rsif.royalsocietypublishing.org

ResearchCite this article: Henss A, Rohnke M, El

Khassawna T, Govindarajan P, Schlewitz G,

Heiss C, Janek J. 2013 Applicability of ToF-SIMS

for monitoring compositional changes in bone

in a long-term animal model. J R Soc Interface

10: 20130332.

http://dx.doi.org/10.1098/rsif.2013.0332

Received: 12 April 2013

Accepted: 24 June 2013

Subject Areas:chemical biology

Keywords:mass spectrometry, chemical imaging, bone

imaging, collagen, calcium distribution in bone

Author for correspondence:Marcus Rohnke

e-mail: marcus.rohnke@phys.chemie.uni-

giessen.de

& 2013 The Author(s) Published by the Royal Society. All rights reserved.

Applicability of ToF-SIMS for monitoringcompositional changes in bone in along-term animal model

Anja Henss1, Marcus Rohnke1, Thaqif El Khassawna2,Parameswari Govindarajan2, Gudrun Schlewitz2,3, Christian Heiss2,3

and Juergen Janek1

1Institute for Physical Chemistry, Justus-Liebig-University of Giessen, Heinrich-Buff-Ring 58, 35392 Giessen,Germany2Laboratory of Experimental Trauma Surgery, University of Giessen, Kerkrader Strasse 9, 35394 Giessen, Germany3Department of Trauma Surgery, University Hospital of Giessen-Marburg, Rudolf-Buchheim-Strasse 7, 35385Giessen, Germany

Time-of-flight secondary ion mass spectrometry (ToF-SIMS) is a well-

established technique in material sciences but has not yet been widely explored

for implementation in life sciences. Here, we demonstrate the applicability and

advantages of ToF-SIMS analysis for the study of minerals and biomolecules in

osseous tissue. The locally resolved analysis of fragment ions deriving from the

sample surface enables imaging and differentiation of bone tissue and facili-

tates histology on non-stained cross sections. In a rat model, bilateral

ovariectomy combined with either a multi-deficiency diet or steroid treatment

was carried out to create osteoporotic conditions. We focused our study on the

Ca content of the mineralized tissue and monitored its decline. Calcium mass

images of cross sections show the progressive degenerative changes in the

bone. We observed a decreased Ca concentration in the edge region of

the trabeculae and a decline in the Ca/P ratio. Additionally, we focused on

the non-mineralized matrix and identified fragment ions that are characteristic

for the collagen matrix. We observed trabeculae with wide ranges of non-

mineralized collagen for the diet group owing to an impaired mineralization

process. Here, the advantage of coeval monitoring of collagen and minerals

indicated an osteomalacic model rather than an osteoporotic one.

1. IntroductionFor analysing osseous tissue, a wide variety of in and ex vivo methods are well

established and unquestionable in their ability to assess bone quality. Standard

techniques such as dual-energy X-ray absorptiometry (DEXA) measurements,

quantitative computed tomography and X-ray micro-tomography (mCT) are

widely used to determine bone mineral density (BMD) [1–3]. BMD represents

a calculated mean value of bone mineral content in a given area (usually given

as area density in mg mm22). Clinically, a reduced BMD assessed by X-ray

absorptiometry is considered to be the first indicator of osteoporosis [1]. However,

the BMD does not account for the bone structure and composition in detail [2–4].

A more significant and accurate fingerprint value for the characterization of

bone quality is the bone mineral density distribution (BMDD), which was

introduced by Roschger and co-workers [5,6]. The BMDD is determined by quan-

titative backscattered electron imaging and is validated for clinical use [7]. The

macro- and micro-structure of bone can be determined non-destructively by

using mCT with a lateral resolution in vitro of 10 mm [2,3,8]. Microbeam X-ray dif-

fraction and the more recent synchrotron-based small and wide-angle X-ray

scattering (SAXS and WAXS) are well-established diffraction methods for the

structural and textural analysis of biological tissue such as bone with a spatial

resolution of 1 mm [9,10]. Supplementary information includes the elemental

analysis and mapping by X-ray fluorescence imaging (XRF) and m-XRF for

primary ion gun

primary ions

yx

secondary ions

mass/charge

mass/charge

inte

nsity

inte

nsity

local mass spectrumstage scan, vertebra

x,y raster

500 × 500 mm2, trabeculae

49.5 × 49.5 mm2, ROI

Ca+

C4H8N+

CaxPOy+

Bi3+

Figure 1. Principles and schematic of time-of-flight secondary ion mass spectrometry.

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

2

higher Z-elements. By means of focusing capillary optics, a

spatial resolution of 10 mm can be achieved [11]. In 2010, the

value of the combination of SAXS, WAXS and m-XRF for the

analysis of the structural and compositional properties of

bone was successfully demonstrated by Mahamid et al. [12].

Furthermore, transmission electron microscopy (TEM) also

provides structural information on the atomic level of small

sample areas in vitro [13]. Raman microscopy and Fourier

transform–infrared microspectroscopy (FT-IR) are used to

characterize both the organic and inorganic matrix of bone

with a spatial resolution of 1 and 10 mm, and FT-IR also

provides information about the collagen cross-linking [14–16].

In addition, histochemical and immunohistochemical tech-

niques such as stains and indicators have been optimized over

decades and enable the identification of biological tissue, cell

types and activity. However, staining and marker reagents

are limited in number and combination, and the well-known

compounds can be applied only in a targeted manner to

visualize structures in biological tissues. Hence, each of the

different methods provides essential information for bone

characterization and is indispensable in clinical use (DEXA)

or for biologists (histology).

To understand the complexity of the material composi-

tion better and to assess chemical and structural changes in

diseased bone, methods are required that offer detailed chemi-

cal information and enable us to define tissue types, mineral

content and distribution with high sensitivity and a high

spatial resolution. In the present paper, we demonstrate that

time-of-flight secondary ion mass spectrometry (ToF-SIMS)

provides this information and that its supplementary use

helps to bridge the gap between tissue and mineral analysis.

ToF-SIMS was originally developed as a technique for the

analysis of inorganic solids and in particular for the evaluation

of concentration differences with very high depth resolution.

In recent years, several studies have been performed to explore

and prove the capability of ToF-SIMS for surface analysis and

structural investigations in biological samples [17]. The tech-

nique has been used to characterize phospholipids, peptides

and other biomolecules on the cellular level [18–22]. The

value of two-dimensional imaging as well as three-dimen-

sional analytical reconstruction has been demonstrated [17].

In medicine, ToF-SIMS can be used to localize drugs and to

explore their distribution in different types of tissue. Based

on earlier studies on pure hydroxyapatite, for example, Malm-

berg et al. and Eriksson et al. characterized bone tissue by ToF-

SIMS and analysed the incorporation behaviour of titanium

implants [23–26]. Furthermore, ToF-SIMS studies on bone

and implant materials have been published by Ni et al. [27]

and Lu et al. [28]. In a more recent study, Zoehrer et al. [29]

applied ToF-SIMS as a supplementary method to characterize

human femoral bone. Up to now, ToF-SIMS has not been

applied to study affected bone in a manner as detailed as we

present herein.

In the following, we report systematic results on the use of

ToF-SIMS mass spectra and images of high spatial resolution

(less than 300 nm) for the analysis and characterization of dis-

eased and normal bone in a long-term small animal rat

model, where pathological changes are induced by ovari-

ectomy combined with either a special multi-deficiency diet

or a steroid medication. We reveal the advantages of ToF-

SIMS using embedded but non-stained cross sections and

use the chemical information to image and characterize sim-

ultaneously the inorganic and organic matrix.

The main objective of this paper is to demonstrate

the wealth of information that we can obtain by using ToF-

SIMS in addition to well-established techniques.

2. Results2.1. Bone tissue analysis2.1.1. Mass spectraThe procedure of ToF-SIMS analyses of bone samples carried

out in our study is shown schematically in figure 1. A typical

positive mass spectrum of trabecular area is depicted in

Ca+

(a) (b)

(c) (d)

(e) ( f )

5 × 104 8 × 103

6 × 103

4 × 103

2 × 103

7 × 101

6 × 101

5 × 101

4 × 101

3 × 101

2 × 101

1 × 101

6 × 10–3 3.0 × 104

12 months sham

12 months diet

5 × 10–3

4 × 10–3

3 × 10–3

2 × 10–3

1 × 10–3

0

39.95 39.96 39.97

0

4 × 104

3 × 104

2 × 104

1 × 104

0

1500

1250

1000

750

500

250

0200 250 300 350

10 20 30 40 50 60 70 80 90 100

10 20 30 40 50 60 70 80 90 100

100 110 120 130 140 150 160 170 180 190 200

C4H8N+CH4N+

CNO–

CN–

PO2– PO3

–

inte

nsity

(co

unts

)

inte

nsity

(co

unts

)

inte

nsity

(co

unts

)

inte

nsity

(co

unts

)in

tens

ity (

coun

ts)

norm

. int

ensi

ty (

coun

ts)

m/z m/z

C4H6N+

CaF+

C4H8NO+

CaPO2+

Ca2PO+3

Ca2PO+4

Ca3PO+5

Ca4PO+6

Ca5PO+6

Ca5PO+7

Ca6PO+8

Ca6PO+8

Ca6P3O13

Ca5P3O12Ca7P3O14

Ca8P3O15Ca9P3O16

400 400450 500 500 600 700

reference

2.5 × 104

2.0 × 104

1.5 × 104

1.0 × 104

5.0 × 103

0

+

++

+

+

Figure 2. Positive ion mass spectra from mineralized bone tissue of a rat obtained with Bi3þ primary ions. Mass spectra are shown in the ranges (a) 1 , m/z , 100, (b)

100 , m/z , 200, (c) 200 , m/z , 500, (d ) 350 , m/z , 750. (e) Caþ signal for the reference—12 months sham—and 12 months diet groups and ( f ) negativeion mass spectrum 1 , m/z , 100.

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

3

figure 2a–d. Several fragment ions were found to be derived

from a Ca10(PO4)6(OH)2 precursor species (ion, m/z value, in

the following always given in Da; Caþ, 39.96; Ca2PO3þ, 158.89;

Ca2PO4þ, 174.88; Ca5P3O12

þ , 484.12; Ca6P3O13þ , 540.04;

Ca7P3O14þ , 595.87; Ca8P3O15

þ , 651.79; Ca9P3O16þ , 707.65). Separ-

ation of the hydroxyl ion OH2 from Ca5(PO4)3OH yields the

mass fragment Ca5(PO4)3þ, 484.12. The fragment CaFþ, 58.96,

could also be identified in the mass spectrum. The fragments

CH4Nþ, 30.03; C4H6Nþ, 68.05; C4H8Nþ, 70.07, are known resi-

dues of the amino acid proline, one of the main components of

collagen type 1 [30]. C4H8NOþ, 86.06, can be assigned as a

hydroxyproline fragment. Figure 2e shows representative

mass spectra for the reference—12 months sham—and 12

months diet groups for the Ca signal. An intense Ca signal

was found for the 12 months sham group. An excerpt from a

typical negative mass spectrum is depicted in figure 2f. PO32,

79.96; PO22, 62.94, were found to derive primarily from the

mineralized areas, CN2, 26.10, from non-mineralized organic

compounds in spectra obtained in the negative ion mode.

2.1.2. Chemical imagingUsing the chemical information, we were able to visualize

the bone specimen morphology and its composition similar

to histology, but without the need for staining. In figure 3,

the microscopic view of the bone tissue (figure 3a) and the

corresponding ToF-SIMS image (figure 3b) are shown. The

cross section, which has been prepared to visualize bone

quality in a histological investigation, shows part of an osteo-

porotic rat vertebra. Trabeculae (Tb), osteoid (Os), bone

microscopic view(a)

At

At

75 mm 40 mm

40 mm100 mm300 mm

ToF-SIMS image 1positive

2positive 2negative

1negative

At

At At

Tb

Tb

Tb

Tb Tb

OsOs

300 mm

100 mm 100 mm

100 mm 100 mm

1

2

AtTb

Tb

TbOs

Tb

1

2

(b)

Figure 3. Light optical view (a) and corresponding ToF-SIMS image (b) of impaired bone quality in the spine of the ovariectomized þ diet group at 12 months (Tb,trabecular bone; At, adipose tissue; Os, osteoid). (a) Microscopic view: toluidine blue histological staining on EPON-embedded samples used to visualize bone quality.Adipose tissue (At) appears in large regions (red boxes). Wide areas of non-mineralized tissue within and around the borders of trabecular bone indicate increased osteoidregions (Os; red arrow, black boxes). (b) ToF-SIMS image: histology by mass spectrometry. ToF-SIMS image (overview) recorded in the positive ion mode, Bi3

þ, showing theoverlay of Caþ (green), C4H8Nþ (red) and C4H9

þ (blue) mass signals. The dashed line indicates the corresponding microscopic view. Detailed images recorded in thepositive ion mode, Bi1

þ, showing the overlay of 1positive Caþ (green), C4H8Nþ (red), C4H9þ (blue) and 2positive Caþ(green), C4H8Nþ (red). Detailed images recorded

in negative ion mode showing the overlay of 1negative PO32 (green), CN2 (red), sum of C16Hx ¼ 29237O2 (blue) and 2negative PO3

2 (green), CN2 (red).

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

4

marrow, connective tissue and adipose tissue (At) can be dis-

criminated. The ToF-SIMS image recorded in the positive ion

mode, collecting all positive SIs, gives an overview using the

information from the lateral distribution of the Caþ signal in

green, the C4H8Nþ signal in red and the C4H9þ signal in blue

(figure 3b). Here, the C4H9þ signal represents residues of the

adipose tissue. These three mass signals suffice to identify

unambiguously the mineralized trabecular area, the non-

mineralized connective tissue as well as the cell membranes

of the adipose tissue. In addition to the overview image, two

smaller regions (1 and 2) have been analysed in the positive

as well as in the negative ion mode. The images recorded in

the positive mode show an overlay of the Caþ (green) signal,

the C4H8Nþ(red) signal and the C4H9þ(blue) signal (1positive

and 2positive). The detail image 2positive allows a distinct deter-

mination of the trabeculae (Tb) in green and the surrounding

non-mineralized border, the so-called osteoid (Os), in red.

The corresponding negative ion mode images show the overlay

of PO32 signals in green (Tb), CN2 signals in red (Os) and sig-

nals of various long-chain fatty acid fragments such as

C16H29O2 in blue. Mapping signals of long-chain fatty acid

fragments identifies the cellular membrane structure of adi-

pose tissue in the negative ion mode. However, the images

recorded in the negative ion mode also allow distinct discrimi-

nation between trabeculae (green), non-mineralized areas such

as osteoid (red) and adipose tissue (blue).

2.1.3. Characterization of boneFrom the positive SIs, we gain information about the local dis-

tribution of minerals as well as of the organic compounds in

the bone. The Ca images shown in figure 4c,g,h exhibit a

quite heterogeneous distribution of Ca all over the trabecular

meshwork. In addition, numerous microfractures within the

trabecular bone are visible. In the immediate vicinity of

some fractures, calcium is accumulated, or the fractures are

located particularly in areas of high Ca content. Furthermore,

the fibril structure of collagen type 1 can be recognized to

some extent. The mass images of Caþ and C4H8Nþ provide

complementary information about the distribution of hydro-

xyapatite and collagen, as shown in figure 4e,f.

2.2. Analysis of diseased bone2.2.1. Scans of vertebraeThe mass images of Caþ and C4H8Nþ depict the structure of

rat vertebrae from the 12 months sham group (figure 5a,b)

(a)

(b)

(c)

(d)

(g)

(h)

(e)

( f )

100 mm

1 2

4 3

100 mm

100 mm

10 mm

100 mm

100 mm

20 mm

20 mm

Figure 4. ToF-SIMS images (a) of an osteoporotic trabecula showing the Caþ distribution (500 � 500 mm2, Bi3þ, high-current bunched mode) and the contour of

the collagen structure (white). (b) Corresponding C4H8Nþ image representing the collagen structure of the trabecula (a). (c) 500 � 500 mm2, Bi3þ, image of the

trabecular meshwork obtained in the unbunched mode for high lateral resolution showing numerous trabecular microfractures. (d ) 49.5 � 49.5 mm2 image, Biþ,obtained in the unbunched burst mode for the determination of the Ca content of the edge and centre regions of the trabecula. (e,f ) Detailed view of (c), 99 �99 mm2, Bi3

þ, obtained in the unbunched mode for high lateral resolution showing the Caþ (e) and C4H8Nþ ( f ) distribution representing the fibril structure ofcollagen. (g,h) Caþ images, 500 � 500 mm2, Bi3

þ, showing numerous trabecular microfractures and accumulation of Ca adjacent to the fractures (white arrow) forreference group (g) and 12 months diet group (h).

Ca+

Ca+

C4H8N+

C4H8N+

1mm 1mm

1mm 1mm

(a) (b)

(c) (d)

Figure 5. ToF-SIMS images of the whole vertebra of (a,b) the 12 months sham group and (c,d) the 12 months diet group showing the Caþ (a,c) and the C4H8Nþ

(b,d) distribution, Bi3þ, high-current bunched mode, stage scans 9 � 9 (a,b) and 6 � 6 mm2 (c,d).

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

5

100 mm 100 mm 100 mm 100 mm

100 mm100 mm100 mm

3 months sham

12 months diet3 months diet 14 months diet

12 months sham 14 months sham0 months reference40

30

20

10

040

30

20

10

0

Figure 6. ToF-SIMS images (500 � 500 mm2, Bi3þ) obtained from cross sections of rat vertebrae. The images show the distribution and the intensity of the Ca

signal across the trabecular meshwork. (Colour scale from black, to red to yellow depending on the Ca content.)

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

6

and the 12 months diet group (figure 5c,d). The characteristic

reduction of the trabecular density and structure in the

affected vertebra can be seen unambiguously (figure 5c,d).

The C4H8Nþ mass image of this vertebra reveals a broaden-

ing of the trabecular collagen matrix and a loss of the

distinct meshwork structure. By contrast, the collagen

matrix of the healthy vertebra shows an intact trabecular

structure which is mineralized homogeneously (figure 5a).

In addition, we found Ca accumulation at the outer part of

the affected vertebra (white rectangle). This accumulation

was observed for osteoporotic vertebra in several cases.

2.2.2. Evaluation of the calcium contentFigure 6 shows Ca images (500 � 500 mm2) which give an

overview of the trabecular meshwork of the vertebra. Com-

paring the images of the reference and sham groups

(healthy group) with the one of the diet groups (diseased

group), the reduction in the trabecular number density and

thickness with time is obvious. We also found a decrease in

the Ca concentration in the affected trabeculae represented

by the reduced brightness of the Caþ mass image. To deter-

mine the decline in the local Ca content, numerous regions

of the trabecular meshwork of each sample were analysed

as described before. In figure 7a, the mean values for the

Ca count rates in each group at different times are shown.

There is an increase in the Ca content for the sham group

compared with the reference group of up to 22%. After

three months, a decline in the Ca content of 48% for the

diet and steroid group compared with the three months

sham group is detected. The decrease amounts to 57% and

60% for the diet group compared with the respective sham

group after 12 and 14 months.

A detailed analysis of the trabecular edge and centre

region revealed a reduced Ca concentration at the trabecular

edge for the 12 months diet group compared with the refer-

ence and 12 months sham group shown in figure 7d. The

decline in the Ca concentration in the edge region compared

with the centre region is 38% for the diet group.

2.3. Statistical analysisThe evaluation of the integrated intensity of the Ca signal con-

firms that the reduction in the Ca content for the three, 12 and

14 months diet groups compared with the reference group is

statistically significant ( p , 0.05). The reduction in the 12 and

14 months diet groups is of even higher significance (p ,

0.001). By contrast, the decline in the Ca content after three

months in the steroid group compared with the reference

group is considerable but is not statistically significant. A com-

parison of the Ca content after three months shows a significant

reduction in the Ca content for the diet group as well as for the

steroid group compared with the sham group. After 12

months, there is also a significant reduction in the Ca content

for the diet group compared with the corresponding sham

group, as expected. In addition, the reduction in the Ca content

after 14 months compared with the three months and 12

months sham groups is statistically significant. In addition,

the decline in the Ca content of the 14 months diet group com-

pared with the three months sham group is again highly

significant ( p , 0.001). All results are summarized in figure 7a.

2.4. Characterization of the trabeculaeIn contrast to the decrease in the Ca concentration, the intensities

of the C4H8Nþ signals representing the collagen matrix remain

constant. A slight increase according to the ongoing grade of

bone deterioration may be inferred. In figure 4a, we sketch the

border of the collagen signal of a diseased bone with white

lines in the Caþ mass image. The C4H8Nþ signal represents

the collagen matrix and therefore the actual area of the trabecu-

lae (figure 4b). Ca is located only in the centre of the trabeculae.

In this figure, we clearly see the extent of the degradation pro-

cess. This significant decline is again accompanied by a

heterogeneous distribution and numerous microfractures.

2.5. Ca/P ratioA further parameter that may be used to assess bone quality is

the variation of the Ca/P ratio within the different groups over

time points (months)

BM

D (

g cm

–2)

M0 M1 M3 M12 M14

Cacentre

Caedge

C4H

8N+

centreC

4H

8N+

edge

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

******** ****

****

****

**

****

************

****

*

*

***

(a) (b)

(d)(c)

1.5 × 106

1.2 × 106

9.0 × 105

6.0 × 105

3.0 × 105

0

1.0 × 106

1.2 × 106

8.0 × 105

6.0 × 105

4.0 × 105

2.0 × 105

00 months 12 months sh 12 months diet

0

1000

2000

3000

4000

5000

0 months 3 months

sh diet st

*

*

**

*

***

**

***

**

sham

diet

norm

. int

ensi

ty (

coun

ts)

of C

a+C

a/P

ratio

norm

. int

ensi

ty (

coun

ts)

**

*

12 months 14 months

sh diet sh dietst

0 months 3 monthssh diet st

12 months 14 monthssh diet sh dietst

Figure 7. (a) Summary of the results for the Ca content of rat vertebrae for the different groups at different times. The mean values and the standard deviation ofthe mean value are depicted. The asterisks indicate the significance level (*p , 0.05, **p , 0.01, ***p , 0.001). (b) Results of DEXA measurements: BMD(g cm22) of sham and diet groups over time (two-way ANOVA and Bonferroni’s multiple comparison test) and longitudinal variation with time (Friedman’stest and Dunn’s multiple comparison test) [31]. (c) Ca/P ratio for the different groups as a function of time. (d ) Ca and C4H8Nþ content of the edge andcentre region of the trabeculae for several groups.

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

7

time. Therefore, the mean values of the Ca and P content and

their ratios were calculated for each sample. The results are

shown in figure 7c. In general, an increase in the Ca/P ratio

with time was observed for the untreated reference and sham

groups. Furthermore, at three, 12 and 14 months a considerable

decrease in the Ca/P ratio was found for the treated groups

compared with the reference or corresponding sham groups.

3. Material and methods3.1. Time-of-flight secondary ion mass spectrometryA good description of the ToF-SIMS method can be found in the

literature [32–34]. ToF-SIMS originates from inorganic elemental

analysis and is one of the most important and versatile surface

analytical techniques. The SIMS process is illustrated schemati-

cally in figure 1. The samples are bombarded with Biþ ions or

Bi3þ clusters; the latter are most appropriate for organic and bio-

logical samples. As a result of the impact of these primary ions,

their energy (typically 25–30 keV) is transferred to the sample

surface, which results in a collision cascade and an emission of

secondary ions (SIs) from the first atomic layers (up to 10 nm).

The SIs are collected and analysed by their ToF according to

their mass to charge ratio. By rastering the primary ion beam, it

is possible to sustain the spatial correlation and distribution of

the desorbed ions, such as in an ion microscope. The amount

of ionized and removed material corresponds to a few atomic/

molecular monolayers in the static mode of ToF-SIMS. Therefore,

the technique can be considered as quasi-non-destructive. Using

a more intensive ion gun (sputter gun), depth profiling is possible

by removing layer after layer in a small defined area.

The development of cluster ion guns has enabled the analysis

of biological samples and polymers in the past decade. Today,

state-of-the-art SIMS analysis of organic samples is performed

with bismuth or carbon clusters such as Bi3þ, Bi5

þ, Bi52þ, C60

þ or

C602þ, which reduce the fragmentation and enhance the secondary

ion yield of larger ion fragments.

ToF-SIMS studies on mineralized tissue revealed hydroxy-

apatite fragments with higher masses using a cluster ion source

[25]. Owing to the high sensitivity of the method, even traces

of chemical compounds are detectable, which may contain valu-

able information about the sample. Mass spectrometric imaging

offers element-specific concentration maps of small and wide

areas to trace, for instance the Ca distribution within the bone.

Therefore, we present mass images of wide-area overviews (up

to 9 � 9 mm2) as well as high-resolution images of regions as

small as 50 � 50 mm2, with a lateral resolution of less than 300 nm.

ToF-SIMS data were acquired using a ToF.SIMS 5-100 machine

(IonTOF Company, Munster, Germany) equipped with a bismuth

cluster primary ion source (25 keV) and Cs, O2 and C60 sputter

guns. Ions were registered within the mass range from m/z ¼ 1 to

800 u for positive and in some cases for negative SIs. Images were

taken in the high-current bunched (hc-bu) mode of the primary

ion gun with high mass resolution of more than 8025 full width

half maximum (FWHM) at m/z ¼ 29.00 u (e.g. the fragment CHO)

and a spatial resolution of less than 10 mm. Images of high lateral

resolution less than 300 nm, but with poorer mass resolution

(95 FWHM at m/z ¼ 29.00 u), were recorded in the non-bunched

burst alignment (bi-ba-i) mode.

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

8

Bi3þ primary ions were supplied with a target current of 0.2–

0.5 pA for the hc-bu mode and of 0.02 –0.1 pA for the bi-ba-i

mode. The fields of view ranged from wide-area overviews of the

whole vertebra of 9 � 9 mm2 (figure 5a–d), medium-sized areas of

500 � 500 mm2 (figures 3a,b, 4a–c,h,i and 6) to small-area images

of 99� 99 mm2 (figure 4e,f) and of 49.5 � 49.5 mm2 (figure 4d).

The small-area images of 49.5 � 49.5 mm2 size were taken in the

hc-bu mode to analyse the Caþ concentration. The resolution was

128 � 128 pixels for small-area images as well as for the 500�500 mm2 images (figure 6). The resolution of the 99� 99 mm2

images (figure 4e,f) and for 500 � 500 mm2 image (figure 4c) was

512 � 512 pixels. The resolution of the 500� 500 mm2 images

shown in figure 3a,b, taken in the bi-ba-i mode, was 1024 � 1024

pixels. All images except for those in figure 3 are presented in a

colour scale ranging from black to red to bright yellow depending

on the mass signal intensity. Bright colours represent high signal

intensities, i.e. large SI yields. During all measurements, charge com-

pensation was applied. All SI yields are normalized to the applied

primary ion current. For data evaluation, the IonTOF software SUR-

FACELAB v. 6.3 was used.

For a semi-quantitative comparison of the Ca concentration of

affected and normal bone, we defined three to five measurement

areas with a size of 49.5� 49.5 mm2 on the trabeculae of each

sample and calculated the average count rates for Ca (figure 7a),

normalized to the primary ion current. The error is given as the

average deviation from the calculated mean value. In addition,

the averaged count rates for P were also calculated to give the

Ca/P ratio for all examined groups of samples (figure 7c).

A semi-quantitative analysis of the Ca concentration of the

edge and centre regions of the trabeculae was carried out for

the sham (12 months), reference and diet (12 months) group of

animals (figure 7d ). Therefore, areas of 49.5 � 49.5 mm2 were

analysed using Biþ, bi-ba-i-burst mode, at a resolution of 256 �256 pixels over 500 scans. Afterwards, the area was divided

into four equal regions of interest (ROIs), and the Ca content of

each ROI was calculated by data reconstruction. For each case,

edge and centre, two ROIs and the corresponding averaged

count rate of Ca were calculated and compared (figure 4d ).

3.2. Statistical analysisStatistical analysis was performed using the one-way ANOVA test

accompanied by Games–Howell pairwise multiple comparison to

determine the variation of the Ca content in each group and among

groups at particular times as well as over the whole duration of the

experiment. The asterisks indicate the significance level (*p , 0.05,

**p , 0.01, ***p , 0.001). The results were expressed as the mean

and standard error (figure 7a). The evaluation was conducted

using the software IBM SPSS STATISTICS v. 20.

3.3. Sample preparation of the cross sectionsThe protocol and procedures used in the animal experiment have

been reported previously in detail [31]. This experimental study

was performed in full compliance with German and our insti-

tution’s animal protection laws and was approved by the ethics

commission of the local governmental institution (‘Regierung-

sprasidium’ Giessen/Germany, permit number: 89/2009 and

20/10-Nr.A31/2009).

Vertebrae were collected after euthanasia at distinct times

(three, 12 and 14 months). Fifty-six samples were embedded

in polymethylmethacrylate (Technovit 9100; Heraeus Kulzer,

Hanau, Germany). The embedded samples were cut and

ground into slices of 20–50 mm thickness. The quality of the sur-

face finish was examined under the optical microscope. One

additional sample was embedded in EPON resin mix, as pre-

viously described [35], sectioned and stained with toluidine

blue. Images were taken using a light microscope (Axioplan 2

imaging with photo module Axiophot 2, Carl Zeiss, Jena,

Germany) using 1 Leica DC500 camera (Leica, Bensheim,

Germany), acquired with Leica IM1000 software and processed

using Adobe PHOTOSHOP CS3 Extended, v. 10.0.1.

For the ToF-SIMS analysis, we used samples from 19 animals

in the sham group (time of euthanasia in months, number of ani-

mals; 3, 7; 12, 6; 14, 6), from 23 animals in the diet group (3, 7; 12,

7; 14, 9), from nine animals in the steroid group (3, 7; 12, 2) and

from eight animals in the reference group at the beginning of the

experiment (0, 8).

4. DiscussionToF-SIMS as a non-routine method in medicine was applied

to analyse osseous tissue and to study the ongoing grade of

bone deterioration in an animal model. We analysed affected

and normal bone and therefore focused on the main

components of the bone, Ca and collagen.

The analysis of mineralized areas reveals typical hydroxy-

apatite fragments characterized by a sequence of masses

separated by the mass of CaO (m/z ¼ 55.96 Da), as recently

reported by Malmberg et al. [25]. Molecular organic frag-

ments, such as CH4Nþ, 30.03; C4H6Nþ, 68.05; C4H8Nþ,

70.07; and C4H8NOþ, 86.06, were detected and identified to

be characteristic residues of glycine, proline and hydroxypro-

line, the main amino acids of collagen [30,36]. We consider

these signals also to be characteristics for the collagen

matrix of the composite material bone. This is supported by

the analysis and interpretation of numerous mass images

of the trabeculae. The spatially resolved detection of these

fragments provides information on the localization of the col-

lagen matrix within the bone. The good lateral resolution

allows imaging of the fibril structure in high-resolution

mass images (figure 4e,f ). Areas of non-mineralized osteoid

at the interface of trabeculae and bone marrow could be

localized (figure 3). By mapping the collagen signals, we

identified regions of high and low collagen content (always

complementary to the Ca distribution). The complementary

characteristics of the collagen matrix and the hydroxyapatite

correlate well with the deposition process of hydroxyapa-

tite along the collagen fibrils [37]. In addition, negatively

charged ions such as PO32, PO2

2, CN2 or long-chain fatty

acid fragments were used to identify cellular structures

within the osseous tissue. In particular, the PO32 ion as well

as the PO22 ion provide rich information: on the one hand,

these ions originate from the mineralized part of the bone,

represented by hydroxyapatite (Ca5(PO4)3OH), whereas, on

the other hand, these ions are characteristic of cellular parts

of the bone tissue being residues from the phospholipids of

the cell membrane and also of the phosphate groups of the

chromosomes in the cell nucleus.

The distribution of Ca within the bone tissue is an impor-

tant parameter to assess bone quality [29,38]. It is known

from the literature that heterogeneity of mineralization affects

the local material properties and therefore bone quality.

Contradictory studies on the distribution of minerals and

the associated fracture risk in osteoporotic bone can be

found in the literature [29,39,40]. We determined the spatial

distribution of calcium, phosphate and several calcium

phosphate ion fragments originating from hydroxyapatite.

Our results confirm a heterogeneous distribution of Ca for

the healthy as well as for the affected bone. However, an

increased heterogeneity for the treated groups could not be

determined. For the reference group, we observed the

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

9

collagen matrix on average being covered homogeneously

with Ca but with local increased concentrations especially

at interfaces of microfractures. These small accumulations

appear to be characteristic for young and healthy bone. The

microfractures, which are scattered all over the trabeculae,

are the result of mechanical stress and have already been

described in the literature [41–43]. Microcracks release

strain and avoid complete fracture. The occurrence of micro-

fractures and the subsequent remodelling are consequences

of the normal functional loading of the skeleton and its adap-

tation to individual physical demand. While an increased

number of microfractures in the case of older age or osteo-

porosis is described for human trabeculae [39], our study

on rat bones does not confirm this. We cannot show unequi-

vocally a significant increase in microcracks for the groups

with osteoporotic phenotype. This might be due to the fact

that physical stress, on the one hand, and altered mechanical

conditions, on the other hand (hyper-mineralization for

example), have a lower impact on microspecimens (rats)

than on macrospecimens (humans).

Nevertheless, the microcracks exhibit different character-

istics for the reference and sham groups than for the treated

groups. For the non-affected trabeculae, a Ca accumulation

directly localized at the microcracks is found (figure 4g).

Apparently, those microcracks arise in the areas of higher

mineral content with reduced elasticity and higher brittle-

ness. The increased Ca concentration suggests increased

remodelling activity at the location of the crack. In the case

of diseased trabeculae, Ca accumulations along the fractures

are more expanded and randomized (figure 4h). Although

the dehydration step during the preparation process also

causes fractures, the exclusive appearance of microcracks as

artefacts can be excluded. The characteristic accumulation

of Ca and the complementary structure of the collagen

matrix right at those fractures allow the distinct differen-

tiation between microfractures and artefacts. A heightened

Ca intensity as a result of edge effects can be excluded for

the same reasons.

Furthermore, a remarkable decrease in the Ca content can

be documented from the centre to the edge region for the diet

group with time in a semi-quantitative analysis. We suggest

that this is caused by the disturbed mineralization process

owing to osteoporotic conditions. An unbalanced remodel-

ling process may account for overactivity of the osteoclasts,

which release hydroxyapatite and the collagen matrix.

We suppose that the enzymatic dissolution of the collagen

matrix is not complete, hence parts of the collagen matrix

remain unchanged. In the following step of the remodelling

process, osteoblasts start to rebuild new bone by expressing

extracellular matrix as a scaffold for the crystallization of cal-

cium phosphate crystals. Owing to the calcium, phosphoros

and vitamin D deficiency caused by the deficit in nutrition,

the mineralization process is extenuated or even impossible

[44]. The result is a heterogeneous distribution of Ca, some-

times leading to a complete absence of minerals in wide

regions. The extent of non-mineralized collagen might be

an argument for an induced osteomalacia rather than osteo-

porosis. This has to be confirmed by further studies. Stage

scans of the whole vertebra clearly show the extent of the

pathological degradation process. The release of the minerals

leads to a trabecular meshwork consisting mainly of collagen,

as shown in figure 5d. Considering the whole vertebra, it

becomes apparent that, in unstressed regions of the inner

part, an extended Ca depletion can be found. By contrast,

an accumulation of calcified tissue is found in the outer

region. This part of the vertebra sustains the largest portion

of physical strain. As remodelling and mineralization is

adapted to stress-related strain, these burdened regions exhi-

bit an increased content of calcium [45,46].

Ca mass images of the trabecular meshwork also showed

a remarkable reduction in the trabecular number and thick-

ness, which would be in accordance with an osteoporotic

disease pattern [47]. In addition, the semi-quantitative evalu-

ation of the Ca content revealed a significant decline in Ca for

the affected groups. Here, the ToF-SIMS data show the same

tendencies as the corresponding DEXA measurements, as

shown in figure 7b. But, the decline of Ca (in terms of per-

centage) clearly exceeds the decrease in BMD obtained by

DEXA measurements. This is probably due to the selectivity

of ToF-SIMS, which allows the exclusive evaluation of the

Ca signal, whereas DEXA integrates all X-ray absorbing

components such as Si, Mg or Sr. Furthermore, the natural

growth of the animals’ skeletons has to be considered as

it results in increasing bone thickness and bone volume. For

the DEXA data evaluation, it was not possible to account

for the parameter bone thickness sufficiently. Therefore, the

calculated BMD values are systematically too high with

ongoing age of the rats. It has to be noted that SIMS

does not a priori offer direct information on absolute concen-

trations. The count rates of the mass signals are proportional

to the surface concentrations of the respective species, but

they also depend on a number of other sample properties,

which influence the ionization process. A substantial

decrease in the Ca content owing to pathological modifi-

cations of the bone influences the composition of the matrix

and therefore the ionization probability for Ca. So, for a

semi-quantitative comparison of the Ca concentration of

affected and normal bone, all samples have been prepared

in the same way, and identical measurement conditions

were applied to give good reproducibility. For a quantitative

interpretation of the measured mass signal intensities, the

matrix effect has to be considered, and calibration with suit-

able standards is necessary. We are currently defining such

standards and are aiming for a reliable quantification of the

Ca content using ToF-SIMS in a future study. Our own pre-

liminary measurements showed a linear correlation of the

Ca signal with the Ca content within the range of 40–80%

Ca in a collagen matrix.

As known from the literature, the building of hydroxy-

apatite Ca10(PO4)6(OH)2 proceeds via several stages with

various calcium phosphate phases as precursors [48]. Amor-

phous calcium phosphate as well as octa calcium phosphate

are discussed as potential biologically relevant precursors

of hydroxyapatite in bone [48]. The different calcium phos-

phate phases are characterized by their chemical formula

and their Ca/P ratio [49]. The local Ca/P ratio of the bone

indicates the stage of mineralization and allows the assess-

ment of bone quality. Therefore, the analysis of the Ca/P

ratio with time can provide supplementary information

about the mineralization process and its alteration in systemi-

cally diseased bone. Diverging results have been published so

far on this subject. While Zoehrer et al. [29] found an

increased Ca/P ratio based on imbalance and decrease of P

in the case of osteoporosis, the study by Basle et al. [50]

showed no differences. By contrast, Tzaphlidou et al. as

well as Kourkoumelis et al. reported a correlation of

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

10

osteoporosis with decreased Ca/P ratio in various studies

[51–53]. We evaluated the Ca/P ratio for the different

groups after defined time intervals. For the untreated and

healthy groups (reference and sham), we observed an increas-

ing Ca/P ratio with time. This is explained by the advancing

age of the animals and crystallization of young to mature

bone, respectively. The density of densely mineralized

regions increases with age [5,6,54]. The study by Zoehrer

et al. [29] reveals that areas of higher mineralization exhibit

an increased Ca/P ratio, and our results are in accord with

this. By contrast, the Ca/P ratio of the corresponding diet

groups shows a decrease in the Ca/P ratio with time. As

we did not observe regions of high calcium content and

areas of low Ca content in particular, as Zoehrer et al. [29]

reported for human bone, the decrease in the Ca/P ratio

appears to be characteristic for the treated groups, indicating

lower bone quality and therefore osteoporotic patterns.

We could confirm, in the present study, that alterations

of the chemical composition of the inorganic and organic

matrix have to be taken into account when estimating bone

quality [16,37,53]. Our study also emphasizes the need for

imaging methods, which facilitate the examination of both

to elucidate disease patterns in more detail. Owing to the

higher spatial resolution than FT-IR [16] (300 nm), good

mass resolution (R . 8000) and high sensitivity (in the ppm

range), ToF-SIMS has been found to be a valuable technique

to address this task. So, the results presented herein show

typical osteoporotic characteristics referred to the minerals

but compelling arguments for osteomalacic patterns when

considering the organic matrix. Therefore, our ToF-SIMS

study revealed that the experimental set-up did not induce

osteoporosis in the diet group unambiguously, but osteoma-

lacia. Clearly, technical improvements will further advance

the application of ToF-SIMS; for example, the development

of standards for the quantitative analysis of bone components

and a distinct mapping of the calcium phosphate phases to

assess and track the mineralization process in more detail.

Using the chemical information obtained by the mass

spectra, it will be possible to learn more about the mechan-

isms taking place in bone such as the complex bone

remodelling process. The parallel detection of all chemical

information deriving from the sample surface offers a wide

range of promising opportunities. Messenger molecules can

be detected, and special biomarkers and drugs can be

traced, once their fragments are identified unequivocally

and if the local concentration of those compounds is high

enough. The characterization of modified implant surfaces

and the analysis of the interface between the implant and

bone tissue will support the development of improved

materials and drugs for patients with systemically diseased

bones. Nevertheless, two important and critical aspects

should be noted. First, sample preparation has to be per-

formed carefully and, second, the low concentration of

biomolecules with high masses, large molar volume and

low ionization probability has to be taken in consideration.

We can overcome this problem if either the concentration of

the species of interest is sufficiently high or we leave the

static SIMS limit. In any case, ToF-SIMS facilitates the collec-

tion of valuable compositional and structural information of

complex biological tissues such as bone as a complement to

other techniques.

Acknowledgements. This study is part of the project M5 and T1 withinthe collaborative research centre SFB/TRR 79—‘Materials for tissueregeneration within systemically altered bone’, funded by theGerman Research Foundation (DFG) and by the State of Hessen.We gratefully acknowledge the financial support within this project.We thank Prof. Dr Ing. habil. H. Worch for fruitful discussions andcareful comments on the manuscript. A.H. performed the ToF-SIMS analysis, data evaluation and interpretation, and wrote andedited the first draft of the manuscript. M.R. advised on the interpret-ation of the data, edited the manuscript together with J.J. and wrotethe grant application that supported the study. T.E.K. performed thelight microscope histomorphometry and edited the manuscript. G.P.did the evaluation of the DEXA measurements. G.S. performed theanimal experiment and edited the manuscript. C.H. developed theanimal experiment design, coordinated the animal experiment per-formance, the sample preparation and with M.R. and J.J. wrote thegrant application that supported the study. J.J. wrote the grant appli-cation that supported the study, coordinated the ToF-SIMS studyperformance and edited the final version of the manuscript. Allauthors state that they have no conflicts of interest.

References

1. Blake GM, Fogelman I. 2007 The role of DXA bonedensity scans in the diagnosis and treatment ofosteoporosis. Postgrad. Med. J. 83, 509 – 517.(doi:10.1136/pgmj.2007.057505)

2. Engelke K, Libanati C, Liu Y, Wang H, Austin M,Fuerst T, Stampa B, Timm W, Genant HK. 2009Quantitative computed tomography (QCT) of theforearm using general purpose spiral whole-body CTscanners: accuracy, precision and comparison withdual-energy X-ray absorptiometry (DXA). Bone 45,110 – 118. (doi:10.1016/j.bone.2009.03.669)

3. Perilli E, Briggs AM, Kantor S, Codrington J, WarkJD, Parkinson IH, Fazzalari NL. 2012 Failure strengthof human vertebrae: prediction using bone mineraldensity measured by DXA and bone volume bymicro-CT. Bone 50, 1416 – 1425. (doi:10.1016/j.bone.2012.03.002)

4. Heuck F, Vanselow K. 2005 Radiologische Analysedes Knochens. Berlin, Germany: Springer.

5. Roschger P, Paschalis EP, Fratzl P, Klaushofer K.2008 Bone mineralization density distribution inhealth and disease. Bone 42, 456 – 466. (doi:10.1016/j.bone.2007.10.021)

6. Ruffoni D, Fratzl P, Roschger P, Klaushofer K,Weinkamer R. 2007 The bone mineralization densitydistribution as a fingerprint of the mineralizationprocess. Bone 40, 1308 – 1319. (doi:10.1016/j.bone.2007.01.012)

7. Roschger P, Fratzl P, Eschberger J, Klaushofer K.1998 Validation of quantitative backscatteredelectron imaging for the measurement of mineraldensity distribution in human bone biopsies.Bone 23, 319 – 326. (doi:10.1016/S8756-3282(98)00112-4)

8. Ito M. 2011 Recent progress in boneimaging for osteoporosis research. J. BoneMiner. Metab. 29, 131 – 140. (doi:10.1007/s00774-010-0258-0)

9. Wagermaier W, Gupta HS, Gourrier A, Paris O,Roschger P, Burghammer M, Riekel C, Fratzl P. 2007Scanning texture analysis of lamellar bone usingmicrobeam synchrotron X-ray radiation. J. Appl.Crystallogr. 40, 115 – 120. (doi:10.1107/S0021889806044888)

10. Zizak I et al. 2000 Investigation of bone andcartilage by synchrotron scanning-SAXS and -WAXDwith micrometer spatial resolution. J. Appl.Crystallogr. 33, 820 – 823. (doi:10.1107/S0021889800001321)

11. Roschger P et al. 2010 Bone material quality intransiliac bone biopsies of postmenopausal

rsif.royalsocietypublishing.orgJR

SocInterface10:20130332

11

osteoporotic women after 3 years of strontiumranelate treatment. J. Bone Miner. Res. 25,891 – 900. (doi:10.1359/jbmr.090702)

12. Mahamid J et al. 2010 Mapping amorphous calciumphosphate transformation into crystalline mineralfrom the cell to the bone in zebrafish fin rays. Proc.Natl Acad. Sci. USA 107, 6316 – 6321. (doi:10.1073/pnas.0914218107)

13. Suvorova EI, Petrenko PP, Buffat PA. 2007Scanning and transmission electron microscopyfor evaluation of order/disorder in bonestructure. Scanning 29, 162 – 170. (doi:10.1002/sca.20058)

14. Kazanci M, Fratzl P, Klaushofer K, Paschalis EP.2006 Complementary information on in vitroconversion of amorphous ( precursor) calciumphosphate to hydroxyapatite from Ramanmicrospectroscopy and wide-angle X-ray scattering.Calcif. Tissue Int. 79, 354 – 359. (doi:10.1007/s00223-006-0011-9)

15. Gamsjaeger S, Buchinger B, Zoehrer R, Phipps R,Klaushofer K, Paschalis EP. 2011 Effects of one yeardaily teriparatide treatment on trabecular bonematerial properties in postmenopausal osteoporoticwomen previously treated with alendronate orrisedronate. Bone 49, 1160 – 1165. (doi:10.1016/j.bone.2011.08.015)

16. Paschalis EP, Glass EV, Donley DW, Eriksen EF. 2005Bone mineral and collagen quality in iliac crestbiopsies of patients given teriparatide: newresults from the fracture prevention trial. J. Clin.Endocrinol. Metab. 90, 4644 – 4649. (doi:10.1210/jc.2004-2489)

17. Fletcher JS, Vickerman JC. 2010 A new SIMSparadigm for 2D and 3D molecular imaging of bio-systems. Anal. Bioanal. Chem. 396, 85 – 104.(doi:10.1007/s00216-009-2986-3)

18. Wu L, Lu X, Kulp K, Knize M, Berman E, Nelson E,Felton J, Wu K. 2007 Imaging and differentiation ofmouse embryo tissues by ToF-SIMS. Int. J. MassSpectrom. 260, 137 – 145. (doi:10.1016/j.ijms.2006.09.029)

19. Debois D, Bralet M-P, Le Naour F, Brunelle A,Laprevote O. 2009 In situ lipodemic analysis ofnonalcoholic fatty liver by cluster TOF-SIMSimaging. Anal. Chem. 81, 2823 – 2831. (doi:10.1021/ac900045m)

20. Heim C, Sjovall P, Lausmaa J, Leefmann T, Thiel V.2009 Spectral characterisation of eight glycerolipidsand their detection in natural samples using time-of-flight secondary ion mass spectrometry. RapidCommun. Mass Spectrom. 23, 2741 – 2753. (doi:10.1002/rcm.4183)

21. Magnusson Y, Friberg P, Malmberg P, Chen Y. 2008Application of multivariate analysis of TOF-SIMSspectra for studying the effect of highglucose intake on aortic lipid profile. Appl. SurfaceSci. 254, 6580 – 6585. (doi:10.1016/j.apsusc.2008.04.035)

22. Brunelle A, Laprevote O. 2008 Lipid imaging withcluster time-of-flight secondary ion massspectrometry. Anal. Bioanal. Chem. 393, 31 – 35.(doi:10.1007/s00216-008-2367-3)

23. Malmberg P, Nygren H. 2008 Methods for theanalysis of the composition of bone tissue, with afocus on imaging mass spectrometry (TOF-SIMS).Proteomics 8, 3755 – 3762. (doi:10.1002/pmic.200800198)

24. Eriksson C, Malmberg P, Nygren H. 2008 Time-of-flight secondary ion mass spectrometric analysis ofthe interface between bone and titanium implants.Rapid Commun. Mass Spectrom. 22, 943 – 949.(doi:10.1002/rcm.3445)

25. Malmberg P, Bexell U, Eriksson C, Nygren H, RichterK. 2007 Analysis of bone minerals by time-of-flightsecondary ion mass spectrometry: a comparativestudy using monoatomic and cluster ions sources.Rapid Commun. Mass Spectrom. 21, 745 – 749.(doi:10.1002/rcm.2890)

26. Eriksson C, Ohlson K, Richter K, Billerdahl N,Johansson M, Nygren H. 2007 Callus formation andremodeling at titanium implants. J. Biomed.Mater. Res. A 83, 1062 – 1069. (doi:10.1002/jbm.a.31433)

27. Ni GX, Lu WW, Xu B, Chiu KY, Yang C, Li ZY,Lam WM, Luk KD. 2006 Interfacial behaviour ofstrontium-containing hydroxyapatite cement withcancellous and cortical bone. Biomaterials 27,5127 – 5133. (doi:10.1016/j.biomaterials.2006.05.030)

28. Lu X, Leng Y, Weng LT. 2004 TOF-SIMS study ofbone mineralization on alkali-treated Ti alloy.J. Mater. Sci. 39, 6809 – 6811. (doi:10.1023/B:JMSC.0000045608.64178.55)

29. Zoehrer R, Perilli E, Kuliwaba JS, Shapter JG,Fazzalari NL, Voelcker NH. 2012 Human bonematerial characterization: integrated imagingsurface investigation of male fragility fractures.Osteoporosis Int. 23, 1297 – 1309. (doi:10.1007/s00198-011-1688-9)

30. Sanni OD, Wagner MS, Briggs D, Castner DG,Vickerman JC. 2002 Classification of adsorbedprotein static ToF-SIMS spectra by principalcomponent analysis and neural networks.Surface Interface Anal. 33, 715 – 728. (doi:10.1002/sia.1438)

31. Heiss C et al. 2012 Induction of osteoporosiswith its influence on osteoporoticdeterminants and their interrelationships in ratsby DEXA. Med. Sci. Monit. 18, 199 – 207. (doi:10.12659/MSM.882895)

32. Vickerman J, Gilmore IS. 2009 Surface analysis:principal techniques. London, UK: John Wileyand Sons.

33. De Souza RA, Manfred M. 2007 Secondary ion massspectrometry (SIMS): a powerful tool for studyingmass transport over various length scales. Phys.Status Solidi C 4, 1785 – 1801. (doi:10.1002/pssc.200675227)

34. Sodhi RNS. 2004 Time of flight secondary ion massspectrometry (TOF-SIMS): versatility in chemical andimaging surface analysis. Analyst 129, 483 – 487.(doi:10.1039/b402607c)

35. Xiang Z, Markel MD. 1995 Bromodeoxyuridineimmunohistochemistry of EPON-embeddedundecalcified bone in a canine fracture healing

model. J. Histochem. Cytochem. 43, 629 – 635.(doi:10.1177/43.6.7769234)

36. Wagner M, Castner DG. 2004 Analysis ofadsorbed proteins by static time-of-flightsecondary ion mass spectrometry. Appl. Surface Sci.231 – 232, 366 – 376. (doi:10.1016/j.apsusc.2004.03.100)

37. Rey C, Combes C, Drouet C, Glimcher MJ. 2009 Bonemineral: update on chemical composition andstructure. Osteoporosis Int. 20, 1013 – 1021. (doi:10.1007/s00198-009-0860-y)

38. Fratzl P, Gupta HS, Paschalis EP, Roschger P.2004 Structure and mechanical quality of thecollagen-mineral nano-composite in bone.J. Mater. Chem. 14, 2115 – 2123. (doi:10.1039/b402005g)

39. Busse B, Hahn M, Soltau M, Zustin J, Puschel K,Duda GN, Amling M. 2009 Increased calciumcontent and inhomogeneity of mineralization renderbone toughness in osteoporosis: mineralization,morphology and biomechanics of human singletrabeculae. Bone 45, 1034 – 1043. (doi:10.1016/j.bone.2009.08.002)

40. Bousson V, Bergot C, Wu Y, Jolivet E, Zhou LQ,Laredo J-D. 2011 Greater tissue mineralizationheterogeneity in femoral neck cortex from hip-fractured females than controls. A microradiographicstudy. Bone 48, 1252 – 1259. (doi:10.1016/j.bone.2011.03.673)

41. Nagaraja S, Lin ASP, Guldberg RE. 2007 Age-relatedchanges in trabecular bone microdamage initiation.Bone 40, 973 – 980. (doi:10.1016/j.bone.2006.10.028)

42. Burr D. 2003 Microdamage and bone strength.Osteoporosis Int. 14, S67 – S72. (doi:10.1007/s00198-003-1476-2)

43. Seeman E. 2008 Bone quality: the material andstructural basis of bone strength. J. BoneMiner. Metab. 26, 1 – 8. (doi:10.1007/s00774-007-0793-5)

44. Sivas F, Gunesen O, Ozoran K, Alemdaroglu E.2007 Osteomalacia from Mg-containing antacid:a case report of bilateral hip fracture. Rheumatol.Int. 27, 679 – 681. (doi:10.1007/s00296-006-0273-6)

45. Goodship AE, Lanyon LE, McFie H. 1979Functional adaption of bone to increased stress. Anexperimental study. J. Bone Joint Surg. Am. 61,539 – 546.

46. Rubin CT, Lanyon LE. 1985 Regulation of bonemass by mechanical strain magnitude.Calcif. Tissue Int. 37, 411 – 417. (doi:10.1007/BF02553711)

47. Rachner TD, Khosla S, Hofbauer LC. 2011Osteoporosis: now and the future. Lancet377, 1276 – 1287. (doi:10.1016/S0140-6736(10)62349-5)

48. Campi G, Ricci A, Guagliardi A, Giannini C,Lagomarsino S, Cancedda R, Mastrogiacomo M,Cedola A. 2012 Early stage mineralization in tissueengineering mapped by high resolution X-raymicrodiffraction. Acta Biomater. 8, 3411 – 3418.(doi:10.1016/j.actbio.2012.05.034)

rsif.royalsocietypublishing.org

12

49. Lu HB, Campbell CT, Graham DJ, Ratner BD. 2000Surface characterization of hydroxyapatite and relatedcalcium phosphates by XPS and TOF-SIMS. Anal.Chem. 72, 2886 – 2894. (doi:10.1021/ac990812h)

50. Basle MF, Mauras Y, Audran M, Clochon P, Rebel A,Allain P. 1990 Concentration of bone elements inosteoporosis. J. Bone Miner. Res. 5, 41 – 47. (doi:10.1002/jbmr.5650050108)

51. Tzaphlidou M, Berillis P, Matthopoulos D. 2005Bone calcium, phosphorus detection by Auger

electron spectroscopy. Micron 36, 706 – 709. (doi:10.1016/j.micron.2005.05.013)

52. Tzaphlidou M, Zaichick V. 2004 Sex and agerelated Ca/P ratio in cortical bone of iliac crest ofhealthy humans. J. Radioanal. Nucl. Chem. 259,347 – 349. (doi:10.1023/B:JRNC.0000017316.20693.45)

53. Kourkoumelis N, Balatsoukas I, Tzaphlidou M. 2012Ca/P concentration ratio at different sites of normaland osteoporotic rabbit bones evaluated byAuger and energy dispersive X-ray spectroscopy.

J. Biol. Phys. 38, 279 – 291. (doi:10.1007/s10867-011-9247-3)

54. Fratzl-Zelman N, Roschger P, Misof BM, Nawrot-Wawrzyniak K, Poetter-Lang S, Muschitz C, Resch H,Klaushofer K, Zwettler E. 2011 Fragility fractures inmen with idiopathic osteoporosis are associatedwith undermineralization of the bone matrixwithout evidence of increased bone turnover. Calcif.Tissue Int. 88, 378 – 387. (doi:10.1007/s00223-011-9466-4)

J

RSocInterface

10:20130332