Oxytocin Controls Differentiation of Human Mesenchymal Stem Cells and Reverses Osteoporosis

Abstract

Osteoporosis constitutes a major worldwide public health burden characterized by enhanced skeletal fragility. Bone metabolism is the combination of bone resorption by osteoclasts and bone formation by osteoblasts. Whereas increase in bone resorption is considered as the main contributor of bone loss that may lead to osteoporosis, this loss is accompanied by increased bone marrow adiposity. Osteoblasts and adipocytes share the same precursor cell and an inverse relationship exists between the two lineages. Therefore, identifying signaling pathways that stimulate mesenchymal stem cells osteogenesis at the expense of adipogenesis is of major importance for developing new therapeutic treatments. For this purpose, we identified by transcriptomic analysis the oxytocin receptor pathway as a potential regulator of the osteoblast/adipocyte balance of human multipotent adipose-derived stem (hMADS) cells. Both oxytocin (OT) and carbetocin (a stable OT analogue) negatively modulate adipogenesis while promoting osteogenesis in both hMADS cells and human bone marrow mesenchymal stromal cells. Consistent with these observations, ovariectomized (OVX) mice and rats, which become osteoporotic and exhibit disequilibrium of this balance, have significant decreased OT levels compared to sham-operated controls. Subcutaneous OT injection reverses bone loss in OVX mice and reduces marrow adiposity. Clinically, plasma OT levels are significantly lower in postmenopausal women developing osteoporosis than in their healthy counterparts. Taken together, these results suggest that plasma OT levels represent a novel diagnostic marker for osteoporosis and that OT administration holds promise as a potential therapy for this disease.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

Author contributions: C.E.: carried out most of the experiments, data analysis and interpretation, collection and assembly of data; A.B., H.B., and C.B.: analysis of bone micro-architecture and biomechanical parameters of mice, manuscript writing; L.Z., M.S., and Z.T.: microarray analysis, data analysis and interpretation; F.M.: contribution to the mice study, data analysis and interpretation; V.B., G.F.C., and L.E.: provision of patients, data analysis and interpretation; N.W.: histological analysis of mice tissues, data analysis and interpretation; E.L.: provision of study material, data analysis and interpretation; C.D.: data analysis and interpretation, manuscript writing; G.A.: conception and design, data analysis and interpretation, manuscript writing; E.A.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing.

The regulation of skeletal homeostasis represents an active research area as the elderly population is steadily rising. Among its physiopathological consequences, osteoporosis represents a major health problem, affecting more than 30% of women above 50 years old [1, 2]. The resistance and integrity of bone depend upon the balance between bone resorption by osteoclasts and bone formation by osteoblasts [3, 4]. The decrease in bone mass that occurs in osteoporosis—for example in association with hypogonadism—involves an acceleration of bone turnover and a disequilibrium between resorption and formation in favor of bone resorption [5, 6]. Whereas most osteoporosis treatments target bone resorption by inhibiting osteoclasts formation and activity, only a few promote bone formation. Identification of such drugs would enable the development of alternative and/or complementary treatments. In addition to bone loss, osteoporosis is associated with an increased bone marrow adipose tissue, leading to the formation of adipocytes at the expense of osteoblasts [7]. Osteoblasts and adipocytes share the same mesenchymal cell precursor, as illustrated by the ability of mesenchymal stem cells (MSC) isolated from different tissues to differentiate into either lineage. MSC have been considered as suitable sources of adult progenitors for cell-based therapy because of easy isolation technique, expandability, and pluripotency. As human adipose-derived MSC represent a more abundant and available source of cells for autologous cell replacement than human bone marrow-derived MSC (hBMSC), interest in human adipose-derived MSC has been growing in the recent years. Recently, we isolated MSC from human adipose tissue termed hMADS (i.e., human multipotent adipose-derived stem) cells. These cells, which exhibit both a normal karyotype and high self-renewal ability, are able to differentiate into various lineages, including adipocytes and osteoblasts, and can also support in vivo regenerative processes [8, [9], [10]–11].

In view of the reciprocal and inverse relationship that exists between osteogenesis and adipogenesis, controlling the fine balance between the two pathways is of clear therapeutic significance [12, 13]. However, the interplay between the two cell types and the decision to commit to either lineage are not well characterized, and a better understanding of the two pathways would be useful for the development of new drug therapies.

To gain better insight into the early steps of osteoblast and adipocyte differentiation in MSC, we performed first a microarray RNA analysis of hMADS cells at different time points of osteogenesis or adipogenesis using a previously developed microarray [14], data deposited in Array Express, accession number A-MARS three and E-MARS 10. As a basis of selection, we focused our attention on genes (i) differentially expressed during osteogenesis and adipogenesis, (ii) encoding for cell surface receptors, and (iii) modulated by estrogens. Oxytocin receptor (otr) emerged as a candidate gene using these criteria.

OTR, known to be a member of the heptahelical G protein-coupled receptor family, is expressed in a variety of cell types, including osteoblasts and adipocytes [15, [16]–17]. Its ligand, oxytocin (OT), belongs to the pituitary hormone family and regulates the function of peripheral target organs. It also modulates a wide range of behaviors, such as social recognition, love, and fear [18, [19], [20]–21]. OT had been suggested to play a role in bone homeostasis and osteoporosis based on the proliferative effects of OT on osteoblasts in vitro and the modulation of blood parameters associated with bone formation of normal rats [22, [23]–24]. As a next step, we examined whether oxytocin could modulate in vitro the osteoblast/adipocyte balance using hMADS cells and hBMSC. Subsequently, the relationship between circulating OT levels and osteoporosis was analyzed as well as the effects of OT injection on bone loss in ovariectomized mice. Finally, plasma OT levels were determined in postmenopausal women suffering or not from osteoporosis and were found to be consistent with animal data. Our results show for the first time that OT signaling is implicated in the regulation of the osteoblast/adipocyte balance and reverses osteoporosis in ovariectomized mice, suggesting the use of OT as a potential therapeutic treatment.

Materials and Methods

Cell Culture

The establishment and characterization of the multipotency and self-renewal of hMADS cells have already been described [8, [9]–10]. In the experiments reported herein, hMADS-2 cells, established from the pubic region fat pad of a 5-year-old male donor, were used at a stage between passages 16 and 35, corresponding to 35–100 population doublings. Cells were seeded at a density of 4,500 cells/cm² in Dulbecco's modified Eagle's medium (DMEM) (Lonza, Levallois-Perret, France, http://www.lonza.com), supplemented with 10% fetal calf serum (FCS), 2.5 ng/ml hFGF2, 60 μg/ml penicillin, and 50 μg/ml streptomycin. The medium was changed every other day and hFGF2 was removed when cells reached confluence. At day 2 post-confluence (designated as day 0), cells were then induced to differentiate toward either the adipocytic or the osteoblastic lineage. For osteoblastic differentiation, cells were maintained in the presence of αMEM containing 1% FCS supplemented with 10 ng/ml epidermal growth factor, 10 nM 1,25-dihydroxyvitamin D3, 100 nM dexamethasone, 50 μg/ml l-ascorbic acid phosphate, and 10 mM β-glycerophosphate [11]. For adipocyte differentiation, cells were cultured in DMEM/Ham's F12 media supplemented with 10 μg/ml transferrin, 0.85 μM insulin, 0.2 nM triiodothyronine, 1 μM dexamethasone, and 500 μM isobutyl-methylxanthine. Three days later, the medium was changed (dexamethasone and isobutyl-methylxanthine were omitted) and 100 nM rosiglitazone was added. The media were then changed every other day and cells were used at the indicated days.

Human mesenchymal stromal cells isolated from bone marrow (hBMSC) were purchased from Cambrex (Paris, http://www.cambrex.com) and used as recommended by the manufacturer. Adipocyte differentiation of hBMSC was performed in DMEM containing 10% FCS, 1 μM rosiglitazone, 0.85 μM insulin, and 0.2 nM triiodothyronine. One micromolar dexamethasone and 500 μM isobutyl-methylxanthine were added over the first three days. Osteogenic differentiation of hBMSC was induced in the presence of αMEM containing 10% FCS supplemented with 10 nM 1,25-dihydroxyvitamin D3, 100 nM dexamethasone, 50 μg/ml l-ascorbic acid phosphate, and 10 mM β-glycerophosphate.

Combined Adipogenesis and Osteogenesis in Culture

hMADS cells or hBMSC were induced to differentiate into adipocytes and osteoblasts in the same dish using for each cell model a dual lineage-promoting (DLP) medium consisting of a mixture of adipogenic and osteogenic media as described above (vol/vol; 1/1).

Enzymatic Activity Measurements and Cell Staining.

Glycerol-3-phosphate dehydrogenase (GPDH) and alkaline phosphatase (ALP) activity measurements were performed as described previously [25, 26]. Oil red O and von Kossa staining were performed as previously described [8, 27].

Isolation and Analysis of RNA.

Total cell RNA was extracted using TRI-Reagent kit (Euromedex, Souffouffelweyersheim, France, http://www.euromedex.com) and RNA from humeri was extracted using the Totally RNA kit (Ambion, Courtaboeuf, France, http://www.ambion.com) according to the manufacturer's instructions, respectively. Two micrograms of total RNA, digested with Dnase I (Promega, Charbonnières-les-Bains, France, http://www.promega.com), was subjected to reverse transcription-polymerase chain reaction (RT-PCR) analysis as described previously [8]. Quantitative PCR assays were run on an ABI Prism 7,000 real-time PCR machine (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com). The final reaction volume was 25 μl, including 300 nM specific primers, 5 ng of reverse-transcribed RNA, and 12.5 μl of SYBR Green master mix (Eurogentec, Angers, France, http://www.eurogentec.com). Quantitative PCR conditions were as follows: 2 minutes at 50°C; 10 minutes at 95°C; and 40 cycles of 15 seconds at 95°C, 1 minute at 60°C. The expression of genes was normalized to the expression of 36B4 or the TATA-binding protein (TBP)-encoding genes. Gene expression was quantified using the comparative-ΔCt method. The oligonucleotides for each target of interest, designed using Primer Express software (PerkinElmer Life and Analytical Sciences), are shown (forward and reverse) in supplemental online Table 1.

Western Blot Analysis and Activated RhoA-GTP Pull-Down.

Whole cell extracts were obtained by cell lysis in 0.5% Nonidet P40, 0.5% Triton X-100, 50 mM Tris-HCl (pH 6.8), 10 mM dithiothreitol, 10 mM β-glycerophosphate, 10 mM NaF, 0.1 mM sodium orthovanadate, and a protease inhibitor cocktail (Roche Molecular Biochemicals, Paris, http://www.roche-diagnostics.fr). Western blot analysis was performed as described previously [10]. Primary antibodies were mouse anti-phospho-ERK1/2 and rabbit anti-ERK1/2 (Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com). Secondary horseradish peroxidase-conjugated antibody was purchased from Promega.

Pull-down assays were performed as described previously [28, 29]. Briefly, hMADS cells were first washed twice with ice-cold Tris-buffered saline and then lysed in RIPA buffer (50 mM Tris, pH 7.5, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 500 mM NaCl, 10 mM MgCl₂, 1 mM β-glycerophosphate, 2 mM sodium orthovanadate, 5 mM dithiothreitol, and 1 mM PMSF). Cell lysates were clarified by centrifugation at 15,000g at 4°C for 30 minutes. After protein quantification, 1 mg of cell lysate was incubated with 40 μg of GST-Rhotekin RBD beads at 4°C for 30 minutes. The beads were then washed three times with lysis buffer. Activated RhoA (RhoA-GTP) were revealed by immunoblotting using monoclonal anti-RhoA. In parallel, 5% of the cell lysate was used to determine the total amount of RhoA. Both RhoA and RhoA-GTP were normalized to the levels of TBP.

Animals.

The experiments were conducted in accordance with French and European regulations for the care and use of research animals and were approved by the experimentation committee. Animals were maintained under constant temperature (21 ± 2°C) and 12:12-hour light-dark cycles, with ad libitum access to standard chow diet and water.

Eight-week-old C57Bl/6J mice were subjected either to bilateral ovariectomies (OVX) from the dorsal approach or to sham surgery in which the ovaries were exteriorized but replaced intact by the supplier (Charles River Laboratories, L'Arbresle, France, http://www.criver.com). OVX and sham mice were divided into two groups (12 mice in each group). Two weeks after ovariectomy or sham surgery, mice were injected daily with vehicle (Ve) or with 1 mg/kg OT for 8 weeks. Plasma OT levels were measured at 1.5 hours, 6 hours, and 24 hours post-subcutaneous OT injection. This dose of OT has been used previously with success in different studies [30, 31]. The level was highest at 1.5 hours and had decreased to control levels at 6 hours and 24 hours (supplemental online Fig. 1). No mortality was observed with vehicle or OT treatment in either sham or OVX mice. Moreover, histological analysis revealed no pathological signs in liver and kidney obtained from the four groups (supplemental online Fig. 2). Animals were anesthetized with isoflurane inhalation and sacrificed by cervical dislocation; blood samples were collected by heart punction. Left and right femurs were removed and frozen at −20°C before bone microarchitecture and biomechanical analysis. Humeri, livers, and kidneys were collected and processed for RNA extraction or histological analysis as previously described [32].

Rat plasma were kindly provided by Dr. Bonnet [33]. Thirty-five-week-old Wistar rats were subjected either to ovariectomy or sham surgery; 10 weeks later blood was collected for plasma preparation and oxytocin determinations.

Determination of Adipocyte Size and Number.

Humeri were paraformaldehyde fixed, decalcified, dehydrated, and paraffin-embedded. Eight micrometer longitudinal sections were hematoxylin and erythrosin stained. Three sections per humerus were performed and three pictures per section were imported into ImageJ software (available at http://rsb.info.nih.gov/nih-image). Adipocyte number and size values were determined and reported as the mean ± SEM.

Biochemical Serum and Plasma Parameters.

Plasma or serum concentration of OT (OT EIA kit, Assay Designs Inc., Ann Arbor, MI, http://www.assaydesigns.com) and estradiol (Estradiol EIA kit, Bayer) were measured using commercially available kits.

Micro-computed Tomography.

Trabecular bone microarchitecture of the distal femoral metaphysis was scanned by micro-computed tomography (μCT, Skyscan 1,072; Skyscan, Kontich, Belgium, http://www.skyscan.be). The x-ray source was set at 70 kV and 100 μA and filtered with a 1-mm-thick aluminum filter. The radiographic projections were acquired with a fixed exposure time of 3.4 seconds per frame and with averaging of four frames to improve the signal-to-noise ratio. The voxel size was isotropic and fixed at 7 μm. The transmitted x-ray beam was recorded by a scintillator coupled to a 1,024 pixel × 1,024 pixel 12-bit digital cooled CCD camera. Four hundred projections were acquired over an angular range of 180° (angular step of 0.45°). After scanning, 900 slices were reconstructed using the manufacturer's reconstruction software based on the Feldkamp algorithm.

Image Analysis.

The recorded data sets were segmented into binary images using simple global thresholding methods. To eliminate the primary spongious bone, we analyzed 160 slices proximal to the growth plate.

The trabecular bone parameters were calculated using the Skyscan software CTan (Skyscan). Trabecular bone volume fraction, trabecular number, trabecular separation, trabecular thickness, structural model index, trabecular bone pattern factor, and the degree of anisotropy were calculated according to methods previously described [34].

Analyses of mouse cortical bone architecture were performed at the mid-diaphysis using micro-computed tomography. Cross-sectional geometric properties, such as cortical thickness (mm) and medullary area (mm²), were measured from image analysis. Cross-sectional cortical area (mm²) and moment of inertia (I, mm⁴) in relation to the horizontal axis were calculated as described previously [35].

Biomechanical Parameters.

The mechanical properties of the left femurs were assessed by anterior-posterior three-point bending test using a Universal Testing Machine (Instron 4,501, Instron, Canton, MA, http://www.instron.us). Each femur was centrally loaded at the mid-diaphysis at a speed of 1 mm/minute. The extrinsic biomechanical parameters stiffness (S, N/mm) and ultimate strength (F_U, N) were determined from load-displacement curves. The intrinsic biomechanical parameters Young's modulus (E, MPa) and ultimate stress (σ_U, MPa) were calculated from the load-displacement curves and geometric properties as described previously [36].

Patients.

Patients were enrolled upon written informed consent according to the hospital's guidelines, approved by the local ethic committee and conducted according to the principles expressed in the Helsinki Declaration. Patients were 55–85-year-old postmenopausal women. Exclusion criteria were patients who had previous or current diseases associated with secondary osteoporosis and who had been previously treated for osteoporosis. Thirty-six postmenopausal women were selected and divided into two groups based on bone mineral density (BMD) values (T-score) as measured by dual-energy x-ray absorptiometry (DXA, Hologic QDR 4,500). The control group (n = 16) corresponded to nonosteoporotic women who had no fractures and normal BMD values (T-score ≥ −1 DS) at the lumbar spine and femoral neck. The osteoporotic group (n = 20) included women with severe osteoporosis defined by at least one osteoporotic fracture and low BMD values (T-score < −2.5 DS) measured at the lumbar spine and/or femoral neck. Patients were examined to obtain height and weight measurements that allowed calculations of body mass index (BMI). The characteristics of both groups are presented in supplemental online Table 2. Blood samples were collected in the morning after overnight fasting. Biochemical assays were performed as described above.

Statistical Analysis.

Results are presented as means ± SE as indicated and are analyzed using the 2-tailed Student's t test. The ANOVA test was used to compare the groups for micro-architectural parameters measured at the distal femurs. For patients, we used Kruskal–Wallis test, Spearman's rank correlation, and logistic regression. Post hoc differences were determined with the Newman–Keuls test with significance defined as p < 0.05.

Results

Effects of OT on the Osteoblast/Adipocyte Balance of MSC

Based upon the selection criteria (see above), quantitative RT-PCR of OTR expression was performed in hMADS cells undergoing adipogenesis versus osteogenesis. As shown in supplemental online Fig. 3, OTR mRNA levels decrease during adipogenesis and increase during osteogenesis. Therefore, we decided to analyze the effect of its ligand on both differentiation processes. Both OT and carbetocin (Cb, a stable OT analog) [37] stimulated osteoblast differentiation and inhibited adipocyte differentiation of hMADS cells in a dose-dependent manner (data not shown). Chronic treatment of hMADS cells with optimal concentrations of 30 nM OT or 300 nM Cb under conditions promoting either adipocyte or osteoblast differentiation stimulated osteogenesis and inhibited adipogenesis. This effect was illustrated by an increase in ALP activity (Fig. 1A) and a decrease in GPDH activity (Fig. 1B) at different time points following the initiation of either differentiation process. Interestingly, Vasopressin, another closely related neuropeptide of the pituitary gland, had no effect on both cell types (data not shown).

Figure Figure 1.. Effect of OT and Cb on osteoblast and adipocyte differentiation. hMADS cells were induced to differentiate toward either the (A) osteogenic or (B) adipogenic lineage in the absence (white columns) or presence of 30 nM OT (black columns) or 300 nM carbetocin (hatched columns). (A): ALP and (B) GPDH activities were determined at the indicated time points and presented as mean ± SE. *, p < .05; **, p < .01; versus untreated cells. Abbreviations: ALP, alkaline phosphatase; Cb, carbetocin; GPDH, glycerol-3-phosphate dehydrogenase; hMADS, human multipotent adipose-derived stem; OT, oxytocin.

To further investigate the effects of OT and Cb on the commitment of MSC to either the osteoblast or adipocyte lineage, we used a DLP medium that allows the differentiation of both cell types in the same dish. hMADS cells differentiated into both osteoblasts and adipocytes under this control culture condition, as shown by von Kossa staining (black color) for mineralized matrix deposition and oil red O staining (red color) for lipid storage (Fig. 2Aa). Strikingly, treatment with OT or Cb enhanced osteoblast differentiation and reduced adipocyte differentiation (Fig. 2Ab, 2Ac) compared to control conditions (Fig. 2Aa). GPDH activity and fatty acid binding protein 4 (FABP4) mRNA as adipogenic markers, and ALP activity and podoplanin (Pdpn) mRNA as osteogenic markers, confirmed these findings (Fig. 2B and 2C).

Figure Figure 2.. Effect of OT and Cb on osteoblast/adipocyte balance of MSC. hMADS cells and hBMSC were induced to differentiate in the dual lineage promoting medium in the (Control: Aa, Ad) absence or (Ab, Ae) presence of 30 nM OT or (Ac, Af) 300 nM Cb. At day 15, (A) cells were fixed and stained with oil red O (red) and von Kossa (black), (B) GPDH and ALP activities were determined, and (C) FABP4 and Pdpn mRNA levels were measured by quantitative reverse transcription-polymerase chain reaction. Results are representative of three independent experiments performed on different series of cells and expressed as percent of cells differentiated in DLP medium alone (dashed lines). Data are presented as mean ± SE. *, p < .05 versus untreated cells. Scale bar represents 100 μm. Abbreviations: ALP, alkaline phosphatase; Cb, carbetocin; FABP4, fatty acid binding protein 4; GPDH, glycerol-3-phosphate dehydrogenase; hBMSC, human bone marrow-derived mesenchymal stem cells; hMADS, human multipotent adipose-derived stem; OT, oxytocin; Pdpn, podoplanin.

To strengthen these observations, human bone marrow-derived mesenchymal stromal cells (hBMSC), which should be the preferential targets in vivo, were induced to differentiate in a DLP medium in the absence or presence of OT or Cb. Results similar to those obtained with hMADS cells were obtained, indicating that hMADS cells represent a valuable model for gaining insights into the osteoblast/adipocyte balance (Fig. 2Ad–2f; Fig. 2B and 2C).

To identify signaling pathways that may mediate OT effects on hMADS cell differentiation, we analyzed RhoA and ERK1/2 pathways. It is known that OT induces rat uterine contraction through the RhoA pathway [38] and stimulates prostaglandin E2 (PGE2) synthesis by triggering the ERK pathway [39]. As both pathways have been involved in osteogenesis and adipogenesis [40, [41]–42], we therefore analyzed whether OT might mediate its effects through RhoA and/or ERK pathways. As shown in Figure 3A, OT treatment of differentiating hMADS cells had no significant effect on the RhoA pathway as determined by RhoA-GTP measurements. In contrast, exposure of hMADS cells to OT at day 0 triggered a transient induction of ERK1 and ERK2 phosphorylation (Fig. 3B). This activation was detected within 2 minutes and the levels of phosphorylated ERK1/2 decreased after 10 minutes, suggesting a role played by the ERK pathway in OT effects.

Figure Figure 3.. Effect of OT on RhoA and ERK1/2 pathways. (A): Levels of total RhoA and active form (RhoA-GTP) were determined in human multipotent adipose-derived stem (hMADS) cells induced to differentiate into osteoblasts or adipocytes in the absence (white bars) or presence (black bars) of 30 nM OT. Total RhoA levels were determined by Western blot and RhoA-GTP levels were determined by pull-down and Western blot. Both levels were normalized to the levels of TATA-binding protein. Data are presented as percentage of RhoA-GTP versus total RhoA by taking 100% for levels observed in proliferating cells. Results are mean ± SE of independent experiments performed on two different series of cells. (B): Western blot versus phospho-ERK1/2 and ERK1/2. hMADS cells at day 0 were maintained for the last 18 hours in a serum-free medium containing 0.2% bovine serum albumin. OT was then added and protein extracts were prepared at the indicated times; 25 μg protein extracts were used for immunoblot. Results are representative of independent experiments performed on two different series of cells. Abbreviation: OT, oxytocin; Pdpn, podoplanin.

Oxytocin Plasma Levels in Ovariectomized Rodents

It is well established that the development of osteoporosis subsequent to ovariectomy or menopause involves disequilibrium of the osteoblast/adipocyte balance, with adipocytes developing at the expense of osteoblasts [43, [44]–45]. Accordingly, to determine whether there might be a correlation between OT levels and osteoporosis, we analyzed plasma OT levels in two ovariectomized (OVX) commonly used rodent models of osteoporosis and in sham-operated controls (sham). Indeed, plasma OT levels were 40% and 35% lower in OVX than in sham mice (Fig. 4A) and rats (Fig. 4B), respectively.

Figure Figure 4.. Plasma OT levels in OVX mice and rats. Ten weeks after ovariectomy or sham surgery, plasma OT levels in C57Bl/6J mice (A, n = 9 each) and Wistar rats (B, n = 10 each) were determined using the OT EIA kit. Data are presented as mean ± SE. *, p < .05; **, p < .01. Abbreviations: OT, oxytocin; OVX, ovariectomized.

Bone Loss Reversal by Oxytocin

To investigate whether OT treatment could reverse ovariectomy-induced osteoporosis, we next analyzed the effects of OT injection on bone microarchitecture, biomechanical parameters, and bone marrow adiposity in OVX mice. The effects of OT on trabecular bone microarchitecture were assessed by micro-CT analysis of distal femurs. The results showed that the trabecular bone volume fraction was significantly higher in the OVX-OT animals than in the OVX-Ve group (Fig. 5A). No significant differences in the volume fractions were observed between the sham groups. In addition, the trabecular number was 65% higher in the OVX-OT group and 26% higher in the sham-OT group than in their respective control groups (Fig. 5B). As a result of this, the trabecular spacing was lower in the OVX-OT animals than in the OVX-Ve group (Fig. 5C). Finally, the trabecular thickness, bone surface to bone volume ratios, trabecular bone pattern factor, structure model index, and degree of anisotropy did not differ significantly between the groups (Fig. 5D and data not shown). A three-dimensional representation of all these parameters illustrates the beneficial effect of OT on the bone microarchitecture of OVX mice (Fig. 5E).

Figure Figure 5.. Micro-computed tomography analysis of distal femur metaphysis of sham and OVX mice injected or not with OT. (A): Trabecular bone volume, (B) trabecular number, (C) trabecular spacing, and (D) trabecular thickness were determined in femurs of 12 mice from each group. (E): A three-dimensional representation of a horizontal analysis of femurs from a representative mouse from each group is shown. Data are presented as mean ± SE. *, p < .05. Abbreviations: BV/TV, trabecular bone volume; OT, oxytocin; OVX, ovariectomized; TbN, trabecular number; TbSp, trabecular spacing; TbTh, trabecular thickness; Ve, vehicle.

The effects of OT injection on cortical bone strength were next analyzed by the three-point bending test. Geometric and biomechanical parameters for the four groups of mice are presented in Table 1. Whereas there was no significant difference in the medullary area between groups, the cortical thickness was significantly higher in the sham-Ve group (control) than in the OVX-OT group. furthermore, the cortical cross-sectional area was 15.5% lower in the OVX-OT group than in the sham-Ve group. The moment of inertia (I) was significantly lower (i.e., 0.103 mm⁴ vs. 0.126 and 0.124 mm⁴) in the OVX-OT group compared to those in the sham-Ve and OVX-Ve groups, respectively. These data indicated that cortical bone geometry is affected only by the combination of ovariectomy and OT treatment.

Table Table 1.. Influence of oxytocin treatment on geometric and biomechanical parameters of bone

In theory, these reduced geometric parameters should lead to a lower bone mechanical strength in OVX-OT mice. However, calculating the extrinsic parameters of the femurs showed that stiffness (S) in the OVX-OT group was 25%, 22%, and 22% higher than in the sham-Ve (control), sham-OT, and OVX-Ve groups, respectively. Moreover, the 8% lower ultimate force (F_U) measured following ovariectomy-induced osteoporosis (OVX-Ve compared to sham-Ve) was back to normal levels in the OVX-OT group, revealing the beneficial effect of OT treatment on cortical bone strength. The Young's modulus and ultimate stress (E and σ_U) were also significantly higher in the OVX-OT group than in the sham-Ve, sham-OT, and OVX-Ve groups. OT treatment alone (sham-OT) did not modify the extrinsic or intrinsic properties of cortical bone. Therefore, despite the reduced cortical bone geometry of OVX-OT mouse femurs at mid-diaphysis, the combination of OT treatment and ovariectomy improved the extrinsic and intrinsic cortical bone properties, suggesting that the mechanical material properties were modified in a way that resulted in stronger bone.

It is well established that osteoporosis is accompanied by increased bone marrow adiposity. Whereas there were no significant differences in bone marrow adipocyte size between the four groups of mice (Fig. 6A), OVX-Ve mice exhibited a striking increase in adipocyte number and FABP4 mRNA expression (Fig. 6B and 6C). OT treatment completely reversed this effect, restoring normal adipocyte numbers and FABP4 mRNA levels in OVX mice (Fig. 6B and 6C).

Figure Figure 6.. Restoration of adipocyte number and FABP4 mRNA levels by OT treatment of OVX mice. Adipocyte size (A) and number (B) were determined by counting and measuring adipocytes on hematoxylin/erythrosin stained humerus serial sections in a blinded manner by two researchers. At least three sections covering the total length of the humerus were analyzed in three to four mice from each group. (C) FABP4 mRNA levels were measured in humerus from six mice of each group by quantitative reverse transcription-polymerase chain reaction. Data are presented as mean ± SE. *, p < .05; **, p < .01. Abbreviations: FABP4, fatty acid binding protein 4; OT, oxytocin; OVX, ovariectomized; Ve, vehicle.

Link Between Osteoporosis and Serum OT Levels in Postmenopausal Women

Finally, to investigate whether our findings obtained in rodent models of osteoporosis were also valid in humans, we compared circulating OT levels in postmenopausal women classified as osteoporotic or nonosteoporotic by DXA analysis according to World Health Organization criteria [46]. Significantly, circulating OT levels were 55% lower in osteoporotic women than in nonosteoporotic women (50.2 ± 8.8 vs. 110.6 ± 19.8 pg/ml, p = 0.005), in agreement with the data presented above from rodents. Statistical analysis and logistic regression showed no correlation with other parameters, that is, age, BMI, and weight. This result reinforces the possible implication of OT in the physiopathology of postmenopausal osteoporosis.

Discussion

A few reports have suggested the role of OT favoring osteogenesis [22, 23] whereas OT injection in normal rats appears to improve blood parameters associated with bone formation [24]. So far, no evidence has been reported that shows that OT controls stem cell fate and could be envisioned as a therapy for osteoporosis. Herein, our data reveal that OT plays a major role in the osteoblast/adipocyte balance of MSC. Adipose-derived MSC are commonly used to study differentiation processes because of their abundance and availability. Although human adipose-derived MSC exhibit some differences with human bone marrow stromal/stem cells, many similarities have also been reported [47, [48], [49]–50]. Herein, similar results were obtained when hMADS cells and hBMSC were treated with OT or Cb during differentiation in a DLP medium, indicating that both MCS are suitable cell models for studies of osteoblast/adipocyte balance.

The effects of OT on osteoblast and adipocyte differentiation are not mediated through increase in cell number as OT treatment of hMADS cells affected neither cell number nor the doubling time (data not shown). These findings are different from those reported by Petersson et al. who showed a stimulation of DNA and protein synthesis by OT in osteoblast-like cells and some osteosarcoma cells [23, 51]. Most likely, these discrepancies are due to differences in cell models and/or culture conditions.

Several modulators of the osteoblast/adipocyte balance have been described. RhoA signaling, which may induce changes in cell shape, has been reported to be involved in the osteoblast/adipocyte balance of human mesenchymal stem cells [40]. We have observed that OT did not affect the activation of RhoA pathway in hMADS cells differentiating to either the osteogenic or the adipogenic lineage. Recently, it has been shown that liver-enriched inhibitory protein (LIP), an isoform of CCAAT/enhancer binding protein β (C/EBPβ) lacking the transcriptional activation domain, plays an important role in the control of the balance between osteoblast and adipocyte differentiation [52]. We have observed no significant effect of OT on the levels of LIP or both isoforms of C/EBPβ during hMADS cell differentiation toward osteoblasts or adipocytes (data not shown). However, OT induced a transient phosphorylation of ERK1/2, suggesting a potential role of this pathway in the OT effects. These data are consistent with observations showing that activation of ERK1/2 leads on one hand to phosphorylation of CBFA1, the osteoblast differentiation key transcription factor, which enhances osteogenesis, and on the other hand to phosphorylation of PPARγ, the adipocyte differentiation master gene, known to inhibit its adipogenic activity [41, 42]. Further detailed analysis of the ERK and other pathways, such as Wnt pathway, will shed light on the fine mechanisms implicated in the beneficial effects of OT in osteoporosis.

Transcription of the ot and otr genes is under the control of estrogens [53, 54]. Therefore, as the estrogen level is decreased in OVX mice and rats as well as in postmenopausal women, the level of OT is lower as well. It is thus tempting to speculate that hypogonadal-induced bone loss is linked to low OT levels, and that restoring OT levels could therefore reverse osteoporosis. Interestingly, recent reports have shown that sera from postmenopausal, but not premenopausal, women promote adipogenesis of mesenchymal stromal cells at the expense of osteogenesis [55, 56]. Our data strongly suggest that circulating OT is a key hormone that may account for these observations. Although OT-deficient mice display impairments in milk ejection and social recognition, and OTR-deficient mice exhibit disorders in several aspects of social behavior, bone defects remain to be shown in these mice [57, [58]–59].

Furthermore, we show for the first time that OT levels are inversely correlated with the development of osteoporosis in both rodents and humans. Interestingly, injecting OT into OVX mice, an animal model of osteoporosis, reversed the disequilibrium in the osteoblast/adipocyte balance, allowing the restoration of bone loss and the recovery of bone's normal biomechanical properties. As OT favored osteogenesis in vitro, it is tempting to postulate that it improves bone formation in vivo and therefore can be considered as an anabolic hormone. This hypothesis is in agreement with a previous report showing that OT injection in normal rats improves bone remodeling in favor of bone formation as circulating levels of Receptor Activator for Nuclear factor-κB Ligand (RANKL) decreased and those of Osteprotegerin (OPG) increased, leading to a decrease in RANKL/OPG ratio [24]. Furthermore, it has been shown that (i) OT stimulates PGE2 synthesis in both undifferentiated and differentiated human osteoblastic cells, (ii) PGE2 increases bone turnover favoring bone formation, and (iii) PGE2 inhibits adipocyte differentiation. Taken together, these observations are in favor of OT exerting its effects through PGE2 [16, 60, [61]–62].

Increased bone resorption by osteoclasts is considered as the main cause of hypogonadism-induced bone loss. As OTR is expressed and functional in human osteoclasts [63], it cannot be excluded that the increased bone mass observed in OT-treated OVX mice could be mediated through the inhibition of osteoclast differentiation and/or activity.

Recent reports have suggested that bone mass and remodeling are centrally controlled [64, 65]. Consistent with this, elevated levels of follicle-stimulating hormone (FSH) resulting from ovariectomy were reported to enhance bone resorption, and this effect was reversed by hypophysectomy [66, [67]–68]. It remains controversial, however, whether or not this effect of FSH is direct [69]. In any case our results favor the possibility that circulating OT from the pituitary gland plays a role in the control of bone mass but does not exclude the possibility that OT may also act centrally to direct bone homeostasis.

Several drugs have been reported to contribute to the reversal of bone loss, for example, by favoring osteogenesis or through antiresorptive effects [70, [71]–72]. Therapies developed to treat bone diseases in humans are considered to be either antiresorptive (including bisphosphonates, selective estrogen-receptor modulators, calcitonin, and vitamin D) or anabolic agents (parathyroid hormone) [5, 73, 74]. For most of these treatments, if not all, side effects have been reported, that is, osteonecrosis, dysphagia, esophagitis, headache, nausea, arthralgy, dizziness, and others [75, [76]–77]. As OT has already been safely administered to patients for other indications [78], it represents a highly promising molecule for the effective treatment of osteoporosis.

Conclusion

In conclusion, our results reveal that OT plays a major role in the osteoblast/adipocyte balance of MSC. OT levels are inversely correlated with the occurrence of osteoporosis in both rodents and humans. Finally, injecting OT into OVX mice, an animal model of osteoporosis, reversed the disequilibrium in the osteoblast/adipocyte balance, allowing the restoration of bone, the recovery of bone biomechanical properties, and a reduction of marrow adiposity.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

This work was supported by the Centre National de la Recherche Scientifique, by a grant from “Equipe FRM, soutenue par la Fondation pour la Recherche Médicale”, by CHU de Nice “Programme Hospitalier de Recherche Clinique Régional”, by CHR d'Orleans, and by the GEN-AU program from the Austrian Ministry for Science and Research (projects GOLD II and BIN II). We are grateful to Drs. R. Arkowitz and P. Follette for critical review of the manuscript, to Dr. N. Bonnet for providing rat plasma, and to O. Cochet and A. Doye for skilled technical assistance. N.W. is a fellow of Fondation de France; C.E. is a recipient of a fellowship from Fondation pour la Recherche Médicale.