Micron 56 (2014) 29–36

Contents lists available at ScienceDirect

Micron
journal homepage: www.elsevier.com/locate/micron

Effect of strontium ranelate on bone mineral: Analysis of nanoscale

compositional changes
André L. Rossi a , Simona Moldovan b , William Querido c , Alexandre Rossi a ,
Jacques Werckmann d , Ovidiu Ersen b , Marcos Farina c,∗
a
Centro Brasileiro de Pesquisas Físicas, Xavier Sigaud 150, CEP 22290-180, Rio de Janeiro, Brazil
b
Institut de Physique et Chimie des Matériaux de Strasbourg, F-67081 Strasbourg cedex, France
c
Universidade Federal do Rio de Janeiro, Centro de Ciências da Saúde, Instituto de Ciências Biomédicas, Av. Carlos Chagas Filho, 373, bloco F, sala F2-027,
CEP 21941-599, Rio de Janeiro, Brazil
d
Instituto Nacional de Metrologia, Normalização e Qualidade Industrial, Estrada de Xerém 27, CEP 25245-390, Rio de Janeiro, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Strontium ranelate has been used to prevent bone loss and stimulate bone regeneration. Although stron-
Received 5 August 2013 tium may integrate into the bone crystal lattice, the chemical and structural modifications of the bone
Received in revised form when strontium interacts with the mineral phase are not completely understood. The objective of this
28 September 2013
study was to evaluate apatite from the mandibles of rats treated with strontium ranelate in the drink-
Accepted 28 September 2013
ing water and compare its characteristics with those from untreated rats and synthetic apatites with
and without strontium. Electron energy loss near edge structures from phosphorus, carbon, calcium and
Keywords:
strontium were obtained by electron energy loss spectroscopy in a transmission electron microscope. The
Strontium ranelate
Hydroxyapatite
strontium signal was detected in the biological and synthetic samples containing strontium. The relative
Bone mineral quantification of carbon by analyzing the CK edge at an energy loss of E = 284 eV showed an increase
Biomaterial in the number of carbonate groups in the bone mineral of treated rats. A synthetic strontium-containing
Biomineralization sample used as control did not exhibit a carbon signal. This study showed physicochemical modifications
in the bone mineral at the nanoscale caused by the systemic administration of strontium ranelate.
© 2013 Elsevier Ltd. All rights reserved.

1. Introduction of substitution) and cancellous bone (high level of substitution)

The anti-osteoporotic drug strontium ranelate can stimulate
bone regeneration (Dahl et al., 2001; Marie et al., 2001; Marie, 2006;
Boivin et al., 2012) by increasing the number of pre-osteoblast cells
and inhibiting bone resorption by osteoclast cells (Marie, 2006;
Bonnelye et al., 2008). Mineral density and the mechanical proper-
ties of osteoporotic bone tissue could also be positively affected by
strontium ranelate treatment, decreasing the risk of bone fractures
(Ammann et al., 2004, 2007; Reginster et al., 2008). Although sev-
eral positive effects of strontium have been shown in vitro (Barbara
et al., 2004; Bonnelye et al., 2008) and in vivo (Boivin et al., 1996;
Dahl et al., 2001; Ammann et al., 2004, 2007; Reginster et al., 2008;
Meunier et al., 2009; Roschger et al., 2010), the mechanism involved
in bone formation in the presence of strontium and the protection
against resorption remains poorly understood.
The amount of strontium incorporated into bone mineral is dose
dependent over a certain range, up to a plateau level (Dahl et al.,
2001). It is heterogeneously distributed in compact bone (low levels
∗ Corresponding author. Tel.: +55 21 2562 6393.
E-mail address: mfarina@icb.ufrj.br (M. Farina).
with a higher uptake in new bone tissue than old one (Boivin et al.,
1996, 2010; Farlay et al., 2005; Roschger et al., 2010). These dif-
ferences are thought to be related to the faster remodeling rate in
cancellous bone, allowing a higher incorporation of strontium into
the tissue (Dahl et al., 2001).
Two mechanisms of strontium incorporation in bone mineral
are currently accepted (Boivin et al., 1996; Dahl et al., 2001; Marie
et al., 2001; Boivin and Meunier, 2003; Bazin et al., 2011): (I)
a fast mechanism of surface exchange between blood and bone
mineral and (II) a slower mechanism of heteroionic substitution,
which results in the incorporation of strontium into the crystal lat-
tice replacing calcium. Regarding the chemical aspect of the first
mechanism, strontium would be surrounded by oxygen atoms if
adsorbed at the surface of apatite crystallites (Bazin et al., 2011).
Strontium could also be engaged in a hydrated poorly crystalline
apatite phase on the surfaces of the crystals linked to oxygen and
phosphate atoms (Cazalbou et al., 2005; Bazin et al., 2011). No struc-
tural changes in the lattice parameters would be expected for this
mechanism.
The second mechanism is based in the resorption/apposition of
bone and depends on osteoclast and osteoblast cells activity. In this
case, strontium would substitute for calcium in apatite crystals at

0968-4328/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.micron.2013.09.008
30 A.L. Rossi et al. / Micron 56 (2014) 29–36
either crystallographic calcium site I (Ca linked to OH− ) and/or II (Ca
linked to PO4 3− ) during bone remodeling (Terra et al., 2009; Bazin
et al., 2011). Physicochemical changes in the bone mineral would be
expected because strontium atoms are larger than calcium atoms.
In synthetic hydroxyapatites, the substitution of calcium by
strontium in the crystals caused several changes in the mineral
properties, including (1) the expansion of the lattice parameters (Li
et al., 2007; Terra et al., 2009), (2) an increase in the local disorder
of the lattice (Terra et al., 2009), (3) a decrease in the crystallinity
and crystal size (Verberckmoes et al., 2004; Li et al., 2007), (4) a
decrease in the number of (OH− ) sites (Verberckmoes et al., 2004;
Terra et al., 2009), (5) an increase in the CO3 2− content (Li et al.,
2007) and (6) an increase in the solubility (Christoffersen et al.,
1997).
In biological apatite, the substitution of calcium by strontium
has been described in several studies (Boivin et al., 1996; Boivin
and Meunier, 2003; Farlay et al., 2005; Li et al., 2010a). In a recent
work on patients treated with strontium ranelate for 5 years, an
increase was observed in the lattice parameters of apatite crystals
in the regions of bone remodeling (Li et al., 2010a). Based on the
variation of the crystallographic parameter along the c-axis, it was
inferred that up to 5% of the calcium atoms from the crystal lat-
tice had been substituted by strontium atoms. However, few works
have explored the effect of strontium in bone mineral at the level
of the apatite crystallites (Li et al., 2007, 2010c).
The objective of this study was to evaluate biological bone
apatite isolated from rats treated with strontium ranelate, com-
paring its characteristics at the nanoscale (e.g. strontium content,
morphology and shape of nanocrystals, Ca/P ratio, presence of
carbonate, etc.) with those of apatite from untreated rats and
of synthetic apatites with and without strontium. Details on the
bone mineral properties after strontium incorporation will help
to understand the mechanism of biomineralization, bone mineral
quality and the potential of strontium salts for clinical applications.
2. Materials and methods
2.1. Biological samples
Biological apatites were obtained from the mandibles of male
Wistar rats, 6 weeks old, treated with strontium ranelate (HAB -
Sr) and untreated rats (HAB , control group). The mandible bone
was chosen due to its high levels of strontium incorporation
(Oliveira et al., 2012). Animals were treated with strontium ranelate
(PROTOS® 2 g, Servier Laboratories) at a dose of 1800 mg/kg/day
in the drinking water for 90 days. The animals were killed, and
the mandible bones were harvested and cleaned manually. Bone
samples were further ground and rinsed with distilled water. The
mineral phase was extracted by removal of the organic matrix with
a 5% sodium hypochlorite solution for approximately 10 min in
continuous motion in a vortex mixer. After 2 min of centrifugation
at 14,000 rpm, the mineral pellets were washed three times with
Milli-Q water and twice with 100% ethanol, and they were dried
overnight at 40 ◦ C. The in vivo procedures were performed accord-
ing to the guidelines of the ethical committee on the use and care
of animals at Federal University of Rio de Janeiro.
2.2. Synthetic samples
Synthetic hydroxyapatite (HAS ) and strontium-containing syn-
thetic hydroxyapatite (HAS -Sr) samples were obtained in aqueous
solution. HAS and HAS -Sr were precipitated by dropwise addi-
tion of Ca(NO3 )2 or Ca(NO3 )2 + Sr(NO3 )2 solutions to a (NH4 )2 HPO4
solution. The pH was adjusted to 10 by NH4 OH addition and the syn-
thesis reactions were conducted at 90 ◦ C. The suspension was kept
in digestion for 3 h. The precipitate was then separated by filtration
and repeatedly washed with boiling deionized water, and it was
dried at 100 ◦ C for 24 h. The dried powder was manually ground,
and particles <200 ␮m were separated by sieving. Only HAS was
sintered at 1100 ◦ C in order to produce a stoichiometric hydroxy-
apatite sample. The theoretical chemical composition from HAS
and HAS -Sr were Ca10 (PO4 )6 (OH)2 and Ca9.5 (Sr0.5 )(PO4 )6 (OH)2 ,
respectively. Experimental calcium, strontium and phosphorus
concentrations were determined by X-ray fluorescence using a PW
2400 Sequential (Philips) apparatus. The Ca/P and Sr/Ca ratios in
HAS and HAS -Sr were used as references to calibrate the energy
dispersive X-ray (EDX) spectra from the biological samples.
To investigate the carbon signal from the samples by elec-
tron energy loss spectroscopy (EELS) in the transmission electron
microscope (TEM), a B-type carbonated (HAS -Carbo) apatite was
included in this work. This sample contains CO3 groups in
the PO4 position, Ca10−x (PO4 )6−x (CO3 )x (OH)2−x−2y (CO3 )y (with
0 ≤ x ≤ 1.1 and 0 ≤ y ≤ 0.2). The HAS -Carbo sample was produced
by the same procedure by addition of Ca(NO3 )2 solution to
(NH4 )2 HPO4 + (NH4 )2 CO3 solutions at pH 12. The precipitate was
obtained at 90 ◦ C and dried at 100 ◦ C for 24 h. The CO3 content from
HAS -Carbo (1.3 wt.%) was determined by X-ray fluorescence using
a PW 2400 Sequential (Philips) apparatus.
In addition to HAS , HAS -Sr and HAS -Carbo, the commercial drug
strontium ranelate (PROTOS® 2 g, Servier Laboratories) was ana-
lyzed by EELS to evaluate the experimental conditions for the
detection of the M and L ionization edges from strontium.
2.3. Analytical methods and experimental conditions
Powders obtained from the synthetic and biological samples
were deposited onto lacey formvar/carbon TEM grids (Ted Pella,
Inc.). The method consisted of spreading the sample powders
between two light microscope glass slides and putting the TEM
grids previously exposed to a plasma cleaner in contact with the
powder.
The samples were analyzed with a Cs-corrected field emission
gun (FEG) Jeol 2100F (TEM/STEM), operated at 200 kV. The micro-
scope was equipped with an EDX detector (NORAN System) and a
Gatan imaging filter (GIF TRIDIEM) for elemental analysis and EELS,
respectively.
Electron diffraction patterns were obtained in the select area
electron diffraction (SAED) mode from regions with similar surface
areas and using the same camera length. The high resolution trans-
mission electron microscopy (HRTEM) images were digitized with a
2048 × 2048 CCD camera (Gatan). Fast Fourier transform (FFT) was
applied to the HRTEM images using the Digital Micrograph software
(Gatan, Inc.) for a better interpretation of the lattice images.
X-ray microanalyses were obtained in the Jeol 2100F using scan-
ning transmission electron microscopy (STEM) mode, from regions
of 50 nm × 50 nm. In STEM mode, the beam never stops in the
same place more than 0.05 s (high scanning speed of the focused
beam was used), allowing analysis of small areas (approximately
50 nm × 50 nm) with minimal damaging. Ca/P and Sr/Ca relative
quantifications were performed using the CaK , PK and SrK peaks by
considering the thin-film approximation. The HAS and HAS -Sr sam-
ples were used to calibrate the Ca/P and Sr/Ca ratios, respectively.
The energy-loss signal obtained by EELS provides elemental
identification related to the ionization of the core–shell electrons
and information about the electronic structure of the specimen
atoms that reveals details about their bonding state and the
nearest-neighbor atomic structure (Colliex et al., 1991; Garvie et al.,
1994; Williams and Carter, 2009). EELS spectra were obtained in
STEM mode (for energy losses between 150 and 600 eV) and in TEM
mode (for higher energy losses >1800 eV) with a 3 mm spectrom-
eter entrance aperture. The energy resolution was approximately
 A.L. Rossi et al. / Micron 56 (2014) 29–36
Fig. 1. Transmission electron microscopy (TEM) and selected area electron diffraction (SAED) from: (A) synthetic hydroxyapatite (HAS ); (B) strontium-containing synthetic
hydroxyapatite (HAS -Sr); (C) biological apatite (HAB ) and (D) biological apatite from rats treated with strontium ranelate (HAB -Sr). SAED were obtained with the same
microscope operating conditions (high voltage, camera length and apertures). The arrows indicate the (2 1 1) planes of hydroxyapatite. HAs presented large crystals with
defined facets. HAB , HAB -Sr and HAS -Sr had crystallites with plate-like morphology.
1 eV and was measured by the full width at half maximum of the
zero loss peak. The EELS analyses were performed in 50 nm × 50 nm
regions in accumulation spectrum mode and with the beam
scanning speed adjusted to 0.05 s. Each final spectrum was the
summation of the spectra obtained during scanning period from
the 50 nm × 50 nm region. EELS from higher energy losses were
performed in TEM mode to obtain more signal to investigate the
strontium L2,3 edge. The spectra were obtained from thin regions
with (t/) < 0.4 (t, thickness; , inelastic mean free path of the
electron in the material). The carbon/calcium atomic ratios were
performed using Digital Micrograph software (Gatan, Inc.) with the
background signal previously removed by fitting with a power law.
The ionization cross sections model used was the Hartree-Slater.
3. Results
Bright-field TEM images and SAED were obtained from HAB ,
HAB -Sr, HAS and HAS -Sr (Fig. 1). Large crystal sizes with well-
defined facets in the HAS samples were observed (Fig. 1A). This
characteristic is related to the sinterization at 1100 ◦ C. The heating
process induces fusion of apatite crystallites, thus creating larger
grains. The HAS -Sr crystals (Fig. 1B) presented a plate-like mor-
phology similar to bone crystals. The biological samples contained
agglomerates of crystals with plate-like morphology. No remark-
able differences between the HAB (Fig. 1C) and HAB -Sr (Fig. 1D)
samples were noted based on the morphology, crystallites disper-
sion and diffraction pattern (see below).
The SAED patterns from all of the samples were obtained for
similar conditions (beam energy, aperture size and camera length)
to allow a qualitative comparison of sample crystallinities (Fig. 1,
inserts). It was observed that the HAS sample was the most crys-
talline (Fig. 1A, insert). The diffracted patterns observed indicate
a superposition of monocrystals. The presence of individual spots
in the diffraction pattern corresponds to an interference of elec-
trons coming from crystallographically similar scattering regions
31
(lattice regions). No diffuse rings indicative of amorphous phases
and/or the presence of smaller crystallographic domains oriented
randomly were observed for the field areas selected. HAS -Sr was
polycrystalline (Fig. 1B, insert). A weak smooth ring was observed
superimposed to individual spots corresponding to the (2 1 1)
planes (Fig. 1B, arrow). No difference was observed between the
diffraction patterns from HAB and HAB -Sr (Fig. 1C and D), inserts,
respectively. In both diffractograms, complete diffuse rings were
observed, which indicates lower crystallinity when compared with
the synthetic samples.
HRTEM images from the four samples were obtained (Fig. 2) to
compare the sizes of the crystallographic domains. The HAS sam-
ple had large crystallographic domains when observed by HRTEM
(Fig. 2A). The corresponding FFT from HAS (Fig. 2A, insert) showed
a perfect crystal oriented in the 4 3 1 zone axis. In general, besides
HAs, the crystallographic domains from HAS -Sr (Fig. 2B) were also
larger than in the biological samples (Fig. 2C and D).
The Ca/P and Sr/Ca ratios from the samples were obtained by
EDX in STEM mode. Fig. 3 shows superimposed EDX spectra from
the four samples (spectra were normalized to the P peaks; only
one typical spectrum from each sample is presented). The L and K
peaks from strontium were clearly observed in the HAB -Sr and HAS -
Sr samples. The Ca/P ratios obtained are presented in Table 1 (all
data were normalized by the values obtained from the standard).
The HAS sample used as standard had a Ca/P ratio of 1.66. The other
samples had lower, but similar Ca/P ratios (approximately 1.5). The
Sr/Ca ratios were calculated from the spectra. The results showed
a relative Sr/Ca ratio of 0.13 in HAB -Sr considering that in HAS -Sr,
used as standard, the Sr/Ca ratio was 0.05 (Table 1).
The EELS analysis was performed in synthetic and biological
HA samples (Figs. 4–6). Two energy-loss ranges were analyzed: (1)
energy losses between 100 and 600 eV (Fig. 4 and 5), which includes
the phosphorus L1,2,3 ionization edge (PL -edge), the carbon K ion-
ization edge (CK -edge), the calcium L1,2,3 ionization edge (CaL -edge)
and the oxygen K ionization edge (OK -edge) and (2) higher energy
32 A.L. Rossi et al. / Micron 56 (2014) 29–36

Fig. 2. High resolution transmission electron microscopy (HRTEM) and Fast Fourier Transform (FFT) from: (A) HAS ; (B) HAS -Sr; (C) HAB and (D) HAB -Sr. Synthetic samples
presented larger crystalline domains when compared to the biological ones.

(ionization of 2 s electron) at 190 eV. A small additional peak was

observed in the HAB -Sr below the PL1 edge at approximately 180 eV
(Fig. 5A, arrow). The broad peak at 153 eV was straighter in HAS than
the other samples.
To compare the CK near-edge structure between synthetic and
biological apatites, a synthetic carbonate apatite (HAS -Carbo) was
included. The energy loss near-edge structure (ELNES) from 250 eV
to 320 eV was obtained for all of the samples and compared
with standard spectra from a different carbon polymorph (Garvie
et al., 1994), in particular the carbonate polymorph (Fig. 5B). The
carbonate sites are formed by C O bonds (1s → ␲* transition)
and C O bonds (1s → ␴* transition) where 1s → ␲* and 1s → ␴*

Fig. 3. Superimposed energy dispersive X-ray spectra (EDX) from HAB , HAS , HAB -
Sr and HAS -Sr. A clear strontium peak can be observed in the HAB -Sr and HAS -Sr
samples. The calculated Ca/P and Sr/Ca ratios are presented in Table 1.

losses, from 1900 to 2050 eV (Fig. 6) to analyze the strontium L2,3

ionization edge (SrL -edge).
The energy-loss spectrum of phosphorus (Fig. 5A) shows the
PL2,3 -edges (ionization of 2p electrons in the L shell) that are char-
acterized by two well-defined peaks at 132 eV and 140 eV, a broad
peak at 153 eV and a less intense peak related to the PL1 -edge

Table 1
Ca/P and Sr/Ca (at.%) obtained from EDX spectra in TEM and STEM modes. The
integrations were performed with the signals under CaK , PK and SrK . The Ca/P and
Sr/Ca ratios from HAS and HAS -Sr respectively, were previously determined by X-ray
fluorescence and used to calibrate the results.

STEM-EDX

Ca/P Sr/Ca N
Fig. 4. Electron energy-loss spectra (EELS) from HAB , HAS , HAB -Sr and HAS -Sr sam-
HAS 1.66 ± 0.06 – 5
ples in the energy range of 100–650 eV obtained in scanning transmission electron
HAS -Sr 1.49 ± 0.02 0.05 ± 0.01 5
microscopy (STEM) mode. The spectra were spliced to show the full energy-loss
HAB 1.49 ± 0.06 – 5
range studied. In this figure, the background signal was not removed from the
HAB -Sr 1.51 ± 0.10 0.13 ± 0.01 5
spectra.
 A.L. Rossi et al. / Micron 56 (2014) 29–36 33

Fig. 5. EELS spectra from HAB , HAB -Sr, HAS and HAS -Sr samples. The spectra were obtained from 50 nm × 50 nm regions. In all of the spectra, the background signal was
removed. (A) Phosphorus L1,2,3 -edges from 125 eV to 250 eV energy losses. The arrow shows an additional peak at 180 eV in HAB -Sr. (B) Carbon K-edge at 280 eV. Five samples
are presented: HAB , HAB -Sr, HAS -Sr, a synthetic carbonate containing hydroxyapatite (HAS -Carbo) and the strontium ranelate medicine. No evidence of carbonate peaks was
observed in the HAS (not shown) and HAS -Sr samples. Spectra were normalized considering the CaL peak as reference. The arrow shows the SrM3 -edge in the strontium
ranelate. (C) Calcium L2,3 -edge at 335 eV (L3 ) and 365 eV (L2 ) energy losses; (D) Energy-loss near edge structure (ELNES) from CaL2,3 and CaL1 edges. Arrows “a” to “d” indicate
the peaks of the energy losses observed in this energy range.

transitions are represented by two intense peaks at 289 and 299 eV,
respectively. The carbon edge from the carbonate groups cannot be
confused with the carbon edge from amorphous carbon because
the latter presents peaks at 284 and 289 eV and has a different
near edge structure. We also included in this energy loss range
the spectrum from the strontium ranelate medicine because the
strontium M3 edge is located just before the carbon edge (Fig. 5B,
arrow).
Fig. 5B shows the CK -edge from HAB , HAB -Sr, HAS -Sr, HAS -
Carbo and strontium ranelate. No evidence of carbonate peaks were
observed in the HAS (not shown) and HAS -Sr samples. HAB , HAB -Sr
and HAS -Carbo clearly presented two maxima at 289 and 299 eV,
indicative of carbonate sites. The carbon peaks related to the car-
bonate group were more intense and well defined in HAB -Sr. The
carbon/calcium ratio semi-quantification showed more carbon in
HAB -Sr (0.19 ± 0.02) than in HAB (0.12 ± 0.02) and HAS -Carbo (0.11
±0.03). Three particles containing thin regions from each sample
were measured.
A small broad peak at 270 eV corresponding to the SrM3 -edge
(arrow in Fig. 5B) was observed in the strontium ranelate sample
and in the HAB -Sr samples with much lower intensity.
Calcium L3 (346 eV) and L2 (350 eV) are related to L3
2p3/2 → 3d3/2 , 3d5/2 and L2 2p1/2 → 3d3/2 transitions. The two peaks
are also called “white lines” and are well resolved by approximately
3 eV. The ELNES from 365 to 450 eV was found to be richer in details.
At least three broad peaks in addition to the CaL1 edge (arrows
“a” to “d” in Fig. 5B) could be observed. Peaks “a” and “b” were
more defined in HAS and HAB than in HAS -Sr and HAB -Sr. Peak “c,”
which corresponds to CaL1 , was observed in all of the samples. Peak
“d” was observed in HAB -Sr and with lower intensities in HAS and
HAS -Sr. As for the Phosphorus, the curves are not superimposed for
the energies exceeding the main peaks. This is due to the uncer-
tainties given by the extraction of the background. Although the
functions used for the background approximation are the same
and the intervals of interpolation as well, the extrapolation and
subsequent background subtraction lead to slightly different signal
aspects.
The oxygen energy loss at 532 eV is related to the OK edge, which
represents the ionization of the 1s electron in the K shell (1s → 2p
transitions). All of the spectra presented two peaks: the more
intense at 537 eV and a broad peak centered at 557 eV (Fig. 6A). A
small shoulder on the right side of the more intense peak at 543 eV
was observed in all of the OK edges. No remarkable differences
between the samples were observed.
The high energy-loss SrL2,3 -edge of strontium from HAB -Sr, HAS -
Sr and strontium ranelate drug was obtained in TEM mode. The
strontium L3 -edge and L2 -edge (ionization of 2p 1/2 and 3/2 elec-
trons in the L shell) are characterized by two peaks at 1940 eV (SrL3 )
and 2008 eV (SrL2 ) and a broad less intense peak between these two
peaks (Fig. 6B). The samples presented the SrL2,3 -edge with similar
structure.
4. Discussion
Until now, most analytical works on hydroxyapatite doped with
strontium were achieved using techniques that analyzed pow-
dered samples or large mineral regions, which means that the
results obtained are an average from many crystal samples (Farlay
34 A.L. Rossi et al. / Micron 56 (2014) 29–36
Fig. 6. STEM- and TEM-EELS from HA samples and the commercial strontium
ranelate medicine. (A) STEM-EELS from oxygen K-edge at an energy loss of 532 eV.
(B) TEM-EELS at higher energy losses (1900–2050 eV) showing SrL2,3 edges from the
samples containing strontium.
et al., 2005; Ni et al., 2006; Bünger et al., 2010; Bazin et al.,
2011; Oliveira et al., 2012). Since the basic mineral unity of bone
is a plate-like crystallites whose dimensions are on the order of
50 nm × 25 × 4 nm (Weiner and Wagner, 1998), belief that fur-
ther understanding of the physical and chemical characteristics of
bone minerals should be linked to the analysis of crystals at the
nanoscale. Here, we performed a multiple-elements spectroscopy
investigation of biological and synthetic hydroxyapatites at this
scale. Strontium ranelate has been used in the last decade as a
promising drug against bone lost in osteoporosis (Dahl et al., 2001;
Marie et al., 2001; Marie, 2006). However, the influence of cal-
cium substitution by strontium in bone mineral is not completely
known. EDX spectroscopy was used to obtain Ca/P ratios from sam-
ples in scanning transmission mode allowing analysis of a few tens
of nanometers (50 nm × 50 nm). The influence of irradiation and
temperature to quantify Ca/P ratios from calcium phosphates by
STEM-EDX was previously investigated (Benhayoune et al., 2001).
It was demonstrated that by using a short acquisition time, the Ca/P
ratio does not vary, even when using a relatively high radiation
dose.
In the present work, HAB , HAB -Sr and HAS -Sr had similar
Ca/P ratios (approximately 1.5) while HAS presented the highest
ratio (1.66). HAS was previously analyzed by X-ray fluorescence
and was used as a standard to calibrate the Ca/P ratios from
the other samples. The results obtained by EDX indicate that
the strontium ranelate treatment did not alter significantly the
calcium/phosphate stoichiometry in the mandible bone mineral
although relatively high levels of strontium were detected (Table 1).
Regarding the synthetic samples, the Ca/P ratio from HAS -Sr
decreased comparing to HAS . These results support previous works,
which showed that only a small amount of strontium is incor-
porated in the lattice of biological apatite crystals (Boivin et al.,
1996; Boivin and Meunier, 2003; Farlay et al., 2005; Li et al.,
2010a). It is important to mention that HAB -Sr was previously
analyzed by synchrotron X-ray diffraction and presented slight
measurable enlargement of the unit cell (Oliveira et al., 2012), indi-
cating an incorporation of strontium in the lattice in some degree.
However, by the present results, it is suggested that most stron-
tium was probably adsorbed at the surface of apatite crystallites.
Sr/Ca semi-quantifications in the biological bone mineral samples
were consistent with previous works using EDX in which monkeys
treated with strontium ranelate at a lower dose than that used
in this study (1250 and 1800 mg/kg/day, respectively) presented
a Sr/Ca ratio of approximately 0.09 (Farlay et al., 2005). It cannot
be discarded the fact that some changes on the structure of bone
crystallites may occur after total or partial removal of the non-
collagenous material (e.g. the crystallite “ghosts”) during sample
preparation (Carter et al., 2002). In the present work, nitrogen and
amorphous carbon (indicative of organic matter) were not found
by EDX and/or EELS, indicating that the observed changes in the
spectra of isolated mineral phase from HAB -Sr were caused by the
direct association between strontium and the crystallites.
Changes in the crystallinity of synthetic apatites were observed
with the incorporation of strontium in the crystal lattices and were
dependent on the concentration of the dopant (Li et al., 2007). Syn-
thetic HA-Sr doped with 0.3% and 1.5% of strontium had SAED
comparable to pure hydroxyapatite, while 15% HA-Sr presented
a SAED with a diffuse background, suggesting the existence of
an amorphous phase (Li et al., 2007). The results obtained from
SAED in the present work concerning HAB and HAB -Sr did not
indicate changes in the crystallinity of the samples. The electron
diffraction patterns from both samples obtained in the same con-
ditions had similar profiles, indicating that the strontium content
did not cause measurable amorphization of the biological sam-
ple.
The HAS -Sr presented an electron diffraction pattern contain-
ing spots superimposed to a diffuse ring. The presence of intense
diffracted dots compared to the corresponding biological samples
suggests a higher proportion of crystalline phase or the presence
of slightly larger individual crystals even if the Ca/P ratios are sim-
ilar in HAS -Sr, HAB and HAB -Sr samples as observed by EDX. The
diffraction pattern from HAS was indicative of even larger single
crystals, as confirmed by the TEM and HRTEM images. The analysis
of samples by HRTEM and the corresponding FFT were consistent
with the SAED. The synthetic samples presented relatively large
crystallographic domains when compared to the biological sam-
ples.
The use of EELS to analyze hydroxyapatite doped with strontium
is of particular interest because the technique provides informa-
tion about the electronic structure of chemical bonds (Colliex et al.,
1991; Garvie et al., 1994; Williams and Carter, 2009) and can help
to elucidate the apatite structure after strontium incorporation by
comparison with previously known samples. These results have
potential practical applications because drugs and biomaterials
containing strontium are used to prevent bone lost and regenerate
bone defects.
 A.L. Rossi et al. / Micron 56 (2014) 29–36
4.1. Phosphorus L-edge
The analysis of the phosphorus PL1 , PL2 and PL3 ELNES from HAB ,
HAB -Sr, HAS and HAS -Sr showed differences mainly between PL2
(140 eV) and PL1 (190 eV). An additional broad peak was observed
at 187 eV in HAB -Sr (arrow in Fig. 5A). A less expressive peak at the
same position was observed in the HAS -Sr sample. Some explana-
tions could be suggested for the difference observed in the PL ELNES
from the strontium-containing samples (HAB -Sr and HAS -Sr). Pre-
vious works using FTIR (Fourier transform infrared spectroscopy)
showed that the incorporation of strontium into synthetic hydroxy-
apatite increased the local disorder of the crystal lattice around
the phosphate sites (Terra et al., 2009). These structural changes
observed by infrared spectroscopy could account for the differences
observed in the ELNES of the PL edge. Alternatively, the broad peak
at 187 eV could be associated with the energy loss from the SrM4 -
edge at approximately 185 eV, which is partially superposed to the
PL signal. In this case, the near edge structure of the edges would not
necessarily be related to modifications in the phosphate electronic
structure but with the presence of the SrM4 signal.
4.2. Carbon K-edge
The identification of the carbon CK edge by EELS in calcium
phosphates is of great interest for several reasons: (1) EELS associ-
ated to the TEM discriminates carbon polymorphs at the nanoscale
as amorphous, carbonate and graphite because each of these has
a specific near edge structure around CK (Garvie et al., 1994;
Mansot et al., 2003). (2) Carbonate is a natural component of bone
(Dorozhkin, 2009). CO3 2− ions substitute PO4 3− (B-substitution)
and OH− ions (A-substitution) in the apatite lattice of bone min-
eral. The relative element quantification (carbon/calcium atomic
ratio) of the CK -edge observed in the spectra shown in this study
indicates a higher quantity of carbon in the form of carbonate in
the HAB -Sr sample when compared to HAB . The increase in the
carbonate sites in HAB -Sr compared to HAB suggests an increase
in the number of carbonate groups in the lattice structure of the
biological apatite treated with strontium ranelate. The absence of
carbonate in HAs -Sr evaluated by EELS indicates that the carbon-
ate incorporated in bone mineral was probably not co-precipitated
with strontium. An additional mechanism in the process of biomin-
eralization should be present. The increase of carbonate sites could
be related to the strain produced in the lattice structure of apatite by
strontium substitutions, allowing for the accommodation of more
carbonate ions (Li et al., 2007). The substitution of phosphate by the
carbonate in synthetic hydroxyapatite results in a decrease in the
crystalline structure and an increase in solubility of crystals (Yao
et al., 2009; Boanini et al., 2010). The crystal unit cell contracts in
the a-axis (Harries et al., 1987) and increases in the c-axis (Baig
et al., 1999).
Recently, calcium phosphate biomaterials containing strontium
have been used as bone implants due to their specific proprieties in
bone regeneration (Landi et al., 2007; Capuccini et al., 2008; Li et al.,
2010b; Ni et al., 2010). Quantification of carbonate at the nanoscale
by EELS may have practical applications since the technique allows
for the local investigation of regenerated bone tissue surrounding
implants containing strontium.
4.3. Strontium ML-edges
The M-edge from strontium was observed in the strontium
ranelate sample and in the HAB -Sr sample. The reason why this
particular peak was not observed in the HAS -Sr sample was pos-
sibly because the concentration of the dopant in the synthetic
sample was too low (5%) and the M-edge from strontium is rel-
atively small and broad, particularly when compared to the carbon
35
K-edge, which is approximately 10 eV above. Even in the strontium
ranelate spectrum, the intensity of the strontium M-edge is much
lower than the carbon K-edge. However, the SrL2,3 peaks obtained
at a high energy loss (1900 eV) are sharper. In the three samples
analyzed (HAS -Sr, HAB -Sr and strontium ranelate), the SrL2,3 peaks
presented similar structures, indicating the presence of the dopant.
4.4. Calcium L-edge
We analyzed the CaL2 and CaL3 edges from 345 to 355 eV (Fig. 5C)
and the region containing the CaL1 edge from 360 to 400 eV (Fig. 5D).
CaL2 and CaL3 edges did not present remarkable differences among
the HA samples. We have investigated the relative intensity (peak
integration) and the splitting (Mahamid et al., 2008) between the
peaks. Considering the spectra analyzed (approximately 10 for each
sample), we concluded that these features could not be used to dis-
criminate samples. However, a previous study on synthetic calcium
phosphates performed with EELS (Liou et al., 2004) indicated that
apatite samples containing low Ca/P ratio had low CaL2,3 intensity
counts compared to samples with high Ca/P ratios. It was sug-
gested that the structure disorder from the apatite sample, which
is related to the stoichiometry and the Ca/P ratio, plays an impor-
tant role in the excitation counts in EELS. This relationship between
excitation counts and the Ca/P ratio was also observed by X-ray
absorption (XANES) (Liou et al., 2004). Based on these findings,
the same authors suggested that different chemical and biologi-
cal properties between apatite samples could be distinguished by
EELS using CaL2,3 edges. This was not the case in the present work.
ELNES from CaL2,3 was found to be rich in relevant details. At
least two peaks below CaL1 edge and one above CaL1 edge were
observed. For all of the spectra analyzed, we observed some sys-
tematic structural differences between the samples. The first two
peaks above CaL2 (arrows a and b in Fig. 5D) were more resolved
in samples HAB and HAS , which do not contain strontium. Further-
more, the peak “d” observed above CaL1 was more pronounced in
HAB -Sr. These features, in addition to the peak observed below PL1
edge (arrow in Fig. 5A) and the SrM3 edge (Fig. 5D), could be used
to distinguish biological samples doped with strontium.
4.5. Oxygen K-edge
The energy loss near edge structure of the oxygen edge at 532 eV
was studied previously in calcium phosphate samples (Liou et al.,
2004). It was shown that ␤-tricalcium phosphate (␤-TCP) samples
exhibited an asymmetric peak with a maximum at 538 eV, while
stoichiometric hydroxyapatite the same peak had a splitting result-
ing in two distinct peaks (Gregori et al., 2006). It was not possible
in this work to correlate the near edge structure of the ionization
edge with the particular characteristics of the different samples.
5. Conclusion
The Ca/P ratio and the crystallinity of HAB -Sr did not change
when compared with the biological control group suggesting that
the majority of the detected strontium content was not incorpo-
rated in the crystal lattice of the mineral. On the other hand, the
energy loss near edge structure of phosphorus and calcium from
HAB -Sr indicated physicochemical modifications relative to the ref-
erence samples. Semi-quantification of carbon from EELS spectra
showed an increase in the carbonate content in HAB -Sr when com-
pared to HAB . The increase of carbonate may contribute to changes
in the lattice parameters. The influence of strontium in the syn-
thetic sample was different from the biological one. Comparing to
HAS , HAS -Sr presented a decrease in the crystallinity and in the Ca/P
ratio. Carbonate was not detected in HAS -Sr suggesting that stron-
tium induced the carbonate incorporation in crystal lattice of the
36 A.L. Rossi et al. / Micron 56 (2014) 29–36
bone mineral (HAB -Sr) and there was not a co-precipitation with
strontium.
Acknowledgments
We thank CNPq, CAPES, FAPERJ and INMETRO Brazilian institu-
tions for financial support and/or facilities. This work is part of a
French-Brazilian cooperation (CAPES-COFECUB).
References
Ammann, P., Badoud, I., Barraud, S., Dayer, R., Rizzoli, R., 2007. Strontium ranelate
treatment improves trabecular and cortical intrinsic bone tissue quality, a deter-
minant of bone strength. J. Bone Miner. Res. 22, 1419–1425.
Ammann, P., Shen, V., Robin, B., Mauras, Y., Bonjour, J.-P., Rizzoli, R., 2004.
Strontium ranelate improves bone resistance by increasing bone mass
and improving architecture in intact female rats. J. Bone Miner. Res. 19,
2012–2020.
Baig, A.A., Fox, J.L., Young, R.A., Wang, Z., Hsu, J., Higuchi, W.I., Chhettry, A., Zhuang, H.,
Otsuka, M., 1999. Relationships among carbonated apatite solubility, crystallite
size, and microstrain parameters. Calcif. Tissue Int. 64, 437–449.
Barbara, A., Delannoy, P., Denis, B.G., Marie, P.J., 2004. Normal matrix mineralization
induced by strontium ranelate in MC3T3-E1 osteogenic cells. Metabolism 53,
532–537.
Bazin, D., Daudon, M., Chappard, C., Rehr, J.J., Thiaudière, D., Reguer, S., 2011. The
status of strontium in biological apatites: an XANES investigation. J. Synchrotron
Radiat. 18, 912–918.
Benhayoune, H., Charlier, D., Jallot, E., Laquerriere, P., Balossier, G., Bonhomme, P.,
2001. Evaluation of the Ca/P concentration ratio in hydroxyapatite by STEM-
EDXS: influence of the electron irradiation dose and temperature processing. J.
Phys. D: Appl. Phys. 34, 141.
Boanini, E., Gazzano, M., Bigi, A., 2010. Ionic substitutions in calcium phosphates
synthesized at low temperature. Acta Biomater. 6, 1882–1894.
Boivin, G., Deloffre, P., Perrat, B., Panczer, G., Boudeulle, M., Mauras, Y., Allain, P.,
Tsouderos, Y., Meunier, P.J., 1996. Strontium distribution and interactions with
bone mineral in monkey iliac bone after strontium salt (S 12911) administration.
J. Bone Miner. Res. 11, 1302–1311.
Boivin, G., Doublier, A., Farlay, D., 2012. Strontium ranelate–a promising therapeutic
principle in osteoporosis. J. Trace Elem. Med. Biol. 26, 153–156.
Boivin, G., Farlay, D., Khebbab, M.T., Jaurand, X., Delmas, P.D., Meunier, P.J., 2010.
In osteoporotic women treated with strontium ranelate, strontium is located
in bone formed during treatment with a maintained degree of mineralization.
Osteoporosis Int. 21, 667–677.
Boivin, G., Meunier, P.J., 2003. The mineralization of bone tissue: a forgotten dimen-
sion in osteoporosis research. Osteoporosis Int. 14, 19–24.
Bonnelye, E., Chabadel, A., Saltel, F., Jurdic, P., 2008. Dual effect of strontium ranelate:
stimulation of osteoblast differentiation and inhibition of osteoclast formation
and resorption in vitro. Bone 42, 129–138.
Bünger, M.H., Oxlund, H., Hansen, T.K., Sørensen, S., Bibby, B.M., Thomsen, J.S.,
Langdahl, B.L., Besenbacher, F., Pedersen, J.S., Birkedal, H., 2010. Strontium and
bone nanostructure in normal and ovariectomized rats investigated by scanning
small-angle X-ray scattering. Calcif. Tissue Int. 86, 294–306.
Capuccini, C., Torricelli, P., Sima, F., Boanini, E., Ristoscu, C., Bracci, B., Socol, G.,
Fini, M., Mihailescu, I.N., Bigi, A., 2008. Strontium-substituted hydroxyapatite
coatings synthesized by pulsed-laser deposition: in vitro osteoblast and osteo-
clast response. Acta Biomater. 4, 1885–1893.
Carter, D.H., Scully, A.J., Heaton, D.A., Young, M.P.J., Aaron, J.E., 2002. Effect of depro-
teination on bone mineral morphology: implications for biomaterials and aging.
Bone 31, 389–395.
Cazalbou, S., Eichert, D., Ranz, X., Drouet, C., Combes, C., Harmand, M.F., Rey, C., 2005.
Ion exchanges in apatites for biomedical application. J. Mater. Sci.: Mater. Med.
16, 405–409.
Christoffersen, J., Christoffersen, M.R., Kolthoff, N., Bärenholdt, O., 1997. Effects of
strontium ions on growth and dissolution of hydroxyapatite and on bone min-
eral detection. Bone 20, 47–54.
Colliex, C., Manoubi, T., Ortiz, C., 1991. Electron-energy-loss-spectroscopy near-edge
fine structures in the iron-oxygen system. Phys. Rev. B 44, 11402–11411.
Dahl, S.G., Allain, P., Marie, P.J., Mauras, Y., Boivin, G., Ammann, P., Tsouderos, Y.,
Delmas, P.D., Christiansen, C., 2001. Incorporation and distribution of strontium
in bone. Bone 28, 446–453.
Dorozhkin, S., 2009. Calcium orthophosphates in nature, biology and medicine.
Materials 2, 399–498.
Farlay, D., Boivin, G., Panczer, G., Lalande, A., Meunier, P.J., 2005. Long-term strontium
ranelate administration in monkeys preserves characteristics of bone mineral
crystals and degree of mineralization of bone. J. Bone Miner. Res. 20, 1569–1578.
Garvie, L.A.J., Craven, A.J., Brydson, R., 1994. Use of electron-energy loss near-edge
fine structure in the study of minerals. Am. Mineral. 79, 411–425.
Gregori, G., Kleebe, H.-J., Mayr, H., Ziegler, G., 2006. EELS characterisation of ␤-
tricalcium phosphate and hydroxyapatite. J. Eur. Ceram. Soc. 26, 1473–1479.
Harries, J.E., Hasnain, S.S., Shah, J.S., 1987. EXAFS study of structural disorder in
carbonate-containing hydroxyapatites. Calcif. Tissue Int. 41, 346–350.
Landi, E., Tampieri, A., Celotti, G., Sprio, S., Sandri, M., Logroscino, G., 2007. Sr-
substituted hydroxyapatites for osteoporotic bone replacement. Acta Biomater.
3, 961–969.
Li, C., Paris, O., Siegel, S., Roschger, P., Paschalis, E.P., Klaushofer, K., Fratzl, P., 2010a.
Strontium is incorporated into mineral crystals only in newly formed bone dur-
ing strontium ranelate treatment. J. Bone Miner. Res. 25, 968–975.
Li, Y., Li, Q., Zhu, S., Luo, E., Li, J., Feng, G., Liao, Y., Hu, J., 2010b. The effect of strontium-
substituted hydroxyapatite coating on implant fixation in ovariectomized rats.
Biomaterials 31, 9006–9014.
Li, Z., Lu, W.W., Deng, L., Chiu, P.K.Y., Fang, D., Lam, R.W.M., Leong, J.C.Y., Luk, K.D.K.,
2010c. The morphology and lattice structure of bone crystal after strontium
treatment in goats. J. Bone Miner. Metab. 28, 25–34.
Li, Z.Y., Lam, W.M., Yang, C., Xu, B., Ni, G.X., Abbah, S.A., Cheung, K.M.C., Luk,
K.D.K., Lu, W.W., 2007. Chemical composition, crystal size and lattice structural
changes after incorporation of strontium into biomimetic apatite. Biomaterials
28, 1452–1460.
Liou, S.-C., Chen, S.-Y., Lee, H.-Y., Bow, J.-S., 2004. Structural characterization of nano-
sized calcium deficient apatite powders. Biomaterials 25, 189–196.
Mahamid, J., Sharir, A., Addadi, L., Weiner, S., 2008. Amorphous calcium phosphate
is a major component of the forming fin bones of zebrafish: indications for an
amorphous precursor phase. Proc. Natl. Acad. Sci. 105, 12748–12753.
Mansot, J.L., Golabkan, V., Romana, L., Césaire, T., 2003. Chemical and physical char-
acterization by EELS of strontium hexanoate reverse micelles and strontium
carbonate nanophase produced during tribological experiments. J. Microsc. 210,
110–118.
Marie, P.J., 2006. Strontium ranelate: a dual mode of action rebalancing bone
turnover in favour of bone formation. Curr. Opin. Rheumatol. 18, S11–S15.
Marie, P.J., Ammann, P., Boivin, G., Rey, C., 2001. Mechanisms of action and thera-
peutic potential of strontium in bone. Calcif. Tissue Int. 69, 121–129.
Meunier, P.J., Roux, C., Ortolani, S., Diaz-Curiel, M., Compston, J., Marquis, P., Cormier,
C., Isaia, G., Badurski, J., Wark, J.D., Collette, J., Reginster, J.Y., 2009. Effects of long-
term strontium ranelate treatment on vertebral fracture risk in postmenopausal
women with osteoporosis. Osteoporosis Int. 20, 1663–1673.
Ni, G.X., Lin, J.H., Chiu, P.K.Y., Li, Z.Y., Lu, W.W., 2010. Effect of strontium-containing
hydroxyapatite bone cement on bone remodeling following hip replacement. J.
Mater. Sci.: Mater. Med. 21, 377–384.
Ni, G.X., Lu, W.W., Chiu, K.Y., Li, Z.Y., Fong, D.Y.T., Luk, K.D.K., 2006.
Strontium-containing hydroxyapatite (Sr-HA) bioactive cement for primary hip
replacement: an in vivo study. J. Biomed. Mater. Res. B 77, 409–415.
Oliveira, J.P., Querido, W., Caldas, R.J., Campos, A.P., Abracado, L.G., Farina, M., 2012.
Strontium is incorporated in different levels into bones and teeth of rats treated
with strontium ranelate. Calcif. Tissue Int. 18, 18.
Reginster, J.-Y., Felsenberg, D., Boonen, S., Diez-Perez, A., Rizzoli, R., Brandi, M.-
L., Spector, T.D., Brixen, K., Goemaere, S., Cormier, C., Balogh, A., Delmas, P.D.,
Meunier, P.J., 2008. Effects of long-term strontium ranelate treatment on the
risk of nonvertebral and vertebral fractures in postmenopausal osteoporosis:
results of a five-year, randomized, placebo-controlled trial. Arthritis. Rheum.
58, 1687–1695.
Roschger, P., Manjubala, I., Zoeger, N., Meirer, F., Simon, R., Li, C., Fratzl-Zelman, N.,
Misof, B.M., Paschalis, E.P., Streli, C., Fratzl, P., Klaushofer, K., 2010. Bone material
quality in transiliac bone biopsies of postmenopausal osteoporotic women after
3 years of strontium ranelate treatment. J. Bone Miner. Res. 25, 891–900.
Terra, J., Dourado, E.R., Eon, J.G., Ellis, D.E., Gonzalez, G., Rossi, A.M., 2009. The struc-
ture of strontium-doped hydroxyapatite: an experimental and theoretical study.
Phys. Chem. Chem. Phys. 11, 568–577.
Verberckmoes, S.C., Behets, G.J., Oste, L., Bervoets, A.R., Lamberts, L.V., Drakopoulos,
M., Somogyi, A., Cool, P., Dorriné, W., De Broe, M.E., D’Haese, P.C., 2004. Effects
of strontium on the physicochemical characteristics of hydroxyapatite. Calcif.
Tissue Int. 75, 405–415.
Weiner, S., Wagner, H.D., 1998. The material bone: structure-mechanical function
relations. Annu. Rev. Mater. Sci. 28, 271–298.
Williams, D.B., Carter, C.B., 2009. Transmission Electron Microscopy. Plenum Press,
New York.
Yao, F., LeGeros, J.P., LeGeros, R.Z., 2009. Simultaneous incorporation of carbonate
and fluoride in synthetic apatites: effect on crystallographic and physico-
chemical properties. Acta Biomater. 5, 2169–2177.