stem cells could e$ciently initiate

tumors in SCID mice.25 For breast cancer stem cells, both EpCAM positive and EpCAM negative cells

were able to form tumors in NOD/SCID mice, but for the EpCAM positive cells, tumor formation was

detected 2 weeks earlier compared to the EpCAM negative stem cells.35

Based on the above, and on the application of EpCAM as a diagnostic marker and a therapeutic

immune target, it is clear that EpCAM plays an important role in carcinogenesis; however, the exact

biological role is not clear. In some carcinoma types, EpCAM is the ”good guy”, being associated with

21

2


Table I. Protective (-) role of EpCAM in carcinomas.

Carcinoma Carcinogenesis Progression Metastasis Survival References

Lung metastasesof colon ca. -c,m Jojovic et al., 1998 36

Disseminated tumorCells (various ca.s) -p Rao et al., 2005 37

Renal Cell ca. -p -p 0p -p Seligson et al., 2004 20

0p -p 0p Went et al., 2005 16

0p -p -p Klatte et al., 2009 38

Thyroid ca. -p 0p -p Ensinger et al., 2006 17

– is a protecting role or longer survival associated with EpCAM expression. 0 is no (signi!cant) role found.p = in patient material; m = in mice/rats; c = in cell lines ca. = carcinoma

improved survival, whereas in other carcinoma types EpCAM is the ”bad guy” being associated with

decreased survival. Interesting, for several types of carcinoma both roles have been reported, which

makes EpCAM an ”ugly player” for the clinical setting.

1.1 EpCAM: the Good?

The name EpCAM re"ects its function as a homophilic intercellular adhesion molecule as

demonstrated by Litvinov et al in 1994.13 It has been suggested that the adhesive properties of

EpCAM might prevent metastasis, because intercellular adhesion should be reduced to gain the

ability to migrate. Indeed, in lung metastases of a colon cancer tumor in SCID mice, large metastases

displayed an equal level of EpCAM expression as the primary tumor, whereas small metastases in

the same mouse were EpCAM negative.36 Similarly, in patients with metastases from breast, lung,

prostate, colon, ovarian and bladder cancer, one log reduction in EpCAM expression in circulating

tumor cells was found compared with primary tumors and metastases, indicating that high EpCAM

expression prevents metastasis.37 In metastases of renal clear cell carcinomas, EpCAM expression

was diminished compared to primaries16 and high EpCAM expression in the primary was associated

with improved patient survival.20;38 Also, in thyroid carcinoma, the frequency of EpCAM expression

was lower in poorly di#erentiated compared to well di#erentiated thyroid carcinomas, and low

EpCAM expression correlated with a poor survival.17 In undi#erentiated, rapidly metastasizing

anaplastic carcinomas EpCAM was even absent. However, in this series of thyroid carcinoma there

was no association with the presence of metastases nor was the expression in metastases analyzed,

therefore loss of EpCAM might be an e#ect of general tumor dedi#erentiation rather than a cause.17

Studies supporting a protective role in carcinogenesis are summarized in Table I.

22

Chapter 2

2

1.2 EpCAM: the Bad?

In view of the reported actions of EpCAM, many studies report on EpCAM having a promoting

role in carcinogenesis (Table II). In contrast with the observation in thyroid carcinoma, in breast

carcinoma EpCAM was signi�cantly higher expressed in less di�erentiated tumors and associated

with worse survival.39 In a larger study by the same group, EpCAM expression was also associated

with nodal metastasis and larger tumors.11 Subgroup analysis revealed that high EpCAM expression

was an indication for poor prognosis in node-positive but not in node-negative breast cancer

patients, which implies that most of the negative e�ect of EpCAM is exerted when tumors cells

are migrating. Also in patients receiving adjuvant cytotoxic or hormonal therapy vs. untreated

patients, high EpCAM was a poor prognostic factor.11 Furthermore, EpCAM expression was higher

in metastases compared to the matched primary cancers of patients who died of widely metastatic

breast cancer.40 In cervical squamous epithelia EpCAM expression increased from low grade to high

grade intraepithelial neoplasia and correlated with an increased proliferation as demonstrated by

Ki-67 expression.41

Also in squamous cell carcinoma (SCC) of the lung, EpCAM has a tumor promoting role, supported

by the observation that high EpCAM expression was associated with nodal metastasis, high-stage

disease and poor di�erentiation.42 Although in this study EpCAM expression was not correlated

with patient survival, a more recent study found a trend toward a shorter survival in patients with

Table II. Promoting (+) role of EpCAM in carcinomas.


Gastrointestinal:Gallbladder ca. 0p +p Varga et al., 2004 12

Pancreatic ca. 0p +p Fong et al., 2008 10

+m Li et al., 2007 34

-disseminated cells +p +p Scheunemann et al, 2008Hepatocellular +c Yamashita et al., 2007 69

+c,m +c Yamashita et al., 2009 25

Lung:Lung SCC +p +p +p 0p Piyathilake et al., 2000 42

Breast:+p +c +p,c Osta et al., 2004 24

+c+m Al-Hajj et al., 2003 35

+p +p Gastl et al., 2000 39

+p +p +p Spizzo et al., 2004 11

+p Cimino et al., 2009 40

Gynecological:Cervix ca. +p Litvinov et al., 1996 41

Urological:Urothelial bladder ca. +p +p Brunner et al., 2008 9

+ is a promoting role or shorter survival associated with EpCAM expression. 0 is no (signi�cant) role found.p = in patient material; m = in mice/rats; c = in cell lines ca. = carcinoma

23

2


high EpCAM expression.43 The lack of a clear e�ect of EpCAM on survival in lung carcinoma might

re�ect the increased EpCAM-induced cell proliferation which could make EpCAM positive cells more

sensitive to radiation therapy, a modality frequently used in lung carcinoma. To our knowledge, as

yet no study has addressed this subject.

Furthermore, high EpCAM has been associated with decreased overall survival in carcinomas of

the bladder9, gall bladder12 as well as of the pancreas.10

1.3 EpCAM: the Ugly?

For the carcinoma types described thus far, EpCAM has either a protective or a promoting

role. However, in several tumor types, the role of EpCAM is far from clear (Table III). In gastric

adenocarcinoma, Songun et al reported in a highly selected group that patients with high EpCAM

expression had a signi!cantly better 10-year survival and that loss of EpCAM identi!ed aggressive

tumors in early stage disease21, whereas in other studies there were no signi!cant relations with

expression.43;44 In the studies of Du et al, survival was not analyzed, but EpCAM expression was

associated with nodal metastasis.23;45 Moreover, they showed that the proliferation marker PCNA

in gastric cancer tissue with high EpCAM expression was higher than those with low EpCAM

expression. In addition, the presence of EpCAM positive disseminated tumor cells in pathological

tumor-free lymph nodes was an independent prognostic factor for both a signi!cantly reduced

relapse-free survival and overall survival.46

Also for colorectal cancer contradictory results have been reported. A reduced EpCAM

expression at the invasive margin of rectal tumor specimens correlated signi!cantly with higher

extent of tumor budding, tumor grade and risk of local recurrence.15 Interestingly, this !nding was

associated with nuclear localization of β-catenin, consistent with the signal transducer function

of EpCAM. Only in a subgroup of moderately di�erentiated colon cancers a signi!cant positive

correlation of EpCAM expression with survival was found.43 For human colorectal cancer stem cells,

however, the capacity to form tumors in mice was restricted to stem cells expressing high levels

of EpCAM, whereas EpCAM low expressing cells failed to form tumors.47 Interestingly, in colorectal

cancer, the interaction of EpCAM with the cell matrix adhesion molecule CD44v6 and the tight

junction molecule claudin-7 were found in association with the tetraspanin CO-029 in tetraspanin-

enriched membrane microdomains. Co-expression and complex formation of these molecules was

accompanied by a signi!cantly decreased disease free survival.31 In addition, EpCAM proved to be a

good marker for RT-PCR based detection of colorectal cancer metastases in lymph nodes.48

In patients with adenocarcinomas of the lung, a trend towards a longer survival of patients with

high EpCAM expression was observed43, although Kim et al. could not con!rm this in a mixed group

of adenocarcinomas and less aggressive bronchioloalveolar carcinomas.49 Interestingly, a signi!cant

lower frequency of EpCAM positivity was found in the latter which might suggest a promoting role

for EpCAM in tumor progression.49 Moreover, also in non-small cell lung cancer (NSCLC) EpCAM has

been shown to be an accurate diagnostic marker for RT-PCR-based identi!cation of lymph node

24

Chapter 2

2

micro-metastasis.50 In addition, the presence of EpCAM positive tumor cells in lymph nodes of

patients with NSCLC correlated with reduced survival rates.51

In specimens of head and neck squamous cell carcinoma (HNSCC), EpCAM mRNA expression

is increasing from hyperplasia via dysplasia to tumor, which might suggest a role for EpCAM in

carcinogenesis.52 EpCAM mRNA is identi�ed as one of the best markers to detect HNSCC metastases

Table III. Protective (-) or promoting (+) role of EpCAM in carcinomas.


Head & Neck:Oral SCC 0p 0p 0p Laimer et al., 2008 55

+p 0p 0p Shiah et al., 2008 56

-p -p -p -p Hwang et al., 2009 18

Hypopharyngeal SCC +c Munz et al., 2004 22

Tongue SCC +p +p,c +p 0p Yanamoto et al., 2007 26

Gastrointestinal:Esophageal ca.

+p 0p 0p 0p Went et al., 2008 3

+p Stoeklein et al., 2006 60

+p -p -p Kimura et al., 2007 19

-disseminated cells +m +p,m +p Hosch et al., 2000 61

Gastric ca. +c +p,m,c Du et al., 2009 23,45

0p 0p Deveci et al., 2007 44

-p Songun et al., 2005 21

0p 0p Went et al, 2006 43

-disseminated cells +p +p +p Scheunemann et al., 2009 46

Colorectal ca. -p 0p -p/0p Went et al., 2006* 43

-p 0p Gosens et al., 2007 15

-m,c Basak et al., 1998 14

+m Dalerba et al., 2007 47

+p 0p 0p +p Kuhn et al., 2007§ 31

-disseminated cells +p Xi et al., 2006 48

Lung:Adenoca. +p 0p 0p 0p Kim et al., 2009 49

-p,c -p Tai et al., 2007 68

Various types 0p 0p 0p Went et al., 2006 43

-disseminated cells +p +p Kubuschok et al., 1999 51

Gynecological:Ovarian ca. +p 0p Heinzelmann et al,200462

+p 0p Kim et al., 2003 63

+p +p Spizzo et al., 2006 64

+p +p Bellone et al., 2009 65

Urological:Prostate ca.

+p 0p Poczatek et al., 1999 66

+p 0p 0p Zellweger et al., 2005 67

0p 0p 0p Went et al., 2006 43

+ means a promoting role in carcinogenesis (e.g. higher expression in tumor compared to normal), tumor pro-gression (higher in larger tumors), metastasis (higher in metastasized tumors) or shorter survival. – is a protect-ing role or longer survival associated with EpCAM expression. 0 is no (signi�cant) role found.p = in patient material; m = in mice/rats; c = in cell lines; * = e!ect on survival only in subgroup of moderatelydi!erentiated tumors; § = as complex.

25

2


to cervical lymph nodes53 and disseminated tumor cells in patients with HNSCC.54 EpCAM is

expressed de novo in HNSCC, but most studies do not !nd any relation with clinicopathologic

variables, including di"erentiation and survival.55-58 However, in a study looking speci!cally at

tongue SCC, EpCAM expression was associated with larger tumor size, nodal metastasis and tumor

dedi"erentiation.26 Interestingly, recently in a Taiwanese series of oral SCC, EpCAM expression was

reported to decrease from normal via dysplasia to carcinoma, and lower EpCAM labeling index

was associated with, amongst others, tumor size and nodal metastasis.18 The lack of consistent

association between EpCAM expression and HNSCC might be attributable to the heterogeneity

within the HNSCC.59

In a group of esophageal cancer (mainly SCC) patients, high EpCAM expression indicated a

signi!cantly higher survival rate.19 In contrast, Stoecklein et al.60 found in cryostat sections of

esophageal SCC that high EpCAM expression correlated with a signi!cant decreased median

relapse-free survival period and median overall survival. Moreover, multivariate analysis disclosed

high EpCAM expression as an independent prognostic factor. Went et al. found no correlation of

EpCAM expression with grade, stage or disease-speci!c survival in esophageal SCC.3 Furthermore,

the presence of EpCAM positive cells in lymph nodes, classi!ed as tumor free by histopathological

staging, was an independent indicator for a poor prognosis in patients with esophageal cancer.61

In epithelial ovarian cancer, EpCAM is highly overexpressed compared with normal ovarian sur-

face epithelium and no di"erences in EpCAM expression were observed among di"erent histologi-

cal subtypes and grades in two independent studies.62;63 In one of these studies, with almost half of

the tumors being of the borderline type (low malignant potential), FIGO stage III/IV showed lower

EpCAM expression than stage I.63 However, in the other study, FIGO stage III/IV showed signi!cant

higher EpCAM expression than stage I/II disease suggesting that a higher expression of EpCAM

correlates with tumor progression, but no correlation with relapse-free survival or disease-speci!c

survival was found.62 A more recent study reported that EpCAM overexpression was signi!cantly re-

lated to a decreased overall survival of patients with epithelial ovarian cancer, especially in patients

with FIGO stage III/IV.64 Furthermore, in this study they found a di"erence in EpCAM expression

among histological subtypes and a signi!cantly higher rate of EpCAM overexpression was observed

in poorly di"erentiated tumors. In addition, metastatic/recurrent tumors were found to express sig-

ni!cantly higher levels of EpCAM protein when compared with primary ovarian carcinomas.65

In the prostate, EpCAM expression was signi!cantly increased from normal via prostatic

intraepithelial neoplasia, to adenocarcinoma, but expression in adenocarcinoma was not associated

with di"erentiation grade or clinical outcome in pT2 tumors.66 A more recent study, in which

tumors with all T stages were analyzed con!rmed this lack of association.43 Interestingly, hormone-

refractory carcinomas were found to express EpCAM in a signi!cantly higher frequency than

untreated carcinomas2, but this !nding was not con!rmed in a study analyzing EpCAM expression

in hormone-refractory and metastatic tissue compared to localized prostate cancer.67 Overall in

prostate adenocarcinoma, there are no clear indications for a role for EpCAM in tumor progression,

but due to its early upregulation EpCAMmight have an e�ect in carcinogenesis.

Studies regarding the expression of EpCAM su�er from the use of di�erent antibodies, scoring

methods and heterogeneous groups of tumors analyzed, which makes comparing results di"cult.

Nevertheless, it isquiteclear thatEpCAMplaysa role,promotingor/andprotecting, incarcinogenesis,

tumorprogressionandmetastasis in various carcinoma types, providingopportunities fordiagnostic

and future therapeutic interventions.

2. Modulation of EpCAM expression to address the biological role of EpCAM

The function of EpCAM as an adhesion molecule was discovered by induction of EpCAM in non-

EpCAM expressing cells.13 Transfection of EpCAM murine cDNA in +broblast and mammary

carcinoma cell lines resulted in aggregates of cells caused by increasing intercellular adhesion.

Moreover, the EpCAM positive transfectants segregated from the EpCAM negative parental cells

and EpCAM expression inhibited invasive growth in cell colonies.

Modulation of EpCAM expression re,ects the dual role of EpCAM: evidence supporting the

protective role of EpCAM in carcinogenesis has been obtained by either induction of EpCAM

expression in colon or reduction of EpCAM in lung adenocarcinoma cell lines. Murine colorectal

carcinoma cells transfected with cDNA encoding the murine EpCAM showed signi+cant lower

growth rates, colony formation and invasion through matrigel in vitro compared with the vector-

only transfected cells.14 Also cells transfectedwith cDNA encoding humane EpCAM showed reduced

invasion through matrigel.14 In syngeneic immunode+cient and immunocompetent mice, the

EpCAM transfected murine colorectal cells showed a reduction in metastatic potential compared

to the control transfected cells. In a lung adenocarcinoma cell line, reduction of EpCAM expression

using shRNA, showed an elevated cell invasion.68

Modulation of EpCAM expression in gastric carcinoma only addresses the promoting role of

EpCAM, despite the possible dual role described in patients studies.21;23;45 Downregulation of EpCAM

by siRNA signi+cantly suppressed proliferation, colony formation, adhesiveness, invasiveness and

migration of gastric cancer cell lines.23;45 Furthermore, cells with lower EpCAM expression showed

a reduced tumor growth in nude mice45, and a tail vein metastatic assay showed that intravenous

Evidence supporting the promoting role of EpCAM in carcinogenesis has also been reported:

stable transfection of EpCAM cDNA in human embryonic kidney (Hek293) cells and murine

%broblasts cells resulted in a reduced requirement for growth factors, an increased metabolic

activity and formation of larger and more colonies compared to the empty vector transfected cells.

Moreover,EpCAM expressing Hek293 induced the expression of c-myc and cyclins A and E.22

The correlation between high EpCAM expression and poor prognosis in breast cancer patients is

re,ected by inhibition of proliferation, migration and invasion via silencing of EpCAM by siRNA

in breast cancer cell lines.24 In agreement with its promoting role in patients with hepatocellular

carcinoma and SCC of the tongue, siRNA-mediated EpCAM reduction in cell lines decreased the

invasion potential and proliferation of the cancer cells.25;26;69

Chapter 2

2

26

27

2


inoculation of EpCAM siRNA treated gastric carcinoma cells, led to signi�cantly less visible tumors

in liver surface compared to non-treated cells.23 Inhibition of EpCAM expression by antisense mRNA

in a HNSCC cell line showed changes in morphology and reduced proliferation and metabolism22,

indicating a promoting role for EpCAM in HNSCC.

To better understand why EpCAM is overexpressed in carcinomas, more insights in the

regulation of the EpCAM gene itself are required, therefore the (epi)genetic events involved in

EpCAM regulation will now be described.

3. Regulation of EpCAM expression

3.1 Genetics

The EpCAM protein is encoded by the GA733-2 gene70 with a minimal estimated size of approximately

14 kb, and is located on chromosome 2p21.71 The GA733-2 gene consists of a total of nine exons70,

the mRNA is approximately 1.5 kb (NCBI: AH003574); all reported open reading frames of EpCAM

are identical and consists of 942 bases encoding a protein of 314 amino acids.72;73 In a large number

of carcinoma cell lines no splicing variants were found.74 To our knowledge, mutations in the EpCAM

gene have only been identi�ed in patients su!ering from Lynch syndrome or congenital tufting

enteropathy. In Lynch syndrome, di!erent heterozygous germline deletions that disrupt the 3’-end

of the GA733-2 gene lead to inactivation of the adjacent MSH2 gene through methylation induction

of its promoter in tissues expressing EpCAM.75 In congenital tufting enteropathy, two di!erent

homozygous point mutations in exon 4 result in a deletion of exon 4 and a decreased expression

of EpCAM on protein level.76 An additional heterozygous point mutation within exon 3, causing

an amino acid change, was also associated with decreased EpCAM expression. Another recently

described homozygous point mutation in exon 3, even led to an absence of EpCAM expression in

the intestinal tissue from the a!ected patient.77 A homozygous single base pair insertion in exon

5 results in a frameshift introducing 21 novel amino acids followed by premature truncation of

EpCAM protein.78

The GA733-2 promoter region that controls the expression of the EpCAM gene has been cloned

and characterized.79-81 The sequence upstream of the transcription start site (TSS) has been de�ned81

(NCBI: AY148099). A 3.4 kb fragment of this EpCAM 5’-regulatory sequence is capable of directing

heterologous gene expression and the promoter activity is restricted to EpCAM expressing cells.81;82

A complementary study con�rmed that the transcriptional activity of a 1.1 kb EpCAM fragment

starting 770 bp upstream of the TSS directly correlated with the amount of EpCAM expression.80

In silico analysis of the EpCAM promoter revealed several homologies to known transcriptional

regulatory sequences and putative transcription binding sites.81 Although no TATA or CAAT boxes

were found, the position of the consensus initiator element (Inr) matches with the putative TSS

based on 5’-UTR sequencing studies.70;83 By deletion analysis it was established that 177 bp of the

5’ -"anking sequence are su#cient to drive reporter gene expression, whereas the region 687 bp to

341 bp upstream of the TSS, appeared to be responsible for epithelial speci�c expression.81

28

Chapter 2

2

3.2 Transcription factors

Several putative transcription binding sites within the EpCAM promoter have been reported69;70;81;84

(Figure 1). Up till now little biological data supporting a role for these transcription factors in EpCAM

gene expression has been described. Indirect evidence has been reported for ESE-1 (Epithelial

speci�c Ets-1): upregulation of ESE-1 in metastatic lymph nodes from lung, breast and pancreas

cancers correlated well with the expression of EpCAM.4 An indication that Sp1 plays an active role in

EpCAM regulation was demonstrated by reporter gene analysis: after transfection with an EpCAM

promoter fragment (-250 to +90, relatively to TSS) containing putative binding sites for Sp1 (Figure

1), an elevated promoter activity was observed in the presence of Sp1 compared to the activity in

the absence of Sp1.68

+282

ATGTSS

-830Sp1-27

NF B+27

Sp1-231

AP-1-125

Lef-489

Ets-527

Ets-375

p53

Figure 1. Schematic overview of part of the EpCAM gene (not to scale). The vertical bars represent the CpGssensitive for methylation. The transcription start site (TSS) and the translation start site (ATG) are indicated. Thered circles represent published putative binding sites for the indicated transcription factors.69;70;81;84 The basepositions mentioned in the �gure and text are relatively to the TSS.

Recently, it has been shown that β-catenin activation induced EpCAM transcription via binding

of TCF/Lef at 489 bp upstream of the EpCAM transcription start site.69 Interesting, TCF/Lef and

β-catenin are also involved in nuclear signaling by EpCAM itself8: proteolytic cleavage of EpCAM,

releases EpICD which forms a complex with β-catenin and TCF/Lef that contacts DNA at the Lef

consensus sites, the authors suggested that EpICD may impose a positive-feedback loop on EpCAM

expression at the level of gene transcription.8

The transcription factors NF-κB and p53 have been described as transcriptional repressors of

the EpCAM gene: treatment of EpCAM positive squamous cell carcinoma cells with TNFα and IFNα

resulted in a reduced endogenous EpCAM expression.79;85 Inhibition of the activation of NF-κB by

cotransfection of a luciferase reporter plasmid under control of the EpCAM promoter, and a plasmid

coding for the dominant negative of the NF-κB inhibiter IκB, supported a direct role for NF-κB as

a repressor of the EpCAM promoter. A second repressor of EpCAM promoter activity is the tumor

suppressor gene p53.84 Induction of wild type p53 (WT p53) was associated with a dose-dependent

decrease in EpCAM expression, whereas ablation of p53 expression was associated with an increase

in EpCAM expression. Ten putative binding sites for p53 in the EpCAM gene were identi�ed and

by Chromatin ImmunoPrecipitation (ChIP) the binding of WT p53 to a site located within intron

4 was con�rmed.84 Interestingly, concomitant silencing of p53 and EpCAM expression via stable

transduction of shRNA prevented the increase of EpCAM expression caused by ablation of p53

expression and decreased the invasiveness of the breast cancer cells.84

29

2


3.3 Epigenetics

Accessibility of transcription factors to the speci�c binding sites within the EpCAM gene depends

on the chromatin structure, which is a�ected by DNA methylation and histone modi�cations.86

Modi�cations of DNA and histones thus have profound impact on gene expression. Here we will

focus on DNA methylation and histone modi�cations, as these epigenetic events are potentially

reversible by drug treatments.

3.3.1 DNA methylation

Already in 1994, it was described that DNA methylation prevents ampli�cation of the EpCAM gene.87

Loss of DNA methylation in the EpCAM gene, caused by inactivation of the p53 gene, resulted in

EpCAM gene ampli�cation.88 In view of these �ndings, the observation that downregulation of p53

caused upregulation of EpCAM expression is noteworthy.84

In humans, DNA methylation occurs in cytosines within cytosine-guanine dinucleotides (CpGs).

CpGs are relatively rare in the genome but tend to cluster in islands which are usually located in the

5’-regulatory region of many genes. Methylation of CpG islands in promoters leads to transcriptional

silencing of genes. Several studies have reported that EpCAM expression is associated with DNA

methylation (Table IV).56;68;89-91 In cell lines of di�erent origin, high EpCAM expression was associated

with hypomethylation and no EpCAM expression was associated with hypermethylation of the

proximal promoter and part of exon 1.90 Interestingly, the CpG within the putative binding site for

Sp1 (-231) was methylated in EpCAM negative cell lines and not methylated in EpCAM positive

study material region technique

Spizzo et al.,2007 89

breast cancer cell linebreast cancer tissue

-156 to +361-135 to -37 no

Bisul�te sequencing (64)MethyLight (6)

Tai et al., 2007 68 lung adeno, bladder, colon,germ cell ovary carcinoma celllineslung adenocarcinoma tissueacH3K9~ EpCAM pos.H3K9me~ EpCAM neg

-265 to -100

-265 to -100 yes-682 to-540-356 to -140

Mehtylation Speci�c PCR (6)

Methylation Speci�c PCR (6 )ChromatinImmunoPrecipitation

Yu et al., 2008 91 colon, prostate, breast, liver,haematological tumor cell linescolon cancer tissue

-321 to +790

-321 to +790 yes

Bisul�te sequencing (122)

Bisul�te sequencing (122)

Van der Gun et al.2008 90

lung, ovarian, colon carcinomaand human embryonic kidney,glioblastoma cell lines

-830 to +283 Bisul�te sequencing (92)

Shiah et al.,2008 56

oral squamous cell carcinomatissue

-265 to -100 yes Methylation Speci�c PCR (6)

Table IV. Determination of molecular epigenetic marks for indicated regions of the EpCAM gene.Positions are relatively to the transcription start site. Between brackets the number of CpGs analyzed by theindicated technique. The remark ’yes’ or ’no’ indicates the correlation between DNA methylation and EpCAMexpression examined in patient tissue.

30

Chapter 2

2

cell lines, whereas in the area around the putative binding site for AP-1 (-125) the CpGs were

unmethylated in all cell lines analyzed.91 For the Sp1 binding site, we also observed methylation of

the CpG within the Sp1 binding site in EpCAM negative ovarian carcinoma cell lines, whereas this

CpG is unmethylated in EpCAM positive ovarian carcinoma cell lines (manuscript in preparation).

Modulation by epigenetic drugs con�rmed the correlation between EpCAM expression and the

DNA methylation status of the EpCAM gene. Treatment of EpCAM negative cell lines with a DNA

methylation inhibitor agent (5-aza-2’-deoxycytidine) induced EpCAM expression de novo, both on

mRNA and protein level and caused further upregulation of EpCAM expression in EpCAM positive

cell lines.68;89;90 However, in the EpCAM negative leukaemia K562 (hypermethylated) and the liver

HepG2 (CpGs were 50% methylated) cell lines, no EpCAM re-expression was observed after 5-aza

treatment, whereas most methylated CpGs were converted to unmethylated CpGs.91 In addition,

upon 5-aza treatment of the EpCAM negative lung carcinoma cell line GLC-1, of which part of the

EpCAM gene (-830 to +282) is intermediated methylated, no de novo induction of EpCAM expression

was detected.90 Alternatively, we demonstrated that endogenous EpCAM expression can be actively

downregulated in a persistent manner via induced DNA methylation.90 After delivery of the DNA

methyltransferase M.SssI into EpCAM positive ovarian carcinoma cells, methylation of the EpCAM

gene resulted in reduced EpCAM expression, which maintained through successive cell divisions as

the reduced EpCAM expression persisted for at least 17 days.90

The association between DNA methylation of the EpCAM gene and EpCAM expression in patient

samples appears to depend on the tissue type (Table IV). In normal colon tissues 50% of the CpGs

were methylated, whereas in colon cancer tissues most CpGs were unmethylated.91 The expression

level of EpCAM was 1000-fold higher in colon cancers than in normal colon tissue, re!ecting the

observed methylation status. Also in lung adenocarcinoma tissue68 and in oral squamous cell

carcinoma (OSCC)56, EpCAM expression was signi�cantly associated with the methylation status of

the EpCAM promoter. In contrast, in breast cancer tissue no correlation was found between EpCAM

protein expression and EpCAM promoter methylation for 6 CpGs measured.89 However, in the

same study they found the promoter of EpCAM negative breast cancer cell line to be methylated

to a higher degree as compared to an EpCAM positive cell line. The discrepancy found between

breast cancer cell lines versus tissue might be due to the di"erence in technique used: MethyLight

technology, analyzing 6 CpGs and bisul�te sequencing analyzing 64 CpGs, respectively (Table

IV). Since in lung adenocarcinoma as well as in oral squamous cell carcinoma tissue, high EpCAM

expression, indeed correlated with a low DNA methylation level, the location of the CpGs as well as

the tissue type, appears to be of importance.

3.3.2 Histone modi�cations

The nucleosome consists of an octamer of the four histone proteins (H2A, H2B, H3, H4) wrapped

around by ~147 bp of DNA. The histones are subject to a variety of post-translational modi�cations

including phosphorylation, acetylation, ubiquitylation and methylation. Histone modi�cations

31

2


play important roles in chromatin structure and function. With regard to transcriptional regulation,

increases in histone acetylation and trimethylation of lysine 4 of histone 3 (H3K4me3) are generally

associated with increased gene expression, while decreased acetylation and methylation of lysine 9

of histone 3 (H3K9me3) are marks of decreased gene expression.92 Methylation of H3K9 is associated

with activation when it is found on the coding region, but when it is found in the promoter area it is

associated with repression.92

The indirect e�ect of histone acetylation on EpCAM expression has been demonstrated by

treating carcinoma cell lines with trichostatin A (TSA), an inhibitor of histone deacetylase.68 Minimally

elevated EpCAM expression was observed after TSA treatment, but an increased EpCAM expression

was induced by concomitant treatment with 5-aza, showing that both histone modi�cation and

DNA methylation are responsible for EpCAM gene expression. Transfection of a plasmid expressing

the histone acetyl transferase p300/CBP abrogated the reduced EpCAM expression caused by NF-κB

upon treatment with TNFα or IFNα, suggesting that NF-κB competes with p300/CBP for binding to

the EpCAM promoter.85

Compared to DNA methylation, less is known about the histone modi�cations characteristic for

EpCAM expression. To our knowledge, only one study reports on enzymes and histone modi�cations

involved in epigenetic regulation of the EpCAM gene.68 In this study, Chromatin ImmunoPrecipitation

revealed that association of heterochromatin protein 1 (HP1), the H3K9 methyltransferase SUV39H1,

histone deactylase HDAC1, and the DNA methyltransferases DNMT1 and 3b with the EpCAM

promoter, increased gradually as EpCAM expression in lung adenocarcinoma cell lines (n=3)

decreased.68 In agreement with their �ndings concerning SUV39H1 and HDAC1, acetylated H3K9

was associated with the EpCAM promoter in EpCAM positive cell lines, whereas methylated H3K9

was associated with the EpCAM promoter in EpCAM negative cells68 (Table IV). Currently, we are

expanding the panel of histone modi�cations charateristic for EpCAM expression (manuscript in

preparation).

4. Perspectives

Although the exact biological role of EpCAM is not clear yet, the e�ect of EpCAM overexpression

or silencing is established for a list of di�erent tumor types. For these types, modulation of

EpCAM expression provides a promising approach to interfere with oncogenic potential of these

tumor cells. Since the aberrant expression of EpCAM on carcinomas seems to be associated with

epigenetic mutations without underlying genetic defects, modulation of EpCAM expression by

epigenetic interference opens up new possibilities to permanently modify expression levels. An

approach resulting in long-term gene expression modulation, is epigenetic editing. In epigenetic

editing, molecular epigenetic marks are overwritten by an epigenetic e�ector domain targeted to

speci�c genes by a sequence speci�c DNA-binding motif. Three classes of DNA-binding motifs are

available that can direct attached epigenetic e�ector domains to a speci�c sequence.93 These motifs

are either based on synthetic polyamides, on designed recombinant zinc �nger moieties94 or on

32

Chapter 2

2

oligonucleotides which can form triple helices with the target double strand DNA.95

Trimeric and hexameric zinc �nger proteins have been designed to target the EpCAM promoter

and when fused to a repressor or an activation domain, these arti�cial transcription factors have been

shown to modulate the EpCAM promoter activity.96 Recently, we also designed an EpCAM speci�c

Triple helix Forming Oligonucleotide, which when coupled to a mutant methyltransferase is able

to target methylation predominantly to a speci�c DNA sequence in the EpCAM promoter without

signi�cant background methylation (manuscript in press). Alternatively, histone modi�ers like the

histone methyltransferase SUV39H1 and G9a have been successfully used as epigenetic e�ector

domains to silence genes. A minimal catalytic domain of the histone methyltransferase linked to

a zinc �nger targeting the VEGF gene showed enrichment of H3K9 methylation associated with

the VEGF promoter, resulting in transcriptional repression of the VEGF gene.97 On the other hand,

epigenetic editing can be used to re-express silenced genes: for instance, the hypermethylated

tumor suppressor gene maspin was reactivated by engineered zinc �nger targeting the maspin

promoter fused to VP64.98 Replacement of the transient activation domain VP64 by epigenetic

enzymes might reactivate epigenetically silenced genes in a permanent way. In view of the progress

in gene targeting using TFOs or zinc �ngers, epigenetic modulation of the EpCAM gene can be

envisioned as a promising new tool in unraveling the role of EpCAM, opening up novel approaches

in exploiting EpCAM as an anti-carcinoma therapeutic. Due to the inheritable nature of epigenetic

marks, revised epigenetic marks are expected to be stable through subsequent cell divisions, this

advantage allows us to induce durable changes of gene expression after transient expression of

the targeted epigenetic rewriters. Epigenetic editing will overcome the use of potentially harmful

integrating vectors thereby avoiding a permanent genetic change in the cell.

33

2


References

1. Herlyn M, Steplewski Z, Herlyn D, Koprowski H. Colorectal carcinoma-speci�c antigen: detection by meansof monoclonal antibodies. Proc Natl Acad Sci U.S.A 1979;76:1438-42.


3. Went P, Dirnhofer S, Schopf D, Moch H, Spizzo G. Expression and prognostic signi�cance of EpCAM. J.CancerMol 2008;3:169-74.

4. Baeuerle PA, Gires O. EpCAM (CD326) �nding its role in cancer. Br.J Cancer 2007;96:417-23.


6. Winter MJ, Nagelkerken B, Mertens AEE, Rees-Bakker HAM, Briaire-de Bruijn IH, Litvinov SV. Expression ofEp-CAM shifts the state of cadherin-mediated adhesions from strong to weak. Experimental Cell Research2003;285:50-8.

7. Nubel T, Preobraschenski J, Tuncay H, Weiss T, Kuhn S, Ladwein M et al. Claudin-7 regulates EpCAM-mediated functions in tumor progression. Molecular Cancer Research 2009;7:285-99.

8. Maetzel D, Denzel S, Mack B, Canis M, Went P, Benk M et al. Nuclear signaling by tumor-associated antigenEpCAM. Nat Cell Biol 2009;11:162-71.

9. Brunner A, Prelog M, Verdorfer I, Tzankov A, Mikuz G, Ensinger C. EpCAM is predominantly expressed in highgrade and advanced stage urothelial carcinoma of the bladder. Journal of Clinical Pathology 2008;61:307-10.

10. Fong D, Steurer M, Obrist P, Barbieri V, Margreiter R, Amberger A et al. Ep-CAM expression in pancreatic andampullary carcinomas: frequency and prognostic relevance. Journal of Clinical Pathology 2008;61:31-5.

11. Spizzo G, Went P, Dirnhofer S, Obrist P, Simon R, Spichtin H et al. High Ep-CAM expression is associated withpoor prognosis in node-positive breast cancer. Breast Cancer Res.Treat. 2004;86:207-13.

12. Varga M, Obrist P, Schneeberger S, Muhlmann G, Felgel-Farnholz C, Fong D et al. Overexpression ofepithelial cell adhesion molecule antigen in gallbladder carcinoma is an independent marker for poorsurvival. Clin Cancer Res. 2004;10:3131-6.

13. Litvinov SV, Velders MP, Bakker HA, Fleuren GJ, Warnaar SO. Ep-CAM: a human epithelial antigen is ahomophilic cell-cell adhesion molecule. J Cell Biol. 1994;125:437-46.

14. Basak S, Speicher D, Eck S, Wunner W, Maul G, Simmons MS et al. Colorectal carcinoma invasion inhibitionby CO17-1A/GA733 antigen and its murine homologue. J.Natl.Cancer Inst. 1998;90:691-7.

15. Gosens MJ, van Kempen LC, van d, V, van Krieken JH, Nagtegaal ID. Loss of membranous Ep-CAM inbudding colorectal carcinoma cells. Mod.Pathol 2007;20:221-32.

16. Went P, Dirnhofer S, Salvisberg T, Amin MB, Lim SD, Diener PA et al. Expression of epithelial cell adhesionmolecule (EpCam) in renal epithelial tumors. Am.J Surg.Pathol 2005;29:83-8.

17. Ensinger C, Kremser R, Prommegger R, Spizzo G, Schmid KW. EpCAM overexpression in thyroid carcinomas:a histopathological study of 121 cases. J Immunother. 2006;29:569-73.

18. Hwang EY, Yu CH, Cheng SJ, Chang JY, Chen HM, Chiang CP. Decreased expression of Ep-CAM protein issigni�cantly associated with the progression and prognosis of oral squamous cell carcinomas in Taiwan. JOral Pathol.Med. 2009;38:87-93.

34

Chapter 2

2

19. Kimura H, Kato H, Faried A, Sohda M, Nakajima M, Fukai Y et al. Prognostic signi�cance of EpCAM expressionin human esophageal cancer. Int.J Oncol. 2007;30:171-9.

20. Seligson DB, Pantuck AJ, Liu X, Huang Y, Horvath S, Bui MH et al. Epithelial cell adhesion molecule (KSA)expression: pathobiology and its role as an independent predictor of survival in renal cell carcinoma. ClinCancer Res. 2004;10:2659-69.

21. Songun I, Litvinov SV, van d, V, Pals ST, Hermans J, van Krieken JH. Loss of Ep-CAM (CO17-1A) expressionpredicts survival in patients with gastric cancer. Br.J.Cancer 2005;92:1767-72.

22. Munz M, Kieu C, Mack B, Schmitt B, Zeidler R, Gires O.The carcinoma-associated antigen EpCAM upregulatesc-myc and induces cell proliferation. Oncogene 2004;23:5748-58.




26. Yanamoto S, Kawasaki G, Yoshitomi I, Iwamoto T, Hirata K, Mizuno A. Clinicopathologic signi�cance ofEpCAM expression in squamous cell carcinoma of the tongue and its possibility as a potential target fortongue cancer gene therapy. Oral Oncology 2007;43:869-77.

27. Beltran AS, Sun X, Lizardi PM, Blancafort P. Reprogramming epigenetic silencing: arti�cial transcriptionfactors synergize with chromatin remodeling drugs to reactivate the tumor suppressor mammary serineprotease inhibitor. Mol Cancer Ther. 2008;7:1080-90.

28. van der Gun B.T.F and Stolzenburg S, Jurkowska R Arendzen A McLAughlin P Jeltsch A Rots MG. Modulationof epigenetic marks controlling the expression of the pancarcinoma antigen Epithelial Cell AdhesionMolecule . Human Gene Ther 19, 1115. 10-1-2008. Ref Type: Abstract

29. Trzpis M, McLaughlin PMJ, de Leij LMFH, Harmsen MC. Epithelial Cell Adhesion Molecule: More than aCarcinoma Marker and Adhesion Molecule. American Journal of Pathology 2007;171:386-95.

30. Le NF, Zoller M. The tumor antigen EpCAM: tetraspanins and the tight junction protein claudin-7, newpartners, new functions 7. Front Biosci. 2008;13:5847-65.

31. Kuhn S, Koch M, Nubel T, Ladwein M, Antolovic D, Klingbeil P et al. A Complex of EpCAM, Claudin-7, CD44Variant Isoforms, and Tetraspanins Promotes Colorectal Cancer Progression. Molecular Cancer Research2007;5:553-67.

32. Munz M, Fellinger K, Hofmann T, Schmitt B, Gires O. Glycosylation is crucial for stability of tumor and cancerstem cell antigen EpCAM. Front Biosci 2008;13:5195-201.

33. Pauli C, Munz M, Kieu C, Mack B, Breinl P, Wollenberg B et al. Tumor-speci�c glycosylation of thecarcinoma-associated epithelial cell adhesion molecule EpCAM in head and neck carcinomas20. Cancer Lett. 2003;193:25-32.

34. Li C. Identi�cation of pancreatic cancer stem cells. Cancer Res. 2007;67:1030-7.

35. Al-Hajj M, Wicha MS, ito-Hernandez A, Morrison SJ, Clarke MF. Prospective identi�cation of tumorigenicbreast cancer cells. Proc.Natl Acad.Sci.USA 2003;100:3983-8.

36. Jojovic M, Adam E, Zangemeister-Wittke U, Schumacher U. Epithelial glycoprotein-2 expression is subjectto regulatory processes in epithelial-mesenchymal transitions during metastases: an investigation ofhuman cancers transplanted into severe combined immunode�cient mice. Histochem.J 1998;30:723-9.

35

2


37. Rao CG, Chianese D, Doyle GV, Miller MC, Russell T, Sanders RA, Jr. et al. Expression of epithelial cell adhesionmolecule in carcinoma cells present in blood and primary and metastatic tumors. Int.J Oncol. 2005;27:49-57.

38. Klatte T, Pantuck AJ, Said JW, Seligson DB, Rao NP, LaRochelle JC et al. Cytogenetic and Molecular TumorPro!ling for Type 1 and Type 2 Papillary Renal Cell Carcinoma. Clinical Cancer Research 2009;15:1162-9.

39. Gastl G, Spizzo G, Obrist P, Dunser M, Mikuz G. Ep-CAM overexpression in breast cancer as a predictor ofsurvival. Lancet 2000;356:1981-2.

40. Cimino A, Halushka M, Illei P, Wu X, Sukumar S, Argani P. Epithelial cell adhesion molecule (EpCAM) isoverexpressed in breast cancer metastases. Breast Cancer Res Treat. 2009.

41. Litvinov SV, van DW, van Rhijn CM, Bakker HA, van KH, Fleuren GJ et al. Expression of Ep-CAM in cervicalsquamous epithelia correlates with an increased proliferation and the disappearance of markers forterminal di"erentiation. Am.J.Pathol. 1996;148:865-75.

42. Piyathilake CJ, Frost AR, Weiss H, Manne U, Heimburger DC, Grizzle WE. The expression of Ep-CAM (17-1A)in squamous cell cancers of the lung. Hum.Pathol. 2000;31:482-7.

43. Went P, Vasei M, Bubendorf L, Terracciano L, Tornillo L, Riede U et al. Frequent high-level expression of theimmunotherapeutic target Ep-CAM in colon, stomach, prostate and lung cancers. Br.J Cancer 2006;94:128-35.

44. Deveci MS, Deveci G. Prognostic value of p53 protein and MK-1 (a tumor-associated antigen) expression ingastric carcinoma. Gastric.Cancer 2007;10:112-6.

45. Wenqi D, Li W, Shanshan C, Bei C, Yafei Z, Feihu B et al. EpCAM is overexpressed in gastric cancer and itsdownregulation suppresses proliferation of gastric cancer 6. J Cancer Res Clin Oncol 2009;135:1277-85.

46. Scheunemann P, Stoecklein NH, Hermann K, Rehders A, Eisenberger CF, Knoefel WT et al. Occultdisseminated tumor cells in lymph nodes of patients with gastric carcinoma. A critical appraisal ofassessment and relevance. Langenbecks Arch.Surg. 2009;394:105-13.

47. Dalerba P, Dylla SJ, Park IK, Liu R, Wang X, Cho RW et al. Phenotypic characterization of human colorectalcancer stem cells. Proceedings of the National Academy of Sciences 2007;104:10158-63.

48. Xi L, Gooding W, McCarty K, Godfrey TE, Hughes SJ. Identi!cation of mRNA Markers for Molecular Stagingof Lymph Nodes in Colorectal Cancer. Clinical Chemistry 2006;52:520-3.

49. Kim Y, Kim HS, Cui ZY, Lee HS, Ahn JS, Park CK et al. Clinicopathological implications of EpCAM expressionin adenocarcinoma of the lung. Anticancer Res 2009;29:1817-22.

50. Mitas M, Cole DJ, Hoover L, Fraig MM, Mikhitarian K, Block MI et al. Real-Time Reverse Transcription-PCRDetects KS1/4 mRNA in Mediastinal Lymph Nodes from Patients with Non-Small Cell Lung Cancer. ClinicalChemistry 2003;49:312-5.

51. Kubuschok B, Passlick B, Izbicki JR, Thetter O, Pantel K. Disseminated tumor cells in lymph nodes as adeterminant for survival in surgically resected non-small-cell lung cancer. J Clin.Oncol. 1999;17:19-24.

52. Andratschke M, Hagedorn H, Luebbers CW, Schmitt B, Lang S, Zeidler R et al. Limited suitability of EpCAMfor molecular staging of tumor borders in head and neck cancer. Anticancer Res 2006;26:153-8.

53. Ferris RL, Xi L, Raja S, Hunt JL, Wang J, Gooding WE et al. Molecular Staging of Cervical Lymph Nodes inSquamous Cell Carcinoma of the Head and Neck. Cancer Research 2005;65:2147-56.

54. Chaubal S, Wollenberg B, Kastenbauer E, Zeidler R. Ep-CAM--a marker for the detection of disseminatedtumor cells in patients su"ering from SCCHN. Anticancer Res 1999;19:2237-42.

55. Laimer K, Fong D, Gastl G, Obrist P, Kloss F, Tuli T et al. EpCAM expression in squamous cell carcinoma of theoral cavity: Frequency and relationship to clinicopathologic features. Oral Oncology 2008;44:72-7.

36

Chapter 2

2

56. Shiah SG, Chang LC, Tai KY, Lee GH, Wu CW, Shieh YS. The involvement of promoter methylation andDNA methyltransferase-1 in the regulation of EpCAM expression in oral squamous cell carcinoma. OralOncology 2008;45:e1-e8.

57. Takes RP, Baatenburg de Jong RJ, Schuuring E, Hermans J, Vis AA, Litvinov SV et al. Markers for assessmentof nodal metastasis in laryngeal carcinoma. Arch.Otolaryngol.Head Neck Surg. 1997;123:412-9.

58. Takes RP, Baatenburg de Jong RJ, Alles MJ, Meeuwis CA, Marres HA, Knegt PP et al. Markers for nodalmetastasis in head and neck squamous cell cancer. Arch.Otolaryngol.Head Neck Surg. 2002;128:512-8.

59. Rossen K, Thomsen HK. Ber-EP4 immunoreactivity depends on the germ layer origin and maturity of thesquamous epithelium. Histopathology 2001;39:386-9.

60. Stoecklein NH, Siegmund A, Scheunemann P, Luebke AM, Erbersdobler A, Verde PE et al. Ep-CAMexpression in squamous cell carcinoma of the esophagus: a potential therapeutic target and prognosticmarker. BMC.Cancer 2006;6:165.

61. Hosch S, Kraus J, Scheunemann P, Izbicki JR, Schneider C, Schumacher U et al. Malignant potential andcytogenetic characteristics of occult disseminated tumor cells in esophageal cancer. Cancer Research2000;60:6836-40.

62. Heinzelmann-Schwarz VA, Gardiner-Garden M, Henshall SM, Scurry J, Scolyer RA, Davies MJ et al.Overexpression of the cell adhesion molecules DDR1, Claudin 3, and Ep-CAM in metaplastic ovarianepithelium and ovarian cancer. Clinical Cancer Research 2004;10:4427-36.

63. Kim JH, Herlyn D, Wong KK, Park DC, Schorge JO, Lu KH et al. Identi!cation of epithelial cell adhesionmolecule autoantibody in patients with ovarian cancer. Clinical Cancer Research 2003;9:4782-91.

64. Spizzo G, Went P, Dirnhofer S, Obrist P, Moch H, Baeuerle PA et al. Overexpression of epithelial cell adhesionmolecule (Ep-CAM) is an independent prognostic marker for reduced survival of patients with epithelialovarian cancer. Gynecol.Oncol. 2006;103:483-8.

65. Bellone S, Siegel ER, Cocco E, Cargnelutti M, Silasi DA, Azodi M et al. Overexpression of epithelial celladhesion molecule in primary, metastatic, and recurrent/chemotherapy-resistant epithelial ovariancancer: implications for epithelial cell adhesion molecule-speci!c immunotherapy. Int.J Gynecol.Cancer2009;19:860-6.

66. Poczatek RB, Myers RB, Manne U, Oelschlager DK, Weiss HL, Bostwick DG et al. Ep-CAM levels in prostaticadenocarcinoma and prostatic intraepithelial neoplasia. J Urol. 1999;162:1462-6.

67. Zellweger T, Ninck C, Bloch M, Mirlacher M, Koivisto PA, Helin HJ et al. Expression patterns of potentialtherapeutic targets in prostate cancer. Int.J Cancer 2005;113:619-28.

68. Tai KY, Shiah SG, Shieh YS, Kao YR, Chi CY, Huang E et al. DNA methylation and histone modi!cation regulatesilencing of epithelial cell adhesion molecule for tumor invasion and progression. Oncogene 2007;26:3989-97.

69. Yamashita T, Budhu A, Forgues M, Wang XW. Activation of Hepatic Stem Cell Marker EpCAM by Wnt {beta}-Catenin Signaling in Hepatocellular Carcinoma. Cancer Research 2007;67:10831-9.

70. Linnenbach AJ, Seng BA, Wu S, Robbins S, Scollon M, Pyrc JJ et al. Retroposition in a family of carcinoma-associated antigen genes. Molecular and Cellular Biology 1993;13:1507-15.

71. Calabrese G, Crescenzi C, Morizio E, Palka G, Guerra E, Alberti S. Assignment of TACSTD1 (alias TROP1,M4S1) to human chromosome 2p21 and re!nement of mapping of TACSTD2 (alias TROP2, M1S1) tohuman chromosome 1p32 by in situ hybridization. Cytogenet.Cell Genet. 2001;92:164-5.

72. Szala S, Froehlich M, Scollon M, Kasai Y, Steplewski Z, Koprowski H et al. Molecular Cloning of cDNA for theCarcinoma-Associated Antigen GA733-2. Proceedings of the National Academy of Sciences 1990;87:3542-6.

37

2


73. PerezMS,WalkerLE. IsolationandcharacterizationofacDNAencodingtheKS1/4epithelialcarcinomamarker32. J Immunol. 1989;142:3662-7.

74. Balzar M, Winter MJ, de Boer CJ, Litvinov SV. The biology of the 17-1A antigen (Ep-CAM). J Mol Med.1999;77:699-712.

75. Ligtenberg MJ, Kuiper RP, Chan TL, Goossens M, Hebeda KM, Voorendt M et al. Heritable somaticmethylation and inactivation of MSH2 in families with Lynch syndrome due to deletion of the 3’ exons ofTACSTD1. Nat.Genet. 2009;41:112-7.

76. Sivagnanam M, Mueller JL, Lee H, Chen Z, Nelson SF, Turner D et al. Identi!cation of EpCAM as the gene forcongenital tufting enteropathy. Gastroenterology 2008;135:429-37.

77. Sivagnanam M, Schaible T, Szigeti R, Byrd RH, Finegold MJ, Ranganathan S et al. Further evidence forEpCAM as the gene for congenital tufting enteropathy. Am J Med.Genet A 2010;152A:222-4.

78. Al-Mayouf SM, Alswaied N, Alkuraya FS, Almehaidib A, Faqih M. Tufting enteropathy andchronic arthritis: a newly recognized association with a novel EpCAM gene mutation3. J Pediatr.Gastroenterol.Nutr. 2009;49:642-4.

79. Gires O, Esko!er S, Lang S, Zeidler R, Munz M. Cloning and characterisation of a 1.1 kb fragment of thecarcinoma-associated epithelial cell adhesion molecule promoter. Anticancer Res. 2003;23:3255-61.

80. Gires O, Pockl S, Chapman RD, Munz M. Targeted gene expression using a 1.1 kilobase promoter fragmentof the tumor-associated antigen EpCAM. Anticancer Res. 2004;24:3715-21.

81. McLaughlin PM, Trzpis M, Kroesen BJ, Helfrich W, Terpstra P, Dokter WH et al. Use of the EGP-2/Ep-CAMpromoter for targeted expression of heterologous genes in carcinoma derived cell lines. Cancer Gene Ther.2004;11:603-12.

82. Gommans WM, van Eert SJ, McLaughlin PM, Harmsen MC, Yamamoto M, Curiel DT et al. The carcinoma-speci!c epithelial glycoprotein-2 promoter controls e"cient and selective gene expression in anadenoviral context. Cancer Gene Ther 2006;13:150-8.

83. PerezMS,WalkerLE. IsolationandcharacterizationofacDNAencodingtheKS1/4epithelialcarcinomamarker32. J Immunol. 1989;142:3662-7.

84. Sankpal NV, Willman MW, Fleming TP, May!eld JD, Gillanders WE. Transcriptional Repression of EpithelialCell Adhesion Molecule Contributes to p53 Control of Breast Cancer Invasion. Cancer Research 2009;0008-5472.

85. Gires O, Kieu C, Fix P, Schmitt B, Munz M, Wollenberg B et al. Tumor necrosis factor alpha negativelyregulates the expression of the carcinoma-associated antigen epithelial cell adhesion molecule. Cancer2001;92:620-8.

86. Esteller M. Epigenetics in cancer. N.Engl.J Med. 2008;358:1148-59.

87. Alberti S, Nutini M, Herzenberg LA. DNA Methylation Prevents the Ampli!cation of TROP1, a Tumor-Associated Cell Surface Antigen Gene. Proceedings of the National Academy of Sciences 1994;91:5833-7.

88. Nasr AF, Nutini M, Palombo B, Guerra E, Alberti S. Mutations of TP53 induce loss of DNA methylation andampli!cation of the TROP1 gene. Oncogene 2003;22:1668-77.

89. Spizzo G, Gastl G, Obrist P, Fong D, Haun M, Grunewald K et al. Methylation status of the Ep-CAM promoterregion in human breast cancer cell lines and breast cancer tissue. Cancer Lett. 2007;246:253-61.

90. van der Gun BT, Wasserkort R, Monami A, Jeltsch A, Rasko T, Slaska-Kiss K et al. Persistent downregulationof the pancarcinoma-associated epithelial cell adhesion molecule via active intranuclear methylation.Int.J.Cancer 2008;123:484-9.

38

Chapter 2

2

91. Yu G, Zhang X, Wang H, Rui D, Yin A, Qiu G et al. CpG island methylation status in the EpCAM promoterregion and gene expression. Oncol.Rep. 2008;20:1061-7.


93. Uil TG, Haisma HJ, Rots MG. Therapeutic modulation of endogenous gene function by agents withdesigned DNA-sequence speci!cities. Nucleic Acids Res 2003;31:6064-78.

94. Sera T. Zinc-!nger-based arti!cial transcription factors and their applications. Advanced Drug DeliveryReviews 2009;61:513-26.

95. Duca M, Vekho" P, Oussedik K, Halby L, Arimondo PB. The triple helix: 50 years later, the outcome. NucleicAcids Research 2008;36:5123-38.

96. Gommans WM, McLaughlin PM, Lindhout BI, Segal DJ, Wiegman DJ, Haisma HJ et al. Engineering zinc!nger protein transcription factors to downregulate the epithelial glycoprotein-2 promoter as a novelanti-cancer treatment. Mol Carcinog. 2007;46:391-401.

97. Snowden AW, Gregory PD, Case CC, Pabo CO. Gene-speci!c targeting of H3K9 methylation is su#cient forinitiating repression in vivo. Curr Biol 2002;12:2159-66.

98. Beltran A, Parikh S, Liu Y, Cuevas BD, Johnson GL, Futscher BW et al. Re-activation of a dormant tumorsuppressor gene maspin by designed transcription factors. Oncogene 2007;26:2791-8.

39

Transcription factors and molecular epigenetic marks

associated with the EpCAM gene in ovarian cancer

Bernardina T.F. van der Gun1, Hinke G. Kazemier1, Alice J. Arendzen1, Peter Terpstra1,Marcel H.J. Ruiters1, 2, Pamela M.J. McLaughlin1 and Marianne G. Rots1

1 Epigenetic Editing, Dept. of Pathology and Medical Biology, University Medical Center Groningen, University ofGroningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands

2 Synvolux Therapeutics Inc., L.J. Zielstraweg 1, 9713 GX Groningen, The Netherlands

Submitted for publication

Chapter 3

40

Chapter 3

3

ABSTRACT

The Epithelial Cell Adhesion Molecule (EpCAM) is a transmembrane glycoprotein that is highly

overexpressed on most carcinomas. The mechanisms responsible for this dysregulation are

largely unknown. In the present study, we describe the DNA methylation status as well as histone

modi�cations underlying the EpCAM expression for a panel of �ve ovarian carcinoma cell lines. High

EpCAM expression correlated with DNA hypomethylation, while no EpCAM expression correlated

with DNA hypermethylation of the promoter. No correlation between EpCAM expression and DNA

methylation was found in patients with advanced-stage serous ovarian cancer. In EpCAM positive

cells, the promoter was associated with acetylated histones H3 and H4 and trimethylated lysine

4 of histone H3. In EpCAM negative cells, the promoter was associated with lysine 9 or lysine 27

trimethylation of histone H3. In EpCAM expressing cells, ten out of sixteen tested transcription

factors were found to be associated with the EpCAM gene, while in EpCAM negative cells no

association of any of the transcription factors was detected. For Sp1, gel retardation experiments

showed a preferential binding of nuclear proteins to the unmethylated CpGs located at -231 bps

and -226 bps upstream of the transcription start site. No di�erence in binding was observed for

methylation of CpG at position -32 in the Sp1 putative binding site, nor for the CpG at position +27

next to putative NFκB site. Because siRNA-mediated downregulation of EpCAM expression has been

shown to decrease oncogenic potential of certain types of carcinomas, regulatory transcription

factors and insights in epigenetic regulation mechanisms may lead to additional novel strategies to

e�ectively silence EpCAM expression.

41

3


INTRODUCTION

The Epithelial Cell Adhesion Molecule (EpCAM) is a transmembrane glycoprotein that is highly

overexpressed on most carcinomas.1 Recently, EpCAM has gained renewed interest as a signal

transducer2 and has been identi�ed as a marker of cancer-initiating stem cells.3 Transient silencing

of EpCAM expression leads to a reduced oncogenic potential in breast4, gastric5, hepatocellular6 and

oral squamous cell7 carcinoma of the tongue.

In normal ovary and in benign ovarian tumors, EpCAM mRNA and protein expression is

signi�cantly lower compared to malignant ovarian tumors.8 In addition, ovarian cancer patients

presented higher amounts of natural anti-EpCAM antibodies in the serum as compared to healthy

donors.8 Several studies have con�rmed the EpCAM overexpression in the majority of ovarian

carcinomas.9-11 Regarding tumor progression, one study reported that FIGO stage III/IV showed

lower EpCAM expression than stage I8, while in another study FIGO stage III/IV showed higher

EpCAM expression than stage I/II disease.9 The latter study suggest that a higher expression of

EpCAM correlates with tumor progression, although no correlation with survival was found.9

However, a more recent study showed that EpCAM overexpression correlated signi�cantly with

decreased overall survival of patients with Epithelial Ovarian Cancer (EOC).11 In addition, metastatic/

recurrent tumors were found to express signi�cantly higher levels of EpCAM protein compared with

primary ovarian carcinomas.12 Besides its possible prognostic role in ovarian cancer, EpCAM proofs

to be an e�ective therapeutic immunotarget in several clinical trials.13 For example, catumaxomab

is a trifunctional monoclonal antibody (anti-EpCAM X anti-CD3) used to treat ovarian cancer

patients with malignant ascites.14;15 A phase I/II study revealed e�ective tumor cell elimination in

malignant ascites and substantially decreased ascites accumulation.16 Recently, it has been reported

that catumaxomab treatment might also have an e�ect on tumor cells in blood of ovarian cancer

patients.17 Similarly, the human monoclonal antibody MT201 could e�ectively eliminate malignant

cells in metastic tumor specimens from patients with ovarian cancer.18

The overexpression of EpCAM has been associated with the degree of DNA methylation.

In cell lines of di�erent tissue types, high EpCAM expression has been shown to correlate with

hypomethylation and no EpCAM expression was associated with hypermethylation of the

promoter.19-22 Treatment of EpCAM negative cell lines with a DNA methylation inhibitor (5-aza-2’-

deoxycytidine) induced EpCAM expression de novo, both on mRNA and protein level and caused

further upregulation of EpCAM expression in EpCAM positive cell lines.19-21 Alternatively, we also

demonstrated that endogenous EpCAM expression can be actively downregulated by treating

cells with the DNA methyltransferase M.SssI.21 In tumor material obtained from patients with

lung adenocarcinoma20, oral squamous cell carcinoma23 and colon cancer tissue22, high EpCAM

expression was associated with hypomethylation of the promoter. Although in breast cancer cell

lines the correlation between EpCAM expression and DNA methylation was con�rmed, in breast

cancer patient tissue no such correlation was found.19 Inhibition of both DNA methylation and

42

Chapter 3

3

histone deacetylation showed a signi�cant increase in EpCAM expression, indicating that next to

DNA methylation also histone modi�cations regulate EpCAM gene expression.20 Indeed in EpCAM

positive cell lines, acetylated histone H3 lysine 9 was associated with the EpCAM promoter, whereas

association of methylated histone H3 lysine 9 was found in EpCAM negative cell lines.20

In the present study, we set out to provide more insights into the underlying mechanisms for

the observed overexpression of EpCAM in ovarian cancer. We determined DNA methylation in

ovarian cancer cell lines as well as in patient material. Moreover, we investigated histone marks

associated with active as well as silenced EpCAM promoters. In addition, we screened over a dozen

of transcription factors for their association with the EpCAM promoter. Since EpCAM overexpression

appears to provide an e�cient therapeutic target in carcinomas, a better understanding of the

mechanisms controlling the expression of this gene may provide new opportunities for cancer

therapy.

MATERIAL AND METHODS

Cell lines and patient samples

TheCaOV3,OVCAR3, SKOV3,andH134Sovariancancercell lineswereculturedinDMEM(BioWhittaker

Inc, Walkersville, MD) supplemented with 50 μg/ml gentamicin sulfate, 2 mM L-glutamine, 10% FBS

(BioWhittaker). The A2780 cell line was cultured in RPMI-1640 medium (BioWhittaker) supplemented

with 50 μg/ml gentamicin sulfate, 2 mM L-glutamine, 10% FBS, 1mM Na-pyruvaat and 0.05 mM

β-mercapto-ethanol. Cells were maintained at 37°C in a humidi�ed 5% CO2-containing atmosphere.

Ten DNA samples of patients with advanced-stage serous ovarian cancer were kindly provided by

Prof. A.G.J. van der Zee and Dr. B. Wisman (Department of Gynecologic Oncology, University Medical

Center Groningen, Groningen, The Netherlands).24

EpCAM expression of cell lines and patient samples

EpCAM protein detection on cell lines was performed with mouse Mab MOC31 hybridoma

supernatant, followed by R M-F(ab)2-FITC (DAKO, Glostrup, Denmark). The Mean Fluorescence

Intensity (MFI) was measured on a BD FACS Calibur $ow cytometer (Beckton Dickenson Biosciences,

San Jose, CA). For EpCAM mRNA expression of patient samples, RNA microarray experiments using

the MIAME-compliant microarray were performed as previously described.24

DNA methylation analysis

Genomic DNA extracted from the cell lines and tumor samples was treated with sodium bisul�te

to convert unmethylated cytosines to uracils. The EZ DNA Methylation-Gold Kit (Zymo, Baseclear

Lab Products, Leiden, Netherlands) was used to modify 1 μg of DNA. Bisul�te speci�c primers void

of any CpG were used in order to obtain ampli�cation products unbiased for the methylation

status. Primer sequences for region B (Figure 1A) were 5’-AGTGTTTTGGAAGGTTTTTTGT-3’

(forward), 5’-AAATTAAAAAAATAAATAAACTCCC-3’ (reverse) and for region A (Figure 1A)

5’-GGAGGGGAGTTTATTTATTTTT-3’ (forward) and 5’-CACAACTCTACTCCAATC-3’ (reverse). PCR

conditions: 95°C for 15 min followed by 40 cycles of 95°C for 60 s, 55°C for 45 s, 72°C for 60 s and

"nished with 72°C for 10 min. PCR fragments were gel extracted using the DNA Extraction Kit

(Qiagen, Benelux B.V., Venlo, The Netherlands) and cloned into pCR 2.1-TOPO TA vector (Invitrogen,

Breda, The Netherlands). Following transformation, plasmids from individual bacterial colonies

were isolated using the Qiaprep Spin Miniprep Kit (Qiagen) and validated for insert by restriction

analysis before sequencing. For the cell lines and ovarian cancer tissues, 10 and 24 clones each were

sequenced, respectively.

Chromatin immunoprecipitation

Histone modi"cations were determined by Chromatin ImmunoPrecipitation (ChIP) using 5 μg of

the following antibodies purchased from Upstate Biotechnology laboratories (Lake Placid, NY, USA):

normal rabbit IgG (PP64B), acH4 (06-598), acH3 (06-599), H3K4me1 (07-436), H3K4me3 (05-745

or 04-745), H3K9me3 (17-625), H3K27me3 (07-449) and H3K36me2 (07-274). ChIP was performed

according to the protocol of Upstate Biotechnology laboratories with minor modi"cations. Brie$y,

cells were used at 70-80% con$uency, treated with 1% formaldehyde (Merck, Darmstadt, Germany)

10 min at 37°C, washed twice with ice-cold PBS, lysed and sonicated using a Bioruptor (High, 15

cycles of 30’’ on, 30’’ o&, total time 15 min) (Diagnode, Liège, Belgium), cell debris was removed

by centrifugation. Chromatin fragments were diluted 2.5-fold and precleared with protein A/G-

agarose/Salmon Sperm DNA (Upstate Biotechnology) for 2 h at 4°C. Supernatant was collected (part

of supernatant was kept at -20°C for use as input), the appropriate antibody was added and samples

were rotated O/N at 4°C followed by 2 h incubation with 60 μl protein A/G-agarose beads. After

washing, immune complexes were eluted, treated with RNAse (Roche, Mannheim, Germany), cross-

linkes reversed and proteins were digested with proteinase K (Roche, Mannheim, Germany). DNA

was puri"ed using QiaQuick DNA spin columns (Qiagen). For histone modi"cations 2 x 106 cells per

Immune Precipitate (IP) were used. Chromatin preparations were performed at least three times,

independently.

To detect association of transcription factors with the EpCAM promoter ChIP was performed

as described by Farnham et al25 (http://www.genomecenter.ucdavis.edu/expression_analysis/

chip.html) using the following antibodies: normal mouse IgG (12-371) (Millipore, Amsterdam, The

Netherlands), LEF-1(REMB6)TCF (Millipore MAB3752), Sp1 (Upstate 17-601), STAT 3 (Upstate 06-

596); the following antibodies were purchased from Santa Cruz Biotechnology Inc. (Heidelberg,

Germany) NF-κB p50(NLS) (sc-114X), NF-κB p65(A) (sc-109X), ESE-1 (H-270) (sc-28683X), SNAI 1 (E-

130) (sc-28199), SLUG (H-140) (sc-15391X), Ets-1 (C-20) (sc-350X), Ets-2 (C-20) (sc-351X), AP2-α (C-18)

(sc-184X); PEA3 (H-120) (sc-22806X), PDEF (H-250) (sc-67022X), E2F-2 (C-20) (sc-633X), E2F-4 (C-20)

(sc-866X) and p53 (sc-126X). Brie$y, cells were cross-linked with formaldehyde as described above,

3

43


44

Chapter 3

3

Figure 1. Part of the EpCAM gene under investigation. A) Schematic overview: nucleotide position -610to +282 relative to the transcription starting site (TSS). The ATG start codon is shown; CpGs are depicted byvertical bars. Region A and B were analyzed for DNA methylation by bisul�te sequencing. Histone modi�cationsassociated with the EpCAM gene were characterized for the regions C, B2 and A1. Association of transcriptionfactors with the EpCAM gene was determined by ChIP followed by PCR with primers for the regions B1 andA1. Interference of Sp1 and NFκB binding by DNA methylation of CpGs located within or close to putativebinding sites (indicated with an open circle) was analyzed by Electromobility Shift Assay (EMSA). B) Nucleotidepositions -697 to +282 relative to the transcription starting site (TSS) are shown; the ATG start codon is depictedin bold. The start of region A and B are indicated in blue. Putative transcription factor binding sites analyzed byin silico analysis (Genomatix, MatInspector version 7.7.3.1) are indicated in red. EMSAs were performed for theunderlined Sp1 (Sp1a and Sp1b) and NFκB sites. The epithelial speci�c region27 is located between !.

Region ARegion B

Region B2

ChIP

Bisulfite sequencing

-610

ATG

+282

TSSSp1b NF BSp1a

-159 +282-130-443

Region C-564 -376 Region A1

-138 +65

-432 -253

-332 -185Region B1

-332 -185Region B1

-697.CGGCCTCCC AAAGTGCTAGG ATTACAGGCG TGAGCCACCG CGCTCAGCCT GGGAACACCT TTTCTTACAT CTTCAAGTGC

-617.TAGAAATGCT TATGAAAACG AAAAAAGAAT TATTAAGAGT AATTATAAAG AAACACTCAT TTTCTTCCCA AGAGAGCCAAPU1.01/Ets LEF1 LEF1

-537.GATTTCTTCT TTCCTCTTCT TTCTTTTTTT TTTCTTTCTA ATTTCAAAGG AGTATAATTA AATTGCCAGG TAAAAGCTCAStart region B PU1.01/Ets STAT1

-457.AAGGTCTTTT TTATAGTGTT CTGGAAGGTT CTCTGCCTGT GTTTGTATTT CCTTTAGCCT CCACGTTCCT CTATCCAGTTE2F4 AP-2 PEA3

-377.CCCGCACCCT TCCCCCCAGG CCCCATTCTT CAAGGC " TTCAGAGCAGCGCT CCTCCGGTTA AAAGGAAGTC TCAGCACAGALEF1 Sp1/Sp1a

-297.ATCTTCAAAC CTCCTCGGAG GCCACCAAAG ATCCCTAACG CCGCCATGGA GACGAAGCAC CTGGGGCGGG GCGGAGCGGGRNApolIIB Sp1 Start region A

-217.GCGCGCGGGC CCACACCTGT GGAGAGGGCC GCGCCCCAAC TGCAGCGCCG GGGCTGGGGG AGGGGAGCCT ACTCACTCCCSp1 AP-1 STAT1/3/Ets

-137.CCAACTCCCG GGCGGTGACT CATCAACGAG CACCAGCGGC CAGAGGTGAG CAGTCCCGGG AAGGGGCCGA GAGGCGGGGCSp1b TSS

-57.CGCCAGGTCG GGCAGGTGTG CGCTCCGCCC CGCCGCGCGC ACAGAGCGCT AGTCCTTCGG CGAGCGAGCA CCTTCGACGCNFκB-p50 HIF1

+23.GGTCCGGGGA CCCCCTCGTC GCTGTCCTCC CGACGCGGAC CCGCGTGCCC CAGGCCTCGC GCTGCCCGGC CGGCTCCTCG

+103.TGTCCCACTC CCGGCGCACG CCCTCCCGCG AGTCCCGGGC CCCTCCCGCG CCCCTCTTCT CGGCGCGCGC GCAGCATGGC

+183.GCCCCCGCAG GTCCTCGCGT TCGGGCTTCT GCTTGCCGCG GCGACGGCGA CTTTTGCCGC AGCTCAGGAA GGTGAGGCGC

+263.GGATTGGAGC AGAGTTGTG

A

B

10 x 106 cells per IP were resuspended in swelling bu!er, nuclei were collected by centrifugation

and lysed in nuclei lysis bu!er. Chromatin was sonicated (High, 15 cycles of 15’’on, 1’ o!, total time

15 min) and precleared with Staphylococcus aureus protein A-positive cells (StaphA) (Merck) for

15 min at 4°C. Supernatant was collected, the appropriate antibody was added and samples were

rotated O/N at 4°C. IP’s with mouse antibodies were incubated for an additional hour with Rabbit-

45

3


anti-Mouse IgG (DAKO), followed by incubation with StaphA for 15 min at RT. Supernatant of IP with

rIgG or mIgG was collected as input. Immune complexes were washed, eluted, cross-links reversed

and treated with RNAse.

Real Time PCR

DNA recovered from ChIP of histone modi!cations was subjected to quantitative real-time PCR using

AbsoluteTM QPCR SYBR Green ROX Mix (Abgene, Surrey, UK), ABI7900HT. The following formula was

calculated using % input = AE(Ctinput-CtChIP) * Fd *100%, where Fd is a dilution compensatory factor to

balance the di"erence in amounts of ChIP and input chromatin taken for immunoprecipitation. AE

represents the primer e#ciency, as determined for every primer set. The primers used for detection

of the EpCAM promoter (Figure 1A) were as follows: region C 5’-CACTCATTTTCTTCCCAAGAG-3’

(forward), 5’-GAACTGGATAGAGGAACGTG-3’ (reverse); region B2 5’-AGGTTCTCTGCCTGTGTTTG-3’

(forward), 5’-CGGCGTTAGGGATCTTTGGT-3’ (reverse); region A1 5’-CCCAACTCCCGGGCGGTGAC-3’

(forward) 5’-GGGTCCGCGTCGGGAGGACA-3’ (reverse). PCR conditions: 95°C for 15 min, followed by

40 cycles of 95°C for 30 s, 56°C (region C) or 59°C (region B2 and A1) for 30 s and 72°C for 40 s

and !nished with a dissociation curve to determine if the correct fragment is ampli!ed. A Q-PCR

calibration line freshly made from a DNA stock has been included for every primer set used.

PCR DNA recovered from ChIP of transcription factors was subjected to standard PCR using: 1 μM

primers (Region A and B1), 1.5 mM MgCl2, 0.2 μM dNTP (Fermentas GmbH, St. Leon-Rot, Germany),

1x Taq bu"er with (NH4)2SO4 (Fermentas) and 1.25U Taq DNA polymerase (Fermentas) in a total

volume of 25 μl. Next to region A1, we also screened for region B1 in the EpCAM promoter using

the primers: 5’-GCGCTCCTCCGGTTAAAAGGAAGTC-3’ (forward) 5’-GCGGCCCTCTCCACAGGTG-3’(

reverse). PCR conditions: 95°C for 5 min., followed by 30 cycles of 95°C for 30 s , 61°C for 30 s and

72°C for 45 s and !nished with 72°C for 5 min. PCR products were run on a 1.5% agarose gel and

visualized using ethidium bromide.

Electro Mobility Shift Assay

Nuclear extracts from OVCAR3 cells were prepared using a NE-PER kit (Pierce Biotechnology,

Thermo Fisher Scienti!c, Etten-Leur, The Netherlands) and quanti!ed using the Bradford

assay (Bio-Rad Laboratories, Veenendaal, The Netherlands). The sequences from

RDY681 labeled double-stranded oligonucleotides EpCAM probes (Isogen, De Meern

The Netherlands) were for Sp1a 5’-CTGGGGCGGGGCGGAGCGGG-3’; Sp1b

5’-CGCTCCGCCCCGCCGCGCGC-3’; NF-κB 5’-GGTCCGGGGACCCCCCTCGTC-3’; The core sequence

of the putative binding sites for Sp1 and NF-κB as analyzed with Genomatix/Matinspector 7.7.3.1

is underlined (Figure 1B). Probes were incubated with 4 μg of nuclear extracts in a 20 μl binding

bu"er (10 mM Tris pH 7.5, 50 mM KCl, 1 mM dithiothreitol, 1 μg of poly dI-dC) for 20 min at R.T. For

competition assays, a 100-fold excess amount of unlabeled competitor was premixed with RDY681

labeled probe before being added to the binding mixture. For competition with a methylated

46

Chapter 3

3

probe, the probes were in vitro methylated by M.SssI (Biolabs, New England) according to the

manufacturer’s recommendations. To adjust binding conditions, the unmethylated probe was

treated similarly but in the absence of methyl donor. The reaction products were resolved on a 4%

nondenaturing polyacrylamide gel at 100 V for 35’. The gels were visualized on the Odyssey Scanner

(Westburg, Leusden, The Netherlands).

RESULTS

EpCAM expression is associated with DNA methylation in ovarian cancer cell lines

To investigate whether the level of EpCAM expression of ovarian cancer cell lines correlates with the

DNA methylation level of the EpCAM promoter, 5 cell lines were subjected to bisul!te sequencing

for the area -443 to +282 relative to the transcription start site (Figure 1A). Ovarian cancer cell lines

were selected based on their EpCAM expression levels: two EpCAM negative lines A2780 and H134S

(MFI of 2.6±0.14 and 4.6±0.05, respectively), SKOV3 with an intermediate EpCAM expression level

(MFI of 104±3) and two cell lines (CaOV3 and OVCAR3) with a high EpCAM expression level (MFI of

461±30 and 496±24, respectively) (Figure 2). The overall methylation level was calculated for the

18 CpGs present in region B as the percentage of methylated CpGs of the total CpGs for 10 clones

sequenced, for each cell line.

Region A 0% (6)100% (4) 0.5% (6)1% (9)89% (6)

Region B 1% (10)99% (11) 3% (10)9% (10)56% (10)

0

100

200

300

400

500

600

A2780 H134S SKOV3 CaOV3 OVCAR3

EpC

AM

expr

essi

on(M

FI)

0

100

200

300

400

500

600

A2780 H134S SKOV3 CaOV3 OVCAR3

EpC

AM

expr

essi

on(M

FI)

DNA methylation

Figure 2. EpCAM expression correlates with DNA methylation in ovarian carcinoma cell lines. EpCAMexpression was measured by #ow cytometry. The Mean Fluorescence Intensity (MFI) of the average (±SD) of onerepresentative staining performed in triplicate is shown. The methylation status of the CpGs covering region Aand B of the EpCAM gene was determined by bisul!te sequencing. The % of DNA methylation represents thenumber of methylated CpGs divided by the total number of CpGs present in the region (Region A: 61, Region B:18). For each cell line the number of clones analyzed is indicated between brackets.

47

3


The EpCAM negative A2780 and H134S cell lines were hypermethylated (99±2% and 56±17%),

compared to the intermediate EpCAM expressing SKOV3 (9±13%), whereas the high EpCAM

expressing CaOV3 and OVCAR3 cell lines were hypomethylated (3±6% and 1±5%) (Figure 2).

Analysis of region A covering 61 CpGs showed 100±0% and 89±23% overall methylation for A2780

and H134S compared to 1±3%, 1±3% and 0±2% for SKOV3, CaOV3 and OVCAR3, respectively.

DNA methylation pro�le of the EpCAM gene in ovarian cancer patients

Next, we examined the methylation status of the EpCAM gene in tumor samples from patients with

advanced-stage serous ovarian cancer. Out of a patient cohort of 157 patients24, ten patients were

selected: 5 samples with the highest EpCAM mRNA levels (15.6-15.9) and 5 patients with lowest

levels (7.0-9.7) (Figure 3). DNA samples, retrieved from the same tumor samples as the mRNA, were

used to correlate changes in methylation status of the EpCAM gene with its expression in cancer

tissue. For each patient sample, 24 clones were analyzed and only 1.2-3.8% of the 61 CpGs located

in region A were methylated, except for one sample in which 10.3% of the CpGs was methylated. No

correlation between EpCAM mRNA expression and the methylation status of the EpCAM gene was

observed (p = 0.97 two tailed).

Figure 3. Correlation between EpCAM expression and DNA methylation in ovarian cancer samples. RNAand DNA were isolated from exactly the same tumor cells and EpCAM mRNA levels were determined by MIAME-compliant micro-array as described before.24 Out of the cohort of 157 patients with advanced-stage serousovarian cancer 10 samples were selected based on their EpCAM mRNA levels: 5 samples with the highest levels(15.6-15.9) and 5 samples with the lowest levels (7.0-9.7). Of each patient sample 24 clones were analyzed bybisul!te sequencing. The % of DNA methylation represents the number of methylated CpGs divided by thetotal number of CpGs present in region A (61). No correlation between EpCAM mRNA of ovarian cancer samplesand the DNA methylation status of region A in the EpCAM gene was found (X-axis is on log2 scale showing themicroarray expression signal and Y-axis on linear scale).

4 8 16 320.0

2.5

5.0

7.5

10.0

12.5

p = 0.97Spearman r = 0.02

EpCAM mRNA

%D

NA

met

hyla

tion

De�ning histone modi�cations associated with active and silent EpCAM promoter

To de!ne which histone modi!cations are involved in EpCAM gene expression, Chromatin Immu-

noPrecipitation was performed on our panel of ovarian cancer cell lines (Figure 2). In EpCAM posi-

tive cell lines, region C and B2 were associated with acetylated histone 4 (acH4), acetylated histone

3 (acH3) and to a lesser extent with trimethylation of lysine 4 of histone 3 (H3K4me3) (Figure 4).

48

Chapter 3

3

Figure 4. Characterization of histone modi!cations associated with EpCAM expression. Histonemodi!cations associated with region C (left) and region B2 (right) within the EpCAM gene in EpCAM negative(-) and positive (+) cells. Quantitative ChIP analysis was performed with the indicated antibodies, the absence ofantibody (no Ab) and rIgG were used as negative controls. The % of input DNA represents the relative amountof immunoprecipitated DNA compared to input DNA after Q-PCR analysis. For every EpCAM promoter regionanalyzed, each immunoprecipitate was quanti!ed in triplicate. The bars represent the mean of 3 or moreindependent ChIP experiments ± the SEM.

Figure 5. Enrichment of active or repressive histone marks encompassing the transcription start siteregion of the EpCAM gene. Histone modi!cations associated with region A1 within the EpCAM gene in EpCAMpositive (left) and negative cells (right) cells. Quantitative ChIP analysis was performed with the indicatedantibodies, the absence of antibody (no Ab) and rIgG were used as negative controls. The % of input DNArepresents the relative amount of immunoprecipitated DNA compared to input DNA after Q-PCR analysis. Eachimmunoprecipitate was quanti!ed in triplicate. The bars represent the mean of 3 or more independent ChIPexperiments ± the SEM.

Similar data were observed for region A1 covering the transcription start site, where for the three

EpCAM expressing cell lines enrichment of the histone modi!cations acH4, acH3 and H3K4me3 was

up to 6, 19 and 9% of input DNA, respectively (Figure 5A). In EpCAM negative cells, association of

these histone modi!cations was not detected, except for low levels of acH3 up to 1% of input DNA

at region A1 (Figure 5B).The repressive histone modi!cations trimethylation of lysine 9 of histone

3 (H3K9me3) as well as of lysine 27 of histone 3 (H3K27me3) were not detected in EpCAM positive

EpCAM positive cells EpCAM negative cells

A B

noAbrIg

Gac

H4ac

H3

H3K4m

e1

H3K4m

e3

H3K9m

e3

H3K27

me3

H3K36

me20

5

10

15

20

25

30

35

%of

inpu

tDN

A

noAbrIg

Gac

H4ac

H3

H3K4m

e1

H3K4m

e3

H3K9m

e3

H3K27

me3

H3K36

me20

1

2

3

4

5

6

%of

inpu

tDN

A

Region A1 Region A1

CaOV3 ++

SKOV3 +

H134S -

A2780 -

OVCAR3 ++

noAbrIg

Gac

H4ac

H3

H3K4m

e1

H3K4m

e3

H3K9m

e3

H3K27

me3

H3K36

me20

1

2

3

4

5

6

%of

inpu

tDN

A

Region C Region B2A B

no AbrIg

Gac

H4ac

H3

H3K4m

e1

H3K4m

e3

H3K9m

e3

H3K27

me3

H3K36

me20

1

2

3

4

5

6

%of

inpu

tDN

A

A2780 -

H143S -

SKOV3 +

CaOV3 ++

OVCAR3 ++

49

Figure 6. Transcription factors associated with the EpCAM gene. ChIP analysis on EpCAM positive (OVCAR3,CaOV3) and EpCAM negative (A2780, H134S) cells performed with the indicated antibodies, IgG was used as anegative control, PCR was performed with primers for region B1 (A) and region A1 (B). PCR products were run onan agarose gel and visualized using ethidium bromide.

3


cell lines for region A1, nor in the regions C and B2 in the EpCAM negative cell lines. Interestingly, in

the EpCAM negative cell lines the region covering the transcription start site (region A1) was associ-

ated with repressive marks: in A2780, but not in H134S, region A1 was associated with H3K9me3,

whereas in H134S, but not in A2780, the promoter was associated with H3K27me3 (Figure 5B).

In vivo EpCAM gene occupancy by transcription factors

To investigate which transcription factors might be involved in the regulation of the EpCAM gene,

we performed in silico analysis, using Genomatix MatInspector to determine putative binding

sites for transcription factors in the EpCAM promoter. The consensus sequences for a number of

transcription factors identi!ed by Matinspector are shown in Figure 1B. Next to previously published

putative sites26-29, additional binding sites and transcription factors were identi!ed. The transcription

factors to be screened, were selected based on these putative binding sites in the EpCAM gene and

on a) evidence for a biological role in EpCAM regulation2;20;28-30 and b) their potential role in ovarian

cancer.31-34

In the EpCAM positive cell line OVCAR3, the promoter was associated with Sp1, NFκB, LEF-1,

E2F2, Ets-1 and Ets2 for all regions tested (Figure 6). The transcription factors E2F4, p53, AP-2α and

STAT3 were only associated with region B1. In the EpCAM positive cell line CaOV3, the promoter

Region A1 (203bp)B

10% IgG Sp1 NF- B LEF-1 E2F2 E2F4 Ets1 Ets2 p53 AP2 STAT3 H2O bp

CaOV3

OVCAR3

10% IgG Sp1 NF- B LEF-1 E2F2 E2F4 Ets1 Ets2 p53 H2O bp

H134S

A2780200

100

B


10% IgG Sp1 NF- B LEF-1 E2F2 E2F4 Ets1 Ets2 p53 H2O bp

200

100

Region B1 (147bp)A

CaOV3

H134S

A2780

OVCAR3200

100


A

200

100

A

200

100


50

Chapter 3

3

was associated with the same transcription factors as for OVCAR3, except that for p53 and STAT3

no association was detected. The transcription factors Sp1, E2F2, Ets2 and again AP-2α were only

associated with region B1, whereas association of LEF-1 and Ets1 were only found in region A1. In

the EpCAM negative cell lines A2780 and H134S, no association of any of the transcription factors

with region B1 nor with region A1 was detected. In addition, no association of ESE-1, SNAI 1, SLUG,

PEA3, PDEF and NF-κBp65 was detected in EpCAM positive nor in EpCAM negative cells (data not

shown).

Interference on binding of transcription factors by DNA methylation

The above data suggest a role for Sp1 and NFκB in regulating EpCAM gene expression. Our present

bisul!te sequencing revealed that the two CpGs present in two putative binding sites for Sp1

(located at -231 and -226, Sp1a in Figure 1B), were both methylated in the EpCAM negative cell line

(A2780 11/11 clones, H134S 7/10 clones), whereas in the EpCAM positive cell lines these two CpGs

were not methylated (of the 10 clones for each cell line sequenced, one methylation event was

observed for SKOV3 and CaOV3, none for OVCAR3). Also for the CpG present in another putative

binding site for Sp1 (located at -32 , Sp1b in Figure 1B), complete methylation in all clones was

observed in the two EpCAM negative cell lines, whereas in the three EpCAM positive cell lines this

particular CpG was not once methylated. Similarly, for a putative NFκB binding site in the EpCAM

gene, the CpG next to the putative binding site of NFκB (+27) was methylated in all clones of the

EpCAM negative cell lines and never methylated in the EpCAM positive cell lines.

To investigate whether the observed DNA methylation indeed interferes with binding of the

transcription factors to the EpCAM promoter, EMSA competition studies were performed. Nuclear

protein extracts of OVCAR3 cells were incubated with unmethylated and methylated probes

containing 20 base pairs of the endogenous EpCAM sequence encompassing the putative binding

site for the transcription factors. Shift assay with probe Sp1a and OVCAR3 nuclear protein extracts

revealed two bands (a+b) with the unmethylated probe (Figure 7A). Both shifts were also observed

when the protein extract was incubated with the labeled (hot) methylated Sp1a probe, but to a

much lesser extent and in a di"erent ratio than observed for binding to the unmethylated Sp1a

probe (lane 2 compared with 6). Moreover, binding of the extract to the unmethylated Sp1a

probe showed competition with an excess of cold unmethylated probe, but not with an excess of

methylated probe, indicating that Sp1 binds preferentially to the unmethylated Sp1a binding site

within the EpCAM promoter.

Shift assay with the other probe Sp1b and OVCAR3 nuclear protein extracts revealed two bands

with the unmethylated probe as well as with the methylated probe (Figure 7B). One of the bands

is not speci!c (N.S.) since the band intensity was not reduced with an excess of competitor. The

other band indicated with a S, showed competition with both an excess of unmethylated as well as

an excess of methylated probe, indicating that for this particular sequence the transcription factor

binds to the Sp1b probe regardless of DNA methylation status of the CpGs within this probe. Also

51

3


for the NF-κB probe, no di�erence in binding patterns to the methylated and unmethylated NF-κB

probes was observed (data not shown).

Hot-Methylated

A B

Sp1a Sp1b

Cold-Unmet-P100X

Cold-Met-P100X

Hot-Unmethylated Hot-Methylated

competitors

N.S.

S

Hot-Unmethylated

Cold-Unmet-P100X

Cold-Met-P100X

competitors

ab

1 2 5 61 2 5 6 1 2 5 61 2 5 6

Figure 7. Interference of DNA methylation on binding of Sp1. Competition EMSA’s were performed withprobes containing the endogenous EpCAM sequence encompassing binding sites for Sp1. Nuclear extracts(NE) of OVCAR3 cells were incubated with labeled (hot) unmethylated or methylated probe. A) probe Sp1aand B) probe Sp1b. The speci!city and methylation sensitivity of the band of interest were shown by using theunlabeled competitors (lane 1,5: probe; 2,6: probe with NE; 3,4: probe with NE in the presence of 100-fold excessof indicated competitor).

DISCUSSION

Epigenetic aberrations, including DNA methylation and histone modi!cations are well established

in the development and progression of ovarian cancer.35;36 Silencing of tumor suppressor genes

by DNA methylation and global hypomethylation of repetitive sequences are frequent events in

ovarian cancer.35;36 In addition, a number of protein coding genes are overexpressed in ovarian

cancer due to loss of DNA methylation including maspin, claudin-337 and claudin-4.36 In addition,

overall loss of the repressive histone mark H3K27me3 has been associated with poor prognosis

in ovarian cancer.38 For ovarian cancer, the possible epigenetic dysregulation in relation to the

observed EpCAM overexpression is unknown. This study describes DNA methylation and histone

modi!cations regulating EpCAM gene expression in ovarian cancer. In addition, we identi!ed

several transcription factors with a potential role in EpCAM gene regulation.

As reported for several cell lines of di�erent tissue types19-22, also in ovarian cancer cell lines,

EpCAM expression was inversely correlated with DNA methylation of the promoter. Cell lines

that express high levels of EpCAM exhibited low methylation of the EpCAM promoter, and the

opposite was true for non-expressing cells. We performed a pilot study, including 5 samples with

52

Chapter 3

3

the highest and 5 samples with the lowest EpCAM mRNA expression out of a group of 157 patients

with advanced-stage serous ovarian cancer.24 All samples showed low levels of DNA methylation

and no correlation between EpCAM mRNA and DNA methylation was found. This absence of

correlation might be explained by the fact that in all the 157 tumor samples, EpCAM was relatively

highly expressed, as is generally observed for this type of ovarian cancer.8-10 Hence, the di�erence in

EpCAM expression levels might be too small to �nd a correlation with DNA methylation. To address

the importance of DNA methylation controlling EpCAM expression in ovarian cancer, EpCAM

negative tumor samples should be included (= 9-17% of all subtypes).10;11 Also for breast cancer

(n=20), no correlation between EpCAM expression and DNA methylation levels of the EpCAM gene

was found, while in this study EpCAM negative samples were included.19 Similarly to ovarian cancer,

a correlation between EpCAM expression and DNA methylation was demonstrated for breast cancer

cell lines. In contrast, in patients with lung adenocarcinoma20, oral squamous cell carcinoma23 and

colon cancer tissue22, high EpCAM expression was associated with hypomethylation of the promoter.

Since epigenetic regulation a�ects tissue speci�c gene expression, it might be that epigenetic

dysregulation of EpCAM expression is tumor type dependent. For example, loss of repressive

histone marks and induction of active histone marks might be of more importance than DNA

hypomethylation in EpCAM dysregulation in breast and ovarian cancer.38;39 Since the signi�cance

of DNA methylation on transcription factor binding is well established40;41, we investigated whether

Sp1 binding is methylation dependent. Interestingly, our observation that the CpG located at -231

within the Sp1 binding site was methylated in EpCAM negative ovarian cell lines and never in the

EpCAM positive lines was also reported for several other types of tumors.22 Reporter gene analysis in

which a EpCAM promoter fragment of -250 to +90 was used, showed an elevated promoter activity

in the presence of Sp1 compared to the activity in the absence of Sp1. We con�rmed association of

Sp1 with the endogenous EpCAM promoter in this region (B1 in Figure 1A). Moreover, association

appears to be restricted to the hypomethylated promoter in the EpCAM positive lines. Together

with our �nding that methylation of this particular CpG a�ects Sp1 binding, it is plausible that Sp1

plays an important role in EpCAM gene regulation.

Apart from Sp1, we also con�rmed association of LEF-1 with the EpCAM promoter. Recently,

association of LEF-1 was shown for the region -564 to -376 in a liver cancer cell line.29 Our in silico

analysis revealed an additional putative binding site for LEF-1 located at -270, and we indeed found

association of LEF-1 with region B1 (-332 to-185) in OVCAR3, but not in CaOV3 cells. In addition,

we found association of LEF-1 with region A1 (-138 to+65) in OVCAR3 and CaOV3 cells. Several

mechanisms have been proposed to explain recruitment of a speci�c transcription factor to occur in

the absence of a consensus motif.40 Cross-linking by formaldehyde creates also protein-protein cross-

links, allowing LEF-1 to be recruited via such interaction independent of its DNA binding abilities.40

Importantly, because it has been shown that β-catenin activation induced EpCAM transcription

through binding of LEF-1 to the EpCAM promoter29, interfering with this binding might lead to a

novel approach to inhibit EpCAM expression.

53

3


The transcription factors p5328 and NF-κB30 have been described as repressors of EpCAM expression.

Induction of wild-type p53, but not mutant p53, has been shown to decrease EpCAM expression.28

In our panel of cell lines, the EpCAM positive cell lines are mutated for p53, whereas the non-

expressing cell lines are not.42-45 Therefore, we expected to �nd association of p53 with the promoter

in the EpCAM negative cell lines. However, we only observed association of p53 with the promoter

in EpCAM positive OVCAR3 cells. In ovarian cancer, p53 is frequently mutated and recent evidence

showed that mutant p53 is required for the recruitment of the histone acetyl transferase p30046,

which is in agreement with the acetylated histones associated with the promoter in OVCAR3 cells.

Interestingly, for NF-κB the repressive action on promoter activity was abolished in the presence of

p300, suggesting that NF-κB competes with p300 for binding to the EpCAM promoter.30 Association

of Nf-κBp50 and acetylated histones with the EpCAM promoter was only detected in EpCAM

positive cells, it might be that the suggested competition between NF-κB and p300 is restricted

to the NF-κB subunit p65 which was not found to be associated with the promoter. In conclusion,

besides p53 and LEF-1, we are the �rst to show an association of other transcription factors with the

EpCAM gene.

The absence of the repressors p53 and NF-κBp50 in the negative cell lines could be explained by

the closed chromatin conformation. Indeed, for the silent promoters, we observed very low levels

of active histone marks (acetylation of histones H3 and H4, trimethylation of histone H3/lysine 447

and e�cient association of one of the repressive marks (H3/lysine 9, H3/lysine 27 trimethylation).

The reported insights into epigenetic mechanisms associated with EpCAM gene expression

may provide new opportunities for therapy. In this respect, silencing of gene expression can be

achieved by fusing a DNA methyltransferase or a histone modi�er to an EpCAM targeting DNA

binding domain. Recently, we showed that an EpCAM speci�c Triple helix Forming Oligonucleotide

coupled to a mutant methyltransferase is able to target methylation predominantly to a speci�c

DNA sequence in the EpCAM promoter (manuscript in press). Alternatively, zinc �ngers targeting

the EpCAM promoter when fused to a repressor or an activation domain have been shown to

modulate EpCAM promoter activity.48 In addition, DNA methyltransferases genetically fused to

zinc �ngers have shown to e�ciently repress reporter gene expression.49 Recently, it has been

demonstrated that this approach is also applicable to speci�cally methylate a genomic integrated

target promoter.50 Because epigenetic marks are inheritable, the forced epigenetic change in the

EpCAM gene is expected to be stable and therefore only one e�ective event might be e�cient to

induce a permanent change in EpCAM expression.

Acknowledgements

We thank Prof. A.G.J. van der Zee and Dr. B. Wisman of the Department of Gynecologic Oncology,

University Medical Center Groningen, Groningen, The Netherlands for providing the patients

samples. We thank Jelleke Dokter (UMCG) for culturing the cell lines.

54

Chapter 3

3

References


2. Maetzel D, Denzel S, Mack B, Canis M, Went P, Benk M et al. Nuclear signaling by tumor-associated antigenEpCAM. Nat Cell Biol 2009;11:162-71.






8. Kim JH, Herlyn D, Wong KK, Park DC, Schorge JO, Lu KH et al. Identi!cation of epithelial cell adhesionmolecule autoantibody in patients with ovarian cancer. Clinical Cancer Research 2003;9:4782-91.

9. Heinzelmann-Schwarz VA, Gardiner-Garden M, Henshall SM, Scurry J, Scolyer RA, Davies MJ et al.Overexpression of the Cell Adhesion Molecules DDR1, Claudin 3, and Ep-CAM in Metaplastic OvarianEpithelium and Ovarian Cancer. Clinical Cancer Research 2004;10:4427-36.

10. Kobel M, Kalloger SE, Boyd N, McKinney S, Mehl E, Palmer C et al. Ovarian carcinoma subtypes are di"erentdiseases: implications for biomarker studies. PLoS.Med. 2008;5:e232.


12. Bellone S, Siegel ER, Cocco E, Cargnelutti M, Silasi DA, Azodi M et al. Overexpression of epithelial celladhesion molecule in primary, metastatic, and recurrent/chemotherapy-resistant epithelial ovariancancer: implications for epithelial cell adhesion molecule-speci!c immunotherapy. Int.J Gynecol.Cancer2009;19:860-6.

13. Baeuerle PA, Gires O. EpCAM (CD326) !nding its role in cancer. Br.J Cancer 2007;96:417-23.

14. Sebastian M, Kuemmel A, Schmidt M, Schmittel A. Catumaxomab: A bispeci!c trifunctional antibody.Drugs Today (Barc.) 2009;45:589-97.

15. Seimetz D, Lindhofer H, Bokemeyer C. Development and approval of the trifunctional antibodycatumaxomab (anti-EpCAMá+áanti-CD3) as a targeted cancer immunotherapy. Cancer TreatmentReviewsIn Press, Corrected Proof.

16. Burges A, Wimberger P, Kumper C, Gorbounova V, Sommer H, Schmalfeldt B et al. E"ective relief ofmalignant ascites in patients with advanced ovarian cancer by a trifunctional anti-EpCAM x anti-CD3antibody: a phase I/II study. Clinical Cancer Research 2007;13:3899-905.

17. Wimberger P, Heubner M, Lindhofer H, Jager M, Kimmig R, Kasimir-Bauer S. In$uence of catumaxomab ontumor cells in bone marrow and blood in ovarian cancer. Anticancer Res 2009;29:1787-91.

18. Xiang W, Wimberger P, Dreier T, Diebold J, Mayr D, Baeuerle PA et al. Cytotoxic activity of novel humanmonoclonal antibody MT201 against primary ovarian tumor cells. J Cancer Res Clin.Oncol. 2003;129:341-8.


55

3


20. Tai KY, Shiah SG, Shieh YS, Kao YR, Chi CY, Huang E et al. DNA methylation and histone modi�cation regulatesilencing of epithelial cell adhesion molecule for tumor invasion and progression. Oncogene 2007.




24. Crijns AP, Fehrmann RS, de JS, Gerbens F, Meersma GJ, Klip HG et al. Survival-related pro�le, pathways, andtranscription factors in ovarian cancer. PLoS.Med. 2009;6:e24.

25. Weinmann AS, Farnham PJ. Identi�cation of unknown target genes of human transcription factors usingchromatin immunoprecipitation. Methods 2002;26:37-47.

26. Linnenbach AJ, Seng BA, Wu S, Robbins S, Scollon M, Pyrc JJ et al. Retroposition in a family of carcinoma-associated antigen genes. Molecular and Cellular Biology 1993;13:1507-15.


28. Sankpal NV, Willman MW, Fleming TP, May�eld JD, Gillanders WE. Transcriptional Repression of EpithelialCell Adhesion Molecule Contributes to p53 Control of Breast Cancer Invasion. Cancer Research 2009;0008-5472.

29. Yamashita T, Budhu A, Forgues M, Wang XW. Activation of Hepatic Stem Cell Marker EpCAM by Wnt {beta}-Catenin Signaling in Hepatocellular Carcinoma. Cancer Research 2007;67:10831-9.

30. Gires O, Kieu C, Fix P, Schmitt B, Munz M, Wollenberg B et al. Tumor necrosis factor alpha negativelyregulates the expression of the carcinoma-associated antigen epithelial cell adhesion molecule. Cancer2001;92:620-8.

31. Anttila MA, Kellokoski JK, Moisio KI, Mitchell PJ, Saarikoski S, Syrjanen K et al. Expression of transcriptionfactor AP-2alpha predicts survival in epithelial ovarian cancer. Br J Cancer 2000;82:1974-83.

32. Hall J, Paul J, Brown R. Critical evaluation of p53 as a prognostic marker in ovarian cancer. Expert.Rev.Mol.Med. 2004;6:1-20.

33. Reimer D, Sadr S, Wiedemair A, Stadlmann S, Concin N, Hofstetter G et al. Clinical relevance of E2Ffamily members in ovarian cancer--an evaluation in a training set of 77 patients. Clinical Cancer Research2007;13:144-51.

34. Silver DL, Naora H, Liu J, Cheng W, Montell DJ. Activated Signal Transducer and Activator of Transcription(STAT) 3: Localization in Focal Adhesions and Function in Ovarian Cancer Cell Motility. Cancer Research2004;64:3550-8.

35. Balch C, Fang F, Matei DE, Huang TH, Nephew KP. Minireview: epigenetic changes in ovarian cancer.Endocrinology 2009;150:4003-11.

36. Barton CA, Hacker NF, Clark SJ, O’Brien PM. DNA methylation changes in ovarian cancer: Implications forearly diagnosis, prognosis and treatment. Gynecologic Oncology 2008;109:129-39.

37. Honda H, Pazin MJ, D’Souza T, Ji H, Morin PJ. Regulation of the CLDN3 gene in ovarian cancer cells. CancerBiol.Ther. 2007;6:1733-42.

38. Wei Y, Xia W, Zhang Z, Liu J, Wang H, Adsay NV et al. Loss of trimethylation at lysine 27 of histone H3 is apredictor of poor outcome in breast, ovarian, and pancreatic cancers. Mol Carcinog. 2008;47:701-6.

39. Lu TY, Lu RM, Liao MY, Yu J, Chung CH, Kao CF et al. Epithelial cell adhesion molecule regulation is

56

Chapter 3

3

associated with the maintenance of the undi�erentiated phenotype of human embryonic stem cells1. J Biol.Chem. 2010;285:8719-32.

40. Farnham PJ. Insights from genomic pro!ling of transcription factors. Nat Rev Genet 2009;10:605-16.

41. Vincent A, Perrais M, Desseyn JL, Aubert JP, Pigny P, Van Seuningen I. Epigenetic regulation (DNAmethylation, histone modi!cations) of the 11p15 mucin genes (MUC2, MUC5AC, MUC5B, MUC6) inepithelial cancer cells. Oncogene 2007;26:6566-76.

42. Bagnoli M, Balladore E, Luison E, Alberti P, Raspagliesi F, Marcomini B et al. Sensitization of p53-mutatedepithelial ovarian cancer to CD95-mediated apoptosis is synergistically induced by cisplatin pretreatment.Molecular Cancer Therapeutics 2007;6:762-72.

43. Horiuchi A, Wang C, Kikuchi N, Osada R, Nikaido T, Konishi I. BRCA1 Expression is an Important Biomarkerfor Chemosensitivity: Suppression of BRCA1 Increases the Apoptosis via Up-regulation of p53 and p21During Cisplatin Treatment in Ovarian Cancer Cells. Biomark.Insights. 2007;1:49-59.

44. Kolfschoten GM, Hulscher TM, Schrier SM, van Houten VMM, Pinedo HM, Boven E. Time-DependentChanges in Factors Involved in the Apoptotic Process in Human Ovarian Cancer Cells as a Response toCisplatin. Gynecologic Oncology 2002;84:404-12.

45. Tonini T, Gabellini C, Bagella L, D ÇÖAndrilli G, Masciullo V, Romano G et al. pRb2/p130 DecreasesSensitivity to Apoptosis Induced by Camptothecin and Doxorubicin but not by Taxol. Clinical CancerResearch 2004;10:8085-93.

46. Strano S, Dell’Orso S, Di AS, Fontemaggi G, Sacchi A, Blandino G. Mutant p53: an oncogenic transcriptionfactor. Oncogene 2007;26:2212-9.



49. Jeltsch A, Jurkowska RZ, Jurkowski TP, Liebert K, Rathert P, Schlickenrieder M. Application of DNAmethyltransferases in targeted DNA methylation. Appl Microbiol.Biotechnol. 2007;75:1233-40.

50. Smith AE, Hurd PJ, Bannister AJ, Kouzarides T, Ford KG. Heritable Gene Repression through the Action of aDirected DNA Methyltransferase at a Chromosomal Locus. Journal of Biological Chemistry 2008;283:9878-85.

57

Serum insensitive, intranuclear protein delivery by themultipurpose cationic lipid SAINT-2

Bernardina T.F. van der Gun1, Amélie Monami2, Sven Laarmann3, Tamás Raskó4,Krystyna Ślaska-Kiss4, Elmar Weinhold2, Reinhold Wasserkort5, Lou F.M.H. de Leij1,

Marcel H.J. Ruiters1, Antal Kiss4, Pamela M.J. McLaughlin1

1 Department of Pathology and Laboratory Medicine, Section Medical Biology, University Medical CenterGroningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands. GUIDE, Groningen University Institute for DrugExploration, Oostersingel 59, 9713 EZ Groningen, The Netherlands

2 Institute of Organic Chemistry, RWTH Aachen University, Landoltweg 1, D-52056 Aachen, Germany3 EUREGIO Biotech Center, Fachhochschule Münster, Stegerwaldstrasse 39, 48565 Steinfurt, Germany4 Institute of Biochemistry, Biological Research Center of the Hungarian Academy of Sciences, Temesvári krt. 62,

6726 Szeged, Hungary5 Epigenomics AG, Kleine Präsidentenstrasse 1, D-10178, Berlin, Germany

J Control Release 2007;123(3):228-38

Chapter 4

58

Chapter 4

4

ABSTRACT

Cationic liposomal compounds are widely used to introduce DNA and siRNA into viable cells, but

none of these compounds are also capable of introducing proteins. Here we describe the use of a

cationic amphiphilic lipid SAINT-2:DOPE for the e�cient delivery of proteins into cells (profection).

Labeling studies demonstrated equal delivery e�ciency for protein as for DNA and siRNA.

Moreover, proteins complexed with SAINT-2:DOPE were successfully delivered, irrespective of the

presence of serum, and the profection e�ciency was not in�uenced by the size or the charge of the

protein:cationic liposomal complex. Using -galactosidase as a reporter protein, enzymatic activity

was detected in up to 98% of the adherent cells, up to 83% of the suspension cells and up to 70%

of the primary cells after profection. A delivered antibody was detected in the cytoplasm for up to 7

days after profection. Delivery of the methyltransferase M.SssI resulted in DNA methylation, leading

to a decrease in E-cadherin expression. The lipid-mediated multipurpose transport system reported

here can introduce proteins into the cell with an equal delivery e�ciency as for nucleotides. Delivery

is irrespective of the presence of serum, and the protein can exert its function both in the cytoplasm

and in the nucleus. Furthermore, DNA methylation by M.SssI delivery as a novel tool for gene

silencing has potential applications in basic research and therapy.

59

4

Serum insensitive, intranuclear protein delivery by the multipurpose cationic lipid SAINT-2

INTRODUCTION

E!cient intracellular delivery of protein, DNA or siRNA is crucial in achieving the desired biological

function of the molecule of interest in vivo. As protein delivery does not require processing by the

host cell’s endogenous machinery, the interest in intracellular delivery of functionally active proteins

has emerged. The most intensively studied approaches of protein delivery are the so-called protein

transduction domains (PTDs).1;2 PTDs deliver the protein of interest by fusing it to peptide sequences

of 10-35 amino acids long that are capable of penetrating the cell’s membrane. The major drawback

of all PTD-mediated protein delivery systems is that the transduction-domain must be covalently

attached to the cargo protein, either by creating a DNA construct in a specially designed vector or

by chemically cross-linking the protein and the PTD via functional groups on each molecule.1 This

makes it very di!cult to investigate the e"ect of the cargo protein itself because the attached PTD

can in#uence the characteristics of the cargo protein.3

Another strategy to deliver proteins is similar to the method widely used for DNA delivery. In

this case the cargo protein is complexed with a delivery reagent, which subsequently delivers the

protein into the cell (profection). Zelphati et al. tested 7 widely used DNA transfection reagents

for protein delivery, and found that the profection-rate never exceeded 5%.4 Out of 25 newly

developed cationic lipid formulas they identi$ed one cationic liposomal compound, which had the

capacity to deliver proteins into the cell. This compound, named BioPORTER (BP), was used to deliver

-galactosidase and several caspases into the cytoplasm where they remained fully active.4 In

contrast to PTDs, liposomes form a non-covalent complex with the cargo-protein, and release the

cargo-protein after fusion with the cell membrane.

Ye et al. evaluated both strategies and concluded that protein delivery using PTDs derived from

the HIV Tat protein or HPV derived VP22 were not nearly as e"ective as a cationic lipid formulation.3

Optimal e!ciency of delivery with the BP cationic lipid formulation was reached after 4 h of

incubation.4 However, longer incubation periods led to decreases in both the number of positive

cells and the intracellular #uorescence intensity.4 This rapid decrease in protein content is a major

disadvantage, as the protein can only exerts its function for a short period of time. Dalkara et al

developed a cysteine-based cationic amphiphile capable to deliver protein into cells over a longer

period of time5 but this agent requires serum-free conditions. The same requirement applies for

the Bp formulation which limits the potential of these agents for in vivo delivery applications. In

combination with its toxicity that was occasionally observed for Bp6, novel in vivo protein delivery

methods are desirable.

In this study, the widely used nucleotide delivery agent SAINT-27 was tested for its ability to deliver

proteins. The cationic amphiphilic lipid SAINT-2 (N-methyl-4(dioleyl)methyl-pyridinium-chloride) is

composed of a pyridinium group which bears the positive charge with two hydrophobic alkyl chains

attached at the para-position. Although SAINT-2 is capable of delivering DNA or siRNA by itself8, it

is usually applied in a 1:1 ratio with the neutral helper-lipid dioleoylphosphatidylethanolamine

60

Chapter 4

4

(DOPE). SAINT-2:DOPE (SD) has been successfully used to deliver nucleotides in vitro and in vivo where

it did not elicit any immune response or showed any other sign of toxicity after injection via various

routes into mice.9;10 To be able to compare the SD-mediated protein delivery with the previously

described BP-mediated protein delivery, the same model proteins e.g. β-galactosidase and labeled

and unlabeled poly- and monoclonal antibodies were used.

Moreover, we established nuclear activity of the introduced protein, by the delivery of the

prokaryotic DNA (cytosine-5) methyltransferase (MTase) M.SssI, which methylates cytosines in

CpG dinucleotides11 Since M.SssI has the same base and sequence speci"city as mammalian DNA

MTases, this enzyme appears to be an excellent tool to study the role of DNA methylation in healthy

and diseased eukaryotic cells provided that it can be delivered into the cell nucleus. Methylation of

promoters and their genes usually leads to loss of gene expression, a process termed silencing.12

Therefore, M.SssI delivery would provide a novel tool for gene silencing. Gene therapy strategies

require DNA incorporation into genomes, and transcription into biologically active proteins. The

major disadvantage of siRNA treatment is the need for sequential deliveries, since the down-

regulation is transient. Protein therapy circumvents such problems and the here reported delivery

agent will contribute to progression of the protein therapy "eld.


Cell culture

The murine B16-F10 (CRL-6475) melanoma, the simian kidney COS-7 (CRL 1651), the human

embryonic kidney HEK293A (CRL-1573), the glioblastoma U373MG (HTB-17), the ovarian

adenocarcinoma SKOV-3 (HTB-77) and the Jurkat T (TIB-152TM) cell lines were obtained from the

ATCC (Manassas, VA) and cultured according to ATTC recommendations. Primary human skin

"broblasts (HSF) were grown out of a skin biopsy obtained under informed consent and cultured in

DMEM supplemented with 50 μg/ml gentamicin sulfate, 2 mM L-glutamine, 10% FBS (BioWhittaker

Inc, Walkersville, MD) and trypsinized (10X trypsin; ICN Biomedicals, Irvine, CA). Endothelial cells

(EC) were isolated from human umbilical veins (HUVEC) and cultured in RPMI1640 supplemented

with 20% FBS, 2 mM L-glutamine, 5 U/ml heparin (Leo, Weesp, The Netherlands), 50 μg/ml crude

endothelial cell growth factor (home isolated), 100 μg/ml streptomycin and 100 U/ml penicillin in

1% gelatin-coated tissue culture $asks. The human bronchialepithelial cell line 16HBE was a gift from

Dr. Gruenert, University of California, San Francisco and was grown in EMEM supplemented with

L-glutamine, 10% FBS,penicillin/streptomycin (100 U/ml) in collagen-(bovine; Inamed, Fremont, CA)

and BSA-coated tissue culture $asks. All cell lines were cultured at 37°C in a humidi"ed 5% CO2

atmosphere.

61

4


Plasmids

The plasmids pBHNC-MSssI and pBHNC-MSssI(C141S) were described previously.13 The variant

encoded by pBHNC-MSssI has MTase activity comparable to the wild-type enzyme, and will be

referred to as M.SssI. Its mutant derivative, M.SssI(C141S) encoded by pBHNC-MSssI(C141S), in

which the active site cysteine is replaced by serine, has a greatly reduced (2-5%) activity relative to

the wild-type enzyme.13

Proteins

E. coli ER1821 cells, harboring pBHNC-MSssI or pBHNC-MSssI(C141S), were grown at 37°C in LB

containing 100 µg/ml ampicillin. At OD600 ~0.6, M.SssI or M.SssI(C141S) production was induced

by adding 1.0% arabinose. After 4 h incubation at 30°C, cells were harvested by centrifugation,

resuspended in breaking bu#er (50 mM Na2HPO4, pH 8.0, 300 mM NaCl, 1 mM imidazole), sonicated

and cell debris removed by centrifugation. For puri$cation a His-Select Nickel A&nity gel column

(1 ml, Sigma) was used according to the manufacturers instructions. The eluate was diluted with

cation exchange bu#er (6.7 mM MES, 6.7 mM Hepes, 6.7 mM NaOAc, pH 7.5, 1 mM EDTA, 10 mM

-mercaptoethanol, 10% glycerol) and applied to a HS POROS 50 column (Applied Biosystems,

Fostercity, CA). After washing with 100 ml cation exchange bu#er containing 0.2 M NaCl, proteins

were eluted with a linear NaCl gradient (0.2–1 M) in cation exchange bu#er. DNA MTase containing

fractions were pooled, concentrated by ultra$tration, mixed with an equal volume of glycerol and

stored at -20°C. All puri$cation steps were performed at 4°C.

Recombinant β-galactosidase was purchased from Sigma. Polyclonal Swine Anti-Rabbit

Immunoglobulins/FITC (S!R-FITC) was purchased from DAKO (Glostrup, Denmark) and Rabbit

Anti-Mouse Alexa Fluor488-labeled antibody (RαM488) from Invitrogen (Carlsbad, CA). The mouse

monoclonal antibody MOC31 (IgG1), which recognizes the human Epithelial Cell Adhesion Molecule

(EpCAM), was produced by a hybridoma cell line and puri$ed by protein A column chromatography

(Prosep A high capacity, Millipore) in our laboratory. MOC31 was labeled with the Alexa Fluor488

+uorescent dye using the Alexa Fluor488 (MOC31488) Labeling Kit (Molecular Probes Inc., Eugene, OR).

Delivery experiments

SAINT-2:DOPE (SD; 0.75 mM) and SAINT-2 (S; 0.75 mM) were purchased from Synvolux Therapeutics

Inc. (Groningen, The Netherlands). For transfection cells were grown subcon+uent in 6-well plates

(Costar, San Diego, CA) or Lab-Tek Chamber Slides (Nunc, Rochester, NY). For one 6-well: 1 µg DNA

or siRNA was complexed with 20 µl SD. The complex was adjusted to 1 ml with serum free medium.

Within 15 min, the complexes were added to the cells. After 3 h, 2 ml serum-containing medium was

added to one 6-well. For 24 well plates and Lab-Tek Chamber slides the amount of DNA/siRNA and

SD was adjusted according to the surface.

For profection cells were seeded 0.5-1 x 106/6 well, 2-3 x 105/24 well or 5-10 x 104/chamber slide

well. Protein delivery was performed when cells were 50-80% con+uent. For one 6-well plate 0.05-

62

Chapter 4

4

20 µg MOC31 or 5 µg M.SssI/C141S in 100 µl PBS were complexed respectively with 20 µl S or 20

µl SD in an equal volume of PBS, for one 24-well: 2 or 10 µg β-galactosidase, 0.0625-5 µg RαM488 or

5 µg MOC31 diluted in 25 µl PBS were complexed with 10 µl SD diluted with equal volume of PBS,

and for one chamber slide well 0.625 µg M.SssI/C141S in 12.5 µl PBS was complexed with 2.5 µl SD

in an equal volume of PBS. The SD:protein complex was incubated for 5 min at RT, 100, 200 and 800

µl serum-free medium was added to respectively one chamberslide well, one 24 or one 6 well and

the complete mixture was pipetted directly onto the cells. As a control, protein, S or SD alone were

added to the cells. After 3 h media were completed with FBS (10%). The cells were incubated for 4,

24, 48, 72, 96 or 168 hours at 37ºC in humidi�ed 5% CO2.

To investigate the in�uence of serum on profection e�ciency the SD-protein complex was

directly diluted in 250 µl medium with 0%, 10%, 20% or 50% FBS. BioPORTER (BP)4 cationic lipid

mixture system was purchased from Gene Therapy Systems Inc. (San Diego, CA) and delivery was

performed according to the manufacturers protocols. Brie�y, the BP dry �lm was resuspended in 250

µl methanol and vortexed for 10-20 s. Then 2.5 µl of BP was transferred into an Eppendorf tube and

the solvent was evaporated at RT. β-Galactosidase was diluted in PBS and used to hydrate the dried

BP formulation. Finally, medium was added to the complexes.

Imaging of enzyme/antibody import

The RαM488 antibody import was measured by �ow cytometric analysis 48 h after profection.

Cells were washed with PBS, detached from the plate with trypsin and resuspended in medium,

the percentage RαM488 positive cells was measured on a Calibur �ow cytometer. To determine

β-galactosidase activity, cells were washed with PBS, trypsinized, harvested in cold medium and

put on ice. Cytospots were generated and �xed for 5 min in acetone/methanol 1:1 and stained for

β-galactosidase activity at 37ºC with 0.2% BluoGal overnight (Invitrogen). Counter-staining was

performed with 0.1% neutral red (Sigma). Via bright-�eld microscopy, the percentage cells displaying

β-galactosidase activity was determined by counting the number of active cells per 100 cells on a

slide, three times in di!erent �elds. For MOC31 detection cells were harvested similarly. Cytospots

were �xed for 10 min in acetone and stained with Rabbit anti-Mouse peroxidase conjugate (R M-PO,

DAKO). After washing with PBS, cells were incubated with 3-amino-9-ethyl-carbazole (AEC, Sigma) in

combination with 0.01% H2O2. Counter-staining was performed with Mayers heamatoxylin solution

(Merck, Haar, Germany). MOC31 detection was also performed by staining with R M-F(ab)2-FITC

(DAKO). For E-cadherin detection, cells were washed with PBS containing 0.9 mM CaCl2 and �xed

for 30 min in 4% PFA, incubated for 10 min in 0.2% TritonX-100 and for 30 min in 10% human pool

serum. E-cadherin was detected by staining with Rabbit-anti-E-cadherin (Santa Cruz Inc., Santa Cruz,

CA) followed by Goat Anti-Rabbit-Alexa488 (G R488, Invitrogen). Nucleoli were stained with DAPI for

�xed cells and with Hoechst 33342 (Molecular Probes) for living cells. Evaluation took place by the

Quantimet (Leica, 600S). To follow MOC31488 antibody in living cells, cells were monitored with the

Leica DM-IL �uorescence microscope equipped with a digital camera from 4 h for every consecutive

63

4


24 h till 168 h and the !uorescent intensity was evaluated. To visualize the intracellular localization

of MOC31488, SKOV-3 cells were "xed 24 h after profection with 4% PFA and the plasmamembrane

was counterstained with WGA-Biotin (Vector Labs)/streptavidin-RRX (Jackson Immunoresearch).

Confocal images were obtained using a LSM510ETA confocal laser scanning microscope, optimal

pin hole (1 airy unit) 0.4 µm step size.14

Western blotting

Cells were washed with PBS containing 0.9 mM CaCl2. After addition of 400 μl ice-cold lysis bu&er (20

mM Tris-HCL, 5.0 mM EDTA, 2.0 mM EGTA, 100 mM NaCl, 0.05% SDS, 0.5% NP-40, 1 mM PMSF, 10 μg/

ml Aprotein, 10 μg/ml Leupeptin and loading bu&er with β-mercaptoethanol) to one 6-well and 10

min of incubation on ice, the cell lysate was transferred to a tube and heated to 100°C for 5-7 min.

Samples separated by SDS-PAGE were semi-dry electroblotted onto 0.45 μm nitrocellulose transfer

membranes (Schleichel&Schuell, Dassel, Germany). After blotting, the "lters were blocked overnight

with 5% BSA in PBS supplemented with 0.1% Tween-20 (PBS-T). After washing with PBS-T, the blot

was incubated with Rabbit anti-E-cadherin, followed by detection with G R-HRP. Visualization was

done using the ECL chemoluminoscence detection kit (Pierce, Rockford, IL).

Particle size and zeta-potential measurement

The size and zeta-potential of the SD:protein complexes were measured with a Nicomp 380/ZLS

apparatus from Particle Sizing Systems Inc. (SANYA, Barbara, CA). For size-measurements SD and

protein were mixed in a 1:1 ratio in the solvent used for delivery and were incubated for 5 min

to allow complex-formation. For the measurement of the zeta-potential the SD:protein complexes

were diluted in H2O and an electric "eld strength of 2.75 V/cm was applied.

Methylation analysis

The untranslated 5’-region of the E-cadherin promoter plus part of exon 1 (amplicon 3027) was

selected to assess the methylation status in SKOV3 cells (Figure 4B). DNA extracted from the cells was

subjected to bisulphite treatment as previously described.15 Bisulphite speci"c primers void of any

CpG were used in order to obtain ampli"cation products unbiased for the methylation status. The

primers were GATTTTAGTAATTTTAGGTTAGAGGG (forward) and AAATACCTACAACAACAACAACAA

(reverse). PCR was done in a "nal volume of 25 µl using 1 U of Hot Star Taq DNA polymerase (Qiagen,

Germany), 1x Taq bu&er, 0.2 mM dNTPs (Fermentas, Germany), 1.5 mM MgCl2, 0.24 µM of each primer

and the following cycling conditions: 95°C for 15 min, followed by 40 cycles of 95°C for 60 s, 55°C for

45 s and 72°C for 60 s, "nished with 72°C for 10 min. Puri"ed PCR products were sequenced by an

ABI3730-capillary sequencer using the ABI Prism Big Dye Terminator V3.1 sequencing chemistry. The

obtained trace "les were subsequently analysed using the ESME software as previously described.16

64

Chapter 4

4

Table 1. Intracellular delivery of β-galactosidase with SAINT-2:DOPE in multiple cell types.

Cell type SD + 2 μg β-gal SD + 10 μg β-gal

B16F10 80 ± 10 nd

COS-7 80 ± 10 nd

SKOV-3 98 ± 2 98 ± 2

U373MG 95 ± 4 95 ± 5

HEK293A 75 ± 8 71 ± 4

HUVEC 70 ± 10 70 ± 6

HSF nd 60 ± 8

Jurkat 83 ± 1 74 ± 16

Profection-e$ciency in percentages of di%erent cell types, 4 h after profection with 2 or 10 µg β–galactosidaseand 10 µl 0.75 mM SAINT-2:DOPE (SD) per one 24 well. Cells were stained for β-galactosidase activity and thepercentage of cells displaying β-galactosidase activity determined via microscopy by counting enzymaticallyactive cells per 100 cells on a slide, three times in a di%erent 'eld (nd = not done).

RESULTS

SAINT-2:DOPE-mediated intracellular delivery of DNA, siRNA and proteins

To investigate whether SD is able to deliver functional proteins into cells, COS-7 cells were profected

with the β-galactosidase enzyme. As a control, cells were transfected with the CMV-β-galactosidase

plasmid. With profection, optimal β-galactosidase activity was observed 4 h after delivery, whereas

with transfection optimal enzymatic activity was reached after 48 h (Figure 1A). No β-galactosidase

activity was detected after addition of SD or β-galactosidase alone.

To compare e$ciency of protein delivery with that of DNA and siRNA delivery, we delivered

Cy3-labeled PuC18, Alexa Fluor546-labeled siRNA and Alexa Fluor488-labeled MOC31 antibody with

SD into EpCAM-negative U373MG cells. No signi'cant di%erence in the number of positive cells was

observed 48 h after DNA, siRNA or protein delivery (Figure 1B). Similar results were obtained after

24 and 72 h (data not shown).

A broad range of cells can be profected with SAINT-2:DOPE

To establish the general applicability of protein delivery with SD, we performed SD-mediated

delivery of β-galactosidase into di%erent cell types. Table 1 depicts the profection e$ciency

of β-galactosidase by SD into various cell types in percentages as determined by counting the

number of enzymatically active cells per 100 cells on a slide. The median profection e$ciency of

β-galactosidase delivered into the adherent cell lines HEK293, B16.F10, COS-7, SKOV-3 and U373MG

varied from 75 to 98%, whereas the median delivery e$ciency for the Jurkat cell line, growing

in suspension, was 83%. In the di$cult to transfect, primary skin 'broblasts (HSF) and umbilical

cord derived endothelial cells (HUVEC) a median profection e$ciency of approximately 70% was

observed.

65

4


SD + b-G alactosidase

SD

SD + C MV -b-G alactosidase

b-G alactosidaseA


SD


b-G alactosidase


SD


b-G alactosidase

SD +ß-Galactosidase

SD

SD + C MV-ß-Galactosidase

ß-Galactosidase

B

Figure 1. SAINT-2:DOPE (SD) mediates intracellular delivery of DNA, siRNA and protein. A) Transfectionand profection of COS-7 cells with SD of CMV-β-galactosidase plasmid or β-galactosidase enzyme resultedin a positive staining for β-galactosidase activity, 48 h or 4 h respectively, after delivery (magni"cation: x10).The enlargement shows the characteristic dotted pattern found after protein delivery.29 B) Comparison oftransfection e#ciency of 0.125 µg Cy3-labeled DNA (red), 0.125 µg Alexa546 labeled siRNA (red) with profectione#ciency of 1 µg MOC31488 (green) by 2.5 µl SD into U373MG cells (upper with, and lower pictures without SD).Nuclei of the living cells were stained blue with Hoechst33342 and images were taken with the Leica, 600SQuantimet (magni"cation: x40).

66

Chapter 4

4

Determination of optimal conditions for protein delivery

The profection e�ciency depends on the amount of SD used in relation to the amount of protein

it is complexed with. To determine the optimal amount of SD, 2 µg of RαM488 was complexed with

a volume of 0.75 mM SD ranging from 2.5 to 20 µl (Figure 2A). Flow cytometric analysis showed

that optimal delivery was achieved with a volume of 10 µl SD. More than 10 µl SD did not lead

to an increase in profection e�ciency for U373MG or for SKOV3 cells. To determine the optimal

amount of protein, 10 µl SD was subsequently complexed with di�erent amounts of RαM488 ranging

from 0.0625 to 5 µg (Figure 2B). The number of RαM488 positive cells increased with an increasing

concentration of antibody. A similar correlation was observed between the increase of protein

concentration and the mean �uorescence intensity of the cells (data not shown). Hardly any positive

staining was detected when U373MG or SKOV3 cells were treated with corresponding amounts of

RαM488 without SD (data not shown).

To monitor protein delivery in time in living cells, we delivered the antibody MOC31488 into

EpCAM-positive SKOV-3 and EpCAM-negative U373MG cells (Figure 2C). Within 4 h the antibody

localized intracellular. In the EpCAM-negative cells a strong uptake was observed after 48 h, which

lasted for at least 168 h after profection. At each time point, the EpCAM-negative U373MG cells

treated with antibody in the absence of SD displayed no intracellular �uorescence (data not shown).

Since DNA delivery via SAINT-2 does not require the helper lipid DOPE8 we investigated if this

also applies to protein delivery. We delivered the antibody MOC31 into EpCAM-negative COS-7

cells (Figure 2D). The number of MOC31 positive cells increased with an increasing concentration of

antibody. A similar linear correlation was observed between the increase of protein concentration

and the amount of introduced antibodies per cell. Hardly any positive staining was detected when

COS-7 cells were treated with corresponding amounts of MOC31 without S.

In!uence of serum on SAINT-2:DOPE mediated protein delivery

Transfection with SD is not signi!cantly a�ected by the presence of serum17;18, whereas profection

using the cationic lipid formulation BP was reported to be unachievable in the presence of serum4,

which is a serious limitation in the applicability of the method. Therefore, we investigated whether

SD can mediate protein delivery in the presence of serum and compared this to protein delivery

mediated by BP under similar conditions. To this end, β-galactosidase was complexed either with SD

or BP. In the absence of serum, both reagents were able to deliver 2 µg active β-galactosidase into

SKOV-3 cells (Figure 3A); however BP was less e�cient when 10 µg β-galactosidase was used. To

investigate the e�ect of serum, we delivered β-galactosidase to SKOV-3 cells in the presence of FBS

ranging from 0 to 50% (Figure 3B). Even in the presence of 50% FBS, SD was capable of delivering 10

µg β-galactosidase with a profection e�ciency of 70%, whereas delivery with BP failed.

To investigate if SD-mediated delivery in the presence of serum is applicable to proteins in

general, MOC31 was delivered to U373MG cells in the presence and absence of 10% FBS (Figure 3C).

Immunohistochemical staining of the !xed cells showed that MOC31 was successfully delivered in

the presence of serum.

67

4

Figure 2. Optimization of SAINT-2:DOPE (SD) mediated intracellular protein delivery. A) SD dose-deliverycurve; 2 µg of RαM488 antibody was delivered into U373MG and SKOV3 cells with various amounts of SD. After48 h, the % RαM488 positive cells was measured by #ow cytometry. The % RαM488 positive cells was plottedagainst the dose applied. B) Protein dose-delivery; 10 µl SD was complexed with increasing amounts of RαM488

ranging from 0.0625 to 5 µg and delivered to U373MG and SKOV3 cells. The number of RαM488 positive cellsincreased with an increasing concentration of antibody. C) To determine the intracellular stability of the proteinafter delivery, 2 µg MOC31488 complexed with 10 µl SD was profected into SKOV-3 (black bars) and U373MG (greybars) cells. Analysis of the #uorescence intensity of MOC31488 in the living cells was evaluated with the Leica DM-IL #uorescence microscope, after 4, 24 h and every consecutive 24 h till 168 h. D) Profection with SAINT-2 only;increasing concentration ranging from 0.05 to 20 µg/ml MOC31 was delivered to COS-7 cells without (upper rowof pictures) or with a constant volume of 20 µl S (lower row of pictures; magni$cation: x40). After 48 h MOC31detection was performed by immunoperoxidase staining with R M-PO.

C.

0

1

2

3

4

5

6

7

8

9

10

4 24 48 72 96 168

Incubation time (hours)Incubati on ti me in hours4 24 48 72 96 168

0

5

10

4

3

2

1

7

6

9

8

Rela

tive

!uor

esce

ntin

tens

ity

g RAM

C.

0

1

2

3

4

5

6

7

8

9

10

4 24 48 72 96 168


0

5

10

4

3

2

1

7

6

9

8

Rela

tive

!uor

esce

ntin

tens

ity

C.

0

1

2

3

4

5

6

7

8

9

10

4 24 48 72 96 168


0

5

10

4

3

2

1

7

6

9

8

Rela

tive

!uor

esce

ntin

tens

ity

0

1

2

3

4

5

6

7

8

9

10

4 24 48 72 96 168

Incubation time (hours)

0

1

2

3

4

5

6

7

8

9

10

4 24 48 72 96 168


Incubation time in hours

4 24 48 72 96 1684 24 48 72 96 1680

5

10

4

3

2

1

7

6

9

8

Rela

tive

!uor

esce

ntin

tens

ity

0

5

10

4

3

2

1

7

6

9

8

0

5

10

4

3

2

1

7

6

9

8

µg RAMC

0.05mgRaM -PO 0.5mg 2.0mg 5.0mg 10.0mg 20.0mg

0.05mgS 0.5mg 2.0mg 5.0mg 10.0mg 20.0mg


0.05mgS 0.5mg 2.0mg 5.0mg 10.0mg 20.0mg


0.05mgS 0.5mg 2.0mg 5.0mg 10.0mg 20.0mg

0.05 gRaM -PO 0.5 g 2.0 g 5.0 g 10.0 g 20.0 g0.05 gRaM -PO 0.5 g 2.0 g 5.0 g 10.0 g 20.0 g

0.05 gS 0.5 g 2.0 g 5.0 g 10.0 g 20.0 g0.05 gS 0.5 g 2.0 g 5.0 g 10.0 g 20.0 g

D µ µ µ µ µ µ

µ µ µ µ µ µ


C

B

0 5 10 15 200

25

50

75

100SKOV3U373MG

µµµµl SD%

RAM

488

pos.

cell s

A

0 1 2 3 4 50

25

50

75

100SKOV3U373MG

µµg RAM488

%RA

M48

8po

s.ce

lls

C

68

Chapter 4

4

Figure 3. SAINT-2:DOPE (SD) mediates intracellular protein-delivery irrespective of the presence of serum.A) Profection of SKOV-3 cells with β-galactosidase complexed with 10 µl SD or 2.5 µl BioPORTER (BP), in the absenceof serum. After 4 h cells were stained for β-galactosidase activity (magni�cation: x20). B) E�ect of serum onprofection with SD as compared to BP. Percentage of SKOV-3 cells displaying β-galactosidase activity per 100cells, 4 h after delivery of 2 or 10 µg β–galactosidase with either SD or BP in the presence of 0, 10, 20 or 50% FBS.C) Delivery of 5 μg MOC31 into U373MG cells with 10 µl SD in the absence (upper panel) or presence (lowerpanel) of 10% FBS during profection. After 48 h, MOC31 detection was performed by immunostaining with R M-F(ab)2-FITC (green), nuclei were stained with DAPI (blue). Images were taken with the Leica, 600S Quantimet(magni�cation: x20).

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50

SD + 2 µg β -Galactosidase

SD + 10 µg β -Galactosidase

Bp + 2 µg β -Galactosidase

Bp + 10 µg β -Galactosidase

SD + 10 ug b-Galactosidase


Bp + 10 ug b-Galactosidase

Bp + 2 ug b-GalactosidaseA





SD + 10 ug ß-Galactosidase

SD + 2 ug ß-Galactosidase

Bp + 10 ug ß-Galactosidase

Bp + 2 ug ß-Galactosidase

B

C

%ß-

Gal

acto

sida

sepo

sitiv

ece

lls

% FBS

Blank MOC31 SD +MOC31

SD MOC31/serum SD +MOC31/ serum

69

4


In�uence of biochemical features of the SAINT-2:DOPE:protein complex on delivery capacity

Physical properties, such as size and charge of a protein have been described to in!uence the ability

to interact with cationic liposomes.4 Table 2 shows the size and zeta-potential of the SD:protein

complexes used for delivery. We measured relatively large di"erences in size and zeta-potential

between the used SD:protein complexes. Although M.SssI is a small protein of ~45 kDa, complexed

with SD it has the same size as an antibody (~145 kDa) complexed with SD, whereas β-galactosidase

(~116 kDa) complexed with SD has the same size as SD alone. Only complex-formation with

β-galactosidase leads to a negatively charged complex, whereas complex-formation of all other

proteins with SD resulted in an almost neutral or positively charged complex. Despite these

di"erences, all proteins analyzed were successfully delivered into the cytoplasm of a variety of cell

types.

B.

SAINT-2 :DOPE : protein zeta-potential (mV) ± stdev

0.075 mM SD 48 ± 4

0.075 mM SD + 20 μg/ml β-gal 47 ± 3

0.075 mM SD + 100 μg/ml β-gal -10 ± 5

0.075 mM SD + 20 μg/ml MOC31 8 ± 4

0.075 mM SD + 50 μg/ml MOC31 1 ± 1

0.075 mM SD + 20 μg/ml SαR-FITC 1 ± 1

0.075 mM SD + 25 µg/ml M.SssI 12 ± 1

0.075 mM SD + 50 µg/ml M.SssI 7 ± 4

(A) SAINT-2:DOPE (SD) protein-ratio was adjusted to the value for delivery, exactly 5 min after incubation at RT theSD:protein complexes were measured. (B) Zeta-potential measurements were performed in H2O with an electric*eld strength of 2.75 V/cm. Size and zeta-potential measurements were performed with the NICOMPTM Model380ZLS.

Table 2. Size and zeta-potential of the SAINT-2:DOPE:protein complexes.

A.

SAINT-2 :DOPE : protein size (nm) ± stdev

0.15 mM SD 127 ± 37

0.15 mM SD + 40 μg/ml β-gal 278 ± 18

0.15 mM SD + 200 μg/ml β-gal 138 ± 22

0.15 mM SD + 40 μg/ml MOC31 1277 ± 67

0.15 mM SD + 100 μg/ml MOC31 1420 ± 78

0.15 mM SD + 40 μg/ml SαR-FITC 1284 ± 29

0.075 mM SD 165 ± 37

0.075 mM SD + 25 µg/ml M.SssI 1456 ± 11

0.075 mM SD + 50 µg/ml M.SssI 1753 ± 45

70

Chapter 4

4

Ga R SD+C141S SD+M.SssIBlank C141S M.SssI Ga R SD+C141S SD+M.SssIG R SD+C141S SD+M.SssIBlank C141S M.SssIBlank C141S M.SssI

Amp 3027-38 +330

010100 40 30 2070 60 5090 80

Methylation

C141S

M.SssI + SD

M.SssI

Blank

C141S + SD

SD

GA

PDH

-38 +330

010100 40 30 2070 60 5090 80-010100 40 30 2070 60 5090 80 010100 40 30 2070 60 5090 80

%

Eca

dher

in

A

B

C

Figure 4. SAINT-2:DOPE (SD) mediates functional active protein delivery into the nucleus. A) Delivery ofMOC31488 with SD into SKOV-3 cells. Confocal laser scanning microscopy images represent x-z projections,0.4 µm step size cross-section, starting at 0.00 and ending at 4.09 µm optical section. The images display thepresence of MOC31488 (green) throughout the entire cell (plasma membrane staining (red)). B) Active silencingof E-cadherin expression by arbitrary methylation of the E-cadherin gene via 48 h profection with M.SssI intothe low E-cadherin expressing SKOV-3 cells. Bisulphite sequencing data of the CpG island in amplicon 3027encompassing 368 bp of the 5’regulatory region of the E-cadherin gene. The bend arrow indicates the ATG codon.Each row corresponds to one experimental treatment of the cell line and each rectangle represents one CpG, ofwhich the methylation status is indicated as a color code (blue: methylated to yellow: unmethylated) (SD: SAINT-2:DOPE, C141S: low-activity mutant of M.SssI). Western blot analysis showed a reduced E-cadherin expression48 h after delivery of M.SssI (GAPDH = loading control). C) Delivery of M.SssI or C141S with SD into the highE-cadherin expressing 16HBE cells. E-cadherin detection was performed 48 h after delivery by immunostainingwith Rabbit-anti-E-cadherin followed by G R488 (green). The nuclei were stained with DAPI (blue). The images,taken with the Leica, 600S Quantimet (magni"cation: x20), clearly show a reduced expression of E-cadherin afterdelivery of M.SssI via SD (G R488 = conjugate control).

71

4


SAINT-2:DOPE mediated delivery of functional protein into the nucleus

Confocal pictures con!rmed that MOC31488 was delivered successfully into the cytoplasm of SKOV-3

cells (Figure 4A). Since protein delivery as a novel tool for gene silencing has potential applications

in basic research and therapy, we chose the DNA methyltransferase M.SssI as a model protein to

demonstrate functional intranuclear delivery. As E-cadherin expression can be silenced by aberrant

methylation of the E-cadherin gene19-21, we tested whether we could silence E-cadherin expression

after M.SssI delivery in the E-cadherin-expressing SKOV-3 cells. Cells were profected with M.SssI and

with its mutant C141S, which has approximately 2-5% catalytic activity of the wild type enzyme.13

After profection with M.SssI, western blot analysis showed reduced E-cadherin expression relative

to the control samples (Figure 4B). Cells profected with C141S displayed a small reduction in

E-cadherin expression, which is consistent with the residual activity observed in vitro. As an internal

control, the amount of GAPDH was determined: it was present in similar amounts in all samples

(Figure 4B). Bisulphite analysis of genomic DNA, obtained from SKOV-3 cells 48 h after profection

with M.SssI, showed that certain CpGs situated in the E-cadherin gene were methylated, whereas

the CpGs in the DNA extracted from the cells profected with the C141S mutant or from the controls

were clearly less methylated (Figure 4B). The observed elevated methylation status corresponds

well with the reduced E-cadherin expression. As SKOV-3 cells have a low expression of E-cadherin,

we also profected the high E-cadherin expressing cell line 16HBE. Figure 4C shows a reduction in

E-cadherin expression 48 h after profection with M.SssI as compared to 16HBE cells profected with

C141S or to cells treated with C141S or M.SssI alone.

DISCUSSION

Until recently, the cationic liposomal compounds available for protein delivery, could only deliver

proteins under serum-free conditions.4;5;22 Using the cationic amphiphilic liposomal device SAINT-

2:DOPE (SD) we demonstrate for the !rst time that proteins can be successfully delivered in the

presence of serum. The serum tolerance of SD broadens the range of cells that allow profection and

makes in vivo application possible. Previously, Tinsley et al. also showed protein delivery, but again

reduced serum conditions were required.23 The here established ability of SD to deliver proteins, can

explain the previously reported e"ect of serum on the delivery of DNA by SD.18 As the serum proteins

are probably competing with the DNA molecules to be complexed with SD, a reduced number of

SD:DNA complexes are formed resulting in a reduced transfection e#ciency. This conclusion is in

concordance with the observation that SD-mediated transfection was not signi!cantly a"ected by

serum after the SD:DNA complex was formed.17;18

In addition to serum-insensitivity, we also show that SD is able to deliver proteins for prolonged

periods of time in contrast to BP. The antibody MOC31488 localized in the cytoplasm within 4 h and

no perceptible lowering in antibody load was observed for at least 96 h (Figure 2C). However using

72

Chapter 4

4

SD, the optimal length of time to reach maximum activity was di�erent per protein. Optimal uptake

of MOC31488 and optimal e�ect of M.SssI was achieved 48 h after profection, whereas the optimal

activity for β-galactosidase was reached 4 h after profection but decreased approximately 6-fold at

24 h. Previous gene delivery experiments indicated that the delivery e�ciency is in�uenced by the

size of the complex formed. In this respect, we observed a 10-fold di�erence in size of the complexes

between β-galactosidase complexed to SD and antibodies or M.SssI complexed to SD (Table 2).

Rejman et al. published that microspheres with a diameter of less than 200 nm are internalized

via clathrin-mediated endocytosis and are ultimately delivered to the lysosomes, while particles

exceeding a diameter of 500 nm enter the cells via caveolae, thereby circumventing the lysosomal

compartment.24;25 We therefore hypothesize that the smaller size of the β-galactosidase:SD compared

to the antibodies/M.SssI:SD complexes (Table 2) explains the observed di�erences in the optimal

delivery-time of 4 versus 48 h respectively, since β-galactosidase might be degraded in the lysomes,

whereas the antibodies/M.SssI complexes internalized via caveolae escape this compartment.

Di�erences in profection e�ciency have been ascribed to di�erences in the interaction between

protein and lipid formulation based on the assumption that only negatively charged proteins can

be successfully delivered by a cationic liposomal compound. However, this does not apply to SD as

M.SssI has a strong positive charge, whereas β-galactosidase is negatively charged at physiological

pH. The SD:protein complexes used di�er in their zeta-potential from positively charged to neutral

to negatively charged. Dalkara et al demonstrated that not only the electric charge but rather the

protein surface area, representing the number of lipid molecules per square nanometer of protein

surface, drives the delivery characteristics22, which might explain why we did not notice any

di�erence in profection e�ciency between the di�erently charged proteins.

A major advantage of SD is its capacity to transfer not only protein, but also DNA and siRNA.

Compared to transfection e�ciencies, the profection e�ciency via SD is very high (Table 1), especially

in the di�cult to transfect primary cell lines. Although di�erences in transfection e�ciency per cell-

type are often related to di�erences in delivery capacity per cell-type, limitations in the accessibility

of the host cell’s DNA transcription machinery is more likely to be the bottle-neck in non-viral DNA

transfection e�ciencies. As DNA is as e�ciently delivered to the cells as siRNA or protein (Figure

1B) and nuclear transportation with SD seems not a problem as profection of M.SssI resulted in a

reduced expression of E-cadherin in virtually all cells.

As shown in Figure 1A intracellular active β-galactosidase was observed, both after DNA as well

as after protein delivery with SD. Transfection of the gene however resulted in a di�use blue staining

of the cytoplasm, whereas profection of the protein led to a more profound granular staining in the

cytoplasm. This di�erence in staining pattern might at least be partially explained by the di�erent

routes of entering the cytoplasm: Protein expressed after gene transfection enters the cell via the

cell’s endogenous machinery, whereas direct exogenous protein delivery requires release of the

protein from the liposomes.

Although !xation can produce artifacts, the granular staining observed after profection of

73

4


MOC31 in COS-7 and U373MG cells (Figure 2D and Figure 3C, both acetone !xed) is comparable with

the staining pattern seen in living cells (Figure 1B). The granular staining observed after profection

of MOC31 in SKOV3 cells ( Figure 4A, PFA !xed) is also comparable to the staining pattern in living

cells (Figure 1B).

Another strong argument against !xation artifacts is that SD-mediated delivery of the

methyltransferase M.SssI led to methylation of the methylation-sensitive E-cadherin gene and

reduced E-cadherin expression demonstrating nuclear uptake. Methylation is a potentially powerful

way to silence the transcription of genes involved in diseases such as hereditary diseases (mutated

genes), viral infection (foreign genes) or cancer (mutated, foreign, and oncogenes).26-28 In contrast

to siRNA, only one initial ‘methylation-hit’ is necessary to achieve this silencing e"ect, because

this hit leads to a DNA-methylation pattern that is maintained and transmitted to the daughter

cell by the cell’s endogenous methylation-machinery.28 Nuclear delivery using BP, has also been

demonstrated by Zheng et al. however, the need to perform profection in the absence of serum

limits the therapeutic application of this !nding.29

Using BP as delivery agent, some toxic activity for non-malignant primary cells has been

reported.6 In contrast, Audouy et al. described minimal toxicity after administering SAINT-2 lipoplexes

intravenously in mice.9;10 Here we demonstrate that the same cationic amphiphilic compound is also

useful in delivering proteins into the cell, irrespective of the presence of serum, where they remain

active and exert their function, hence their physiological e"ects can be investigated both in vitro and

in vivo. This is the !rst reported compound which is capable of delivering proteins as well as DNA

or siRNA. Therefore this agent facilitates comparison of the e"ects on cellular processes between

protein, DNA, and siRNA after delivery. This multipurpose delivery agent allows new approaches in

basic research as well as for therapeutic applications.

ACKNOWLEDGMENTS

We thank Bill Jack (New England Biolabs) for the original plasmid with the M.SssI gene and his

advice, Geert Mesander (UMCG) for technical assistance and Jelleke Dokter (UMCG) for culturing

the cell lines. We also thank Marianne G. Rots (UMCG) for her critical advice. This work was !nancially

supported by the European Commission’s Fifth and Sixth Framework Program (Contract, COOP-

CT-2005-017984).

74

Chapter 4

4

REFERENCES

1. Jarver P, Langel U. Cell-penetrating peptides A brief introduction. Biochimica et Biophysica Acta (BBA) -Biomembranes 2006;1758:260-3.

2. Schwarze S, Dowdy S. In vivo protein transduction: intracellular delivery of biologically active proteins,compounds and DNA. Trends in Pharmacological Sciences 2000;21:45-8.

3. Ye D., Xu D., Singer A.U., Juliano RL. Evaluation of Strategies for the Intracellular Delivery of Proteins.Pharmaceutical Research 2002;19:1302-9.

4. Zelphati O, Wang Y, Kitada S, Reed J, Felgner P, Corbeil J. Intracellular Delivery of Proteins with a New Lipid-mediated Delivery System. Journal of Biological Chemistry 2001;276:35103-10.

5. Dalkara D, Chandrashekhar C, Zuber G. Intracellular protein delivery with a dimerizable amphiphile forimproved complex stability and prolonged protein release in the cytoplasm of adherent cell lines. Journalof Controlled Release 2006;116:353-9.

6. Zassler B, Blasig IE, Humpel C. Protein delivery of caspase-3 induces cell death in malignantC6 glioma, primary astrocytes and immortalized and primary brain capillary endothelial cells2. J Neurooncol. 2005;71:127-34.

7. van der Woude I, Wagenaar A, Meekel A, ter Beest M, Ruiters M, Engberts J et al. Novel pyridiniumsurfactants for e!cient, nontoxic in vitro gene delivery. PNAS 1997;94:1160-5.

8. Zuhorn I, Oberle V, Visser W, Engberts J, Bakowsky U, Polushkin E et al. Phase Behavior of CationicAmphiphiles and Their Mixtures with Helper Lipid In"uences Lipoplex Shape, DNA Translocation, andTransfection E!ciency. Biophysical Journal 2002;83:2096-108.

9. Audouy S, Hoekstra D. Cationic lipid-mediated transfection in vitro and in vivo (review). Mol.Membr.Biol.2001;18:129-43.

10. Andouy S., de leij L., Hoekstra D., Molema G. In Vivo Characteristics of Cationic Liposomes as DeliveryVectors for Gene Therapy. Pharmaceutical Research 2002;19:1599-605.

11. Nur I, Szyf M, Razin A, Glaser G, Rottem S, Razin S. Procaryotic and eucaryotic traits of DNA methylation inspiroplasmas (mycoplasmas). The Journal of Bacteriology 1985;164:19-24.

12. Turker MS. Gene silencing in mammalian cells and the spread of DNA methylation. Oncogene 2002;21:5388-93.

13. Rathert P, Rasko T, Roth M, Slaska-Kiss K, Pingoud A, Kiss A et al. Reversible inactivation of the CG speci#cSssI DNA (cytosine-C5)-methyltransferase with a photocleavable protecting group. Chembiochem.2007;8:202-7.

14. Laarmann S, Schmidt MA. The Escherichia coli AIDA autotransporter adhesin recognizes an integralmembrane glycoprotein as receptor. Microbiology 2003;149:1871-82.

15. Tetzner R, Dietrich D, Distler J. Control of carry-over contamination for PCR-based DNA methylationquanti#cation using bisul#te treated DNA. Nucleic Acids Research 2007;35:e4.

16. Lewin J, Schmitt A, Adorjan P, Hildmann T, Piepenbrock C. Quantitative DNA methylation analysis based onfour-dye trace data from direct sequencing of PCR ampli#cates. Bioinformatics 2004;20:3005-12.

17. Audouy S, Molema G, de Leij L, Hoekstra D. Serum as a modulator of lipoplex-mediated gene transfection:dependence of amphiphile, cell type and complex stability. J.Gene Med. 2000;2:465-76.

18. Zuhorn I, Visser W, Bakowsky U, Engberts J, Hoekstra D. Interference of serum with lipoplex-cell interaction:modulation of intracellular processing. Biochimica et Biophysica Acta (BBA) - Biomembranes 2002;1560:25-36.

19. Chen C-L, Liu SS, Ip S-M, Wong LC, Ng TY, Ngan HYS. E-cadherin expression is silenced by DNA methylationin cervical cancer cell lines and tumors. European Journal of Cancer 2003;39:517-23.

75

4


20. Chen Q, Lipkina G, Song Q, Kramer R. Promoter methylation regulates cadherin switching in squamous cellcarcinoma. Biochemical and Biophysical Research Communications 2004;315:850-6.

21. Corn P, Heath E, Heitmiller R, Fogt F, Forastiere A, Herman J et al. Frequent Hypermethylation of the 5’ CpGIsland of E-Cadherin in Esophageal Adenocarcinoma. Clinical Cancer Research 2001;7:2765-9.

22. Dalkara D, Zuber G, Behr J. Intracytoplasmic Delivery of Anionic Proteins. Mol Ther 2004;9:964-9.

23. Tinsley J, Hawker J, Yuan Y. E!cient protein transfection of cultured coronary venular endothelial cells. AJP- Heart and Circulatory Physiology 1998;275:H1873-H1878.

24. Rejman J, Oberle V, Zuhorn I, Hoekstra D. Size-dependent internalization of particles via the pathways ofclathrin- and caveolae-mediated endocytosis. Biochemical Journal 2004;377:159-69.

25. Rejman J, Bragonzi A, Conese M. Role of Clathrin- and Caveolae-Mediated Endocytosis in Gene TransferMediated by Lipo- and Polyplexes. Mol Ther 2005;12:468-74.

26. Plass C. Cancer epigenomics. Human Molecular Genetics 2002;11:2479-88.

27. Robertson K. DNA methylation and human disease. Nature Reviews Genetics 2005;6:597-610.

28. Xu G, Bestor T. Cytosine methylation targetted to pre-determined sequences. Nat Genet 1997;17:376-8.

29. Zheng X, Lundberg M, Karlsson A, Johansson M. Lipid-mediated Protein Delivery of Suicide NucleosideKinases. Cancer Research 2003;63:6909-13.

77

Persistent down-regulation of the pancarcinoma-associated Epithelial Cell Adhesion Molecule via active

intranuclear methylation

Bernardina T.F. van der Gun1 , Reinhold Wasserkort2, Amélie Monami3, Albert Jeltsch4, Tamás Raskó5

Krystyna Ślaska-Kiss5, Rene Cortese2, Marianne G. Rots1, Lou F.M.H. de Leij1, Marcel H.J. Ruiters1

Antal Kiss5, Elmar Weinhold3, Pamela M.J. McLaughlin1

1 Department of Pathology and Laboratory Medicine, Section Medical Biology, University Medical CenterGroningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands

2 Epigenomics AG, Kleine Praesidentenstrasse 1, D-10178, Berlin, Germany3 Institute of Organic Chemistry RWTH Aachen University, Landoltweg 1, D-52056 Aachen, Germany4 School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, D-28725 Bremen, Germany5 Institute of Biochemistry, Biological Research Center of the Hungarian Academy of Sciences, Temesvári krt. 62,

6726 Szeged, Hungary

Int J Cancer 2008;123(2):484-9

Chapter 5

78

Chapter 5

5

ABSTRACT

The Epithelial Cell Adhesion Molecule (EpCAM) is expressed at high levels on the surface of most

carcinoma cells. SiRNA silencing of EpCAM expression leads to reduced metastatic potential of

tumor cells demonstrating its importance in oncogenesis and tumor progression. However, siRNA

therapy requires either sequential delivery or integration into the host cell genome. Hence, we

set out to explore a more de�nite form to in�uence EpCAM gene expression. The mechanisms

underlying the transcriptional activation of the EpCAM gene, both in normal epithelial tissue as

well as in carcinogenesis, are poorly understood. We show that DNA methylation plays a crucial

role in EpCAM expression, and moreover, active silencing of endogenous EpCAM via methylation

of the EpCAM promoter results in a persistent down-regulation of EpCAM expression. In a panel

of carcinoma derived cell lines, bisul�te analyses showed a correlation between the methylation

status of the EpCAM promoter and EpCAM expression. Treatment of EpCAM-negative cell lines

with a demethylating agent induced EpCAM expression, both on mRNA and protein level, and

caused up-regulation of EpCAM expression in an EpCAM-positive cell line. After delivery of the DNA

methyltransferase M.SssI into EpCAM-positive ovarian carcinoma cells, methylation of the EpCAM

promoter resulted in silencing of EpCAM expression. SiRNA-mediated silencing remained for 4

days, after which EpCAM re-expression increased in time, while M.SssI-mediated down-regulation

of EpCAM maintained through successive cell divisions as the repression persisted for at least 17

days. This is the �rst study showing that active DNA methylation leads to sustained silencing of

endogenous EpCAM expression.

79

5

Persistent down-regulation of EpCAM

INTRODUCTION

Since its discovery the human pancarcinoma-associated Epithelial Cell Adhesion Molecule (EpCAM),

also referred to as 17-1A, EGP-2, TROP1 or CD326, has become a major target for carcinoma-directed

immunotherapy. However, evidence for its direct involvement in carcinogenesis has only been

given recently. EpCAM expression has a direct impact on the cell cycle via c-myc and cyclin A/E, and

inhibition of EpCAM expression with antisense mRNA reduces the proliferation and metabolism in

human carcinoma cells.1 Similarly, silencing of EpCAM expression with siRNA reduces the migration

and invasive potential of breast cancer cells by 90%.2 Moreover, EpCAM over-expression in breast,

ovarian and gallbladder cancer correlates with a strong negative prognosis.3-5 For human colorectal

cancer it has been shown that the ability to engraft in vivo in immunode!cient mice, was restricted

to a minority subpopulation of epithelial cells with high EpCAM expression.6 This direct involvement

of EpCAM in the development of carcinomas quali!es EpCAM as an important target for therapy.

The EpCAM regulatory sequences have been cloned and characterized,7;8 and the basic proximal

promoter region still able to confer epithelial-speci!c expression was de!ned.8 It has been described

that DNA methylation prevents the ampli!cation of the EpCAM gene.9 Furthermore, recent studies

provide evidence that DNA methylation is involved in the regulation of the EpCAM gene.10;11

Although currently siRNA is most commonly used to down-regulate gene expression, a major

drawback of siRNA is that down-regulation is transient. SiRNA treatment requires either sequential

deliveries or integration of shRNA (small hairpin RNA) expressing plasmid DNA into the target

cell’s genome, encompassing the same limitations as encountered with gene therapy. Hence, we

set out to explore active DNA methylation as a tool to silence EpCAM gene expression. One major

advantage of gene silencing by DNA methylation compared to siRNA-mediated silencing is that the

cellular DNA methylating system will maintain the new methylation pattern in the absence of the

methyltransferase and long-term presence of the methylating agent is not required.12 Moreover,

DNA methylation a"ects the initiation of transcription, whereas siRNA acts in general on the mRNA

level, where the target pool is much larger. In principle only one initial event is required for DNA

methyltransferases as the DNA methylation pattern is epigenetically imprinted13 and inherited to

the daughter cells. To actively silence endogenous EpCAM expression we used the prokaryotic DNA

(cytosine-5) methyltransferase (MTase) M.SssI, which methylates cytosines in CpG dinucleotides.14

Since M.SssI has the same base and sequence speci!city as mammalian DNA MTases, this enzyme

appears to be an excellent tool to study the role of DNA methylation in healthy and diseased

eukaryotic cells provided that it can be delivered into the cell nucleus.

80

Chapter 5

5


Cell culture and 5-AZAC treatment

The HEK293A/T (CRL-1573), U373MG (HTB-17), SKOV3 (HTB-77) and SW948 (CCL-237) cell lines

were purchased from ATCC (Manassas, VA) and cultured according to ATTC recommendations. The

HEKOGM cell line was kindly provided by Dr. O. Gires (Ludwig-Maximilians-University, Munich,

Germany). The lung carcinoma cell lines GLC8 and GLC1 were maintained in RPMI-1640 medium

(BioWhittaker Inc, Walkersville, MD). The fetal lung �broblasts (FLF) were isolated in 1992 under

informed consent and cultured in DMEM (BioWhittaker Inc, Walkersville, MD). Cells were cultured

at 37°C and 5% CO2.

For methylation inhibition studies, U373MG, SKOV3, FLF and GLC1 cells were cultured in their

appropriate media with a �nal concentration of 2 µg/ml 5-aza-2’-deoxycytidine (5-AZAC; Sigma, St

Louis, MO) during days 1, 3 and 5. At day 2 and 4 medium was refreshed and on day 6 cells were

harvested for extraction of total mRNA and EpCAM expression.

Protein expression

EpCAM detection was performed with 1 µg/ml mouse Mab MOC31 (protein A puri�ed) or

supernatant, followed by R M-PO or R M-F(ab)2-FITC (DAKO, Glostrup, Denmark). The Mean

Fluorescence Intensity (MFI) was measured on a BD FACS Calibur !ow cytometer (BD Biosciences,

San Jose, CA). For Western blotting cells were lyzed in 200 μl bu#er, 10 μg total protein separated

and blotted as previously described.15 As loading control GAPDH (Abcam, Cambridge, UK) was used,

detection was accomplished with GαR-AF ( Jackson ImmunoResearch, Su#olk, England) and BCIP/

NBT substrate.

Reverse-transcriptase PCR

RNA was isolated using RNeasy Mini Kit (Qiagen Inc., Valencia, CA) according to the manufacturer

recommendations. Prior to cDNA synthesis on 2 µg of puri�ed total RNA with an oligo(dT18) primer

and M-MuLV Reverse Trancriptase (Fermentas Inc., Hanover, MD), RNA samples were treated with

rDNaseI (Ambion Ltd, Cambridgeshire, UK). cDNA was ampli�ed using primers for EpCAM: exon 3

5’-GAACAATGATGGGCTTTATG-3’(sense), exon 7 5’-TGAGAATTCAGGTGCTTTTT-3’(antisense), β-actin

5’-TCACCAACTGGGACGACATG-3’ (sense), 5’-ACCGGAGTCCATCACGATG-3’ (antisense), purchased

from Biolegio Inc. (Malden, The Netherlands). The predicted size of the PCR product was 500 bp for

EpCAM and 242 bp for β-actin.

Quantitative gene expression analysis by real-time RT-PCR

RT-PCR was performed as previously described.16 In short, 1 µg RNA was reverse-transcribed using

SuperScript III reverse transcriptase (Invitrogen, Breda, The Netherlands) and random hexamer

primers (Promega, Leiden, The Netherlands). Quantitatieve PCR ampli�cations were performed

81

5


according to manufacturers protocol on an ABI Prism 7900HT Sequence Detection System (Applied

Biosystems, Applera Nederland, Nieuwekerk a/d Ijssel, The Netherlands). Primers and probes for

EpCAM (Hs00158980_m1) and the housekeeping gene GAPDH (Hs99999905_m1) were purchased

as customized assays from Applied Biosystems. All PCR reactions were carried out in triplicate

on duplicate profections. Relative quantitation of gene expression was calculated based on the

comparative cycle treshold (Ct) method (ΔCt = Ct EpCAM - Ct GAPDH). Comparison of EpCAM

expressions in di�erent samples was performed based on the di�erences in ΔCt of individual

samples (ΔΔCt).

Methylation analysis

DNA extracted from the cells was subjected to bisul!te treatment as previously described.17 Bisul!te

speci!c primers void of any CpG were used in order to obtain ampli!cation products unbiased for the

methylation status. Two overlapping amplicons were selected to cover a 700 bp region (A225830,

Figure 1). Primer sequences for the !rst amplicon were 5’-ACCTCCCCAATAACTAAAATTAC-3’

(forward), 5’-TTGAAGATTTTGTGTTGAGATTT-3’ (reverse), and for the second amplicon

5’-AGTGTTTTGGAAGGTTTTTTGT-3’ (forward), 5’-AAATTAAAAAAATAAATAAACTCCC-3’ (reverse).

A neighboring region extending into the CpG island (A225850, Figure 1) was covered with

an amplicon of 441 bp. Primers were 5’-GGAGGGGAGTTTATTTATTTTT-3’ (forward) and

5’-CACAACTCTACTCCAATC-3’ (reverse). PCR conditions: 95°C for 15 min, followed by 40 cycles of

95°C for 60 s, 55°C for 45 s and 72°C for 60 s and !nished with 72°C for 10 min. Puri!ed PCR products

were used directly for cycle sequencing on an ABI3730-capillary sequencer using the ABI Prism

BigDye Terminator V3.1 sequencing chemistry. The obtained trace !les were subsequently analyzed

using the ESME software as previously described.18

Plasmids

The plasmids pBHNC-MSssI and pBHNC-MSssI(C141S) were described previously.19 The variant

encoded by pBHNC-MSssI has MTase activity comparable to the wild-type enzyme, and will be

referred to as M.SssI. Its mutant derivative, M.SssI(C141S) encoded by pBHNC-MSssI(C141S), in

which the active site cysteine is replaced by serine, has a greatly reduced (2-5%) activity relative to

the wild-type enzyme.19

DNA MTases

E. coli ER1821 cells, harbouring pBHNC-MSssI or pBHNC-MSssI(C141S), were grown at 37°C in LB

containing 100 µg/ml ampicillin. At OD600 ~0.6, M.SssI or M.SssI(C141S) production was induced

by adding 1.0% arabinose. After 4 h incubation at 30°C, cells were harvested by centrifugation,

resuspended in breaking bu�er (50 mM Na2HPO4, pH 8.0, 300 mM NaCl, 1 mM imidazole), sonicated

and cell debris removed by centrifugation. For puri!cation a His-Select Nickel A$nity gel column

(1 ml, Sigma) was used according to the manufacturers instructions. The eluate was diluted with

82

Chapter 5

5

cation exchange bu�er (6.7 mM MES, 6.7 mM Hepes, 6.7 mM NaOAc, pH 7.5, 1 mM EDTA, 10 mM

-mercaptoethanol, 10% glycerol) and applied to a HS POROS 50 column (Applied Biosystems,

Fostercity, CA). After washing with 100 ml cation exchange bu�er containing 0.2 M NaCl, proteins

were eluted with a linear NaCl gradient (0.2–1 M) in cation exhange bu�er. DNA MTase containing

fractions were pooled, concentrated by ultra!ltration, mixed with an equal volume of glycerol and

stored at -20°C. All puri!cation steps were performed at 4°C.

DNA MTase and siRNA delivery

SAINT-2:DOPE (SD; 0.75 mM) was purchased from Synvolux Therapeutics Inc. (Groningen, The

Netherlands).20 SKOV3 cells were seeded 0.5 x 106/6 well or 12.5 x 104/chamber slide well. MTase

or siRNA delivery was performed at 50-80% con#uency. Ten µg MTase, 1µg siRNA-EpCAM (sense

5’-GGAGAUCACAACGCGUUAUUU-3’ and anti-sense 5’-AUAACGCGUUGUGAUCUCCUU-3’) (Qiagen)

or 1 µg irrelevant siRNA (AM4611, Applied Biosystems) in 100 µl PBS was complexed with 20 µl SD

in an equal volume of PBS and the SD-MTase or SD-siRNA complex was pipeted directly onto the

cells. In one chamber slide well 0.625 μg M.SssI/C141S was complexed with 2.5 µl SD. As controls,

MTase, siRNA or SD alone were added. Cells were split, and EpCAM expression was measured at day

2, 6, 10, 14 and 17.

RESULTS

To investigate the relation between EpCAM expression and DNA methylation, we assessed the

methylation status of the EpCAM promoter in a panel of cell lines with di�erent EpCAM expression

levels by bisul!te sequencing (Figure 1). In the EpCAM-negative cell lines U373MG and HEK293OGM,

the promoter region was extensively methylated. Whereas the most upstream portion of the

analyzed region was methylated in all cell lines analyzed, hypomethylation in the promoter region

adjacent to the coding region was characteristic for those cell lines that do express EpCAM. In this

latter region, more CpGs were methylated in the low EpCAM expressing HEK293T cell line, compared

to the higher expressing HEK293A. The EpCAM-negative GLC1 cell line displayed an intermediate

methylation status.

The observed correlation between EpCAM expression and the methylation status of the

EpCAM promoter suggests that the EpCAM gene is regulated by DNA methylation. Indeed, after

addition of the demethylating agent 5-AZAC for 3 non-consecutive days, de novo induction of

EpCAM expression was observed in U373MG and up-regulation in SKOV3 cells (Figure 2A). RT-PCR

con!rmed the presence of mRNA in the 5-AZAC treated EpCAM-negative U373MG and FLF cells

(Figure 2B), although in the latter cells no EpCAM protein was detected. 5-AZAC treatment of the

GLC1 cell line did not result in EpCAM expression on both protein and mRNA level.

As EpCAM expression is clearly associated with promoter methylation, we investigated whether

Figure 1. Analysis of methylation status of part of the EpCAM promoter and exon 1 in relation toEpCAM expression. CpGs in the analyzed region are depicted by vertical bars, the transcription initiation sitecorresponds to position +1, the untranslated (UTR) as well as translated (ATG) region of exon 1 are shown byarrrows. A CpG-dense region (A225850) and a less dense region (A225830) spanning together about 1100 bpof the EpCAM gene are covered by three amplicons. Each row corresponds to one cell line and each rectanglerepresents one CpG, of which the methylation status is indicated as a color code (blue: methylated to yellow:unmethylated). White areas indicate CpGs for which no reliable data were retrieved. EpCAM expression wasmeasured by !ow cytometry.

5


83

Figure 2. EpCAM expression on protein and mRNA level before and after 5-AZAC treatment. A)Immunohistochemical staining with the EpCAM-speci"c antibody MOC31: after AZAC treatment de novoinduction of EpCAM was observed in the EpCAM-negative U373MG cells, but not in the EpCAM-negative FLFand GLC1 cells. SKOV3 cells showed up-regulation of EpCAM compared with non-treated cells (magni"cation:×40). B) Reverse-transcriptase PCR analysis displayed induction of EpCAM mRNA (exon 3 and 7) in EpCAM-negative FLF and U373MG cells and up-regulation in SKOV3 cells after AZAC treatment. The gel has been loadedwith 15 and 5 µl of each PCR-product obtained from the U373MG, FLF and GLC1 cells. For SKOV3 cells, due to thehigh expression found on these cells 3, 2 and 1 µl PCR product has been loaded. For the loading control β-actin5 and 2 µl PCR product has been loaded (- = without AZAC, + = with AZAC).

MOC31

5-AZACMOC31

U373MG GLC1 FLF SKOV3A

EpCAM3/7

- - -U373MG

5-azacFLFGLC1 SKOV3

H2O- -+ + + + + + +B

β-actin

U373MG5-azac

FLFGLC1 SKOV3 H2O- -+ + + + + + - - + +- - - -

++- - --

U373MG

SW948

SKOV3

HEK293T

HEK293OGM

HEK293A

GLC8

GLC1

+282058522A038522A

0 100 200 300 400 500 600

-830

% Methylation40 30 20 1070 60 5090 80100

ATGUTR

U373MG

SW948

SKOV3

HEK293T

HEK293OGM

HEK293A

GLC8

GLC1

+282058522A038522A

0 100 200 300 400 500 600

-830

% Methylation40 30 20 1070 60 5090 80100

ATG

U373MG

SW948

SKOV3

HEK293T

HEK293OGM

HEK293A

GLC8

GLC1

U373MG

SW948

SKOV3

HEK293T

HEK293OGM

HEK293A

GLC8

U373MG

SW948

SKOV3

HEK293T

HEK293OGM

HEK293A

GLC8

GLC1

+282058522A038522A

0 100 200 300 400 500 600

-830

% Methylation40 30 20 1070 60 5090 80100 40 30 20 1070 60 5090 80100

ATGUTRUTR

M

Chapter 5

we could actively silence EpCAM expression by induced methylation of the EpCAM promoter. To

this end, we delivered M.SssI19 directly as protein via a cationic amphiphilic compound SAINT-2:DOPE

(SD)20 into SKOV3 cells (profection). As a control, cells were profected with the mutant M.SssI protein

C141S, which has approximately 2-5% catalytic activity of the wild-type enzyme.19 Analysis of

genomic DNA, obtained from SKOV3 cells 48 h after profection with M.SssI, demonstrated increased

methylation of CpGs located in the EpCAM promoter and the !rst exon of the gene, whereas the

cells treated with the MTases without SD, were not, or much less methylated (Figure 3A). Cells

profected with C141S showed an intermediate methylation status, which is in agreement with the

residual activity observed in vitro.19

Next, we assessed whether the induced methylation was associated with repression of gene

and protein expression. Quantitative Real-Time PCR displayed reduced EpCAM mRNA levels after

5

84

Figure 3. Active silencing of EpCAM expression in SKOV3 cells by induced methylation of the EpCAMpromoter via profection with M.SssI (48 h after profection). A) Bisul!te sequencing data obtained from the441 bp fragment (A225850) within the CpG island spanning part of the promoter and exon 1 of the EpCAMgene. Each row corresponds to one experimental treatment of the cell line (SD: SAINT-2:DOPE, C141S: low-activitymutant of M.SssI) and each rectangle represents one CpG, of which the methylation status is indicated as a colorcode (blue: methylated to yellow: unmethylated). White areas indicate CpGs for which no reliable data wereretrieved. Profection of SKOV3 cells with M.SssI resulted in increased methylation levels, delivery of its less activemutant C141S showed an intermediate methylation status. Immunohistochemical staining of EpCAM (3A, middlepanel) displayed a reduced EpCAM expression after profection with M.SssI compared to the controls, whichwas con!rmed by "ow cytrometric analysis (3A, right panel). B) Quantitative Real-Time PCR analysis showed areduced EpCAM mRNA level compared to the controls, expression levels of mRNA in untreated SKOV3 (blank)cells were arbitrarily set at 1. C) Western blot analysis with the EpCAM-speci!c antibody MOC31 demonstrated aclear reduction of EpCAM expression after profection with M.SssI compared to the controls. The two bands aredue to di#erential glycolysation of EpCAM, GAPDH is shown as loading control.

400400

+204

C141S + SD

C141S

M.SssI + SD

M.SssI

Blank

-

% Methylation

A

100 200 300

UTR

0 20 40 60 80 100 120% Relative Mean Fluorescense

EpCAM

CB

C141S+ SD

C141SM.SssI+ SD

M.SssISDBlank

GAPDH

40 30 20 1070 60 5090 80100 0

+204- 159

ATG

100 200 300

0 20 40 60 80 100 120% Relative Mean Fluorescense Intensity

40 30 20 1070 60 5090 80100 040 30 20 1070 60 5090 80100 0

0,0

0,3

0,6

0,9

1,2

Blank SDM.SssI

SD+M.SssI

C141S

SD+C141Srela

tive

gene

exp

ress

ion

85

5


profection with M.SssI (Figure 3B). Moreover, immunohistochemical staining (Figure 3A middle

panel), �ow cytometric analysis (Figure 3A right panel) and Western blot analysis (Figure 3C) showed

reduced EpCAM protein expression after profection with M.SssI compared to the cells profected

with C141S or treated with the enzymes without SD. Altogether, these results demonstrate that

methylation is involved in the regulation of EpCAM expression, and even more important, EpCAM

expression can be actively down-regulated by DNA methylation.

In contrast to siRNA-mediated silencing, which generally requires sequential deliveries

to maintain silencing, only one initial exposition of the genome to the MTase is required as the

resultant DNA methylation pattern is inherited through successive cell divisions.12;13 To investigate if

the down-regulation of EpCAM via methylation is lasting, we delivered, only once, the protein M.SssI

or siRNA directed against EpCAM, into SKOV3 cells on day 0, and cultured them for 17 days (Figure

4). Within 2 days after siRNA-fection, EpCAM expression was reduced to 20% of the expression levels

observed in non-transfected EpCAM expressing SKOV3 cells. This down-regulation remained up

to day 6, where after EpCAM re-expression increased with time. Irrelevant siRNA had no e�ect on

EpCAM expression (data not shown). In contrast, profection with M.SssI resulted in a 40% reduction

of the EpCAM expression, which persisted at least up to 17 days after profection. Although the

EpCAM speci�c siRNA-mediated down-regulation of EpCAM was initially more e�ective than

profection with the non-targeted M.SssI, down-regulation via profection was enduring. Profection

with the mutant C141S showed a gradually decrease in EpCAM expression, which stabilized at day

14 at 75% of the expression levels observed in non-profected cells (Figure 4). This residual activity

0

20

40

60

80

100

120

0 2 6 10 14 17

Days after profection/siRNA-fection

EpCA

Mex

pres

sion

as%

ofco

ntro

ls

C141SM.SssIsiRNA EpCAMC141S+SDM.SssI+SDsiRNA EpCAM+SD

Figure 4. Persistent down-regulation of EpCAM via profection with M.SssI as compared to transient down-regulation of EpCAM after siRNA-fection. At day 0, SKOV3 cells were profected with M.SssI or transfected withEpCAM-speci�c siRNA, cells were cultured for 17 days and EpCAM expression was measured by �ow cytometryat the days indicated (SD: SAINT-2:DOPE, C141S: low-activity mutant of M.SssI). Because of auto-�uorescence ofSD, the EpCAM expression after profection or siRNA-fection was expressed as percentage of the SD control. ForMTases or siRNA without SD, the blank was set as 100%. The reduction in EpCAM expression at day 2 after siRNA-fection remains up to day 6 after which re-expression is increasing. The 40% reduction in EpCAM expression atday 6 after profection with M.SssI persisted up to day 17.

86

Chapter 5

5

of C141S, as also shown by the intermediate methylation status (Figure 3A), was not detectable on

the level of protein expression at day 2 after profection (Figure 3A middle and right panel, 3C), but

this gradually decrease does correlate with the observed slight decrease in mRNA level (Figure 3B).

DISCUSSION

This study demonstrates for the !rst time that endogenous EpCAM expression can be actively

down-regulated in a persistent manner via induced DNA methylation. Previously Tai et al reported

inhibition of EpCAM promoter activity by ex vivo DNA methylation of the promoter. Indeed, a

signi!cant association was demonstrated between EpCAM expression and methylation status of

the EpCAM promoter (-265 to -100) in microdissected tumor tissue.11 Also Spizzo et al found part of

the promoter and exon 1 (-156 to +361) to be methylated to a higher degree in an EpCAM negative

breast cancer cell line as compared to an EpCAM positive cell line. Interestingly, in this study no

correlation was observed between methylation status and EpCAM expression in primary breast

cancer tissue.10 These paradoxical observations concerning the methylation status of the EpCAM

gene and its expression in clinical tumor specimens might be due to the di"erent tumor types

analyzed, but also to di"erences in the region examined.

In this study, we therefore investigated a larger region of the EpCAM promoter (-830 to +282),

and showed a correlation between EpCAM expression and the methylation status of the promoter

region in EpCAM expressing and non-expressing cell lines. As previously demonstrated8, this region

includes part of the promoter (-687 to +93) which is su#cient to confer epithelial speci!city. In

the EpCAM-negative cell lines U373MG and HEK293OGM this promoter region was extensively

methylated, whereas in EpCAM expressing cells this was not the case. The importance of methylation

in regulating EpCAM promoter activity is further demonstrated by 5-AZAC treatment, which indeed

led to de novo induction of EpCAM expression in U373MG and FLF and a further up-regulation in

SKOV3 cells. Only the EpCAM-negative GLC1 cell line, which displayed an intermediate methylation

status, showed no induction of EpCAM expression upon 5-AZAC treatment. This !nding might be

due to genomic deletions or mutations in the EpCAM gene in this particular cell line.

Controversial results regarding EpCAM expression in correlation with cancer invasiveness

and tumor progression have been reported. Several studies showed that EpCAM overexpression

correlates with a poor patient survival.3-5 Futhermore, inhibition of EpCAM expression with antisense

mRNA or siRNA reduces the oncogenic potential of carcinoma cells.1;2 Moreover, the capacity to

form tumors out of human colorectal xenograft lines has been shown to be restricted to EpCAM

high expressing stem cells, whereas EpCAM low expressing cells failed to form tumors.6 These

!ndings supported our aim to develop a potential strategy to down-regulate EpCAM expression

in a persistent manner. However, an other study found an inverse correlation between EpCAM

expression and cancer invasiveness in cancer cell lines.11 Dalerba et al demonstrated that tumors

87

5


grown from EpCAM high expressing colon carcinoma cells in immunode�cient mice contained

both EpCAM high and low expressing populations in proportions similar to those of their parent

lesions.6 This heterogeneity in EpCAM might explain the contrary results concerning the prognostic

value of EpCAM.

Tools including active and sustained modulation of endogenous EpCAM expression should

be developed to provide insights in the precise role of EpCAM in tumorigenesis and tumor

progression of di�erent origin and might eventually lead to potent therapeutics. Therefore we set

out to explore active DNA methylation as a tool to silence EpCAM gene expression. DNA MTase-

and siRNA-mediated down-regulation of EpCAM expression showed di�erent kinetics (Figure 4).

The initial down-regulation via siRNA was higher compared to DNA methylation. This di�erence in

e�ciency can be explained by the fact that the siRNA used is speci�c for EpCAM, whereas M.SssI is

not. The resulting overall increase of methylated CpGs in the genome is toxic. Targeting of M.SssI to

the EpCAM promoter will allow to optimize the dose of M.SssI thereby increasing the reduction of

EpCAM expression. Nevertheless, in this study even by using non-targeting M.SssI, we could show

e�cient down-regulation of EpCAM which was enduring and more pronounced than siRNA after

17 days.

Now that we have shown that active methylation of the EpCAM promoter results in sustained

silencing of gene expression, the next step is to target M.SssI to the EpCAM promoter speci�cally to

reduce its toxicity. Targeting of the EpCAM promoter by engineered zinc �nger protein transcription

factors (ZFPs) has already been demonstrated by us.21 Methyltransferases fused to zinc �nger proteins

targeting predetermined sites in the DNA to repress gene expression have been reported.12;12;22;23

Engineering ZFPs targeted to the EpCAM promoter21 fused to M.SssI as an e�ector domain, provides

a powerful tool to achieve targeted methylation. The ZFP binds speci�c to the promoter where

after the enzyme will methylate only the CpGs close to the ZFP target sequence. Another approach

is the use of a Triple helix-Forming Oligonucleotide (TFO) targeted to the EpCAM promoter. Such

targeting devices will enable e�cient and sustained gene silencing which has potent applications

for basic research and therapy. Considering the dynamic change of EpCAM expression in di�erent

tumor stages, active regulation of the EpCAM gene is a powerful tool to explore the function of

EpCAM. Because of the contributory role of cancer-linked genomic hypomethylation of oncogenes

to tumorigenesis or tumor progression24 active silencing of speci�c genes via DNA methylation can

provide a novel approach in anti-cancer treatment.

ACKNOWLEDGMENTS

We thank Bill Jack (New England Biolabs) for the original plasmid with the M.SssI gene and his

advice, Geert Mesander, Henk Moes (UMCG) for technical assistance with the Quantimet and !ow

cytometer, and Jelleke Dokter (UMCG) for culturing the cell lines. This work was �nancially supported

88

Chapter 5

5

by the European Commission’s Fifth and Sixth Framework Program (Contract QRLT-2001-0448 and

COOP-CT-2005-017984). MHJ Ruiters is Associated Professor at the section Medical Biology, next to

his involvement in the process of marketing the liposomal compound SAINT-2:DOPE via the company

Synvolux. R. Cortese and R. Wasserkort are employees of Epigenomics AG, and R. Wasserkort is also

a shareholder of this company.

REFERENCES



3. Spizzo G, Went P, Dirnhofer S, Obrist P, Simon R, Spichtin H et al. High Ep-CAM expression is associated withpoor prognosis in node-positive breast cancer. Breast Cancer Res.Treat. 2004;86:207-13.


5. Varga M, Obrist P, Schneeberger S, Muhlmann G, Felgel-Farnholz C, Fong D et al. Overexpression ofepithelial cell adhesion molecule antigen in gallbladder carcinoma is an independent marker for poorsurvival. Clin.Cancer Res. 2004;10:3131-6.

6. Dalerba P, Dylla SJ, Park IK, Liu R, Wang X, Cho RW et al. Phenotypic characterization of human colorectalcancer stem cells. Proc Natl Acad Sci USA 2007;104:10158-63.

7. Gires O, Esko!er S, Lang S, Zeidler R, Munz M. Cloning and characterisation of a 1.1 kb fragment of thecarcinoma-associated epithelial cell adhesion molecule promoter. Anticancer Res. 2003;23:3255-61.


9. Alberti S, Nutini M, Herzenberg LA. DNA Methylation Prevents the Ampli!cation of TROP1, a Tumor-Associated Cell Surface Antigen Gene. Proc Natl Acad Sci USA 1994;91:5833-7.


11. Tai KY, Shiah SG, Shieh YS, Kao YR, Chi CY, Huang E et al. DNA methylation and histone modi!cation regulatesilencing of epithelial cell adhesion molecule for tumor invasion and progression. Oncogene 2007.

12. Xu G, Bestor T. Cytosine methylation targetted to pre-determined sequences. Nat Genet 1997;17:376-8.

13. Smith AE, Ford KG. Speci!c targeting of cytosine methylation to DNA sequences in vivo. Nucleic Acids Res.2007;35:740-54.

14. Nur I, Szyf M, Razin A, Glaser G, Rottem S, Razin S. Procaryotic and eucaryotic traits of DNA methylation inspiroplasmas (mycoplasmas). The Journal of Bacteriology 1985;164:19-24.

15. McLaughlin P, Harmsen M, Dokter W, Kroesen B, van der Molen H, Brinker M et al. The Epithelial Glycoprotein2 (EGP-2) Promoter-driven Epithelial-speci!c Expression of EGP-2 in Transgenic Mice: A New Model toStudy Carcinoma-directed Immunotherapy. Cancer Res. 2001;61:4105-11.

16. Asgeirsdottir SA, Kamps JAAM, Bakker HI, Zwiers PJ, Heeringa P, van der Weide K et al. Site-Speci!c Inhibitionof Glomerulonephritis Progression by Targeted Delivery of Dexamethasone to Glomerular Endothelium.Mol Pharmacol 2007;72:121-31.

89

5


17. Tetzner R, Dietrich D, Distler J. Control of carry-over contamination for PCR-based DNA methylationquanti�cation using bisul�te treated DNA. Nucleic Acids Res. 2007;35:e4.

18. Lewin J, Schmitt A, Adorjan P, Hildmann T, Piepenbrock C. Quantitative DNA methylation analysis based onfour-dye trace data from direct sequencing of PCR ampli�cates. Bioinformatics 2004;20:3005-12.

19. Rathert P, Rasko T, Roth M, Slaska-Kiss K, Pingoud A, Kiss A et al. Reversible inactivation of the CG speci�cSssI DNA (cytosine-C5)-methyltransferase with a photocleavable protecting group. Chembiochem.2007;8:202-7.

20. vanderWoude I, Wagenaar A, Meekel A, ter Beest M, Ruiters M, Engberts J et al. Novel pyridinium surfactantsfor e!cient, nontoxic in vitro gene delivery. Proc Natl Acad Sci USA 1997;94:1160-5.

21. Gommans WM, McLaughlin PM, Lindhout BI, Segal DJ, Wiegman DJ, Haisma HJ et al. Engineering zinc�nger protein transcription factors to downregulate the epithelial glycoprotein-2 promoter as a novelanti-cancer treatment. Mol.Carcinog. 2006.

22. Li F, Papworth M, Minczuk M, Rohde C, Zhang Y, Ragozin S et al. Chimeric DNA methyltransferases targetDNA methylation to speci�c DNA sequences and repress expression of target genes. Nucleic Acids Res.2007;35:100-12.

23. Minczuk M, Papworth MA, Kolasinska P, Murphy MP, Klug A. Sequence-speci�c modi�cation ofmitochondrial DNA using a chimeric zinc �nger methylase. PNAS 2006;103:19689-94.

24. Shukeir N, Pakneshan P, Chen G, Szyf M, Rabbani S. Alteration of the Methylation Status of Tumor-Promoting Genes Decreases Prostate Cancer Cell Invasiveness and Tumorigenesis In vitro and In vivo.Cancer Res. 2006;66:9202-10.

91

Chapter 6

Targeted DNA methylation by a DNA methyltransferasecoupled to a Triple helix Forming Oligonucleotide todownregulate the Epithelial Cell Adhesion Molecule

Bernardina T.F. van der Gun1, Maria Maluszynska-Ho�man2, Antal Kiss3, Alice J. Arendzen1,Marcel H.J. Ruiters4, Pamela M.J. McLaughlin1, Elmar Weinhold2 and Marianne G. Rots1

1 Epigenetic Editing, Dept. of Pathology and Medical Biology, University Medical Center Groningen, University ofGroningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands

2 Institute of Organic Chemistry, RWTH Aachen University, Landoltweg 1, D-52056 Aachen, Germany3 Institute of Biochemistry, Biological Research Center of the Hungarian Academy of Sciences, Temesvári krt. 62, 6726

Szeged, Hungary4 Synvolux Therapeutics Inc., L.J. Zielstraweg 1, 9713 GX Groningen, The Netherlands

Bioconjugate Chem. 2010; in press

92

Chapter 6

6

ABSTRACT

The Epithelial Cell Adhesion Molecule (EpCAM) is a membrane glycoprotein that has been identi�ed

as a marker of cancer-initiating cells. EpCAM is highly expressed on most carcinomas and transient

silencing of EpCAM expression leads to reduced oncogenic potential. To silence the EpCAM gene

in a persistent manner via targeted DNA methylation, a low activity mutant (C141S) of the CpG-

speci�c DNA methyltransferase M.SssI was coupled to a Triple helix-Forming Oligonucleotide

(TFO-C141S) speci�cally designed for the EpCAM gene. Reporter plasmids encoding the green

�uorescent protein under control of di�erent EpCAM promoter fragments were treated with the

TFO-C141S conjugate to determine the speci�city of targeted DNA methylation in the context of

a functional EpCAM promoter. Treatment of the plasmids with TFO-C141S resulted in e�cient and

speci�c methylation of the targeted CpG located directly downstream of the Triple helix Forming

Site (TFS). No background DNA methylation was observed, neither in a 700 bp region of the EpCAM

promoter nor in a 400 bp region of the reporter gene downstream of the TFS. Methylation of

the target CpG did not have a detectable e�ect on promoter activity. This study shows that the

combination of a speci�c TFO and a reduced activity methyltransferase variant can be used to target

DNA methylation to predetermined sites with high speci�city, allowing determination of crucial

CpGs for promoter activity.

93

6

Gene speci�c DNA methylation

INTRODUCTION


on most carcinomas. Recently, EpCAM has gained renewed interest as a signal transducer in

carcinogenesis1, and has been identi�ed as a marker of cancer-initiating cells in breast2, pancreatic3,

hepatocellular4 and colon cancer.5 In the clinical setting, EpCAM has become a target for carcinoma

directed immunotherapy.6 Transient silencing of EpCAM expression utilizing antisense or siRNA

led to reduced proliferation, migration and invasiveness7-10, illustrating the therapeutic potential of

EpCAM inhibition. However, as mRNA molecules are constantly produced, RNA-based approaches

require repeated administration of the inactivating reagent. This study aims to develop a tool to

silence the EpCAM gene in a more persistent manner via targeted DNA methylation.

IthasbeenshownthattheEpCAMpromoterisdi�erentiallymethylatedinlungadenocarcinoma11,

oral squamous cell carcinoma12 and colon cancer13, and that EpCAM expression is associated with

the methylation status of the promoter.11;13;14 For several cell lines, we and others have demonstrated

that treatment with DNA methyltransferase inhibitors like 5-aza-2-deoxycytidine upregulates the

expression of EpCAM.11;14;15 Moreover, after delivery of the CpG-speci�c prokaryotic DNA (cytosine-5)

methyltransferase M.SssI into EpCAM positive ovarian carcinoma cells, methylation of the EpCAM

gene resulted in sustained repression of EpCAM expression.14 However, only if DNA methylation can

be targeted with high speci�city to predetermined sites in the genome, DNA methylation-mediated

gene silencing can be fully exploited as a research tool and further developed as a therapeutic

approach. Application of targeted DNA methylation to silence EpCAM expression would have many

advantages over transient silencing by siRNA. Firstly, the maintenance DNA methyltransferases in

the cell will copy the new methylation mark in the absence of the exogenous methyltransferase, so

a single hit is expected to be su�cient to silence the EpCAM gene in a permanent way. Secondly,

targeted DNA methylation needs to a�ect just two copies of the EpCAM gene rather than the

numerous copies of mRNA present in each cell.

Targeted DNA methylation, pioneered by Xu and Bestor16, traditionally employed DNA

methyltransferases genetically fused to sequence speci�c DNA binding proteins, zinc �nger

proteins, which acted as targeting domains.16-19 As an alternative to zinc �nger proteins, Triple

helix Forming Oligonucleotides (TFO) can be used as targeting domains. TFOs bind by sequence-

speci�c Hoogsteen hydrogen bonds in the major groove of double-stranded DNA and have been

used to target cleaving20, cross-linking reagents21 or anticancer agents22 to unique target sequences.

The advantages of TFOs relative to zinc �nger proteins are the easy synthesis and low cost. The

drawbacks are the requirements for an in vitro coupling step to covalently link the e�ector protein

to the TFO and the limitation of binding to oligopurine-oligopyrimidine sequences. However, the

latter limitation is mitigated by the over-representation of oligopurine-oligopyrimidine stretches in

promoter regions of human genes.23-25 Recently, we and others have shown that coupling of a TFO to

the restriction enzymes scPvuII20 and Mun-I is feasible (Geel et al, manuscript in preparation).

94

Chapter 6

6

To explore the possibility of targeting methylation to speci�c DNA sequences using a TFO as

targeting domain, we coupled a variant of the methyltransferase M.SssI to aTFO speci�cally designed

for the EpCAM gene.26 This M.SssI variant, named C141S throughout this paper, carried the C141S

and C368A substitutions and a C-terminal 6xHis-Cys tag. In this variant the internal cysteines were

thus replaced and the C-terminal Cys was introduced to allow coupling of the methyltransferase

to the TFO. Although one of the replaced cysteines (C141S) is the active site cysteine, the C141S

mutation does not fully abolish the methyltransferase activity, the mutant enzyme has 2-5% of

the wild type (WT) activity.27 DNA binding a�nity of the C141S variant is similar to that of the WT

enzyme.28 Coupling of C141S to the TFO did not a�ect activity of the enzyme, and binding speci�city

of the TFO-C141S conjugate to the DNA was dominated by the TFO. Site-speci�c methylation by

the TFO-C141S conjugate was demonstrated using a plasmid containing a 43 bp segment of the

EpCAM promoter encompassing the Triple helix Forming Site (TFS) and a target CpG.26 To investigate

targeted methylation in the context of a functional EpCAM promoter, reporter plasmids encoding

the Green Fluorescent Protein (GFP) under control of di�erent EpCAM promoter fragments29 were

treated with the TFO-C141S conjugate and assayed for DNA methylation and gene expression.

In this study we demonstrate that a TFO coupled to a reduced activity DNA methyltransferase

can be directed to a pre-determined site to induce targeted methylation, allowing determination of

crucial CpGs for promoter activity.


Methylation of plasmids with TFO-C141S

The p39E plasmid and its promoter deletion derivatives29 are schematically depicted in Figure 1A.

These plasmids encode GFP under control of a 3.4 kb fragment of the EpCAM promoter. Construc-

tion of the plasmids expressing WT M.SssI or C141S, puri�cation of the enzymes14, and coupling

of the 5’-TTTTTTTTTTTTTTTCTCTCTTTT-3’ TFO to M.SssI(C141S) was done as described.26 Plasmids

were incubated with �ve-fold molar excess of TFO-C141S, M.SssI, C141S or TFO in a bu�er contain-

ing 20 mM Tris, 50 mM NaCl, 10 mM MgCl2, pH 7.9 with or without 640 μM S-adenosylmethionine

(SAM) (New England Biolabs, Ipswich, MA) at 30°C. The reaction was terminated after 20 h by heat

inactivation at 65°C for 20 minutes, and plasmids were puri�ed by Qiagen PCR puri�cation kit (Qia-

gen, Benelux B.V., Venlo, The Netherlands).

To test the involvement of enzyme activity in TFO-C141S-induced relaxation of supercoiled plasmid

DNA, C141S and TFO-C141S were heat-inactivated by incubation at 65°C for 20 min. Plasmid p39E

was incubated with active or inactivated C141S or TFO-C141S for 15 or 30 minutes, 1, 2.5, 5 or 20

hours at 30°C, then the reaction was stopped by 10% SDS and digestion with proteinase K.

95

6


Transfection

The SKOV3 (HTB-77) cell line was purchased from ATCC (Manasas, VA) and cultured according to

ATCC recommendations. SKOV3 cells were seeded 100.000/well in 24-wells plates, and transfection

was performed at 60-80% con�uency using SAINT-2:DOPE (SD; 0.75mM) (Synvolux Therapeutics,

Groningen, The Netherlands).30 Pretreated plasmid DNA (250 ng) in 25 μl Hanks Balanced Salt

solution was added to 5 μl SD �lled up with 20 μl HBS. Within 15 minutes the complex was diluted in

200 μl serum free culture medium and added directly to the cells. After 3 h incubation at 37°C, in 5%

CO2, serum containing medium was added. After 48 h, cells were harvested and GFP expression was

measured by �ow cytometry (Beckton Dickenson Bioscience Calibur, San Jose, CA).


Plasmid DNA (250 ng) methylated in vitro by TFO-C141S was treated with sodium bisul�te to convert

unmethylated cytosines to uracils using the EZ DNA Methylation-Gold Kit (Zymo, Baseclear Lab

Products, Leiden, The Netherlands). Bisul�te speci�c primers void of any CpG were used to obtain

Figure 1. Schematic overview of EpCAM promoter fragments in the GFP reporter plasmids. A) Plasmidp39E encompasses the GFP gene under control of 3.4 kb of the EpCAM promoter. The Triple helix FormingSite (TFS) as well as sequences di!ering from the targeted TFS by only 3 or 4 mismatches (3MM or 4MM) areindicated. CpGs are depicted by vertical bars. The Transcription Start Site (TSS) corresponds to position +1, thelocation of the amplicons analyzed for DNA methylation are indicated (not drawn to scale). B) Sequence of onestrand of bisul�te converted, fully methylated DNA of the TFS amplicon. All cytosines in CpG combination areassumed to be methylated and therefore not converted by bisul�te treatment. The CpGs are numbered anddepicted in green. The TFS is shown in red, containing three Cs which have been converted to Ts, CpG number7 is the targeted CpG.

TSS

GFP

Second amplicon

TFSp39E (-3340)

First ampliconTFS amplicon GFP amplicon

TFS (4MM) TFS(3MM)TFS(3MM) TFS (4MM)

TSS

p7-2 (-1023) TFS (4MM)TFS

TSS

p15-2 (-2088) TFS (4MM)TFSTFS(3MM)

TSS

p4-1 (-688) TFS (4MM)

GFP

GFP

TSS

p11-1 (-341) GFP

GFP

aataatatagtgtgttgtgatttgaatttatttgtacg1gaaatcg2attattgttttttttttttatttttttatatttttttttcg3aaggcg4ttattaatat

tttggttttttaatagtaattaaaattcg5aaattatttcg6gtttttagtatttggttttatgggaatatttttttttttttttttttttttttttttttgagacg7

gagttttgtttttgtcg8tttaggttggagtgtaatggtacg9atttttgtttattgtaattttagtttttttagtagttgggattatagg

A

B

96

Chapter 6

6

ampli�cation products (amplicons) unbiased for the methylation status. Primer sequences for the

Triple helix Forming Site (TFS) amplicon were 5’-AATAATATAGTGTGTTGTGATTT-3’ (forward) and

5’-CCTATAATCCCAACTACTAA-3’ (reverse) (The PCR product is shown in Figure 1B). Two overlapping

amplicons were selected to cover a 700 bp region directly downstream of the TFS in the EpCAM

promoter. Primer sequences for the �rst amplicon were 5’-ACCTCCCCAATAACTAAAATTAC-3’

(forward), 5’-TTGAAGATTTTGTGTTGAGATTT-3’ (reverse), and for the second one

5’-AGTGTTTTGGAAGGTTTTTTGT-3’ (forward), 5’-AAATTAAAAAAATAAATAAACTCCC-3’ (reverse).

Primers used for the GFP amplicon were 5’-GGGGTGGTGTTTATTTTG-3’ (forward) and

5’-CTCCAACTTATACCCCAAAAT-3’(reverse). The location of the amplicons in the plasmids is indicated

in Figure 1A. PCR conditions: 95°C for 15 min, followed by 35 cycles of 95°C for 45 s, 53-56°C for 45

s, 72°C for 45 s and �nished with 72°C for 10 min. PCR fragments were puri�ed from gels using

the DNA Extraction Kit (Qiagen) and cloned into pCR 2.1-TOPO TA vector (Invitrogen, Breda, The

Netherlands). Following transformation, plasmids were isolated from individual bacterial colonies

using the Qiaprep Spin Miniprep Kit (Qiagen) and subjected to restriction analysis. Clones with the

expected structure were sequenced.

RESULTS

Targeted DNA methylation of the EpCAM promoter in p39E

To test the targeting speci�city of the TFO-C141S conjugate in the context of a functional EpCAM

promoter, p39E containing the 3.4 kb EpCAM promoter was treated with TFO-C141S. Three

independent TFO-C141S treatments resulted in 57% (n=7), 89% (n=9) and 75% (n=8) methylation

of CpG7 (Figure 1B) located directly downstream the TFS (Table 1A). Of the 24 clones analyzed, 18

showed methylation of CpG7; only three clones were devoid of methylation, one clone showed

methylation only of CpG9 and two clones showed methylation only of CpG5 located ~40 bp

upstream of the TFS. Four of the 18 positive clones showed methylation of CpG5 and CpG7, and

one of these clones also showed methylation of CpG6. CpG1, 2, 3, 4 and 8 within the TFS amplicon

(Figure 1B) were not methylated in the clones sequenced. Importantly, apart from one sporadic

methylation event, no methylation was observed downstream of the TFS neither in the EpCAM

promoter (EpCAM �rst and second amplicon, Table 1A) nor in the part of the GFP gene (GFP

amplicon). In the control samples, obtained by treating p39E with WT M.SssI, all CpGs analyzed were

methylated (Table 1B), whereas treatment with the less active C141S variant resulted in random

methylation.

GFP expression from the EpCAM promoter after pretreatment with TFO-C141S

After observing that the TFO-C141S conjugate induced targeted DNA methylation, we tested

if methylation of this single CpG (CpG7) is su!cient to inhibit EpCAM promoter activity. Hence,

97

6Table 1. Targeted DNA methylation of p39E by TFO-C141S. A) Methylation status of bisul!te amplicons forp39E treated with TFO-C141S and the methyl donor SAM. Results of bisul!te sequencing. B) Methylation statusof bisul!te amplicons for p39E treated with the controls as indicated: no treatment, treated with TFO-C141Swithout methyl donor (-SAM), with untargeted C141S or with untargeted WT M.SssI. (n = number of clonesanalyzed, below the lollypop: the percentage of clones of which the CpG was methylated; open lollypop = 0%methylated CpG, black lollypop = 100% methylated CpG).

all treated plasmids were transfected in SKOV3 cells to determine GFP expression. The three

independent treatments of p39E with TFO-C141S showed 46%±20 (n=3), 48%±16 (n=3) and

57%±6 (n=2) GFP downregulation (see Figure 2A for a representative experiment). Transfection

of plasmids treated with untargeted M.SssI or untargeted C141S resulted in 91%±1 and 27%±1

GFP downregulation, respectively, whereas treatment with just the TFO did not in#uence GFP

expression. Transfection of p39E treated with a 100-fold excess of TFO in the presence of TFO-C141S

(competition sample in Figure 2A) showed 24%±17 downregulation of GFP, which is approximately

33% less then the value obtained with TFO-C141S only, indicating that targeting was dependent on

the TFO. These observations suggested that methylation of a single CpG was su$cient to reduce

EpCAM promoter activity.


GFP 33 CpGsEpCAM second ampliconoverlapEpCAM first ampliconTFS ampliconn

none3

2

6

24

25 4 75 4

50

4M.SssI

3M.SssI

4M.SssI

9C141S

10C141S

3TFO-C141S-SAM

2no

3no

EpCAM second ampliconoverlapEpCAM first ampliconTFS ampliconntreatment

4M.SssI

3M.SssI

4M.SssI

9C141S

10C141S

3TFO-C141S-SAM

2no

3no

EpCAM second ampliconoverlapEpCAM first ampliconTFS ampliconntreatment

100

100

10 10 20 10 40 10 10

11 11 11

75

100 67 67100 100

A

B

98

Chapter 6

6

Downregulation of GFP gene expression by TFO-C141S in the absence of TFS or methyl donor

To investigate if the reduction in GFP expression was indeed due to targeted DNA methylation,

deletion derivatives of p39E29 (p15-2, p7-2, p4-1 and p11-1) (Figure1A) were treated with TFO-C141S

and transfected into SKOV3 cells to determine GFP expression. Plasmids p15-2 and p7-2 contained

the TFS, whereas plasmids p4-1 and p11-1 lacked this sequence. In addition to the targeted TFS, the

EpCAM promoter region contains four sites, which di�er from the targeted TFS only by three or four

mismatches (3MM or 4MM, as indicated in Figure 1A). Plasmids p7-2 and p4-1 treated with TFO-

C141S showed 51%±4 and 46%±8 GFP downregulation, respectively (Figure 2B). The latter result

was unexpected as plasmid p4-1 lacked the TFS. To exclude the possibility that downregulation was

caused by binding of TFO-C141S to the TFS 4MM site (Figure 1A) present in p4-1, plasmids p15-2 and

p11-1 were treated with TFO-C141S and transfected into SKOV3 cells. Plasmid p15-2 contains the

Figure 2. E!ect of TFO-C141S treatment on GFP expression in EpCAM positive SKOV3 cells. A) Relative GFPexpression measured 48 h after transfection of pretreated p39E. Plasmid p39E was treated as indicated: p39E= treatment without TFO-C141S, treated with TFO only, with untargeted M.SssI or C141S, with the TFO-C141Sconjugate or with 100-fold excess of TFO and TFO-C141S (=competition). The value obtained with p39E withoutTFO-C141S was taken as 100%. Average GFP expression (±SD) of one representative transfection performed intriplicate. B) Relative GFP expression measured 48 h after transfection of pretreated deletion derivatives p7-2and p4-1. For each derivative, the values obtained with samples treated without TFO-C141S were taken as100%. Average GFP expression (±SEM) of the mean of three independent transfections performed in triplicate.C) Relative GFP expression measured 48 h after transfection of pretreated p39E or p39C. Treatments were asindicated: (+) or (-) indicate the presence or absence of the methyl donor (SAM). Average GFP expression (±SEM)of the mean of three independent transfections performed in triplicate.


TFS and two TFS-like sites, whereas from p11-1 all potential binding sites had been deleted (Figure

1A). Both pretreated plasmids showed approximately 45% reduction of GFP expression compared

to their untreated controls (data not shown).

To exclude that the reduction in GFP expression was not caused by aspeci�c DNA methylation,

the methylation status of the deletion derivatives of p39E was analyzed. Treatment of p15-2 with

TFO-C141S resulted in 36% methylation of CpG7 (5/14 clones) and one clone showed methylation

of CpG5 (Table 2). Apart from one sporadic event, deletion derivatives p7-2 and p4-1 and p11-1

did not show methylation in the downstream amplicons. These results con�rm the speci�city of

targeted methylation by TFO-C141S.

99

6

Table 2. Targeted DNA methylation of p39E deletion derivatives by TFO-C141S. Methylation status ofbisul�te amplicons for the deletion derivatives p39E treated with TFO-C141S and the methyl donor SAM. Resultsof bisul�te sequencing (n = number of clones analyzed, below the lollypop: the percentage of clones of whichthe CpG was methylated; open lollypop = 0% methylated CpG, black lollypop = 100% methylated CpG).

To exclude that reduction in GFP expression was the result of the conjugate still being attached

to the plasmid, p39E was also treated with TFO-C141S in the absence of S-adenosyl-methionine

(SAM). No di!erence in reduction of GFP expression was observed between samples treated with or

without the methyl donor (Figure 2C). To con�rm that reduction in GFP expression was not caused

by aspeci�c DNA methylation, plasmid p39C encoding GFP under control of the CMV promoter

without a TFS like sequence, was subjected to TFO-C141S treatment. Also for p39C, GFP expression

was reduced while no di!erence was observed between treatments in the presence or in the

absence of SAM (Figure 2C).

TFO-C141S induced relaxation of supercoiled plasmids

Because GFP expression from the transfected pretreated plasmids seemed to be independent of

their methylation status, we examined the conformation of the pretreated plasmids. Treatment of

p39E with TFO-C141S with or without methyl donor led to relaxation of the supercoiled plasmid

(Figure 3A lane 6, 7). The deletion plasmid p11-1 behaved in the same way, treatment with TFO-

C141S caused relaxation (lane 12), whereas treatment with only the TFO, M.SssI or C141S did not

a!ect conformation of the plasmid. Interestingly, the presence of excess TFO seemed to protect to

5p4-1

1CpG in 1clone

6p11-1

3P7-2

6p7-2

14p15-2

GFP 33 CpGsEpCAM second ampliconoverlapEpCAM first ampliconTFS ampliconnplasmid

5p4-1

1CpG in 1clone

6p11-1

3P7-2

6p7-2

14p15-2

GFP 33 CpGsEpCAM second ampliconoverlapEpCAM first ampliconTFS ampliconnplasmid

7 36

33

100

6

Figure 3. E�ect of TFO-141S treatment on plasmid conformation. A) Agarose gel electrophoresis of plasmidsp39E and p11-1 treated with the TFO, WT M.SssI, C141S or the TFO-C141S conjugate: lane 1: untreated, 2:without TFO-C141S, 3: TFO, 4: M.SssI, 5: C141S, 6: TFO-C141S, 7: TFO-C141S without SAM, 8: 100-fold excess ofTFO and TFO-C141S, 9: marker, 10: without TFO-C141S, 11: C141S, 12: TFO-C141S. B) Agarose gel electrophoresisof plasmid p39E treated with active and heat-inactivated C141S and TFO-C141S. The supercoiled plasmidwas incubated at 30˚C for di�erent time points as indicated above the lanes (hours), then the samples weredeproteinized before electrophoresis as described in Material and Methods C) Agarose gel electrophoresis ofplasmid p39C and p39E treated with the TFO-C141S conjugate only or in the presence of 100-fold excess of TFO.Plasmids were incubated as in B, lane a is puri!ed plasmid.

some extent the plasmid from conversion into the relaxed form (lane 8).

The phenomenon of conformation change appeared to be general: treatment of p39C, p7-2 and

p4-1 with TFO-C141S also caused a conversion to the relaxed form (data not shown). To exclude the

possibility that the slower electrophoretic mobility was caused by TFO-C141S still being attached

to the plasmid, the TFO-C141S treated samples were digested with proteinase K. No di�erence in

conformation was observed between the plasmids with and without proteinase K treatment (data

not shown), indicating that C141S was not attached to the plasmid.

Because it has been reported that M.SssI can display topoisomerase activity31, TFO-C141S was

heat-inactivated before treatment of p39E. Treatment of the plasmid p39E with heat-inactivated

TFO-C141S did not result in conformation change, whereas treatment with active TFO-C141S led to

conversion from the supercoiled into the relaxed form (Figure 3B). This change in conformation was

dependent on the duration of the treatment i.e. longer treatment resulted in a less supercoiled and

Chapter 6

1091 2 3 4 876 21115

supercoiledrelaxed

p39E p11-11091 2 3 4 876 21115 1091 2 3 4 87651 2 3 4 876 21115

-A

BC141S

C141Sinactivated TFO-C141S

TFO-C141S

inactivated

0.25 0.5 1 2.5 5

supercoiled

relaxed

0.25 0.5 1 2.5 5 0.25 0.5 1 2.5 5 0.25 0.5 1 2.5

-

-

0.25 0.5 1 2.5 50.25 0.5 1 2.5 5 0.25 0.5 1 2.5 50.25 0.5 1 2.5 5 0.25 0.5 1 2.5 50.25 0.5 1 2.5 5 0.25 0.5 1 2.50.25 0.5 1 2.5

C

p39C

TFO-C141S TFO-C141S + excess TFO

0 5 202.51a 0 5 202.51

p39E

TFO-C141S TFO-C141S + excess TFO

0 5 202.51a 0 5 202.51

supercoiled

relaxed

- - - -

1

- -- - -- -

5

101

6


a more relaxed conformation. Treatment of p39E with C141S or heat-inactivated C141S showed no

conformation change.

Based on the observation that excess TFO could protect the p39E plasmid from TFO-C141S

mediated relaxation (Figure 3A lane 6, 7, 8), we investigated this protection in more detail.

Treatment of p39E as well as p39C with TFO-C141S only, showed e�cient relaxation already after

2.5h of treatment (Figure 3C). In the presence of excess TFO, again both plasmids showed relaxation

although the e�ciency was somewhat less (Figure 3C). To further prove that the presence of the TFS

is not necessary for relaxation, plasmids with a di�erent backbone were subjected to TFO-C141S

treatment. Although the backbones did not contain any TFS-like sequences we again observed an

increase in the relaxed conformation when the incubation time with TFO-C141S was increasing

(data not shown).

DISCUSSION

In this study, we demonstrated that a DNA methyltransferase can be directed to a pre-determined

site by a covalently attached TFO, and that TFO-mediated targeting can be used to induce targeted

methylation. Because the TFO binds antiparallelly to the sense strand of the Triple helix Forming

Site (TFS) in the EpCAM promoter and C141S is coupled to the 5’-end of the TFO, C141S should

orient downstream of the TFS. We therefore expected CpG7 (Figure 1B) to be the main target for

C141S. Indeed, treatment of reporter plasmids with TFO-C141S resulted in e�cient (18/24 clones)

and speci!c (14/21 clones) methylation of CpG7 located directly downstream the TFS. These results

indicate that binding speci!city of TFO-C141S is dominated by the TFO.

The 24 nucleotide longTFO used in this study contains three cytosines. Speci!c binding of theTFO

requires protonation of the three cytosines to form the C+.GC triplets (pH<6).21 However, treatment

of the plasmids with TFO-C141S was performed under conditions optimal for methylation (pH 7.9),

i.e. at a pH higher than required for e�cient formation of C+.GC triplets. Despite the suboptimal

annealing conditions, e�cient DNA methylation was observed at the target site.

The 3.4 kb EpCAM promoter encompasses four sites that are similar to the targeted TFS (i.e. di�er

only at three or four positions); hence they would be expected to be potential sites of nontargeted

binding and methylation. To test this possibility, deletion derivatives of p39E containing (p15-

2, p7-2) or lacking the TFS (p4-1, p11-1) were treated with TFO-C141S. Despite the presence of a

possible binding site in amplicon 1, CpG7 was preferentially methylated (18/24 clones), whereas

no methylation was observed for amplicon 1 in all treated plasmids. Moreover, deletion derivatives

lacking the TFS (p4-1, p11-1) displayed no methylation in the analyzed areas. We thus conclude that

the TFO directs C141S only to its predetermined site.

Next, we investigated if this site-speci!c DNA methylation in the EpCAM promoter is su�cient to

induce inhibition of gene expression. It has been shown that methylation of one or a few CpGs within

102

Chapter 6

6

a promoter might be su�cient to repress transcription. Transfection of a reporter plasmid under

control of the p53 promoter in which a single CpG was methylated in vitro by HhaI, showed 85%

downregulation of the reporter gene.32 Unexpectedly, in our experiments TFO-C141S treatment of

all GFP reporter plasmids resulted in lower GFP expression and this phenomenon was independent

of the methylation status of the transfected plasmid. Gel electrophoresis revealed that TFO-C141S

treatment, in contrast to treatment with the enzyme or TFO only, led to relaxation of the supercoiled

plasmid explaining the observed GFP repression.33 Since plasmid treatment with heat-inactivated

TFO-C141S did not cause a conformation change of the plasmid, we might conclude that the

observed relaxation is probably caused by topoisomerase activity of C141S. Matsuo et al described

that M.SssI contains both methylase and topoisomerase activities.31 However, the observed

topoisomerase activity was only displayed by the TFO-C141S conjugate and not after treatment

with the TFO only or with C141S only. Independent of the presence of the TFS, all plasmids treated

with the TFO-C141S conjugate showed relaxation of the plasmid: the observed ratio of supercoiled

versus relaxed plasmid conformation was decreasing when the treatment time was prolonged.

Somehow, the chemical coupling of the TFO with the enzyme seems to change the conformation of

the enzyme, thereby uncovering the catalytic topoisomerase domain. It requires further research to

determine which amino acid should be replaced to abolish this activity.

A puzzling observation of this study was that there was no signi!cant di"erence in GFP

downregulation between plasmids treated in the presence or absence of the methyl donor.

A possible explanation might be that the CpGs targeted in this study do not play a role in the

epigenetic regulation of the EpCAM gene. Indeed, recent observations11;14;15 suggest that the CpGs,

which display di"erential methylation are located further downstream (approximately -400 to +280

bps) of the TFS which is located around -917. Currently, we are investigating which CpGs are crucial

in regulating EpCAM gene expression. The important CpGs can then be targeted for induced DNA

methylation.

In summary, the data obtained with the TFO-C141S conjugate o"er a novel approach for

targeted DNA methylation. The combination of a speci!c TFO and the reduced methyltransferase

activity of the M.SssI mutant C141S allowed us to target methylation predominantly to a speci!c

DNA sequence without signi!cant background methylation. Because of the #exibility provided by

the use of TFOs as targeting domain, this approach appears to be a promising tool in both research

and therapeutic areas.

ACKNOWLEDGEMENTS

This work was supported by grants QRLT-2001-0448 and COOP-CT-2005-017984 of the European

Commission’s Fifth and Sixth Framework Programmes, respectively. We thank dr. Bill jack (New

England Biolabs) for the original plasmid with the M.SssI gene and Burcu Duycu for technical

assistance.

103

6


REFERENCES

1. Maetzel, D., Denzel, S., Mack, B., Canis, M., Went, P., Benk, M., Kieu, C., Papior, P., Baeuerle, P. A., Munz, M., andGires, O. (2009) Nuclear signaling by tumor-associated antigen EpCAM, Nat Cell Biol 11, 162-171.

2. Al-Hajj, M., Wicha, M. S., ito-Hernandez, A., Morrison, S. J., and Clarke, M. F. (2003) From the Cover:Prospective identi!cation of tumorigenic breast cancer cells, PNAS 100, 3983-3988.

3. Li, C., Heidt, D. G., Dalerba, P., Burant, C. F., Zhang, L., Adsay, V., Wicha, M., Clarke, M. F., and Simeone, D. M.(2007) Identi!cation of Pancreatic Cancer Stem Cells, Cancer Res 67, 1030-1037.

4. Yamashita, T., Forgues, M., Wang, W., Kim, J. W., Ye, Q., Jia, H., Budhu, A., Zanetti, K. A., Chen, Y., Qin, L. X.,Tang, Z. Y., and Wang, X. W. (2008) EpCAM and {alpha}-Fetoprotein Expression De!nes Novel PrognosticSubtypes of Hepatocellular Carcinoma, Cancer Res 68, 1451-1461.

5. Dalerba, P., Dylla, S. J., Park, I. K., Liu, R., Wang, X., Cho, R. W., Hoey, T., Gurney, A., Huang, E. H., Simeone, D.M., Shelton, A. A., Parmiani, G., Castelli, C., and Clarke, M. F. (2007) Phenotypic characterization of humancolorectal cancer stem cells, PNAS 104, 10158-10163.

6. Baeuerle, P. A. and Gires, O. (2007) EpCAM (CD326) !nding its role in cancer, Br. J Cancer 96, 417-423.

7. Munz, M., Kieu, C., Mack, B., Schmitt, B., Zeidler, R., and Gires, O. (2004) The carcinoma-associated antigenEpCAM upregulates c-myc and induces cell proliferation, Oncogene 23, 5748-5758.

8. Osta, W. A., Chen, Y., Mikhitarian, K., Mitas, M., Salem, M., Hannun, Y. A., Cole, D. J., and Gillanders, W. E.(2004) EpCAM is overexpressed in breast cancer and is a potential target for breast cancer gene therapy,Cancer Res. 64, 5818-5824.

9. Yanamoto, S., Kawasaki, G., Yoshitomi, I., Iwamoto, T., Hirata, K., and Mizuno, A. (2007) Clinicopathologicsigni!cance of EpCAM expression in squamous cell carcinoma of the tongue and its possibility as apotential target for tongue cancer gene therapy, Oral Oncology 43, 869-877.

10. Du, W., Ji, H., Cao, S., Wang, L., Bai, F., Liu, J., and Fan, D. EpCAM: A Potential Antimetastatic Target for GastricCancer, Digestive Diseases and Sciences, 1033-1038.

11. Tai, K. Y., Shiah, S. G., Shieh, Y. S., Kao, Y. R., Chi, C. Y., Huang, E., Lee, H. S., Chang, L. C., Yang, P. C., and Wu, C. W.(2007) DNA methylation and histone modi!cation regulate silencing of epithelial cell adhesion moleculefor tumor invasion and progression, Oncogene 26, 3989-3997.

12. Shiah, S. G., Chang, L. C., Tai, K. Y., Lee, G. H., Wu, C. W., and Shieh, Y. S. The involvement of promotermethylation and DNA methyltransferase-1 in the regulation of EpCAM expression in oral squamous cellcarcinoma, Oral Oncology 45, e1-e8.

13. Yu, G., Zhang, X., Wang, H., Rui, D., Yin, A., Qiu, G., and He, Y. (2008) CpG island methylation status in theEpCAM promoter region and gene expression, Oncol. Rep. 20, 1061-1067.

14. van der Gun, B. T., Wasserkort, R., Monami, A., Jeltsch, A., Rasko, T., Slaska-Kiss, K., Cortese, R., Rots, M. G.,de Leij, L. F., Ruiters, M. H., Kiss, A., Weinhold, E., and McLaughlin, P. M. (2008) Persistent downregulation ofthe pancarcinoma-associated epithelial cell adhesion molecule via active intranuclear methylation, Int. J.Cancer 123, 484-489.

15. Spizzo, G., Gastl, G., Obrist, P., Fong, D., Haun, M., Grunewald, K., Parson, W., Eichmann, C., Millinger, S., Fiegl,H., Margreiter, R., and Amberger, A. (2007) Methylation status of the Ep-CAM promoter region in humanbreast cancer cell lines and breast cancer tissue, Cancer Lett. 246, 253-261.

16. Xu, G. and Bestor, T. (1997) Cytosine methylation targetted to pre-determined sequences, Nat Genet 17,376-378.

17. Li, F., Papworth, M., Minczuk, M., Rohde, C., Zhang, Y., Ragozin, S., and Jeltsch, A. (2007) Chimeric DNAmethyltransferases target DNA methylation to speci!c DNA sequences and repress expression of targetgenes, Nucleic Acids Res 35, 100-112.

18. Nomura, W. and Barbas, C. F., III (2007) In vivo site-speci!c DNA methylation with a designed sequence-enabled DNA methylase, J Am Chem. Soc. 129, 8676-8677.

104

19. Smith, A. E. and Ford, K. G. (2007) Speci�c targeting of cytosine methylation to DNA sequences in vivo, Nucl.Acids Res. 35, 740-754.

20. Eisenschmidt, K., Lanio, T., Simoncsits, A., Jeltsch, A., Pingoud, V., Wende, W., and Pingoud, A. (2005)Developing a programmed restriction endonuclease for highly speci�c DNA cleavage, Nucl. Acids Res. 33,7039-7047.

21. Duca, M., Vekho!, P., Oussedik, K., Halby, L., and Arimondo, P. B. (2008) The triple helix: 50 years later, theoutcome, Nucl. Acids Res. 36, 5123-5138.

22. Vekho!, P., Halby, L., Oussedik, K., Dallavalle, S., Merlini, L., Mahieu, C., Lansiaux, A., Bailly, C., Boutorine, A.,Pisano, C., Giannini, G., Alloatti, D., and Arimondo, P. B. (2009) Optimized synthesis and enhanced e"cacyof novel triplex-forming camptothecin derivatives based on gimatecan, Bioconjug. Chem 20, 666-672.

23. Wu, Q., Gaddis, S. S., Macleod, M. C., Walborg, E. F., Thames, H. D., DiGiovanni, J., and Vasquez, K. M. (2007)High-a"nity triplex-forming oligonucleotide target sequences in mammalian genomes, Mol Carcinog. 46,15-23.

24. Goni, J. R., de la Cruz, X., and Orozco, M. (2004) Triplex-forming oligonucleotide target sequences in thehuman genome, Nucl. Acids Res. 32, 354-360.

25. Goni, J., Vaquerizas, J., Dopazo, J., and Orozco, M. (2006) Exploring the reasons for the large density oftriplex-forming oligonucleotide target sequences in the human regulatory regions, BMC Genomics 7, 63.

26. Maluszynska-Ho!man, M. (2009) Super-speci�cDNA methyltion by a DNA methyltransferase coupled witha triple helix-forming oligonucleotide, http://darwin.bth.rwth-aachen.de/opus3/volltexte/2009/2854/ .

27. Rathert, P., Rasko, T., Roth, M., Slaska-Kiss, K., Pingoud, A., Kiss, A., and Jeltsch, A. (2007) Reversibleinactivation of the CG speci�c SssI DNA (cytosine-C5)-methyltransferase with a photocleavable protectinggroup, Chembiochem. 8, 202-207.

28. Darii, M. V., Cherepanova, N. A., Subach, O. M., Kirsanova, O. V., Rasko, T., Slaska-Kiss, K., Kiss, A., ville-Bonne, D., Reboud-Ravaux, M., and Gromova, E. S. (2009) Mutational analysis of the CG recognizing DNAmethyltransferase SssI: Insight into enzyme-DNA interactions, Biochim. Biophys. Acta 1794, 1654-1662.

29. McLaughlin, P. M., Trzpis, M., Kroesen, B. J., Helfrich, W., Terpstra, P., Dokter, W. H., Ruiters, M. H., de Leij, L.F., and Harmsen, M. C. (2004) Use of the EGP-2/Ep-CAM promoter for targeted expression of heterologousgenes in carcinoma derived cell lines, Cancer Gene Ther. 11, 603-612.

30. van der Gun, B. T., Monami, A., Laarmann, S., Rasko, T., Slaska-Kiss, K., Weinhold, E., Wasserkort, R., de Leij, L.F., Ruiters, M. H., Kiss, A., and McLaughlin, P. M. (2007) Serum insensitive, intranuclear protein delivery bythe multipurpose cationic lipid SAINT-2, J. Control Release 123, 228-238.

31. Matsuo, K., Silke, J., Gramatiko!, K., and Scha!ner, W. (1994) The CpG-speci�c methylase SssI hastopoisomerase activity in the presence of Mg2+, Nucleic Acids Res 22, 5354-5359.

32. Pogribny, I. P., Pogribna, M., Christman, J. K., and James, S. J. (2000) Single-Site Methylation within thep53 Promoter Region Reduces Gene Expression in a Reporter Gene Construct: Possible in Vivo Relevanceduring Tumorigenesis, Cancer Res 60, 588-594.

33. Remaut, K., Sanders, N. N., Fayazpour, F., Demeester, J., and De Smedt, S. C. (2006) In#uence of plasmid DNAtopology on the transfection properties of DOTAP/DOPE lipoplexes, J. Control Release 115, 335-343.

105

Chapter 7Sustained downregulation of EpCAM gene expression by

siRNA targeting a coding region

Bernardina T.F. van der Gun1, Hinke G. Kazemier1, Alice J. Arendzen1, Marcel H.J. Ruiters2,Pamela M.J. McLaughlin1,a, and Marianne G. Rots1,a

1 Epigenetic Editing, Dept. of Pathology and Medical Biology, University Medical Center Groningen, Hanzeplein1, 9713GZ Groningen, The Netherlands

2 Synvolux Therapeutics Inc., Groningen, The Netherlandsa Both authors contributed equally

under revision

106

Chapter 7

7

ABSTRACT

Short interference (si)RNAs are commonly used to transiently silence genes by degrading target

mRNA. To achieve sustained gene silencing, continuous presence of siRNA is required. Previously,

we have demonstrated that genome wide DNA methylation leads to permanent downregulation

of the Epithelial Cell Adhesion Molecule (EpCAM). Here, we report on sustained EpCAM silencing

in a subset of cells after a single transfection with siRNA. EpCAM positive SKOV3 cells were

transfected with siRNA targeting exon 5 or with the DNA methyltransferase M.SssI, followed by

two cycles of sorting and sub-culturing the weak EpCAM expressing cells. EpCAM expression was

reduced by siRNA to 14% of untreated SKOV3 cells. After a �rst sort and sub-culturing for 20 days,

EpCAM expression of these siRNA-sort-I cells was restored to 88% of untreated SKOV3 cells. After

a second sort, EpCAM expression of the siRNA-sort-II cell population remained at about 50% of

untreated SKOV3 cells for at least another 44 days. EpCAM expression was reduced by M.SssI to 70%

of untreated SKOV3 cells, which remained about 20% for the sort-II population. The overall DNA

methylation percentages of the 79 tested CpGs in the EpCAM gene were 31% for the siRNA-sort-II

and 67% for the M.SssI-sort-II population compared to 2% for SKOV3 cells. To exclude selection of

cells with a ”spontaneous” high DNA methylation level, untreated SKOV3 cells were also subjected

to two cycles of sorting and subculturing. EpCAM expression of the SKOV3 sort-II population was

restored to 83% of the non-sorted cells. The DNA methylation percentage was 1% for both, the

SKOV3-sort-II as for the non-sorted SKOV3 cells. This study shows a double mode of action for exon

targeting siRNAs: inducing transient silencing of gene expression in all treated cells while inducing

permanent, DNA methylation associated silencing, in a subset of cells.

107

7

SiRNA-mediated transcriptional gene silencing

INTRODUCTION


on most carcinomas. Transient downregulation of EpCAM by siRNA reduced the oncogenic potential

of treated cells1-5, illustrating the therapeutic potential of EpCAM inhibition. We and others6-9 have

previously shown that EpCAM expression is associated with DNA methylation. After delivery of the

DNA methyltransferase M.SssI into EpCAM positive cells, the reduced EpCAM expression persisted

for at least 17 days.8 In this paper, we describe that siRNA targeting a coding region of the EpCAM

gene does not only induce e�cient transient downregulation of EpCAM expression in all cells, but

also resulted in sustained silencing in a subset of cells.

RNA interference pathways silence gene expression either at the post-transcriptional or the

transcriptional level.10 Post-transcriptional gene silencing by siRNAs designed to target mRNA is

based on induction of sequence speci�c cleavage of perfectly complementary messenger RNA.11

Transcriptional gene silencing (TGS) by siRNA designed to target gene promoters is based on

induction of cytosine DNA methylation and histone modi�cations at targeted sites.10 In plants,

dsRNAs designed to target CpG islands within a promoter are known to induce RNA-directed

DNA methylation resulting in sustained gene silencing.12 Also in human cells, siRNA induced

DNA methylation has been reported: siRNA designed to target to the promoters of the urokinase

plasmogen activator13 and elongation factor 1 alpha14 induced DNA methylation and showed

suppression of the genes at the transcriptional level. Similarly, siRNA targeting the E-cadherin

promoter induced DNA methylation and methylation of lysine 9 of histone H3 resulting in repression

of E-cadherin at the transcriptional level.15

In contrast, no DNA methylation of the target regions was observed for siRNA targeting the

Huntingtin gene16 nor in mouse oocytes constitutively expressing long dsRNA targeting the Mos

gene.17 Similarly, Ting et al could not detect any DNA methylation after transfection of siRNA

targeting the CDH1 promoter, although the siRNA e!ectively suppressed CDH1 transcription.18

Interestingly, they did observe the presence of H3K9me2 in the targeted area, a typical marker of

inactive promoters.

Based on the above observations, it is clear that under certain conditions which are still being

de�ned19-21, some siRNAs can trigger DNA methylation or histone modi�cations and induce

gene silencing. In this study, we monitored the long-term e!ect of downregulation caused by

two approaches: a single delivery of a genome wide acting DNA methyltransferase and a single

transfection with exon speci�c siRNA, for up till 3 months. Cells treated only once with siRNA

targeting exon 5 of the EpCAM gene showed a transient strong EpCAM repression in all cells.

Interestingly, a subpopulation of cells obtained by sorting for weak EpCAM expressing cells, showed

a long-term EpCAM repression. This siRNA induced permanent mode of silencing was associated

with an elevated DNA methylation level of part of the EpCAM gene.

108

Chapter 7

7


EpCAM-siRNA and M.SssI delivery

The SKOV3 (HTB-77) cell line was purchased from ATCC (Manassas, VA) and cultured according

to ATCC recommendations. The delivery agent SAINT-2:DOPE (SD; 0.75 mM) was purchased

from Synvolux Therapeutics Inc. (Groningen, The Netherlands) and was shown to introduce

proteins into the cell with an equal delivery e�ciency as for siRNA or DNA.22 EpCAM-siRNA sense

(5’-GGAGAUCACAACGCGUUAUUU-3’ and antisense 5’-AUAACGCGUUGUGAUCUCCUU-3’ was

purchasedfromQiagen(BeneluxB.V.,Venlo,TheNetherlands).TheCG-speci�cDNAmethyltransferase

M.SssI23 was kindly provided by Prof. A Kiss from the Hungarian Academy of Sciences and Prof. E

Weinhold from the RWTH Aachen University. SKOV3 cells were seeded 12.5 x 104/well in a 6-well

plate, siRNA or M.SssI delivery was performed at 50-80% con!uency. One µg EpCAM-siRNA or 5 µg

MSssI in 100 µl HBS or PBS, respectively was complexed with 20 µl SD in 100 µl of HBS/PBS, within

15 minutes the complex was diluted in 800 μl serum free culture medium and added directly to the

cells. After 3 h of incubation at 37°C, 5% CO2 serum containing medium was added.

EpCAM expression and cell sorting

EpCAM detection was performed with mouse Mab MOC31 hybridoma supernatant followed by

R M-F(ab)2-FITC (DAKO, Glostrup, Denmark). The Mean Fluorescence Intensity (MFI) was measured

on a FACS Calibur !ow cytometer (Beckton Dickenson Biosciences, San Jose, CA). At day 6 (sort-I)

and day 34 (sort-II) after siRNA and M.SssI delivery, weak EpCAM expressing cells were sorted by

!ow cytometry (MoFlo XDP Cell Sorter, Beckman Coulter, Woerden, The Netherlands) and counting

was reset to passage 1 (p1) after each sorting. Untreated and treated SKOV3 cells were split 2-3

times a week to maintain similar con!uences. EpCAM expression was measured at day 26, 51, 62, 72

and 78 after treatment. The untreated SKOV3 cells were used as control cells for EpCAM expression,

which was set at 100%. At day 78, EpCAM weak siRNA-sort-II p12 and M.SssI-sort-II p11 were stored

in the liquid N2. After thawing and culturing up to siRNA-sort-II p26 and M.SssI-sort-II p25, EpCAM

expression was measured. In a separate experiment, untreated SKOV3 cells were sorted for weak

EpCAM expressing cells. After 31 days of culturing the weak EpCAM expressing sort-I population, a

second sort was performed (EpCAM weak sort-II). EpCAM expression of the EpCAM weak-sort-I and

sort-II cell populations was measured 19, 48, 59 and 73 days after the �rst sorting and compared to

the non-sorted SKOV3 cells, which were cultured in parallel.


Genomic DNA was isolated and treated with sodium bisul�te to convert unmethylated cytosines

to uracils. The EZ DNA Methylation-Gold Kit (Zymo, Baseclear Lab Products, Leiden, The

Netherlands) was used to modify 1 μg of DNA. Bisul�te speci�c primers void of any CpGs were

used in order to obtain ampli�cation products unbiased for the methylation status. Two regions


were selected to cover a 700 bp region of the EpCAM gene. Primer sequences for region A were

5’-GGAGGGGAGTTTATTTATTTTT-3’ (forward) and 5’-CACAACTCTACTCCAATC-3’ (reverse) and for

region B 5’-AGTGTTTTGGAAGGTTTTTTGT-3’ (forward), 5’-AAATTAAAAAAATAAATAAACTCCC-3’

(reverse). PCR conditions: 95°C for 15 min, followed by 35 cycles of 95°C for 45 s, 53-56°C for 45 s, 72°C

for 45 s and �nished with 72°C for 10 min. PCR fragments were puri�ed using the DNA Extraction Kit

(Qiagen) and cloned into pCR 2.1-TOPO TA vector (Invitrogen, Breda, The Netherlands). Following

transformation, plasmids from individual bacterial colonies were isolated using the Qiaprep Spin

Miniprep Kit (Qiagen) and subjected to restriction analysis. Correct clones were sequenced.

RESULTS

Sustained downregulation of EpCAM expression by siRNA or DNA methyltransferase

EpCAM positive SKOV3 cells were treated once with siRNA targeting exon 5 of the EpCAM gene or

with the DNA methyltransferase M.SssI. At day 6, EpCAM expression of the total siRNA treated cell

population was down to 14% of the expression of the control cells, while the M.SssI treated cells

showed 70% remaining EpCAM expression (Figure 1A). These treated cell populations were sorted

to obtain the weak EpCAM expressing cells (Figure 1B). The number of weak EpCAM expressing cells

7

109

Figure 1. EpCAM expression of control, siRNA- and M.SssI-treated cell populations. A) Histograms showingthe EpCAM expression in relation to the amount of cells, 6 days after siRNA or M.SssI treatment measured by !owcytometry. The percentage above the histograms indicate the EpCAM expression of the total cell populationrelative to the Mean Fluorescence Intensity of the total untreated SKOV3 control cell population, which was setat 100%. B) Dot plots of the cell populations as shown in A. The EpCAM weak sort-I populations sorted out of thesiRNA- and M.SssI-treated cell populations are boxed and indicated with R2 and R3, respectively. R2 = 65% andR3 = 23% of the total cell populations.

B

R3 = sort-I = 23%

siRNA M.SssI

Control: 100% siRNA: 14% of control M.SssI: 70% of controlA

B

R2 = sort-I = 65%

siRNAControlB

R3 = sort-I = 23%

siRNA M.SssI


B

R2 = sort-I = 65%

siRNAControlB

R3 = sort-I = 23%R3 = sort-I = 23%

siRNA M.SssI


B

R2 = sort-I = 65%

siRNAControl

Chapter 7

obtained for siRNA-sort-I was 2.5*106 and for M.SssI-sort-I 1*105 (65% and 23% of the cells before

sorting, respectively). After 20 days of subculturing, EpCAM expression of the siRNA-sort-I cells was

restored to 88%, whereas the EpCAM expression of the M.SssI-sort-I cells was still down to 16% of

the expression of the control cells (Figure 2A).

At day 34, a second sort was performed on the siRNA- and the M.SssI-sort-I cell populations,

resulting in 11% and 60% of the sort-I populations, respectively (Figure 2B and C). The EpCAM

expression of the sort-II cell populations remained stable for at least another 44 days, at about 50%

of the control cells for the siRNA-sort-II and about 20% for the M.SssI-sort-II cell population (Figure

3). Also after freezing, thawing and culturing the cells for another 14 passages, EpCAM expression

7

110

Figure 2. EpCAM expression of the siRNA-sort-I and M.SssI-sort-I cell populations. A) EpCAM expressionof the total sort-I populations 26 days after treatment with siRNA or M.SssI, measured by !ow cytometry (MFI =Mean Fluorescence Intensity). The percentage above the bars indicates the remaining EpCAM expression relativeto the SKOV3 control cells. B) Histograms showing the EpCAM expression in relation to the amount of cells, 34days after treatment. C) Dot plots of the cell populations as shown in B. The EpCAM weak sort-II populationssorted out of the siRNA- and M.SssI-sort-I cell populations are boxed and indicated with R2. R2 is 11% for thesiRNA sort-II and 60% for the M.SssI-sort-II of the total sort-I cell populations, at day 34 after treatment.

M.SssI sort-IsiRNA sort-I

R2 = sort-II = 60%

M.SssI sort-I

R2 = sort-II = 11%

siRNA sort-I

A

control siRNA sort-I M.SssI sort-I0

50

100

150 100%88%

16%

EpC

AMex

pres

sion

MFI

B

C

M.SssI sort-IsiRNA sort-I

R2 = sort-II = 60%R2 = sort-II = 60%R2 = sort-II = 60%

M.SssI sort-I

R2 = sort-II = 11%R2 = sort-II = 11%R2 = sort-II = 11%R2 = sort-II = 11%

siRNA sort-I

A


50

100

150 100%88%

16%

EpC

AMex

pres

sion

MFI


50

100

150 100%88%

16%

EpC

AMex

pres

sion

MFI

B

C

111

7


Figure 3. Sustained downregulation of EpCAM expression after siRNA and M.SssI treatment. EpCAMexpression of the siRNA- and M.SssI treated SKOV3 cells relative to the expression of the SKOV3 control cells,which was set at 100%. At day 0, SKOV3 cells were treated with siRNA targeting exon 5 or with M.SssI. Cellswere sorted twice for the EpCAM weak subpopulation, at day 6 (Sort-I) and day 34 (Sort-II) after treatment,indicated with an arrow. Cells were cultured and EpCAM expression was measured at the days indicated. After78 days, the cell populations were frozen, thawed and recultured up to passage 26 (p26). The initial reductionin EpCAM expression of the total cell population by siRNA treatment is restored at day 26 for the siRNA-sort-Icell population. After a second sort for the weak EpCAM expressing cells, EpCAM expression of the siRNA-sort-II population persisted low for at least another 44 days. The EpCAM expression of the M.SssI-sort-I and -II cellpopulations remained low compared to the expression of the SKOV3 control cells.

of both sort-II populations remained constant (Figure 3). Assuming equal rates of cell division,

7% (0.65*0.11) and 14% (0.23*0.60) of the siRNA and M.SssI treated cell populations, respectively

showed a permanent reduced EpCAM expression.

Reduced EpCAM expression is associated with elevated DNA methylation

To examine if the long-term downregulation of EpCAM expression of the sort-II populations was

associated with induced DNA methylation of the EpCAM gene, bisul!te sequencing was performed.

The methylation level of the CpGs present in region A indicated in Figure 4 was 31% for the siRNA-

sort-II cell population and 67% for the M.SssI-sort-II cell population compared to 1-2% generally

observed for the SKOV3 cell line. The di"erence in methylation level between untreated SKOV3

cells, siRNA-sort-II and M.SssI-sort-II is re#ected by the di"erence in EpCAM expression as shown

in Figure 3. In general, SKOV3 cells display a small percentage weak EpCAM expressing cells (also

shown in Figure 1B). To exclude selection of cells with a ”spontaneous” high DNA methylation level,

untreated SKOV3 cells were also subjected to two cycles of sorting and subculturing. Nineteen days

of culturing after the !rst sorting, EpCAM expression of the EpCAM weak sort-I cell population was

completely restored (Figure 5A, sort-I p6). On day 31, a second sort for weak EpCAM expressing

0

20

40

60

80

100

120

6 26 34 51 62 72 78 p26

days after treatment

Rel

ativ

eE

pCA

Mex

pres

sion control

siRNAM.SssI

0

20

40

60

80

100

120

6 26 34 51 62 72 78 p26

days after treatment

Rel

ativ

eE

pCA

Mex

pres

sion control

siRNAM.SssI

Figure 4. DNA methylation level of part of the EpCAM promoter and exon 1. Bisul�te sequencing of regionA and B of the EpCAM gene. CpGs are depicted by vertical bars, the transcription start site (TSS) corresponds toposition +1, and the ATG site are indicated. The % of DNA methylation of each region was calculated by addingthe methylated CpGs divided by the number of CpGs present in the region (Region A: 61, Region B: 18). For eachcell population the number of clones analyzed is indicated between brackets.

Chapter 7

112

7

cells on the sort-I population was performed. After 28 days of culturing, EpCAM expression of the

EpCAM weak sort-II cell population was 66% of the non-sorted SKOV3 cells (Figure 5A, sort-II p8).

After culturing for another 14 days, EpCAM expression was restored to 83% of the non-sorted SKOV3

cells (Figure 5A, sort-II p12). The weakest EpCAM expressing passage of the EpCAM weak sort-II

population (sort-II p5, Figure 5) was subjected to bisul�te sequencing. The DNA methylation level of

Figure 5. EpCAM expression and DNA methylation status of the untreated non-sorted SKOV3 cellscompared to the untreated EpCAM weak sort-II SKOV3 cells. A) EpCAM expression of the untreated EpCAMweak sort-I and sort-II populations relatively to the untreated non-sorted SKOV3 cells, which was set at 100%.Expression was measured by !ow cytometry after culturing for 19 (sort-I p6), 48 (sort-II p5), 59 (sort-II p8) and 73(sort-II p12) days after the �rst sorting. B) The % of DNA methylation of region A and B of the non-sorted SKOV3cells (p163 and p174 are di"erent passages) and the EpCAM weak sort-II passage 5, is indicated. The % of DNAmethylation of each region was calculated by adding the methylated CpGs divided by the number of CpGspresent in the region (Region A: 61, Region B: 18). Of each cell population 10 clones per region were analyzed.

Region A

Region BDNA methylation

1% 1%

4% 2%

1%

6%

non-sorted p163 non-sorted p174 sort-II p5

0

20

40

60

80

100

120

sort I p6 sort II p5 sort II p8 sort II p12

untreated sorted SKOV3 cells

Rela

tive

EpCA

Mex

pres

sion

A

B

-830

Region A

Region BDNA methylation

2% (6) 31% (18)

6% (19) 11% (10)

67% (19)

17% (14)

SKOV3 siRNA-sort-II M.SssI-sort-II

+282

Region ARegion B

-

Region A

Region B

2% (6) 31% (18)

6% (19) 11% (10)

67% (19)

17% (14)

SKOV3 - - - -

ATGTSS

113

7


the CpGs present in region A indicated in Figure 5B was 1% for EpCAM weak sort-II cells and 1% for

the non-sorted SKOV3 cells grown in parallel.

DISCUSSION

In this study, we show that siRNA designed to target mRNA molecules is able to induce sustained

silencing in a subset of cells, which correlate with an elevated DNA methylation level of part of

the gene. In our experimental set up, one siRNA treatment was followed by two cycles of sorting

and subculturing of weak EpCAM expressing cells. The resulting subpopulation showed a sustained

reduced level of EpCAM expression for up to 44 days. The DNA methylation level of the EpCAM gene

in this cell population was higher compared with the level of untreated sorted cells. Two cycles of

sorting and subculturing of untreated cells did not result in a subset of cells with a higher DNA

methylation level, excluding selection of cells with a ”sponteneous” high DNA methylation level.

The elevated DNA methylation level was most pronounced in the area around the transcriptional

start site, suggesting that the methylation status of the CpGs in this area is of more importance for

EpCAM gene regulation than the analyzed region more upstream.8

Previously, we have demonstrated that siRNA- and MSssI-mediated downregulation of EpCAM

resulted in 20% and 60% remaining EpCAM expression, respectively.8 In the current study, siRNA-

and MSssI-mediated downregulation of EpCAM resulted in 14% and 70% remaining EpCAM

expression. Interestingly, siRNA induced downregulation of EpCAM expression in all cells, whereas

M.SssI induced downregulation of EpCAM expression in a subpopulation of cells. To enrich for

epigenetically silenced cells, weak EpCAM expressing cells were sorted and sub-cultured, twice.

Calculated from the percentages of sorted cells per cycle, the !nal population with sustained EpCAM

repression represent 7% and 14% of the siRNA and M.SssI treated cell populations, respectively.

Within these cells the repression was associated with DNA methylation. Since this percentage of

cells with a permanent reduced EpCAM expression due to DNA methylation is so small, this was only

detectable after two cycles of sorting and subculturing.

Speculating on our !nding, it is tempting to hypothesize that siRNA delivery into cells

mainly results in cytoplasmic localisation targeting only mRNA molecules, and that only a small

percentage is able to reach the nucleus for direct e"ect on DNA level. Indeed, comparison of

delivery agents revealed that delivery of siRNA directed to EF1 promoter induced only silencing

by DNA methylation if the delivery agent was capable of delivery into the nucleus.14 A similar study

con!rmed that nuclear-speci!c delivery is required for histone methylation.21 We have previously

shown that the delivery agent SAINT-2:DOPE is indeed capable of cargo delivery into the nucleus22

explaining the relatively high e#ciency of 10%.

To our knowledge, only one brief communication in Nature reported about long-term gene

silencing by siRNA targeting a coding region.24 C. elegans worms were fed on bacteria expressing

114

Chapter 7

7

dsRNA targeting the transgene GFP. For non-GFP expressing progeny, all siblings showed reduced

GFP expression for over 80 generations. Culturing worms fed with trichostatin A relieved silencing,

indicating that RNAi-induced phenotypes might be induced by histone modi�cations.

The fact that some studies do and others don’t observe epigenetic silencing via RNAi indicates

that the target DNA sequence and its accessibility is of great importance. Turunen et al tested four

shRNAs targeting di!erent areas of the VEGF-A promoter, of which one caused a major decrease and

one caused an increase in VEGF-A production.25 The di!erence in expression levels was associated

with repressive and activating histone modi�cations, respectively. The other two shRNAs did not

show e!ects on VEGF-A production. Similarly, of three shRNAs targeting the RASSF1A gene, only

one was able to induce a low DNA methylation level.26 Moreover, Turunen et al showed that the

epigenetic changes within the VEGF-A promoter were cell speci�c, suggesting that siRNA activity is

dependent on cell type.

We are the �rst to investigate sustained gene silencing in correlation with DNA methylation

in a small subpopulation for such a long period of time as reported here. It might be that in some

studies DNA methylation could not be detected simply to the fact that DNA methylation did not

yet occur in the majority of the cells. Recruitment of the components of the silencing complex and

methylation of histones might precede DNA methylation. Targeting the human ubiquitin C (UbC)

with promoter-associated small RNAs, DNA methylation was observed after histone methylation

of the targeted locus had already taken place.27 Moreover, 7 days of sustained shRNA exposure was

required to establish suppression of UbC for one month. Remarkable, sustained shRNA exposure

correlated with an increase in DNA methylation and reduced expression levels. This long exposure

time is in agreement with our �nding that only in a small percentage of cells, siRNAs appears to

reach the nucleus. This long exposure time could also indicate that shRNAs entering the nucleus

requires a period of time to excert their function. Taken together, our results suggest that upon

delivery of siRNA targeting a coding region, sustained gene silencing can be achieved without

the repeated administration of the siRNA. Since transient downregulation of EpCAM reduces its

oncogenic potential, active silencing of EpCAM expression via DNA methylation can provide a novel

approach in anti-cancer treatment.

Acknowledgements

We thank Prof A Kiss from the Institute of Biochemistry, Biological Research Center of the Hungarian

Academy of Sciences, Szeged, Hungary and Prof. E Weinhold from the Institute of Organic Chemistry,

RWTH Aachen University, Germany for providing the wild type DNA methyltransferase M.SssI.

115

7


References





5. Wenqi D, Li W, Shanshan C, Bei C, Yafei Z, Feihu B et al. EpCAM is overexpressed in gastric cancer and itsdownregulation suppresses proliferation of gastric cancer 6. J Cancer Res Clin Oncol 2009;135:1277-85.


7. Tai KY, Shiah SG, Shieh YS, Kao YR, Chi CY, Huang E et al. DNA methylation and histone modi!cation regulatesilencing of epithelial cell adhesion molecule for tumor invasion and progression. Oncogene 2007;26:3989-97.



10. Hawkins PG, Morris KV. RNA and transcriptional modulation of gene expression. Cell Cycle 2008;7:602-7.

11. de Fougerolles A., Vornlocher HP, Maraganore J, Lieberman J. Interfering with disease: a progress report onsiRNA-based therapeutics. Nat Rev.Drug Discov. 2007;6:443-53.

12. Verdel A, Vavasseur A, Le GM, Touat-Todeschini L. Common themes in siRNA-mediated epigenetic silencingpathways. Int.J Dev.Biol 2009;53:245-57.

13. Pulukuri SM, Rao JS. Small interfering RNA directed reversal of urokinase plasminogen activatordemethylation inhibits prostate tumor growth and metastasis. Cancer Res 2007;67:6637-46.

14. Morris KV, Chan SW, Jacobsen SE, Looney DJ. Small interfering RNA-induced transcriptional gene silencingin human cells. Science 2004;305:1289-92.

15. Kawasaki H, Taira K. Induction of DNA methylation and gene silencing by short interfering RNAs in humancells. Nature 2004;431:211-7.

16. Park CW, Chen Z, Kren BT, Steer CJ. Double-stranded siRNA targeted to the huntingtin gene does notinduce DNA methylation. Biochemical and Biophysical Research Communications 2004;323:275-80.

17. Svoboda P, Stein P, Filipowicz W, Schultz RM. Lack of homologous sequence-speci!c DNA methylation inresponse to stable dsRNA expression in mouse oocytes. Nucleic Acids Res 2004;32:3601-6.

18. Ting AH, Schuebel KE, Herman JG, Baylin SB. Short double-stranded RNA induces transcriptional genesilencing in human cancer cells in the absence of DNA methylation. Nat Genet. 2005;37:906-10.

19. Kawasaki H, Taira K, Morris KV. siRNA induced transcriptional gene silencing in mammalian cells. Cell Cycle2005;4:442-8.

20. Kim DH, Rossi JJ. Strategies for silencing human disease using RNA interference. Nat Rev.Genet. 2007;8:173-84.

21. Weinberg MS, Villeneuve LM, Ehsani A, Amarzguioui M, Aagaard L, Chen ZX et al. The antisense strand ofsmall interfering RNAs directs histone methylation and transcriptional gene silencing in human cells. RNA.

116

Chapter 7

7

2006;12:256-62.

22. van der Gun BT, Monami A, Laarmann S, Rasko T, Slaska-Kiss K, Weinhold E et al. Serum insensitive,intranuclear protein delivery by the multipurpose cationic lipid SAINT-2. J.Control Release 2007;123:228-38.

23. Darii MV, Cherepanova NA, Subach OM, Kirsanova OV, Rasko T, Slaska-Kiss K et al. Mutational analysis of theCG recognizing DNA methyltransferase SssI: Insight into enzyme-DNA interactions. Biochim.Biophys.Acta2009;1794:1654-62.

24. Vastenhouw NL, Brunschwig K, Okihara KL, Muller F, Tijsterman M, Plasterk RHA. Gene expression: Long-term gene silencing by RNAi. Nature 2006;442:882.

25. Turunen MP, Lehtola T, Heinonen SE, Assefa GS, Korpisalo P, Girnary R et al. E!cient Regulation of VEGFExpression by Promoter-Targeted Lentiviral shRNAs Based on Epigenetic Mechanism: A Novel Example ofEpigenetherapy. Circulation Research 2009;105:604-9.

26. Castanotto D, Tommasi S, Li M, Li H, Yanow S, Pfeifer GP et al. Short hairpin RNA-directed cytosine (CpG)methylation of the RASSF1A gene promoter in HeLa cells. Mol Ther 2005;12:179-83.

27. Hawkins PG, Santoso S, Adams C, Anest V, Morris KV. Promoter targeted small RNAs induce long-termtranscriptional gene silencing in human cells. Nucleic Acids Res 2009;37:2984-95.

117

Summary & General discussion and perspectives

Chapter 8

118

Chapter 8

8

EpCAM is highly overexpressed on most carcinoma types and serves as a marker for diagnosis

and as an immunotarget for clinical trials.1 For some carcinoma types, EpCAM overexpression has

been associated with poor clinical outcome, suggesting that downregulation of EpCAM expression

provides a promising approach to interfere with oncogenic potential of the tumor cells. The studies

described in this thesis aim to selectively target and downregulate EpCAM by epigenetic editing.

Epigenetic editing provides a novel approach to overwrite molecular epigenetic marks by an

epigenetic e�ector domain targeted to speci�c genes by a sequence speci�c DNA-binding motif.

Towards this end, we characterized the (epi)genetic regulation of the gene that codes for the EpCAM

protein and designed a novel approach to downregulate EpCAM expression in a permanent way.

Biological role of EpCAM in cancer

In Chapter 2, we summarize current literature regarding the (epi)genetic regulation of the EpCAM

gene itself, and we review the biological role of EpCAM in carcinogenesis, tumor progression and

metastasis in a broad range of carcinoma types. The role of EpCAM in development of cancer and

tumor progression appears to be paradoxical. For example, in breast cancer high EpCAM expression

correlates with poor prognosis2 and downregulation of EpCAM has been shown to decrease the

oncogenic potential.3 In contrast, high EpCAM expression in primary renal cell carcinomas is

associated with improved patient survival.4;5 In other types of carcinoma like ovarian cancer, the

role of EpCAM is not clear and contradictory results have been reported. In one study, FIGO stage III/

IV showed lower EpCAM expression than stage I6, while in another study, FIGO stage III/IV showed

higher EpCAM expression than stage I/II disease.7 The latter study suggests that a higher expression

of EpCAM correlates with tumor progression, although no correlation with survival was found.7

However, a more recent study reported that EpCAM overexpression was signi�cantly related to a

decreased overall survival of patients with epithelial ovarian cancer.8 Importantly, metastatic and

recurrent tumors were found to express signi�cantly higher levels of EpCAM protein when compared

with primary ovarian carcinomas.9 Despite the seemingly contradictory results, these observations

suggest rather a promoting than a protecting role for EpCAM in ovarian cancer.

To reveal more insights in this apparently paradoxical role of EpCAM in ovarian cancer, we e�ectively

downregulated EpCAM expression of ovarian cancer cell lines by siRNA and performed migration

assays. EpCAM siRNA treatment resulted in almost 90% decrease in EpCAM expression compared

with irrelevant siRNA (Figure 1, left). EpCAM siRNA treatment resulted in a reduced migration

potential compared to cells treated with irrelevant siRNA as shown in Figure 1 (right). However,

this reduced migration e�ect was not consistently shown in independent migrations assays, nor in

di�erent ovarian cancer cell lines.

Similarly, proliferation or scratch assays showed no signi�cant di�erence in oncogenic potential

between ovarian cancer cell lines treated with EpCAM siRNA or irrelevant siRNA. To validate our

119

8


assays, we included a breast cancer cell line for which others had shown a decrease in invasion and

migration potential after siRNA-mediated downregulation of EpCAM.3 However, despite an e�ective

downregulation of EpCAM expression, we could not con�rm their results. Further studies need to be

conducted (identifying suitable irrelevant siRNA; optimizing read out systems) before a conclusion

can be drawn on the role of EpCAM in ovarian cancer. So far, for �ve di�erent carcinoma cell lines

(head and neck10, gastric11, hepatocellular12, tongue squamous cell carcinoma13 and breast cancer3)

transient downregulation of EpCAM expression reduced the oncogenic potential, indicating a

powerful role of EpCAM at least in some carcinoma types. For ovarian cancer, the frequent observed

EpCAM overexpression might be the e�ect of dysregulated transcription factors. We therefore

investigated in Chapter 3 whether transcription factors described to play a potential role in ovarian

cancer, are associated with the EpCAM gene in living cells.

Transcription factors in ovarian cancer

It has been suggested that the ß-catenin/TCF/LEF pathway might be an important factor in the

development of ovarian cancer.14 Nuclear localization of β-catenin in the high-grade serous

carcinomas was shown to be signi�cantly higher than in the low-grade carcinoma group15,

indicating that one of the mechanisms for carcinogenesis in high-grade serous epithelial ovarian

cancer might be through the activation of the LEF/β-catenin pathway.14 Activity of ß-catenin/

TCF complex is essential for the transcription of genes that direct proliferation of tumor cells. In

a hepatocellular carcinoma cell line it has been shown that EpCAM is induced upon activation of

β-catenin.16 Moreover, upon proteolytic cleavage of EpCAM’s extracellular domain, the intracellular

part of EpCAM forms a nuclear complex containing LEF-1/β-catenin which upregulates the oncogen

c-myc and cyclin A and E.17 In turn, the extracellular domain of EpCAM can function as a ligand

in EpCAM signaling. In EpCAM positive ovarian cancer cell lines, we indeed found association of

Figure 1. Migration assay performed with OVCAR3 cells after e!ectively siRNA mediated downregulationof EpCAM expression. Cells were transfected with EpCAM speci�c siRNA or irrelevant siRNA. After 72h, partof the cells was harvested for EpCAM expression by "ow cytometry (left). The rest of the cells were used formigration assessed in AP48 Microchemotaxis Chamber in which a FCS gradient served as a chemoattractant(right).

blank

EpCAM siRNA

irr siRNA0

200

400

600

800Ep

CA

Mex

pres

sion

blank

EpCAM siRNA

irr siRNA0

100

200

300

400

abs.

num

ber

mig

rate

dce

lls

120

Chapter 8

8

LEF-1 with the EpCAM promoter, indicating that LEF-1 might play a potential role in the EpCAM

overexpression observed in ovarian cancer.

Other examples of transcription factors playing an important role in ovarian cancer are the

proliferation promoting E2F2 and the inhibiting E2F4 transcription factors. High mRNA levels of

E2F2 and low levels of E2F4 compared to levels in normal tissue are signi�cantly associated with a

poor survival.18 Interestingly, a low E2F2 to E2F4 ratio appears to be the most powerful prognostic

marker for disease free-survival.18 In the EpCAM positive ovarian cancer cell lines, we found both

E2F2 and E2F4 to be associated with the EpCAM promoter. In view of this, it is of great interest to

investigate the in�uence of these transcription factors on EpCAM expression. Especially, since it has

been shown that methylation of CpGs of some promoters within the binding motif of E2F2 and E2F4

e�ectively blocks the binding of the transcription factors.19 Since we indeed found only association

of these transcription factors with the hypomethylated EpCAM promoter, we propose a direct role

for E2F2 and E2F4 in EpCAM regulation.

Epigenetic marks and EpCAM expression

The accessibility of transcription factors to the DNA is dictated by the chromatin structure which is

dependent on among others DNA methylation and histone modi�cations. Since EpCAM is highly

overexpressed in all subtypes of ovarian cancer, loss of DNA methylation might also be one of the

underlying mechanisms in this carcinoma type. For patient samples of lung adenocarcinoma20,

colon21 and oral squamous cell carcinoma22, high EpCAM expression has been shown to correlate

with a hypomethylated EpCAM promoter. In Chapter 3 we show that in a panel of ovarian cancer

cell lines EpCAM expression indeed correlated inversely with DNA methylation. Interestingly, gel

retarding assays with a nuclear extract and a probe containing a putative binding site for Sp1,

showed inhibition of binding when the CpG within this binding site is methylated. Moreover, the

CpG located in this putative binding site was methylated in EpCAM negative ovarian cancer cell

lines and never methylated in EpCAM positive cells. We con�rmed that association of Sp1 with the

endogenous EpCAM promoter was restricted to EpCAM positive cells. Since it has been shown that

the presence of Sp1 increases promoter activity20, it is plausible that methylation of the promoter

impairs the activation by Sp1. This �nding is of great importance in view of our aim to downregulate

EpCAM gene expression, since it gives an indication to which location in the promoter one should

target to achieve e!cient downregulation of EpCAM expression.

In addition, we found general active histone modi�cations to be associated with an active

promoter, whereas a silent promoter was associated with repressive marks. Interestingly, recently

it has been reported that during di�erentiation of human embryonic stem cells EpCAM expression

is not silenced by DNA methylation but by reduction of active histone marks and an enhancement

of repressive marks.23 However, in our panel of ovarian cancer cell lines we found both epigenetic

marks DNA hypermethylation as well as repressive histone marks to be associated with no EpCAM

expression. The combination of DNA hypermethylation of the promoter in EpCAM negative cells

121

8


with the association of repressive histone marks may account for the fact that we did not detect any

transcription factors to be associated with the EpCAM promoter. Hypermethylated DNA associated

with trimethylated lysines 9 or 27 of histone 3, induces a closed chromatin structure which hinders

the accessibility for transcription factors.

Delivery of DNA methyltransferase

Modulation of EpCAM expression on epigenetic level implies that epigenetic modi!ers i.e. DNA

methyltransferase and/or histone modi!ers have to be delivered into the cells. Although DNA

and siRNA can be e"ciently delivered into cells, no agent was available to deliver proteins in the

presence of serum. The advantage of protein delivery over DNA delivery is that the former does not

require transcription by the host cell into a biologically active protein. In this way, the dose of protein

to be delivered is easier to control. In addition, to achieve targeting of epigenetic modi!ers to the

EpCAM gene, the enzyme needs to be coupled to a sequence speci!c synthetic DNA-binding motif.

In Chapter 4 we show that the cationic liposome SAINT-2:DOPE is an excellent protein delivery agent.

Labeling studies demonstrated e"cient delivery for protein (called profection) as well as for DNA

and siRNA. Delivered proteins were still able to exert their function as shown by β-galactosidase

activity and the general applicability was shown by the delivery of this enzyme into adherent or

non-adherent cell lines, as well as into di"cult to transfect primary cells. Of great importance for in

vivo delivery, profection with SAINT-2:DOPE was not signi!cantly a$ected by the presence of serum.

Since epigenetic modi!ers need to enter the nucleus to exert their function, we delivered the

DNA methyltransferase M.SssI and measured its e"ciency in DNA methylation. The E-cadherin

gene was used as a model gene, as the expression of this gene is known to be responsive to DNA

methylation silencing.24 Nuclear activity of delivered M.SssI was established by the observed elevated

methylation status of the E-cadherin gene and con!rmed by reduced E-cadherin expression.

Because M.SssI e"ciently methylates CpGs and acts genome wide, the enzyme is toxic to the

cells. Therefore, M.SssI delivery speci!c to the tumors cells would be an alternative approach to

eliminate tumor cells. Because of its high expression on a broad range of tumor types, EpCAM can be

used as a target antigen to deliver M.SssI speci!c to the tumor cells. The advantage of this approach

compared to currently ongoing immunotherapeutic clinical trials is that once the M.SssI is delivered

into the tumor cells, the cells will be directly killed instead of being dependent on the recruitment

of immune cells to induce elimination of the tumor cells. To this extend, an antibody speci!c for

EpCAM can be coupled to the delivery agent SAINT-2:DOPE which can direct M.SssI speci!c to the

tumor cells. Indeed, an anti-EpCAM Fab’-fragment coupled to liposomes showed more e"cient

delivery of the conjugated enzyme to tumor cells than the corresponding anti-EpCAM-enzyme

conjugate.25 Also for SAINT-2:DOPE, conjugation of an anti-E-selectin antibody e"ciently increased

the siRNA uptake speci!c into activated endothelial cells expressing E-selectin compared to resting

cells, which are E-selectin negative.26 Although, EpCAM is also expressed on healthy epithelia, the

relative inaccessibility of epithelia for antibodies as compared to the good tumor accessibility due

122

Chapter 8

8

to the leaky vasculature of tumors, allows the therapeutic compound to preferentially localize in the

tumor. This idea was con�rmed in a study in transgenic mice in which EpCAM on tumor cells was

much more accessible to antibodies than EpCAM expressed in normal tissues.27

Epigenetic downregulation of EpCAM expression

In Chapter 5 we explored whether active interfering with the DNA methylation status of the

promoter indeed resulted in changes in EpCAM gene expression. Treatment of EpCAM negative cells

with a DNA methylation inhibitor induced de novo EpCAM expression, both on mRNA and protein

level, and caused upregulation of EpCAM expression in a moderately EpCAM expressing ovarian

cancer cell line. Upon delivery of M.SssI, an elevated DNA methylation level of the EpCAM promoter

was observed, which correlated with an e!ciently reduced EpCAM expression. While siRNA-

mediated downregulation remained for 4 days, after which EpCAM re-expression increased in time,

M.SssI-mediated downregulation remained through successive cell divisions as the reduced EpCAM

expression persisted for at least 17 days, illustrating the transferable e"ect of epigenetic modulation.

Compared to siRNA-mediated downregulation, the M.SssI induced EpCAM reduction was initially

less e!cient, but delivery of M.SssI is dose limiting because of toxicity. Although the elevated DNA

methylation of the promoter suggests a direct e"ect of M.SssI on EpCAM expression, the genome

wide methylation by M.SssI will silence many genes. Hence, the reduced EpCAM expression could

be the indirect result of for example silencing of endogenous miRNA181s. Inhibition of endogenous

miRNA-181s has been shown to reduce EpCAM mRNA levels28 and epigenetic control of miRNA

expression has been reported.29 To demonstrate that DNA methylation of the EpCAM promoter

directly a"ects the EpCAM expression, and to increase the DNA methylation e!ciency of the EpCAM

promoter, targeting of the enzyme to the EpCAM promoter is required.

EpCAM speci!c downregulation by DNA methylation

In Chapter 6, we are the �rst to demonstrate that the conjugation of a TFO with a DNA

methyltransferase is able to target methylation predominantly to the target CpG without

any background methylation. The advantage of using a mutant methyltransferase with less

methyltransferase activity is that the targeting is dominated by the TFO and not by the enzyme.

Treatment with the TFO-C141S conjugate caused relaxation of the plasmid, most likely due to

topoisomerase activity reported for M.SssI.30 However, the observed topoisomerase activity was only

displayed by the TFO-C141S conjugate and not by the TFO or C141S only. Somehow, the chemical

coupling of the TFO with the enzyme seems to change the conformation of the enzyme, thereby

uncovering the catalytic topoisomerase domain. The enzyme activity should be easy to abolish by

replacing an amino acid in the active site of the isomerase domain. Since, M.SssI contains a tyrosine

in a similar amino acid context to the catalytic tyrosine of other characterized topoisomerases,

tyrosine 137 might be the most promising candidate.30 Once this is successful, the mutated DNA

methyltransferase C141S is generally applicable, because in this variant the internal cysteines are

123

8


replaced and the introduced C-terminal Cys allows coupling of the methyltransferase to any TFO.

However, the topoisomerase activity displayed by our TFO-C141S may cause no problem in

living cells, because the topoisomerase will reconnect the DNA again. Transfection of the TFO-C141S

with the delivery agent SAINT-2:DOPE in EpCAM positive ovarian cancer cells showed in our hands,

no e�ect on EpCAM expression. To improve the biological e�ect of targeted methylation by TFO-

C141S on endogenous EpCAM gene expression, additional considerations should be taken into

account. Under physiologic conditions, the TFO needs to be nuclease resistant and the pyrimidine-

rich TFO used in our study requires protonation of cytosines at N3 to form proper Hoogsteen

bonds.31 Progress in TFO-technology allows solving of both limitations by chemical modi�cations.

For example, substitution of cytosine with 5-methyl-2’-deoxycytidine increases protonation as

5-methyl-C has a higher pK than does cytosine and resistance to 3’-end nucleases can be increased

by addition of a propanediol tail. TFOs have been used to target cleavage, cross-linking or anticancer

agents to various genes resulting in inhibition of gene expression31, illustrating that modi�ed TFOs

are suitable tools as targeting domains in living cells.

Transfection of the TFO-C141S treated reporter plasmid in the presence or absence of a methyl

donor showed no e�ect on EpCAM promoter activity caused by methylation of a single CpG. The

absence of silencing could be explained by 1) methylation of one single CpG is not enough to

induce silencing of the promoter or 2) the targeted CpG is located in a region which is not important

for regulation. If the targeted CpG is located in a binding site for a transcription factor important

for activation of EpCAM gene transcription, and this binding is DNA methylation sensitive like we

found for Sp1, methylation of this particular CpG might have a direct e�ect on gene expression.

Alternatively, methylation of just a single CpG may recruit repressor proteins like heterochromatin

protein 1, histone methyltransferases, deacetylases and DNA methyltransferase DNMT1, DNMT3a

and 3b and induce DNA methylation spreading. Transfection of a reporter plasmid under control of

the p53 promoter in which a single CpG was methylated in vitro by HhaI, showed 85% downregulation

of the reporter gene.32 Restriction analysis with methylation sensitive enzymes performed on

the plasmids 48h after transfection, revealed indeed methylation of additional CpGs. These data

demonstrate that a single CpG could trigger subsequent spreading of methylation to other CpG

sites. In addition, the location of the targeted CpG seems to be of great importance in epigenetic

modulation, because methylation of a single CpG at another location by FnuDII methylase in the

p53 promoter was not associated with promoter suppression and methylation spreading.32

Apart from the number of CpGs that has to be methylated to induce gene silencing, the acquired

insights regarding the DNA methylation status of the EpCAM promoter in correlation with EpCAM

expression suggest that one should target to another region in the EpCAM promoter than the target

CpG of our TFO-C141S. Since the DNA methylation level in the area upstream of the transcription

starting site (-443 to -130) showed an increase correlating with a decrease in EpCAM expression

(Chapter 3) this might be the most promising region to initiate e�ective downregulation of EpCAM

expression. Moreover, the Sp1 site shown to be sensitive for DNA methylation is also located in this

124

Chapter 8

8

region. However, this region is CG rich which makes it more di�cult to design a TFO with a high

a�nity. An alternative targeting moiety might be deduced from zinc �ngers. Trimeric and hexameric

zinc �ngers, designed to target the region (-171 to -130) and fused to a repressor or an activation

domain, have been shown to modulate EpCAM promoter activity.33 DNA methyltransferases

genetically fused to zinc �nger proteins have been shown to e�ciently repress reporter gene

expression.34 Interestingly, mouse DNMT3a or DNMT3b DNA methyltransferases fused to DNA

binding domains induced dense methylation of DNA regions comprising up to 380 bp on both sites

of the speci�c DNA binding site.35 This �nding suggests that initial methylation with mammalian

DNA methyltransferases might serve as trigger for DNA methylation spreading as described above.

Moreover, a mutant DNA methyltransferase fused to zinc �nger proteins has been shown to induce

targeting methylation leading to gene silencing via initiation of a repressive chromatin signature at

the targeted genomic locus.36

An alternative way to induce targeted methylation is via short interference (si)RNA targeting

promoters.37 In Chapter 7 we report on a siRNA designed to target mRNA molecules inducing

sustained silencing in a subset of cells, which correlated with an elevated DNA methylation level

of the promoter. This observation was unexpected as the aim of this study was to demonstrate the

advantages of the permanent silencing via inheritable DNA methylation compared to the siRNA-

mediated transient silencing. As mRNA molecules are constantly produced, RNA-based approaches

require repeated administration of the inactivating reagent. In contrast, a single administration

of a DNA methyltransferase is expected to be su�cient to silence, because the maintenance DNA

methyltransferases in the cell will copy the new methylation mark in the absence of the exogenous

methyltransferase. In addition, targeted DNA methylation needs to a!ect just two copies of the

EpCAM gene rather than the numerous copies of mRNA present in each cell. Since M.SssI induced

downregulation of EpCAM expression in a subpopulation of cells, we sorted and cultured this

subpopulation to demonstrate that the downregulation of EpCAM expression was indeed enduring

and correlated with an elevated methylation level of the EpCAM promoter. In this experimental

setting the siRNA treated cells functioned as a control. Unexpectedly, after two rounds of selection

for weak EpCAM expressing cells, we found a small percentage of the initially siRNA treated cells

with a permanently reduced EpCAM expression also correlating with an elevated DNA methylation

level. Speculating on this interesting �nding, it appears that in a small percentage of cells the siRNA

is transfected into the nucleus by SAINT-2:DOPE, hereafter spreading of DNA methylation causes the

permanent downregulation. To proof the occurrence of methylation spreading, DNA methylation

analysis of the siRNA targeting region should be analyzed �rst, followed by the possible recruitment

of repressive histone modi�cations. To exclude the selection of cells with a ”spontaneous” high

DNA methylation, two rounds of selection and culturing of weak EpCAM expressing cells out of

the untreated cell line, did not result in a subpopulation of cells with a higher DNA methylation

level than unsorted cells. However, to validate our �nding, the appropriate control is to repeat the

125

8


experiment and include a single transfection with irrelevant siRNA.

Future perspectives

E!ective and speci"c delivery of the EpCAM gene speci"c epigenetic modi"er in vivo to minimize

possible side e!ects is the most challenging remaining hurdle. Combination of a tumor targeted

delivery system with a gene speci"c epigenetic modi"er is expected to increase the speci"city.

This double targeted system consists of an EpCAM speci"c antibody coupled to the delivery agent

SAINT-2:DOPE which can direct the EpCAM speci"c TFO-C141S or siRNA speci"c to the tumor cells

(Figure 2).

Figure 2. Proposed model of double targeting to tumor cells by using the EpCAM protein on the surfaceof the tumor cell to deliver the EpCAM gene speci!c DNA methyltransferase.

TFODNAMethyltransferase

Triple helix Forming Sitein EpCAM promoter

=

=

=

=

=

TFO-M.SssI-(C141S)

EpCAM target antigen

Anti-EpCAM antibody

SAINT molecules

Carcinoma cell with nucleus

An alternative approach to prevent methylation of nontargeted sites is the so called split DNA

methyltransferase strategy as proposed by Kiss and Weinhold.38 This approach is based on splitting

naturally monomeric methyltransferases into two fragments and fusing the fragments to di!erent

DNA bindingdomains like zinc "nger proteins that bind DNA #anking the target site for methylation.

When both fusion products are expressed in the same cell an active methyltransferase is formed

that can only methylate the target site. Delivery of this double targeting device by EpCAM-SAINT-

2:DOPE is expected to silence the overexpressed EpCAM gene only in the EpCAM positive tumor

cells. Because of such triple targeting, possible negative side e!ects are expected to be minimized.

126

Chapter 8

8

In conclusion, the work presented in this thesis revealed more insights in the regulation of the

EpCAM gene. We explored three novel approaches to downregulate EpCAM expression in a

permanent way via DNA methylation. Firstly, nuclear protein delivery of M.SssI resulted in an

elevated DNA methylation level of the EpCAM promoter inducing a persistent downregulation of

EpCAM expression. However, untargeted M.SssI can methylate the whole genome and is therefore

toxic. Secondly, we report on a siRNA designed to target EpCAM mRNA molecules which induced

a sustained downregulation of EpCAM expression in a subpopulation of cells, correlating with

an elevated DNA methylation of the EpCAM promoter. Although this new �nding might expand

the range of potential clinical siRNA applications, this might not be an easy applicable strategy.

Targeted DNA methylation by our TFO-C141S conjugate provides a �exible tool: by the use of

TFOs as targeting domain, and the wide applicability of the mutated DNA methyltransferase, this

approach appears to be a promising tool in both research and therapeutic areas.

References

1. Baeuerle PA, Gires O. EpCAM (CD326) �nding its role in cancer. Br.J Cancer 2007;96:417-23.

2. Spizzo G, Went P, Dirnhofer S, Obrist P, Simon R, Spichtin H et al. High Ep-CAM expression is associatedwith poor prognosis in node-positive breast cancer. Breast Cancer Res.Treat. 2004;86:207-13.


4. Klatte T, Pantuck AJ, Said JW, Seligson DB, Rao NP, LaRochelle JC et al. Cytogenetic and Molecular TumorPro�ling for Type 1 and Type 2 Papillary Renal Cell Carcinoma. Clinical Cancer Research 2009;15:1162-9.

5. Seligson DB, Pantuck AJ, Liu X, Huang Y, Horvath S, Bui MH et al. Epithelial cell adhesion molecule (KSA)expression: pathobiology and its role as an independent predictor of survival in renal cell carcinoma. ClinCancer Res. 2004;10:2659-69.

6. Kim JH, Herlyn D, Wong KK, Park DC, Schorge JO, Lu KH et al. Identi�cation of epithelial cell adhesionmolecule autoantibody in patients with ovarian cancer. Clinical Cancer Research 2003;9:4782-91.

7. Heinzelmann-Schwarz VA, Gardiner-Garden M, Henshall SM, Scurry J, Scolyer RA, Davies MJ et al.Overexpression of the Cell Adhesion Molecules DDR1, Claudin 3, and Ep-CAM in Metaplastic OvarianEpithelium and Ovarian Cancer. Clinical Cancer Research 2004;10:4427-36.

8. Spizzo G, Went P, Dirnhofer S, Obrist P, Moch H, Baeuerle PA et al. Overexpression of epithelial celladhesion molecule (Ep-CAM) is an independent prognostic marker for reduced survival of patients withepithelial ovarian cancer. Gynecol.Oncol. 2006;103:483-8.

9. Bellone S, Siegel ER, Cocco E, Cargnelutti M, Silasi DA, Azodi M et al. Overexpression of epithelial celladhesion molecule in primary, metastatic, and recurrent/chemotherapy-resistant epithelial ovariancancer: implications for epithelial cell adhesion molecule-speci�c immunotherapy. Int.J Gynecol.Cancer2009;19:860-6.

10. Munz M, Kieu C, Mack B, Schmitt B, Zeidler R, Gires O. The carcinoma-associated antigen EpCAMupregulates c-myc and induces cell proliferation. Oncogene 2004;23:5748-58.


12. Yamashita T, Ji J, Budhu A, Forgues M, Yang W, Wang HY et al. EpCAM-Positive Hepatocellular Carcinoma

127

8


Cells Are Tumor-Initiating Cells With Stem/Progenitor Cell Features. Gastroenterology 2009;136:1012-24.


14. Gatcli"e TA, Monk BJ, Planutis K, Holcombe RF. Wnt signaling in ovarian tumorigenesis. Int.J Gynecol.Cancer 2008;18:954-62.

15. Lee CM, Shvartsman H, Deavers MT, Wang SC, Xia W, Schmandt R et al. [beta]-catenin nuclear localizationis associated with grade in ovarian serous carcinoma. Gynecologic Oncology 2003;88:363-8.

16. Yamashita T, Budhu A, Forgues M, Wang XW. Activation of Hepatic Stem Cell Marker EpCAM by Wnt{beta}-Catenin Signaling in Hepatocellular Carcinoma. Cancer Research 2007;67:10831-9.

17. Maetzel D, Denzel S, Mack B, Canis M, Went P, Benk M et al. Nuclear signaling by tumor-associatedantigen EpCAM. Nat Cell Biol 2009;11:162-71.

18. Reimer D, Sadr S, Wiedemair A, Stadlmann S, Concin N, Hofstetter G et al. Clinical relevance of E2Ffamily members in ovarian cancer--an evaluation in a training set of 77 patients. Clinical Cancer Research2007;13:144-51.

19. Campanero MR, Armstrong MI, Flemington EK. CpG methylation as a mechanism for the regulation ofE2F activity. Proc Natl Acad Sci U.S.A 2000;97:6481-6.

20. Tai KY, Shiah SG, Shieh YS, Kao YR, Chi CY, Huang E et al. DNA methylation and histone modi!cationregulate silencing of epithelial cell adhesion molecule for tumor invasion and progression. Oncogene2007.



23. Lu TY, Lu RM, Liao MY, Yu J, Chung CH, Kao CF et al. Epithelial cell adhesion molecule regulation isassociated with the maintenance of the undi"erentiated phenotype of human embryonic stem cells1. J Biol.Chem. 2010;285:8719-32.

24. Chen Q, Lipkina G, Song Q, Kramer R. Promoter methylation regulates cadherin switching in squamouscell carcinoma14. Biochemical and Biophysical Research Communications 2004;315:850-6.

25. Fonseca MJ, Jagtenberg JC, Haisma HJ, Storm G. Liposome-mediated targeting of enzymes to cancercells for site-speci!c activation of prodrugs: comparison with the corresponding antibody-enzymeconjugate. Pharm.Res 2003;20:423-8.

26. Asgeirsdottir SA, Talman EG, de G, I, Kamps JA, Satchell SC, Mathieson PW et al. Targeted transfectionincreases siRNA uptake and gene silencing of primary endothelial cells in vitro--a quantitative study. JControl Release 2010;141:241-51.

27. McLaughlin PM, Harmsen MC, Dokter WH, Kroesen BJ, van der MH, Brinker MG et al. The epithelialglycoprotein 2 (EGP-2) promoter-driven epithelial-speci!c expression of EGP-2 in transgenic mice: a newmodel to study carcinoma-directed immunotherapy. Cancer Research 2001;61:4105-11.

28. Ji J, Yamashita T, Budhu A, Forgues M, Jia HL, Li C et al. Identi!cation of microRNA-181 by genome-widescreening as a critical player in EpCAM-positive hepatic cancer stem cells. Hepatology 2009;50:472-80.

29. Valeri N, Vannini I, Fanini F, Calore F, Adair B, Fabbri M. Epigenetics, miRNAs, and human cancer: a newchapter in human gene regulation. Mamm.Genome 2009;20:573-80.

30. Matsuo K, Silke J, Gramatiko" K, Scha"ner W. The CpG-speci!c methylase SssI has topoisomerase activityin the presence of Mg2+. Nucleic Acids Res 1994;22:5354-9.

31. Duca M, Vekho" P, Oussedik K, Halby L, Arimondo PB. The triple helix: 50 years later, the outcome. Nucleic

128

Chapter 8

8

Acids Research 2008;36:5123-38.

32. Pogribny IP, Pogribna M, Christman JK, James SJ. Single-Site Methylation within the p53 Promoter RegionReduces Gene Expression in a Reporter Gene Construct: Possible in Vivo Relevance during Tumorigenesis.Cancer Research 2000;60:588-94.


34. Jeltsch A, Jurkowska RZ, Jurkowski TP, Liebert K, Rathert P, Schlickenrieder M. Application of DNAmethyltransferases in targeted DNA methylation. Appl Microbiol.Biotechnol. 2007;75:1233-40.

35. Li F, Papworth M, Minczuk M, Rohde C, Zhang Y, Ragozin S et al. Chimeric DNA methyltransferases targetDNA methylation to speci!c DNA sequences and repress expression of target genes. Nucleic Acids Res2007;35:100-12.

36. Smith AE, Hurd PJ, Bannister AJ, Kouzarides T, Ford KG. Heritable Gene Repression through the Action of aDirected DNA Methyltransferase at a Chromosomal Locus. Journal of Biological Chemistry 2008;283:9878-85.

37. Morris KV. siRNA-mediated transcriptional gene silencing: the potential mechanism and a possible role inthe histone code24. Cell Mol.Life Sci 2005;62:3057-66.

38. Kiss A, Weinhold E. Functional reassembly of split enzymes on-site: a novel approach for highlysequence-speci!c targeted DNA methylation. Chembiochem. 2008;9:351-3.

129

Nederlandse samenvatting

130

Samenvatting

In de ontwikkelde landen krijgt een op de drie mensen kanker. Ondanks betere screening en

behandelingsmogelijkheden is kanker in ontwikkelde landen nog steeds de meest voorkomende

doodsoorzaak na hart- en vaatziekten. Bij kanker is er sprake van ongecontroleerde celgroei,

waardoor cellen blijven delen en uitgroeien tot een tumor. De tumor drukt weefsel opzij of kan

binnendringen in een orgaan en daar schade toebrengen. Cellen kunnen losraken van een tumor

en zich via het bloed of lymfevaten verspreiden naar andere organen. Als ze daar uitgroeien tot

tumoren is er sprake van uitzaaiingen (metastasen).

Kanker kan ondermeer ontstaan door mutaties in een bepaald gen. Een gen bestaat uit een

stukje DNA met een speci�eke volgorde dat codeert voor een functioneel RNA-product en/of een

bepaald eiwit. Eerst wordt het DNA van een gen overgeschreven naar RNA, dit wordt transcriptie

genoemd (Figuur 1). Bij het produceren van een eiwit wordt RNA vertaald (translatie) naar een

eiwit. Tot voor kort dacht men dat alléén de genen die voor een eiwit coderen verantwoordelijk

zijn voor allerlei biologische processen in ons lichaam. Het blijkt echter dat RNA-producten die

niet worden vertaald naar een eiwit, ook een belangrijke rol spelen bij genexpressie regulatie. Het

aanzetten van het gen om een RNA-product te maken, wordt mede gereguleerd door binding van

bepaalde eiwitten (bv. transcriptiefactoren) aan een stuk DNA dat direct voor het gen gelegen is, de

zogeheten promoter (Figuur 1). Wanneer er mutaties in de DNA-volgorde hebben plaatsgevonden

kan dit leiden tot een niet functioneel RNA-product of eiwit, of er wordt te veel of te weinig, of in zijn

geheel geen product aangemaakt.

Figuur 1. Het proces van genexpressie. In de kern bevindt zich het DNA dat codeert voor een RNA-productdat vervolgens vertaald kan worden naar een eiwit. De DNA-volgorde van het gen wordt voorafgegaan door depromoter, waaraan regulerende eiwitten kunnen binden die de transcriptie in gang zetten. Tevens bestaan erRNA-produkten die niet naar een eiwit worden vertaald, maar direct het transcriptieproces kunnen beïnvloeden.

DNAstart overschrijving

genpromoter

RNA

transcriptie

translatie

eiwit

regulerende eiwitten

131

Samenvatting

Epigenetica

Het laatste decennium is steeds duidelijker geworden dat naast genetische mutaties, epigenetische

mutaties eveneens een belangrijke rol spelen bij het ontstaan van kanker. Epigenetica is de studie van

overdraagbare veranderingen in genexpressie die niet verklaard kunnen worden door veranderingen

in de DNA-volgorde zelf. Epigenetische factoren die genexpressie reguleren zijn o.a. DNA-methylatie,

histonmodi�caties, RNA-produkten en de dichtheid van de nucleosomen. Het onderzoek beschreven

in dit proefschrift concentreert zich op DNA-methylatie en histonmodi�caties. Het DNA is opgerold

om verpakkingseiwitten (histonen) die samen met andere eiwitten het chromatine vormen waar

de chromosomen uit bestaan (Figuur 2). Factoren die de chromatinestructuur bepalen zijn nauw

verbonden met genexpressie. Wanneer het chromatine ”open” staat kunnen er allerlei regulerende

eiwitten, waaronder transcriptiefactoren, aan de promoter binden waardoor het gen tot expressie

komt. Wanneer het chromatine ”gesloten” is, is de promoter als het ware afgeschermd voor deze

eiwitten, en staat het gen uit. Factoren die de chromatinestructuur bepalen zijn o.a. DNA-methylatie

en de zogeheten histoncode (Figuur 2). Binding van een methylgroep aan een bouwsteen van DNA

(cytosine voorafgaand door de bouwsteen guanine) veroorzaakt een gesloten chromatinestructuur.

Aan histonstaarten kunnen op verschillende plaatsen chemische groepen (methyl, acetyl, fosfaat,

ubiquitine) gekoppeld worden. Dit patroon vormt een histoncode die bepaalt of het chromatine

een open of gesloten structuur aanneemt. Acetylering van histonen is bijvoorbeeld geassocieerd

met een open chromatinestructuur.

Figuur 2.Twee belangrijke componenten van de epigenetische code. Chromosomen bestaan uit chromatine:DNA opgerold rondom histonen. Zowel het methyleren van het DNA als de verschillende moleculen die aan destaarten van de histonen worden gekoppeld bepalen mede of er transcriptie van het DNA plaatsvindt.

chromosoom

histonstaart

DNA methylatie

Qiu, Nature 2006

nucleosoom

132

Samenvatting

Een verstoring van de normale genexpressie door epigenetische mutaties speelt een belangrijke rol

bij kanker. Veranderingen in de methylatiestatus van het DNA en/of in modi�caties van de histonen

kunnen leiden tot een veranderde genexpressie. Het belangrijkste verschil met genetische mutaties

is dat epigenetische mutaties omkeerbaar zijn en dus mogelijk met medicijnen zijn te herstellen.

Epigenetische medicijnen zoals remmers van DNA-methylatie en histon-modi�cerende eiwitten

worden al in de kliniek gebruikt. Ondanks dat de resultaten veelbelovend zijn, is een groot nadeel

dat deze medicijnen werkzaam zijn op het DNA van het gehele genoom. Dit betekent dat zij ook

e�ect hebben op de genen met een correcte expressie. Om te bereiken dat alléén een defecte

genexpressie wordt gecorrigeerd dient het epigenetische medicijn doelgericht naar dat speci�eke

”foute” gen gedirigeerd te worden.

EpCAM en kanker

Een gen dat codeert voor een eiwit dat geassocieerd is met kanker is het Epitheliale Cel Adhesie

Molecuul (EpCAM). Dit eiwit is oorspronkelijk geïdenti�ceerd als een marker voor carcinomen, toe te

schrijven aan de hoge expressie op snel woekerende tumoren van epitheliale oorsprong. Bijna alle

normale epitheelcellen brengen ook EpCAM tot expressie, maar veel minder dan carcinoomcellen.

Aanvankelijk werd EpCAM voorgesteld als een celadhesie molecuul. Recente inzichten laten echter

een meer veelzijdige rol voor EpCAM zien, die niet uitsluitend is beperkt tot celadhesie, maar

diverse processen omvat zoals celmigratie, proliferatie, di�erentiatie en mogelijk kankerinitiatie.

Voor sommige carcinoomtypen zoals borstkanker is gebleken dat remming in de aanmaak van het

EpCAM-eiwit tot minder proliferatie, migratie en invasie van tumorcellen leidt. Tot op heden kan

men de eiwitaanmaak echter alleen maar tijdelijk verlagen.

Doel van het onderzoek

Het onderzoek in dit proefschrift beschreven richt zich op een blijvende verlaagde aanmaak van

het EpCAM-eiwit om tumorgroei en metastasering te voorkomen. Het blijkt dat de overexpressie

van EpCAM op carcinomen niet wordt veroorzaakt door onderliggende genetische defecten. In

dit proefschrift wordt de epigenetische regulatie van het gen dat codeert voor het EpCAM-eiwit

onderzocht. Vervolgens wordt door te interfereren met de epigenetische code van het EpCAM-

gen beoogd de expressie van het EpCAM-gen uit te schakelen. Door een gerichte verandering

aan te brengen in de epigenetische code van alléén het EpCAM-gen, wordt het gen uitgeschakeld

waardoor er geen EpCAM-eiwit meer wordt aangemaakt. De nadruk ligt hierbij op genspeci�citeit.

In hoofdstuk 1 wordt het eiwit EpCAM geïntroduceerd en kort uitgelegd wat epigenetica inhoudt.

Tevens worden verschillende methoden om het EpCAM-gen uit te schakelen uitgelegd. Hoofdstuk

2 geeft een gedetailleerd overzicht van de schijnbaar tegenstrijdige biologische rol van EpCAM

in het ontstaan van kanker, het voortschrijden van de tumor en de metastasering in een breed

spectrum aan carcinoomtypen. In dit hoofdstuk wordt tevens de (epi)genetische regulatie van het

133

Samenvatting

gen dat codeert voor EpCAM beschreven en worden de mogelijkheden om hierop in te grijpen

besproken.

Tot op heden wordt de overexpressie van EpCAM op carcinomen gebruikt als doelwit voor

klinische studies waarin getest wordt of antilichamen gericht tegen het EpCAM-eiwit in staat

zijn om de tumorcellen te doden met behulp van de afweercellen in het lichaam. In sommige

tumortypen laat remming in de aanmaak van EpCAM-eiwit een verminderd oncogeen fenotype

zien. Dit opent de mogelijkheid om EpCAM op gen-niveau uit te schakelen zodat de tumorcellen

minder snel delen, migreren en invaseren. Bovendien is aangetoond dat bepaalde kankercellen in

de tumor, de zogeheten kankerstamcellen, sneller een tumor initiëren wanneer deze cellen EpCAM

tot expressie brengen dan kankerstamcellen die géén EpCAM tot expressie brengen. Ingrijpen in de

epigenetische regulatie van het EpCAM-gen, waardoor tumorcellen minder EpCAM-eiwit maken

kan mogelijk leiden tot een nieuwe therapie tegen kanker.

In hoofdstuk 3 is de epigenetische regulatie van het EpCAM-gen speci!ek in ovariumkanker

onderzocht. Het bleek dat in cellijnen van ovariumcarcinomen, hypermethylatie van de EpCAM-

promoter correleert met de afwezigheid van EpCAM-expressie, en omgekeerd, dat hypomethylatie

een hoge EpCAM-expressie laat zien. Tevens bleken de histonmodi!caties die kenmerkend zijn voor

een open chromatinestructuur geassocieerd te zijn met een EpCAM-promoter die aanstaat, terwijl

de histonmodi!caties karakteristiek voor een gesloten chromatinestructuur correleerden met een

EpCAM-promoter die uitstaat. Bovendien is onderzocht of transcriptiefactoren die een bewezen rol

spelen bij ovariumkanker, mogelijk betrokken zijn bij de regulatie van het EpCAM-gen. In cellijnen

die EpCAM tot expressie brengen bleken tien van de zestien geteste transcriptiefactoren aan de

EpCAM promoter te binden. De transcriptiefactor Sp1 bleek op een speci!eke plaats in de promoter

moeilijker te kunnen binden indien het DNA hier gemethyleerd was.

Het koppelen van een methylgroep aan de cytosine in het DNA wordt uitgevoerd door het

enzym DNA-methyltransferase. Om het EpCAM-gen te kunnen uitschakelen dienen we dit enzym

in de tumorcel te kunnen a"everen. In hoofdstuk 4 laten we zien dat het cationische liposoom

SAINT in staat is eiwitten af te leveren in de cel. De in verschillende type cellen gebrachte eiwitten

waren functioneel actief. Bovendien bleek SAINT als enige transportmiddel in staat om functioneel

actieve eiwitten in de cel af te leveren in de aanwezigheid van serum dat een essentiële voorwaarde

is voor een toekomstige toepassing in de kliniek. Om te veri!ëren of het DNA-methyltransferase in

de kern van de cel kan komen en functioneel is, hebben we gebruik gemaakt van het E-cadherin-

gen waarvan bekend is dat wanneer het wordt gemethyleerd, het niet meer tot expressie komt.

Nucleaire activiteit van het door SAINT afgeleverde DNA-methyltransferase werd bevestigd door

een verhoogde DNA-methylatie van het E-cadherin-gen en een verlaagde E-cadherin eiwitexpressie.

De resultaten beschreven in hoofdstuk 5 laten zien dat actief interfereren met de DNA

methylatiestatus van de EpCAM-promoter inderdaad resulteert in een verandering van

EpCAM-expressie. Behandeling van EpCAM-negatieve cellen met een DNA methylatieremmer

induceerde EpCAM-expressie en veroorzaakte verhoogde EpCAM-expressie in een EpCAM-

134

Samenvatting

positieve ovariumcarcinoom cellijn. A�everen van het DNA-methyltransferase door SAINT liet een

toegenomen DNA-methylatie van de EpCAM-promoter zien, wat correleerde met een e�ciënte

verlaagde EpCAM-expressie.

Eiwitexpressie kan ook worden geremd door zogenaamde RNA-interferentie. Door het a�everen

van korte stukken dubbelstrengs RNA (siRNA) speci�ek voor het EpCAM-RNA, wordt alléén het RNA

afgebroken dat codeert voor het EpCAM-eiwit. Het nadeel van deze methode is dat er voortdurend

siRNA in de cel afgeleverd moet worden omdat de cel EpCAM-RNA moleculen blijft aanmaken. Om

te illustreren dat DNA-methylatie in tegenstelling tot RNA-interferentie een blijvend e�ect heeft

zijn beide methoden met elkaar vergeleken. Terwijl de verlaagde EpCAM-expressie na a�evering

van EpCAM-speci�ek RNA vier dagen aanbleef en daarna weer toenam, bleef de door het DNA-

methyltransferase verlaagde EpCAM-expressie voor maar liefst zeventien dagen constant. Deze

bevinding illustreert dat DNA-methylatie wordt doorgegeven aan de dochtercellen.

Het inbrengen van het DNA-methyltransferase in de cel heeft als gevolg dat het DNA van het

gehele genoom wordt gemethyleerd. Om genspeci�ek te methyleren maken we in hoofdstuk 6

gebruik van een DNA-bindend domein dat ontwikkeld is om slechts aan één positie in het genoom

te binden. Dit DNA-bindend domein, het zogenoemde Triple helix vormend Oligonucleotide (TFO),

is een enkelstrengs stuk DNA dat speci�ek bindt aan het dubbelstrengs DNA in de EpCAM-promoter.

Aan de TFO is een DNA-methyltransferase gekoppeld: de TFO bindt aan de EpCAM promoter, alwaar

alléén op die positie in het genoom DNA-methylatie plaatsvindt.

Hoofdstuk 7 beschrijft dat siRNA, ontworpen om EpCAM-RNA af te breken zoals genoemd in

hoofdstuk 5, ook in staat is een blijvende verlaging van EpCAM-expressie te bewerkstelligen in

een klein deel van de celpopulatie die behandeld is met dit siRNA. De verlaagde EpCAM expressie

correleerde met een toename in DNA-methylatie van de EpCAM- promoter.

Tenslotte wordt in hoofdstuk 8 het onderzoek zoals beschreven in dit proefschrift bediscussieerd

en worden de perspectieven voor de verdere ontwikkeling van EpCAM-gen speci�eke therapieën

besproken.

Samenvattend heeft dit onderzoek naar de regulatie van het gen dat codeert voor het EpCAM-

eiwit, geleid tot drie verschillende methoden waarop de expressie van dit gen langdurig kan

worden verlaagd. Deze methoden zijn allen gebaseerd op methylatie van het DNA waardoor een

gen uitgeschakeld wordt met als gevolg dat er geen RNA, en dus geen eiwit, meer door de cel

wordt aangemaakt. Om een DNA-methylerend enzym in de cellen te brengen is onderzocht of het

transportmiddel SAINT functionele eiwitten in de cel(kern) kan a�everen. Het inbrengen van het

DNA-methylerend enzym in de cellen resulteerde in methylatie van het DNA en een verminderde

EpCAM-eiwit aanmaak. Om DNA-methylatie van andere genen te voorkomen, werd bij de tweede

methode het DNA-methylerende enzym gekoppeld aan een DNA-bindend domein, speci�ek voor

het EpCAM-gen. De geïnduceerde DNA-methylatie bleef inderdaad beperkt tot het gebied waar

het DNA-bindend domein bindt. Bij de derde methode werden korte stukken RNA ingebracht om

135

Samenvatting

speci�ek het EpCAM-RNA af te breken. In een subpopulatie van de cellen bleek echter ook het

EpCAM-DNA gemethyleerd dat resulteerde in een langdurig verlaagde aanmaak van het EpCAM-

eiwit. Genspeci�eke DNA-methylatie is breed toepasbaar: in principe kan op deze manier ieder

willekeurig ”ziek gen” uitgeschakeld worden.

137

Dankwoord

138

Dankwoord

DANKWOORD

Het proefschrift is klaar en ik kan alleen maar zeggen dat ik ontzettend blij ben dat ik hieraan

begonnen ben. Ik heb immens veel geleerd en geniet ontzettend van de verworven kennis, wat

overigens ook hongerig maakt naar meer. Het onderzoek beschreven in dit proefschrift is volbracht

door een fantastisch researchteam, dat ik hier graag ”chronologisch” wil bedanken.

Allereerst Pamela, mede dankzij jou heb ik de kans gekregen om als analist te promoveren. Ik

vind je een enthousiaste en zeer creatieve onderzoekster, van jou heb ik o.a. geleerd om met een

helikopterview te kijken i.p.v. in de details te blijven hangen. De IMEDGEN meetingen waren zeer

verrassend wanneer we (onafgesproken) allebei weer in een gelijkend out�t uit de hotelkamer

kwamen. Pam, voor mij ben en blijf je de EpCAM expert! Je hebt me de eerste beginselen van

het schrijven bijgebracht en Lou eraan herinnerd dat hij al héél lang geen ’eerste versie’ meer had

gelezen. Lou, dankjewel dat je me deze kans gaf, ik was dolgelukkig met jouw commentaar op

de ’eerste versie’: ”Ieneke, niet slecht” (Lou, je bent toch een Limburger, géén Groninger?). Toen je

als hoofd van de Medische Biologie wegging, was ik vereerd dat je mijn tweede promoter wilde

blijven. Marcel, naast Pamela heb jij mij ook gesteund om te kunnen promoveren. Dankjewel voor je

vertrouwen. Ook ben ik erg blij dat je aan mij hebt gedacht met het ProTuMA project en ik verheug

me al op de kick-of meeting met de Italianen.

En toen kwam…….Marianne. Marianne, je kwam bij ons toen Pamela naar Leeuwarden vertrok. Aan

de eerste alinea van dit dankwoord heb jij een heel groot deel bijgedragen, ik vind je zeer kritisch,

opbouwend, enthousiast, inventief, gestructureerd en een prima supervisor. Soms moest ik naar

aanleiding van jouw kritische commentaar even �ink slikken, maar het daagt ook vreselijk uit met

absoluut een beter resultaat. Daarom ben ik trots dat je mijn eerste promotor bent en ben ik blij dat

ik voorlopig nog in jouw groep kan blijven.

I also would like to thank the other members from the IMEDGEN consortium: Antal Kiss, Elmar

Weinhold, Amélie Monami, Maria Maluzynska-Ho�man, Reinhold Wasserkort, Krystyna Ślaska-Kiss

and Tamás Raskó for our fruitful meetings. Dear Antal, thanks a lot for all your scienti�c input, your

quick comments on our manuscripts and your willingness to evaluate my thesis. Hierbij wil ik ook

graag Prof. dr. Hollema en Prof. dr. H.J. Haisma bedanken voor de beoordeling van mijn proefschrift.

Alice, ik ga je vreselijk missen als je straks weggaat. Jouw kennis die tussen de regels in de protocollen

staat wordt door menig mens onderschat. Alice, ontzettend bedankt voor de ontelbare EMSA’s,

cloneringen, bisul�te sequences en nog veel meer. En niet te vergeten onze ko�e uurtjes en je

altijd luisterend oor. Ik ben apetrots dat je mijn paranimf wilt zijn en hoop dat, ook als je niet meer

dagelijks op het lab bent, onze vriendschap standhoudt. Inge, je kwam als studente o.a. de migratie

139

Dankwoord

assay opzetten, weet je nog dat prachtige kunstwerk met het �lter omgekeerd? Inmiddels ben je de

expert in de ChIP en menig andere techniek en steunt de hele groep op jouw vaardigheden. Inge,

bedankt voor al je inzet, hulp, humor en gezelligheid. I also would like to thank the students Anna

and Burcu for all transfections, ChIP and TFO experiments. Jelleke, jij hebt mij de kneepjes van het

celkweken bijgebracht. Momenteel werken we nauw samen wat ik zeer waardeer, bedankt voor die

megahoeveelheden cellijnen en de gezellige uurtjes achter de �ow.

Verder wil ik mijn kamergenoten bedanken voor de prima sfeer: Bram “onze man”, Betty ”onze

perfectionist”, Sabine our american German, Marieke tja komt nog, Marloes komt ook nog en de

”nieuwkomers”: Fahimeh en Cristian. Marieke, onze ko�e/thee momenten waren goud waard, ik

mis je echt, ik hoop dat we samen onze frustraties op de gewichten blijven afreageren. Marloes,

ons ”social beast”, dankzij jouw inbreng op het sociale vlak zien we elkaar ook eens in een andere

context, wat onze groep absoluut hechter maakt. Bovendien waardeer ik je zeer als directe collega.

Akshay, Epigenetic Editing member next door, I am sure future will bring us beautiful green target

cells.

Henriëtte, Annet en Susan heel veel dank voor jullie secretariële ondersteuning, weet dat ik dit zeer

waardeer. En dit geldt ook voor Anita en Linda, zonder jullie geen goedlopend draaiend lab. Linda,

ontzettend bedankt dat je me altijd hielp met mijn last minute acties, dankzij jou kwam menige

bestelling op tijd binnen of werd een urgent probleem opgelost. Marja, dank voor alle tips rond

cloneren, PCR-en en veel meer, en uiteraard voor de uitwisseling van ervaringen als ouder van een

enig kind. Onze FACS-operators, Geert, Henk en Roelof-Jan, heel veel dank voor al jullie hulp, ik hoop

er in de toekomst nog veel gebruik van te kunnen maken. Verder wil ik iedereen op het lab, te veel

om allen bij naam te noemen, ontzettend bedanken voor alle hulp maar vooral ook voor de goede

werksfeer.

De mensen rondom mij hebben mogen genieten van mijn “dipjes” maar ook zeker van de “toppen”.

Familie en vrienden dank voor jullie belangstelling en medeleven, altijd heerlijk om even van je af te

kunnen praten of je enthousiasme te kunnen delen. Maar nog meer dank voor het feit dat niemand

ooit moeilijk deed als ik social events liet schieten omdat er een deadline gehaald moest worden.

Heerlijk, om zo’n onvoorwaardelijke back-up achter je te weten. Gootje, dank dat je mijn paranimf

wilt zijn, ik ben ontzettend gelukkig met onze vriendschap, je bent mijn hartsvriendin! Mijn ouders

wil ik bedanken voor de vrijheid van het maken van mijn eigen ”foute” keuzes, eindelijk heb ik mijn

passie gevonden. Mam, ik weet dat we papa nu allebei missen.

Mijn allerliefsten: Erik en Emma. Lieve Emma, voor jou ben ik blijkbaar een open boek: “Proe�e soms

mislukt, mam”? “Je hebt zeker een goeie dag gehad, mam”? Ems, je bent een prachtdochter (wel

een makkie hé, als enige), je vrolijkheid werkt ontzettend aanstekelijk. Onze gesprekken tijdens de

140

afwas en vooral onze avondjes met een boek voor de kachel, geen woord gewisseld (was gezellig

mam) koester ik. Ems, ik heb me als moeder weleens schuldig gevoeld, bang dat ik je te weinig

aandacht gaf, maar je hebt me menigmaal verzekerd dat jij dat niet zo ervaart. Dankjewel Ems, ik

heb inderdaad veel plezier in mijn werk en ’ben juist daardoor natuurlijk zo’n leuke moeder’. Erik, we

zijn al 26 jaar samen, al veel lief en leed gedeeld, je bent een partner voor het leven. Dankjewel dat

je me gestimuleerd hebt om hieraan te beginnen, vooral jouw relativeringsvermogen helpt me de

dingen in het juiste perspectief te zien. Erik en Emma, jullie maken me gelukkig! DANK

Ieneke

141

Publications

142

Publications

PUBLICATIONS

van der Gun BTF, Maluszynska-Ho�man M, Kiss A, Arendzen AJ, Ruiters MHJ, McLaughlin PMJ,

Weinhold E, Rots MG. Targeted DNA methylation by a DNA methyltransferase coupled to a Triple

helix Forming Oligonucleotide to downregulate the Epithelial Cell Adhesion Molecule. Bioconjugate

Chem. 2010; in press.

Geel TM, Meiss G, van der Gun BTF, Kroesen BJ, de Leij LFMH, Zaremba M, Silanskas A, Kokkinidis

M, Pingoud A, Ruiters MHJ, McLaughlin PMJ, Rots MG. Endonucleases induced TRAIL-insensitive

apoptosis in ovarian carcinoma cells. Exp Cell Res. 2009 Sep 10;315(15):2487-95.

van der Gun BTF, Wasserkort R, Monami A, Jeltsch A, Raskó T, Slaska-Kiss K, Cortese R, Rots MG,

de Leij LFMH, Ruiters MHJ, Kiss A, Weinhold E, McLaughlin PMJ. Persistent downregulation of the

pancarcinoma-associated epithelial cell adhesion molecule via active intranuclear methylation. Int

J Cancer. 2008 Jul 15;123(2):484-9.

van der Gun BTF, Monami A, Laarmann S, Raskó T, Ślaska-Kiss K, Weinhold E, Wasserkort R, de Leij

LFMH, Ruiters MHJ, Kiss A, McLaughlin PMJ. Serum insensitive intranuclear protein delivery by the

multipurpose cationic lipid SAINT-2. J Control Release. 2007 Nov 20;123(3):228-38.

van der Gun BTF, Maluszynska M, McLaughlin PMJ, Gommans W, Arendzen AJ, Wasserkort R, Kiss A,

Ruiters MHJ, Weinhold E, Rots MG. Towards Sustained Gene-Speci#c Silencing of the Epithelial Cell

Adhesion Molecule. Hum Gene Ther. 2007 Oct;18(10):970-1

Koopmans J, de Haan A, Bruin E, van der Gun I, van Dijk H, Rozing J, de Leij L, Staal M. Porcine

fetal ventral mesencephalic cells are targets for primed xenoreactive human T cells. Cell Transplant.

2006;15(5):381-7.

Koopmans J, de Haan A, Bruin E, van der Gun I, van Dijk H, Rozing J, de Leij L, Staal M. Individual

human serum di�ers in the amount of antibodies with a$nity for pig fetal ventral mesencephalic

cells and the ability to lyse these cells by complement activation. Cell Transplant. 2004;13(6):631-7.

Pamela M.J. McLaughlin, Monika Trzpis, Ieneke T. F. van der Gun, Martin C. Harmsen, Marcel H.J.

Ruiters. Cationic Liposome-Based Delivery of DNA Modulating Enzymes to Speci#cally Regulate

Gene Expression in Carcinoma Derived Cell Lines. Mol Ther. 2004;9:S177-8.

143

Publications

De Haan A, van der Gun I, van der Bij W, de Leij LFMH, Prop J. Detection of alloreactive T cells

by !owcytometry: a new test compared with limiting dilution assay. Transplantation. 2002 Aug

27;74(4):562-70.

De Haan A, van den Berg AP, van der Bij W, Hepkema BG, Bruin-van Dijk E, van der Gun I, Lems SP,

Sloo" MJ, Haagsma EB, de Leij LFMH, Prop J. Rapid decreases in donor-speci#c cytotoxic T lymfocyte

precursor frequencies and graft outcome after liver and lung transplantation. Transplantation. 2001

Mar 27;71(6):785-91.

de Haan A, van der Gun I, van Dijk E, Hepkema BG, Prop J, de Leij LF. Activation of alloreactive

T cells by allogeneic nonprofessional antigen-presenting cells and interleukin-12 from bystander

autologous professional antigen-presenting cells. Transplantation. 2000 Apr 27;69(8):1637-44.

de Haan A, van der Gun I, Hepkema BG, de Boer WJ, van der Bij W, de Leij LF, Prop J. Decreased

donor-speci#c cytotoxic T cell precursor frequencies one year after clinical lung transplantation do

not re!ect transplantation tolerance: a comparison of lung transplant recipients with or without

bronchiolitis obliterans syndrome. Transplantation. 2000 Apr 15;69(7):1434-9.

Spronk PE, Horst G, van der Gun BTF, Limburg PC, Kallenberg CGM. Anti-dsDNA production

coincides with concurrent B and T cell activation during development of active disease in systemic

lupus erythematosus (SLE). Clin. Exp. Immunol. 1996 Jun;104(3):446-53.

Spronk PE, van der Gun BTF, Limburg PC, Kallenberg CGM. B cell activation in clinically quiescent

systemic lupus erythematosus (SLE) is related to immunoglobulin levels, but not to levels of anti-

dsDNA, nor to concurrent T cell activation. Clin. Exp. Immunology, 1993;93:39-44.

Grefte JMM, van der Gun BTF, Schmolke S, van der Giessen M, van Son WJ, Plachter B, Jahn G,

The TH. The lower matrix protein pp65 is the principal viral antigen present in peripheral blood

leukocytes during an active cytomegalovirus infection. J Gen Virol. 1992 Nov;73(Pt11):2923-32.

Grefte JMM, van der Gun BTF, Schmolke S, van der Giessen M, van Son WJ, Plachter B, Jahn G,

The TH. Cytomegalovirus Antigenemia Assay: Identi#cation of the Viral Antigen as the Lower Matrix

Protein PP65. J Infect Dis. 1992 Sep;166(3):683-4.

Serum insensitive, intranuclear protein delivery by the multipurpose cationic lipid S aint-2

Documents

Transcript of Serum insensitive, intranuclear protein delivery by the multipurpose cationic lipid S aint-2