Bone loss prevention in ovariectomized rats using stable amorphous calcium carbonate *

In assessing the relationship between calcium supplementation and maintaining bone mass or reducing the risk of fracture, the effectiveness of calcium supplementation has never been decisive. Freshwater crayfish rely on amorphous calcium carbonate (ACC), an instable polymorph of calcium carbonate, as the main mineral in the exoskeleton and in the temporary storage organ, the gastrolith. Inspired by the crayfish model, we have previously shown an increase in calcium bioavailability in rats administered with synthetic stable ACC vs. crystalline calcium carbonate (CCC). The current study compared the effects of amorphous calcium derived from either gastrolith or synthetic ACC with those of crystalline calcium, found in commercial CCC or calciumcitrate supplements, in a bone loss prevention model. Rats were subjected to either sham or ovariectomy (OVX) operation (n~20/ group) followed by administration of food pellets supplemented with 0.5% calcium from either source over 12 weeks. Micro-computed tomography (μCT) and histomorphometric analyses revealed bone loss prevention by both gastrolith and ACC treatments, manifested by an increase in morphometric bone parameters, compared to both CCCand calcium citrate-treated groups. Both gastrolith and ACC treatments resulted in bone formation in the tibia cancellous bone, indicated by dynamic histomorphometry parameters, compared to either the CCC or calcium citrate treatments. Levels of urine deoxypyridinoline (DPD), suggested an anti-resorptive effect of ACC, which was also the only treatment that led to a significant increase in vertebral mechanical strength, as supported by μCT analysis of topology and orientation parameters of the vertebral trabeculae. To our knowledge, such levels of bone loss prevention by calcium supplements have never been reported. These findings thus suggest the potential of both natural (crayfish gastrolith) and, to a greater extent, synthetic ACC sources for the prevention of metabolic bone disorders and possibly of osteoporotic processes.


INTRODUCTION
Strategies for preventing osteoporotic processes include a balanced diet rich in calcium, vitamin-D and regular exercise [1].Calcium supplementation with and without vitamin-D has become a widely recognized strategy for realizing adequate calcium intake.The calcium currently used in dietary supplements is derived from several origins, including natural sources, such as oyster shells, coral calcium and dolomite minerals, and synthetic precipitates, comprising both organic and inorganic calcium salts [2].Calcium from carbonate and citrate salts are by far the most commonly used forms of calcium supplements [2].Still, the effectiveness of calcium supplementation for the prevention of progressive skeletal disease has never been decisive.The relatively limited absorbability of calcium has resulted in an ongoing scientific debate as to the superior bioavailability of different calcium supplements [2].Yet, sensitive isotopic tracer methods revealed comparable bioavailability of all calcium salts [2,3].The effectiveness of calcium supplements with or without vitamin-D in reducing bone loss and preventing fractures has also been the focus of many clinical trials and several meta-analyses [4][5][6][7][8].Most of these studies reported a modest but significant effect of calcium supplementation in reducing bone loss [4,7,9,10], although the effectiveness of calcium supplementation on the prevention of fractures was less conclusive [5,6,[11][12][13] probably due to the fact that it is simply so modest that significance is not found in every study.According to an extensive meta-analysis, calcium combined with vitamin-D was associated with a 12% reduction of fracture risk [4].
Despite the documented benefits of calcium supplementation in combination with vitamin-D as an over the counter first-line management strategy in the prevention of osteoporosis, the effectiveness of available calcium forms is still very moderate.Calcium carbonate, one of the most abundant minerals in nature, has six known polymorphs of which the most thermodynamically stable form is calcite.The least stable polymorph is the amorphous form, termed amorphous calcium carbonate (ACC) [14].Stabilization of synthetic ACC can be a challenging task.Indeed, of the several attempts at stabilizing synthetic ACC reported [14][15][16][17], all used either toxic materials or various organic polymers to enable the stabilization of ACC, yet for no more than 3 days [14,15].
Freshwater crayfish construct specialized transient mineral storage sites, gastroliths, that are composed of stabilized amorphous calcium carbonate embedded into an organic matrix comprising dense chitin fibers and proteins [18].When needed, the thermodynamic instability of ACC is exploited to offer a highly bioavailable calcium source that enables fast and effective transport of the mineral across the intestinal epithelium and into the hardening exoskeleton.Inspired by the molecular mechanism used by crayfish to stabilize ACC, a novel method for synthetic production of stabilized ACC using phospho-amino acids was developed [19].
The superior bioavailability of ACC over crystalline calcium carbonate (CCC) was recently confirmed using an intrinsic radioactive labeling rat model, where in-creased bioavailability of up to 40% and 30% were seen in the serum and bone, respectively, while a 26.5% increase in retention levels was measured [20].In the scope of the current study, we compared the effects of the above synthetic ACC and of the gastrolith to those of leading commercially-available calcium carbonate and calcium citrate supplements on bone loss prevention, bone formation and mechanical strength in an ovariectomized (OVX) estrogen-deficiency bone loss rat model.

Diet Preparation
For administration of the different calcium sources, low calcium rodent pellets containing 0.02% calcium (Rat chow #TD95027, Harlan, Jerusalem, Israel) were finely grounded and separately mixed with the following four different calcium sources (Table 1): Calcium carbonate as contained in a commercial supplement (carbonate; Caltrate, Pfizer, Madison, NJ); gastrolith powder harvested from the Australian crayfish, Cherax quadricarinatus (Gastrolith, Ben's farm, Beer-Tzofar, Israel; see composition in [18,21]), stable ACC (Batch #ETS001, Amorphical, Beer-sheva, Israel; see composition in [19,20]) and calcium citrate from a commercial supplement (citrate; GNC, Pittsburgh, PA).Each mix was separately pelleted and kept in a sealed bag at RT. Pellets from all four mixtures contained ~0.5% elemental calcium (as generally used in EU and correlates to 250 mg calcium/kg body weight/day; a standard calcium requirement [22]), as confirmed by an atomic absorption (Varian AA240, Palo Alto, CA; Table 1) with no added vitamin-D, except the amounts found in the original food pellets (22 IU/g).

Animal Treatment
One hundred and one 16 -17 weeks old female Sprague-Dawley rats (Harlan) with an average body weight of 250 ± 16 g were group-housed (~3 rats/cage) and acclimated under controlled room conditions (21˚C ± 2˚C and a 12-hour dark-light cycle).During the 7-day of acclimation, the rats were fed a standard laboratory diet (#2018SC rat chow, Harlan) and de-ionized water ad libitum.
Following this period, the rats were individually housed in metabolic cages and deprived of food and water for 18 h.Urine was collected in the dark and stored at −80˚C for bone resorption markers analysis at baseline (time 0).Immediately after, the rats were anesthetized by intraperitoneal (IP) injection of ketamine (75 mg/kg body weight; Fort Dodge.Laboratories, Fort Dodge, IA) and xylazine (5 mg/kg body weight; EuroVet AH, Bladel, UK) solution.Blood was sampled from the rats' tails (~180 µl) and centrifuged for 10 minutes at 3000 g (Hettich Zentrifugen, Bach, Switzerland) for serum extraction.Serum samples were stored at −80˚C for bone formation markers analysis at baseline (time 0).
Rats were randomly assigned to five treatment groups, namely a sham-operated group that underwent sham operation and four other groups that were ovariectomized by a bilateral dorsal approach [23] and subsequently divided into OVX-control, OVX-gastrolith, OVX-ACC and OVX-citrate treatment groups (see below).Following surgery, the rats were given a 7-days recovery period, allowing a washout of estrogen, during which they were administered standard laboratory diet ad libitum and deionized water containing 1.5 ml/400ml dipyrone (Vitamed, Binyamina, Israel) and 1 mg/kg enrofloxacin (Bayer, Leverkusen, Germany).
At the end of the recovery period, the rats were grouphoused (2 rats/cage) according to the treatment assignment and fed the following different food mixes (Table 1): Sham (n = 19) and OVX-control (n = 19) rats were fed the CCC-containing mix, OVX-gastrolith rats (n = 18) were fed the gastrolith-containing mix, OVX-ACC rats (n = 20) were fed the ACC-containing mix and OVXcitrate rats (n = 18) were fed the calcium citrate-containing mix.The supplemental pellets and de-ionized water were provided ad libitum throughout the entire treatment period (90 days).Food consumption and body weight were recorded weekly (Table 2).
For dynamic histomorphometric analysis, the fluoro-chromatic dye, calcein (20 mg/kg body weight; Sigma-Aldrich, St. Louis, MO), was IP-injected into all rats on days 76 and 83.All 94 animals had calcein double labeling, demonstrated by two distinguished labeling lines along the trabecular bone.On day 90, the rats were individually housed in metabolic cages and deprived of food and water for 18 h.Urine was collected and stored at −80˚C for deoxypyridinoline (DPD) bone resorption marker analysis at the end of the experiment (day 90).Serum samples were extracted for osteocalcin (OC) bone formation markers evaluation and stored as described for day 0 samples.
All rats were sacrificed through inhalation of CO 2 .Right femurs and 4 th -and 5 th -lumbar vertebrae were dissected, wrapped in gauze pads soaked with saline buffer and stored −80˚C for micro-computed tomography (µCT) scanning or mechanical testing.Right tibias were dissected and immediately placed in 3.7% formaldehyde solution (F1636, Sigma-Aldrich) at 4˚C for 24 h, and processed as elaborated below in the structural and dynamic histomorphometry section.All procedures were approved by Ben-Gurion Animal Care and Use Committee, Ben-Gurion University, Beer-Sheva, Israel.

Micro-Computed Tomography (µCT) Scanning
Trabecular and cortical bone microarchitecture of right femurs and 4 th -lumbar vertebrae were analyzed using a Skyscan 1174 µCT scanner (Skyscan, Kontich, Belgium).Bone mineral density (BMD) calibration was performed according to the manufacturer's instructions, using designated rat phantom rods with densities of 0.25 and 0.75 g/cm 3 as standards (Skyscan).Femurs and vertebrae were removed from storage at −80˚C, transferred to new gauze pads soaked with 70% ethanol and placed in plastic vials containing 70% ethanol.The distal region of each femur was scanned for trabecular and cortical bone parameters at an isotropic resolution of 13.8 μm.Reconstruction was carried out employing a modified algorithm using Nrecon software, ver.1.6.4(Skyscan).The volume of interest (VOI) was set to 0.5 mm below the growth plate (most proximal boundary with the metaphysis) and extended proximally for 1.5 mm [24,25].Trabecular and cortical segments within the VOI were extracted by depicting ellipsoid contours every ~5 slices using CTan analysis software, ver.1.11 (Skyscan) [26].The trabecular and cortical BMD, bone volume/tissue volume (BV/ TV, %), trabecular number (Tb.N, 1/mm), trabecular thickness (Tb.Th, µm), and trabecular separation (Tb.Sp, µm) were calculated by CT analyzer software, ver.1.11 (Skyscan).Vertebrae were scanned at a 13.8 μm isotropic resolution.Reconstruction and analysis were performed as described above for the femurs.Trabecular and cortical bone parameters were evaluated in the vertebral body, set to cover a VOI of a 2-mm thick crosssection [25], starting 0.14 mm below the cranial growth plate.In addition to the same trabecular and cortical bone parameters as studied for the distal femur, the topology parameters: i.e., trabecular bone pattern factor (TBPf) and structural model index (SMI); and the trabecular orientation parameter, degree of anisotropy (DA) were also measured in the vertebral body.

Evaluation of Bone Turnover Biochemical Markers
Serum OC levels were determined using an enzymelinked immunoassay (EIA) kit (#BT-490, Biomedical Technologies, Stoughtone, MA) following a 1:10 dilution of the sample.Urinary DPD levels were determined using an EIA kit (#8007, Quidel, San Diego, CA) following a 1:50 dilution of the sample.DPD levels were normalized to creatinine levels, measured with the appropriate EIA kit (#8009, Quidel).

Evaluation of the Mechanical Strength
of the 5 th -Lumbar Vertebra 5 th -lumbar vertebral biomechanical properties were measured after soft tissue removal.The transverse and posterior processes of the vertebra were removed and the ends of the centrum were rendered parallel using a diamond wafering saw (Buehler Isomet, Lake Bluff, IL).The vertebral bodies were then broken by compression along the S-I axis at a speed of 0.5 mm/min using the Modular Test Machine 100P255 (Test Resources, Shakopee, MN).Force and displacement measurements were collected every 0.1 s.From the force versus displacement curves, structural (extrinsic) properties included ultimate force [maximum force obtained during testing ultimate load (UL, N)], energy (E) to UL (mJ) were calculated using standard procedures [27].Toughness (material level energy to fracture, E to UL/TotBV) was calculated using total bone volume as measured by µCT.

Statistical Analysis
Results are expressed as means (±SD or SEM).Oneway ANOVA followed by Fisher-LSD post-hoc comparison tests were performed using STATISTICA 6.1 software (StaSoft, Tulsa, OK).A p value < 0.01 was deemed significant.

Femoral and Vertebral µCT Morphometric Analysis
Three-dimensional µCT-based schematic microarchitectural reconstructions of a representative right distal femur and 4 th -lumbar vertebra from each group are presented in Figure 1.Femoral and vertebral (Figures 2A  and B, respectively) Tb.BMD of the sham group was significantly higher than that of all four OVX groups (one-way ANOVA; p < 0.001).Within the various OVX groups, the femoral and vertebral Tb.BMD of OVXgastrolith and OVX-ACC groups were significantly higher than those of the OVX-control and OVX-citrate groups.No significant differences were observed in the femoral  and vertebral Tb.BMD of the OVX-citrate group compared to the OVX-control group.Comparing the OVXgastrolith and OVX-ACC groups revealed a 13% increase in the femoral Tb.BMD of the former, relative to the latter (Figure 2A), although no difference in the vertebral Tb.BMD was noted (Figure 2B).In contrast to what was observed with the Tb.BMD, no significant differences were observed in the cortical BMD of either femurs or vertebrae across all groups (Tables 3(A) and (B)).
Femoral and vertebral µCT morphometric analysis (Figures 2C and D and Tables 3(A) and (B)) revealed a significant decrease in BV/TV (p < 0.01) and Tb.N (p < 0.01) in all OVX groups, compared to the sham.Within the femurs and vertebras of the OVX groups, OVX-gastrolith and OVX-ACC group members presented a significantly higher BV/TV and Tb.N, compared to the OVX-control and OVX-citrate groups, while no significant differences were observed in BV/TV and Tb.N between the OVX-citrate and OVX-control groups.Comparison of femurs from OVX-gastrolith and OVX-ACC group members revealed a significant difference in BV/ TV, Tb.Th and Tb.Sp (Figure 2C and Table 3A), while no significant differences in any of the morphometric parameters considered were observed between the vertebrae of these two groups (Figure 2D and Table 3B).

Structural and Dynamic Histomorphometry Analysis of the Proximal Tibia
The above differences observed in microarchitectural structure were further emphasized in the structural histomorphometry analysis of the proximal tibia, as presented in Table 3(C).All OVX groups showed a significantly lower BV/TV (p < 0.01) and Tb.N (p < 0.01) relative to the sham group, and Tb.Sp was significantly higher (p < 0.01) in all OVX groups, compared to the sham group.Within the OVX groups, the OVX-gastrolith and OVX-ACC groups presented a significantly higher BV/TV and Tb.N together with a lower Tb.Sp compared to the OVX-control group, and a significantly higher Tb.Th in the OVX-ACC group, compared to the same value in the OVX-control group.The OVX-citrate group presented significantly higher Tb.Th compared to the OVX-control group.Examination of the OVX-gastrolith and OVX-ACC groups revealed a significantly higher BV/TV and Tb.N along with lower Tb.Sp in the OVXgastrolith group than measured in the OVX-ACC group.
Proximal tibia dynamic histomorphometric analysis (Table 3C) revealed a 13% elevation of BFR/BS between OVX-control and OVX-citrate groups compared to the sham group; however, this did not reach a statistic- cal significance.Nevertheless, the BFR/BS ratios of the OVX-gastrolith and OVX-ACC groups were significantly higher than were the same parameters in the sham group (p < 0.01).MS/BS ratios between the OVX-control and OVX-citrate groups, compared to the sham group were not altered, while the MS/BS ratios of the OVX-gastrolith and OVX-ACC were elevated relative to the sham (p < 0.01).Within the OVX groups, the OVX-gastrolith and OVX-ACC groups presented significantly higher BFR/BS and MS/BS ratios compared to OVX-control.The BFR/ BS and MS/BS ratios of the OVX-citrate group were similar to that measured for the OVX-control group, and significantly lower from the value determined for the OVX-gastrolith and OVX-ACC groups, while MAR showed no significant differences among all treatment groups.

Biochemical Markers Analyses
Evaluation of bone turnover in all treatment groups at day 0 and 90, as reflected by the levels of OC, a formation marker, and urine DPD, a resorption marker, is presented in Tables 3(D) and (E).When both markers were evaluated at baseline (i.e.day 0) prior to randomization, no statistical differences among the treatment groups were found (Table 3

(D)). At day 90 (Table 3(E)
), however, significantly higher DPD levels were observed in the OVX-control, OVX-gastrolith and OVX-citrate groups, compared to the sham group (p < 0.001), while the DPD levels in the OVX-ACC group were not statistically different than that of the sham group, yet were significantly lower than in the OVX-control and OVX-citrate groups.DPD levels in the OVX-gastrolith group were significantly lower than that of OVX-citrate group, yet were not different from that of the OVX-control or OVX-ACC group.Serum OC marker levels showed no significant difference across the various experimental groups.

Vertebral Mechanical Strength Analysis
Mechanical strength analysis of the 5 th -lumbar vertebra exhibited significantly lower values, namely 40% for the OVX-control, OVX-gastrolith and OVX-citrate groups, on average, relative to the values measured with the sham group (Figures 3A-C; p < 0.01).At the same time, most of the mechanical strength values of the OVX-ACC group were lower only by an average of 17% than were the values determined for the sham group.Interestingly, the ultimate force parameter of the OVX-ACC group was not significantly different from that of the sham group.Within the OVX groups, the OVX-ACC group presented significantly higher values in comparison to all other OVX groups in every tested parameter, including ultimate force (Figure 3A), energy to ultimate force (Figure 3B) and toughness (Figure 3C).

Vertebral Topology and Orientation Parameters
Due to the notion that compression tests are not only affected by bone density but also by microarchitecture topology and orientation of the vertebral body [28][29][30], we further sought to determine trabecular topology and orientation parameters within the vertebral body of these rats.Micro-CT analysis of the 4 th -lumbar vertebra showed that TBPf was lower in the sham compared to all OVX groups.Within the OVX groups, the TBPf of the OVX-ACC group was significantly lower than that of the OVX-control, OVX-citrate and OVX-gastrolith groups (Figure 3D; p < 0.001).SMI evaluation points to the same trend, being significantly lower in the sham group, compared to all OVX groups.Here, values obtained with the OVX-ACC group were significantly lower that what was measured for the OVX-control and OVX-citrate groups.At the same time, SMI values obtained with the OVX-ACC group were reduced relative to the index measured with the OVX-gastrolith group (Figure 3E; p < 0.001).The last micro-architectural parameter measured, namely DA, was lower in the sham group, compared to the OVX-control, OVX-gastrolith and OVXcitrate groups, by 22% on average.No significant differences in DA was observed between the sham and OVX-ACC groups, or among all OVX groups, although OVX-ACC showed lower values that measured with the other OVX groups (Figure 3F; p < 0.001).

DISCUSSION
The results of this study point on a significant prevention of OVX-induced bone loss by both natural ACC (i.e.gastrolith powder of crayfish origin) and synthetic ACC treatments compared to that of two commercially available calcium supplements (i.e.calcium carbonate and calcium citrate).A significant increase in morphometric bone parameters, studied by μCT and structural histomorphometry was observed for the gastrolith and ACC compared to CCC and calcium citrate treatments.Both the gastrolith and ACC treatments displayed increased bone formation rate as indicated by dynamic histomorphometry, compared to the other treatments, with no effect on serum OC levels.Intriguingly, urinary DPD levels suggested an anti-resorptive effect for the synthetic ACC compared to the other treatments, and this was also the only treatment that prevented the loss of vertebral mechanical strength following OVX, supported by microarchitectural analysis of topology and orientation parameters.
The results of this study are consistent with our basic hypothesis, that both natural and synthetic ACC have a beneficial effect on bone, derived from the peculiar thermodynamic properties of ACC in both its analytic and natural forms and further support our previous report, showing the enhanced bioavailability of synthetic stable ACC over that of CCC [20].
In the current study we observed a significant bone loss in the trabecular area of the distal femur, proximal tibia and vertebrae in all four OVX groups, relative to the sham-operated group, as reflected in almost all of the μCT and histomorphometry parameters considered.The lack of change in cortical BMD between the sham and any of the OVX groups is presumably due to the fact that this study is shorter than others who relied on longer periods and who reported cortical BMD changes [31].The most significant bone loss was observed in the OVX-control and OVX-citrate groups, demonstrating the limited ability of currently available calcium supplements to prevent osteoporotic processes.In almost all of the parameters evaluated in this study, no significant differences were observed upon treatment with either of these two supplements, supporting the observation that both calcium carbonate and calcium citrate, when administered with food, have similar absorption and effects on bone [2,3].Interestingly, both the gastrolith and synthetic ACC treatments resulted in significantly less bone loss in the femur, tibia and vertebrae, compared to control or citrate treatments.
In most of the μCT-analyzed parameters of the distal femur and in the histo-morphometric analysis of the proximal tibia, the gastrolith presented a slightly enhanced preventative effect than did ACC.On the other hand, μCT analyses of the vertebrae revealed insignificant differences following the two treatments.The variability of bone loss rate induced by estrogen depletion in different bones is well documented [32], with the vertebrae being less affected than are the femur and tibia.
The use of gastrolith in traditional medicines dates back several hundreds of years [33].Both gastrolith and synthetic ACC contain stable ACC; yet, the enhanced performance of the gastrolith observed in some of the μCT and histo-morphometric analyses reported here suggests that organic and inorganic components present in the gastrolith, other than calcium and carbonate, such as the indigestible oligosaccharide chitin in the scaffold, proteins, inorganic phosphate or other metabolites [19,21], may enhance the positive effect of the ACC mineral in this compound.Chitooligosaccharides converted from the derivative of chitin were previously suggested to enhance calcium absorption, suppress bone resorption and affect bone strength in OVX rats [34].Nevertheless, in our previous report [20], such calcium absorption-related advantages were not observed upon treatment with ACC with chitosan.The nanometric particle size of ACC in both the gastrolith and synthetic ACC might also contribute to the increased bioavailability and mineral bone density observed, as was suggested in OVX mice and rat models [16,35].However, further findings suggest that diet calcium levels rather than particle size are responseble for the effects of this mineral on bone density and mechanical properties [36].
The higher trabecular bone fraction measured in the OVX-gastrolith and OVX-ACC groups could be a result of reduced bone resorption, as reflected in urinary DPD marker levels in both groups.Previous reports of the effects of calcium supplements on urinary DPD markers following OVX were inconsistent [37,38].In our study, a significant decrease in DPD levels, relative to those measured in the OVX-control group was only observed with the OVX-ACC group.In the OVX-gastrolith group, the degree of DPD reduction was not significant, relative to the OVX-control group, but only compared to the OVX-citrate group.
Calcium supplements are not thought to induce anabolic effects on bone [39].Dynamic histomorphometric analysis of the proximal tibia trabecular bone revealed only a 13% increase of BFR/BS in the OVX-control and OVX-citrate groups compared to the sham, an effect that did not reach a statistical significance.BFR/BS is almost always elevated in OVX-control rats [25], yet there are studies showing no effect and even reduction of this parameter following OVX intervention [40,41].Interestingly, both the OVX-gastrolith and OVX-ACC groups presented significant increased BFR/BS and MS/BS ratios compared to OVX-control rats, reflecting bone formation during bone remodeling [41,42].The active mineralization that apparently transpires upon treatment with gastrolith and ACC can most probably be attributed to the mineral component of these substances, as well as by the presence of phosphoserine [19,21,43], previously reported to induce bone formation in-vitro and in-vivo [44].Nevertheless, MAR and the OC serum marker levels following both gastrolith and the synthetic ACC treatments did not coincide with a possible anabolic effect, showing insignificant differences from the sham, OVX-control and OVX-citrate groups.Yet, a trend of reduction in the Sham and OVX-ACC group is seen in the serum OC.
In contrast to most of the morphometric data obtained in this study, treatment with the synthetic ACC resulted in significantly higher bone strength compared to all other OVX groups, including the OVX-gastrolith group, and in some of the parameters considered, such as ultimate force, even presented no statistical differences from the sham.Bone strength can only be partially predicted using BMD and morphometric parameters, and discrepancies between mechanical strength and parameters such as BMD and trabecular bone morphology were reported by others [45].We, therefore, further analyzed specific topology and orientation microarchitectural parameters previously shown to be useful for the discovery of minimal modifications to the spongy lattice, affecting mechanical properties of bones [28].Intriguingly, in contrast to the bone volume morphometric parameters and in good agreement with our mechanical strength findings, ACC had a beneficial effect on the topology and orientation parameters.Studies presenting changes in topology parameters with no effect on BV/TV were already reported by others [28], suggesting that the surface contour of the trabecular bone structure, plays a critical role in the loss of bone strength [29].These results might propose a different mechanism of action for the gastrolith versus synthetic ACC that might be attributed to elements from the extracellular matrix of the gastrolith and that are absent in the synthetic ACC [19,21].
To conclude, our findings point to the superior effects of ACC-containing compounds over two of the most common calcium supplements in terms of bone loss prevention.In particular, this study highlights the advantage of synthetic ACC with respect to mechanical bone strength.In future, the ability of ACC-containing compounds to reduce, prevent or even treat bone deterioration in human metabolic bone disorders should be tested.

Figure 1 .
Figure 1.Three-dimensional reconstruction of representative distal femural and 4 th -lumbar vertebral cross sections.The trabecular bone region from the (A) sham-operated group treated with crystalline CaCO 3 (sham).OVX-treated groups treated with: (B) crystalline CaCO 3 (control), (C) gastrolith powder, (D) amorphous calcium carbonate (ACC) or (E) calcium citrate (Ca-citrate).Scale bar represents 1.5 mm.For the illustration, the femoral volume of interest (VOI) was set to 1 mm below the growth plate (most proximal boundary with the metaphysis) and extended proximally for 0.4 mm.The vertebral VOI was set to a 0.4 mm thick cross-section starting 0.14 mm below the cranial growth plate.

Table 1 .
Elemental calcium content in calcium sources and specialized pellets.

Table 2 .
Food consumption and total body weight.
* Weight gain (%) bMean food consumption (g/day) 15.2 (0.07) 15.2 (0.07) 15.2 (0.07) 15.1 (0.02) 15.2 (0.05) Weight gain represents the rate of weigh increase over the course of the experiment.Mean daily food consumption was calculated per animal, by averaging the weekly food intake per cage over the treatment period and dividing it by 14.Values correspond to means (±SD).One-way ANOVA: * p < 0.01.Letters represents Fisher's LSD post-hoc comparison tests.

Table 3 .
Bone morphometric parameters and biochemical markers.