Introduction

Osteoporosis is a major public health problem not only in postmenopausal women (type 1-primary osteoporosis) and people over 70 years old (type 2-primary osteoporosis) but also in people under 70 years and premenopausal women (secondary osteoporosis). In these latter cases, osteoporosis is a consequence of treatments and/or diseases such as diabetes mellitus (DM) [1]. Both primary and secondary osteoporosis are a significant contributor to morbidity, mortality, and lost life years globally. In the USA, osteoporosis and low bone mass combined affected approximately 53.6 million (54%) older adults in 2010 [2].

In the last decades, there has been a remarkable progress in our understanding of the mechanisms behind the development of osteoporosis and new drugs have been discovered to treat the disease. However, there is still an increasing treatment gap for patients with osteoporosis and a high risk of fracture. This is in part due to concerns about rare side effects such as atypical fractures and osteonecrosis of the jaw (e.g. bisphosphonates) and lack of long-term (> 5 year) efficacy (e.g. bisphosphonates and denosumab) [3]. Even the recently discovered osteoporosis drugs such as parathyroid hormone-related peptide analog, sclerostin antibody and cathepsin K inhibitor have been reported to induce serious adverse effects in clinical phase 3 trials (injection-site reaction, arthralgia, osteonecrosis of the jaw, atypical femoral fracture and stroke [3]). As the life expectancy of humans is increasing, the prevalence of osteoporosis itself and diseases leading to osteoporosis (e.g. DM) will continue to rise in the world [1, 4, 5]. Thus, there is a medical need to develop safer and more efficacious anti-osteoporotic therapies.

Bone remodeling is a highly regulated and dynamic process involving the orchestration of osteoclastic bone resorption and osteoblastic bone formation [6, 7]. An alteration between these anabolic and catabolic events can result in a decreased bone quality and an increased fracture risk. For example, the estrogen loss in women after menopause frequently results in excessive osteoclast formation and bone resorption activity leading to osteoporosis. This process is termed postmenopausal osteoporosis and is the most common type of osteoporosis [8,9,10].

Apart from the known vascular complications, patients with DM have pathological changes in bone remodeling leading to an increased risk of bone fracture [11]. Several studies have demonstrated that in patients with type 1 diabetes (T1D), there is a ~ 6–17 fold increased risk of hip fractures compared to age-matched people without T1D [12, 13]. While the mechanisms underlying bone loss in T1D are not completely understood, some mouse studies have shown that streptozotocin (STZ)-induced hyperglycemia positively regulates osteoclast differentiation and function through suppression of osteoblast maturation and increases in bone adiposity in long bones [14, 15].

Colony-stimulating factor-1 [CSF-1, also known as macrophage colony-stimulating factor (M-CSF)], released by different cell types including fibroblasts, endothelial cells, and osteoblasts, induces homodimerization and subsequent activation of its receptor CSF-1R (also known as c-Fms or CD115) [16, 17]. CSF-1R is expressed predominantly on monocytes, macrophages, myeloid dendritic cells, microglia, and osteoclasts. CSF-1R regulates the survival, proliferation, differentiation, and function of monocyte lineage cells, including macrophages and osteoclasts (bone resorption cells) [18]. Several studies have shown a pivotal role of the CSF-1/CSF-1R pathway in pathological bone resorption [19]. Mice lacking either CSF-1 or CSF-1R display osteopetrosis and reduced osteoclastogenesis [20, 21]. Moreover, genetic deletion or pharmacological blockade of CSF-1 prevents ovariectomy (OVX)-induced bone loss. Additionally, increased levels of CSF-1 protein and mRNA have been reported in bone marrow cells isolated from mice with OVX and levels of CSF-1R mRNA expression are significantly increased in cultured macrophages under hyperglycemia stages [22]. Finally, levels of CSF-1 in bone marrow cells positively correlate with number of osteoclasts [23] and very recently, it has been reported that a CSF-1R inhibitor (PLX3397) significantly abolished the LPS-induced bone microstructure damage [24]. Altogether, these results suggest that CSF-1/CSF-1R axis is involved in the pathogenesis of bone loss associated with OVX and T1D. Therefore, we aimed to determine the effect of a blocking anti-CSF-1R antibody on both cortical and trabecular bone in two accepted murine models of primary (OVX-induced bone loss) and secondary (STZ-induced T1D) osteoporosis using micro-computed tomography analysis.

Materials and methods

Animals

Experiments were performed on a total of 54 female ICR mice obtained from Envigo Laboratories (Mexico City, Mexico) weighing 25–30 g (10 weeks old) at the beginning of the experiments. The sample size was based on previous studies in our laboratory [25, 26]. Animals were randomly selected and divided into two experimental groups (OVX and T1D). OVX group (n = 24) was subdivided into four groups, naïve (n = 6), sham (n = 6), OVX + vehicle (Veh) (n = 6), OVX + 2G2 (n = 6), while T1D group (n = 23) was subdivided into three groups, citrate (n = 7), STZ + Veh (n = 7), STZ + 2G2 (n = 9). Additionally, a group of age-matched naïve mice was treated with 2G2 (n = 7). Efforts were made to minimize the number of animals used. Mice were housed in groups of four per cage at a constant temperature of 22 °C ± 2 °C and a 12-h light/dark cycle, with access to food (standard diet, catalog no. 2016S) and water ad libitum. Procedures involving mice and their care were conducted in conformity with the Mexican Official Norm for Animal Care and Handling (NOM-062-ZOO-1999) and the Guide for the Care and Use of Laboratory Animals. All animal procedures were approved by Animal Ethics Committee of the Universidad Autónoma de Tamaulipas (CEI-UAMRA-2016-0015).

Model of primary osteoporosis

Primary osteoporosis was induced in mice by performing a bilateral OVX at 10 weeks of age. Briefly, bilateral OVX or sham surgery (Sham) under ketamine/xylazine anesthesia (50/5 mg/kg, sc) were performed. Mice were placed in prone position and shaved between the last rib and hip. One-centimeter incision was made in the back on each side, then the ovaries were located and removed together with their capsule and part of the oviduct. The uterine horns were then sutured with absorbable suture, and the skin was closed with nylon 4–0 suture (Atramat, Nylon). At the end of the recovery (2 weeks after OVX), the success of the surgery was confirmed by vaginal cytology [27].

Model of secondary osteoporosis

Type-1 diabetes-induced osteoporosis was induced at 10 weeks of age. Female mice were daily administrated by intraperitoneal injection for 5 consecutive days with STZ (50 mg/kg in 0.1 M citrate buffer pH 4.5) as previously reported [25]. Control group was injected only with citrate buffer. Two weeks post-first injection, mice were fasted for 12 h and glucose levels were measured in a blood drop obtained from a 1 mm cut made in the distal tail tip using a glucometer (Accutrend Plus CTL., Roche Diagnostic). Mice with blood glucose levels higher than 200 mg/dL were considered as diabetic [25].

Micro-computed tomography analysis

Trabecular bone was analyzed at the level of the distal femur, proximal tibia, and femoral neck. Cortical bone was evaluated at the mid-diaphysis level of the femur and tibia, using micro-computed tomography system (Skyscan 1272, Bruker, Belgium). The scanning process was made at a 10 µm voxel size, an X-ray power of 60 kVp and 166 µA with an integration time of 627 ms, according to the guidelines for microCT analysis for rodent bone structure [28]. Obtained images were reconstructed using NRecon software (Bruker, Belgium). The trabecular region of interest (ROI) at distal femur and proximal tibia metaphysis was evaluated by selecting 2 mm in the vertical axis, subsequent to 0.5 mm from the growth plate (reference point). For the cortical ROI analysis, the sample level was selected a band of 1 mm thick at 4 mm distally from the growth plate. For the femoral neck analysis, ROI was selected using a 0.5 mm2 cylinder that was positioned in the region of the neck in all analyzed samples. The CT analyzer program (Bruker, Belgium) was used to determine trabecular bone parameters; an automatic segmentation algorithm (CT analyzer) was applied to isolate the trabecular bone from the cortical bone. The parameters used for the trabecular bone were trabecular bone volume rate (BV/TV), trabecular separation (Tb.Sp), trabecular number (Tb.N). Cortical bone parameters in 3D were cortical thickness (Ct.Th) and 2D mean cross-sectional bone area (B.Ar), mean cross-sectional tissue area (T.Ar) and mean cortical area fraction (B.Ar/T.Ar). Finally, hydroxyapatite calibration phantoms (250 and 750 mg/cm3) were used to calibrate trabecular and cortical bone mineral density values (BMD).

Administration of 2G2

The 2G2 is a chimera hamster and mouse IgG1 antagonistic antibody that binds to mouse CSF-1R with high affinity and enables specific and long-term CSF-1R inhibition in immune-competent mice. The dose of anti-CSF-1R antibody administered in this study was based on previous studies and adjusted to non-tumor bearing mice [29]. Four weeks post-surgery, OVX or sham mice, 14 weeks post-first injection of STZ, or age-matched naïve (14 weeks of age at the beginning of treatment) animals were administered 2G2 (anti-CSF-1R) or vehicle (ip) at the dose of 10 mg/kg once weekly for 6 consecutive weeks (Fig. 1). Based on our pilot studies, anti-CSF-1R was initially given when there was significant bone loss in the distal femur (OVX study) and neck of the femur (T1D study). At the end of the administration, mice were sacrificed using a CO2 chamber and hind limbs were harvested and stored at 4 °C in 0.1 M PBS (pH 7.4) until their analysis.

Fig. 1
figure 1

Experimental design for the evaluation of the anti-resorptive effect of anti-CSF-1R antibody. Murine models of (a) primary osteoporosis (ovariectomy, OVX) and (b) secondary osteoporosis (streptozotocin-induced type-1 diabetes; STZ-T1D)

Full size image

Reagents

The compounds used in the present study were Streptozotocin (STZ) (Sigma-Aldrich, catalog #S0130), Citric acid monohydrate (J.T. Baker, catalog #0110-01), Sodium citrate dihydrate (J.T. Baker, catalog #3646-01), antibody against CSF-1R (2G2, kindly provided by Roche Diagnostics GmbH, Germany), L-Histidine (Sigma-Aldrich, catalog #H8000), L-Histidin monohydro-chloride monohydrate (Sigma-Aldrich, catalog #H8125) and sodium chloride (Sigma-Aldrich, catalog #S5886).

Statistical analysis

All values are presented as mean ± standard error of the mean (SEM). A one-way analysis of variance (ANOVA) followed by Student–Newman–Keuls post hoc test was used to compare each parameter between the different experimental groups. Significance level was set at p < 0.05. Statistical comparisons were performed using Sigma Plot software V12.0.

Results

Effect of 2G2 antibody on body weight in an OVX model

Chronic administration (10 mg/kg; once per week for 6 weeks) of anti-CSF-1R (2G2) did not significantly affect the body weight between different experimental groups (37.31 ± 2.03 g in OVX mice treated with vehicle vs 39.76 ± 2.10 g in OVX mice treated with 2G2 antibody). At sacrifice time (10 weeks after OVX), the body weight was not statistically different between sham mice vs OVX mice (36.04 ± 1.59 g in sham mice vs 37.31 ± 2.03 g in OVX mice). The reason why OVX did not induce a body weight gain in ICR mice is unclear; however, it is possible that strain of mice may explain this observation. In support of this, a significant increase in body weight has been reported after OVX in C3H/HeJ mice, but not in BALB/cByJ, CAST/EiJ, DBA2/J, and C57BL/6 J mice [30].

Effect of 2G2 antibody on trabecular bone of distal femur and proximal tibia in an OVX model

Figure 2 shows representative images of longitudinal views of the femur from naïve group (Fig. 2a), sham group treated with vehicle (Veh) (Fig. 2b), OVX group treated with vehicle (Fig. 2c) and OVX mice treated with 2G2 antibody (Fig. 2d) at 10 weeks post-OVX surgery. Sham surgery itself did not modify the trabeculae as compared to age-matched naïve group. In contrast, OVX induced a significant loss of trabecular bone as compared to sham group. Chronic treatment for 6 weeks with a 2G2 antibody (10 mg/kg) reversed the OVX-induced trabecular bone loss in the femur. Quantitative trabecular bone analysis at the distal femoral metaphysis and proximal tibial metaphysis revealed that sham group had similar values in magnitude of tBMD (Figs. 2e, 3a), BV/TV (Figs. 2f, 3b), Tb.N (Figs. 2g, 3c), and Tb.Sp (Figs. 2h, 3d) as compared to age-matched naïve group. In contrast, mice with OVX had a significantly smaller tBMD (Fig. 2c, e), BV/TV (Fig. 2c, f), Tb.N (Fig. 2c, g) and greater Tb.Sp (Fig. 2c, h) as compared to sham group at the femoral metaphysis. At the tibial metaphysis, mice with OVX had a significant smaller BV/TV (Fig. 3b) and greater Tb.Sp (Fig. 3d) as compared to mice with sham surgery. At this anatomical level, mice with OVX had smaller values of tBMD (Fig. 3a) and Tb.N (Fig. 3c) but these parameters were not statistically significant vs those values in sham mice (p = 0.1728 and 0.0637, respectively). Interestingly, chronic treatment with 2G2 antibody reversed both the OVX-induced reduction of tBMD (Fig. 2e), BV/TV (Fig. 2f), Tb.N (Fig. 2g) and OVX-induced increase of Tb.Sp (Fig. 2h) at the distal femoral metaphysis. The anti-CSF-1R treatment also reversed the OVX-induced reduction of BV/TV (Fig. 3b) and OVX-induced increase of Tb.Sp (Fig. 3d) at the proximal tibial metaphysis. Additionally, we evaluated changes in bone remodeling in the femoral neck between OVX and sham groups and there were no statistical differences in tBMD, BV/TV and Tb.N (data not shown).

Fig. 2
figure 2

The 2G2 antibody reversed the loss of trabecular bone of the distal femur in an OVX model. Representative 3D images of longitudinal distal femur sections obtained by microCT. Ovariectomy + Vehicle group (OVX + Veh) (c) showed significant bone loss compared to naïve (a) and sham groups (b). After the administration of anti-CSF-1R (2G2), bone loss caused by OVX was reversed (d). After OVX surgery, the animals showed significant bone loss in trabecular bone mineral density (tBMD) (e), trabecular bone volume rate (BV/TV) (f), trabecular number (Tb.N) (g), and increases in trabecular separation (Tb.Sp) (h) compared to naïve group. Administration of 2G2 antibody (10 mg/kg) reversed bone loss caused by the OVX model according to tBMD (e), BV/TV (f), Tb.N (g), and Tb.Sp (h) parameters. Data are presented as mean ± SEM (n = 6 for each group). *p < 0.05, data were analyzed by one-way ANOVA, followed by Student–Newman–Keuls post hoc test

Full size image
Fig. 3
figure 3

The 2G2 antibody partially reversed the trabecular bone loss at the proximal tibia in an OVX model. There were no significant statistical differences among the groups in the trabecular bone mineral density (tBMD) (a), and trabecular number (Tb.N) (c) parameters. In the Ovariectomy + Vehicle (OVX + Veh) group, a significant decrease in trabecular bone volume rate (BV/TV) (b), and an increase in trabecular separation (Tb.Sp) (d) were found in comparison to naïve group. The administration of an anti-CSF-1R (2G2) antibody resulted in a significant increase in BV/TV (b), and a significant decrease in Tb.Sp (d) compared to the OVX + Veh group. Data are presented as mean ± SEM (n = 6 for each group). *p < 0.05, data were analyzed by one-way ANOVA, followed by Student–Newman–Keuls post-hoc test

Full size image

Effect of 2G2 antibody on cortical bone at the femoral and tibial mid-diaphysis in an OVX model

MicroCT analysis revealed that cBMD, Ct.Th, B.Ar, T.Ar and B.Ar./T.Ar of the mid-diaphyseal femur and tibia did not show significant differences between sham and age-matched naïve groups. Additionally, OVX did not induce any significant decrease on these parameters. Chronic treatment with 2G2 antibody in OVX mice did not significantly alter these parameters (Table 1).

Table 1 Structural parameters of cortical bone in femur and tibia in mice with primary osteoporosis
Full size table

Effect of 2G2 antibody on body weight and glucose levels in a T1D model

In our experimental conditions, two mice out of nine mice in the group of T1D plus vehicle and one mouse out of ten mice in the group of T1D + 2G2 showed signs of distress. Then, blood glucose levels were measured, and they were above the upper limit of detection in our glucometer (600 mg/dL), thus, these mice were sacrificed. Chronic treatment with 2G2 antibody did not modify the body weight of mice after 6 weeks of treatment; mice with STZ treated with vehicle weighed 33.19 ± 1.51 g and mice with STZ treated with 2G2 antibody weighed 34.42 ± 1.31 g. Regarding serum glucose levels, chronic treatment with 2G2 antibody did not significantly modify the fasting glucose levels (526.1 ± 22.38 mg/dL) as compared to mice with T1D treated with vehicle (476.9 ± 47.85 mg/dL).

Effect of 2G2 antibody on cortical and trabecular bone of the femur in a T1D model

Analysis of the mid-diaphysis cortical bone of the femur was performed at 19 weeks after induction of T1D. Figure 4a shows a longitudinal section to illustrate where the analysis was performed. Mice with STZ-induced T1D had a thinner cortical thickness and cortical area (Fig. 4c) as compared to citrate group (control) (Fig. 4b). Chronic treatment with 2G2 antibody in the STZ-induced T1D for 6 weeks reversed the deterioration of cortical bone (Fig. 4d). Quantitative analysis revealed that STZ-induced T1D significantly reduced the cBMD (Fig. 4e), Ct.Th (Fig. 4f), B.Ar (Fig. 4g), T.Ar (Fig. 4h) but not the B.Ar/T.Ar (Fig. 4i) as compared to citrate group. The 2G2 antibody treatment only significantly reversed the T1D-induced reduction of Ct.Th, B.Ar and T.Ar but not the cBMD. When the trabecular bone was analyzed at the distal metaphysis (Fig. 5a), STZ-induced T1D did not reduce the tBMD (Fig. 5c, e), BV/TV (Fig. 5c, f), or Tb.N (Fig. 5c, g). Surprisingly, mice with STZ-induced T1D and treated with 2G2 antibody showed a greater tBMD (Fig. 5d, e), BV/TV (Fig. 5f), and Tb.N (Fig. 5g) as compared to control mice (citrate group, Fig. 5b).

Fig. 4
figure 4

The 2G2 antibody reversed cortical bone loss in a T1D model. Three-dimensional longitudinal image section of distal femur (a), transverse sections of femoral cortical bone (b, c, d) obtained by microCT analysis. The quantifications show that mice with type-1 diabetes (T1D) induced by streptozotocin (STZ) had significant loss in cortical bone mineral density (cBMD) (e), cortical thickness (Ct.Th) (f), mean cross-sectional bone area (B.Ar) (g), and mean total cross-sectional tissue area (T.Ar) (h), but not in the mean cortical area fraction (B.Ar/T.Ar) (i) as compared to citrate group. The administration of an anti-CSF-1R (2G2) antibody reversed the loss of cortical thickness (f), mean cross-sectional bone area (g), and mean cross-sectional tissue area (h) in secondary osteoporosis (STZ-induced T1D) model. Data are presented as mean ± SEM (n = 7–9 for each group). *p < 0.05, data were analyzed by one-way ANOVA, followed by Student–Newman–Keuls post hoc test

Full size image
Fig. 5
figure 5

Effect 2G2 antibody on femoral trabecular bone of mice with T1D. Three-dimensional longitudinal image section of distal femur (a), and high power magnifications of the areas analyzed (b, c, d). In the trabecular analysis, no significant differences among trabecular bone mineral density tBMD (e), trabecular bone volume rate (BV/TV) (f) or trabecular number (Tb.N) (g) were found in the streptozotocin (STZ) group compared to the citrate group. The group administered anti-CSF-1R (2G2) showed a significant increase in tBMD (e), BV/TV (f), and Tb.N (g) compared to vehicle (Veh) group. Data are presented as mean ± SEM (n = 7–9 for each group). *p < 0.05, data were analyzed by one-way ANOVA, followed by Student–Newman–Keuls post-hoc test

Full size image

Effect of 2G2 antibody on trabecular bone of femoral neck in a T1D model

Figure 6a shows a longitudinal section to illustrate where the analysis was performed. Mice with STZ-induced T1D and treated with vehicle (Veh) had a loss of trabecular bone as reflected by a smaller number and greater separation of trabeculae (Fig. 6c) as compared to control group (Fig. 6b). Six weeks of treatment with 2G2 antibody partially reversed the deterioration of trabecular bone at femoral neck (Fig. 6d). Quantitative analysis revealed that mice with STZ-induced T1D had a significantly smaller tBMD (Fig. 6e), BV/TV (Fig. 6f), and greater Tb.Sp (Fig. 6h) as compared to citrate group. In the Tb.N parameter, only a light trend was observed in the STZ-induced T1D. Chronic treatment with 2G2 antibody partially reversed the T1D-induced decrease of tBMD and BV/TV. However, this antibody did not reverse the T1D-induced increase of Tb.Sp.

Fig. 6
figure 6

The 2G2 antibody partially reversed bone loss at the femoral neck in a mouse T1D model. Longitudinal section of proximal femur (a), transverse sections of femoral neck (b–d) obtained by microCT analysis. Mice with streptozotocin (STZ)-induced T1D had a significant decrease in trabecular bone mineral density (tBMD) (e), trabecular bone volume rate (BV/TV) (f), and an increase in the trabecular separation (Tb.Sp) (h), but no significant changes in the trabecular number (Tb.N) (g) compared to citrate group. The administration of an anti-CSF-1R (2G2) antibody partially relieved the loss of tBMD (e), and BV/TV (f) but it did not reduce the increase in Tb.Sp (h) compared to vehicle (Veh) group. Data are presented as mean ± SEM (n = 7–9 for each group). *p < 0.05, data were analyzed by one-way ANOVA, followed by Student–Newman–Keuls post hoc test

Full size image

Effect of 2G2 antibody on cortical and trabecular bone of naïve mice

The chronic administration of 2G2 antibody in age-matched naïve mice tended to increase the parameters of trabecular bone of the distal femur and proximal tibia (with exception of tibial tBMD and femoral Tb.N, in where there was a significant increase, Table 2) as compared to naïve mice. The antibody 2G2 did not significantly alter the parameters of cortical bone of the distal femur or proximal tibia or the trabecular bone at the femoral neck (Table 2).

Table 2 Structural parameters of trabecular and cortical bone in femoral neck, distal femur and proximal tibia in intact mice
Full size table

Discussion

It has been recently reported that regulation of the CSF-1/CSF-1R signaling pathway is more complex than previously thought, as CSF-1 has at least three different isoforms; membrane-bound (mCSF-1), soluble (sCSF-1) and “shed” CSF-1, all which bind to CSF-1R [19, 31,32,33]. Furthermore, recent studies have identified IL-34 as a novel ligand of the CSF-1R [33, 34]. IL-34 and CSF-1 bind to partially different binding sites on CSF-1R and have a complementary, but not redundant activation of CSF-1R [18]. Thus, blockade of the receptor, instead of ligands, could offer an alternative and likely a more effective way to inhibit the CSF-1/CSF-1R pathway. While a previous study showed that neutralization of CSF-1 prevented the loss of cortical bone density (DEXA evaluation) induced by OVX [23], the present study shows that blockade of the CSF-1R reverses not only the loss of cortical bone density but also restores the microarchitecture of trabecular bone in two different mouse models of osteoporosis (primary and secondary) using highly sensitive microCT analysis. Our study is in agreement with a very recently study in which a CSF-1R inhibitor can protect against LPS-induced bone loss and deterioration of bone biomechanical properties by inhibiting osteoclast formation [24].

Our data show that chronic treatment with antibody against CSF-1R (2G2) reverted the OVX-induced loss of the trabecular bone at the distal femur and proximal tibia. While the mechanism behind this effect was not examined in the current study, we suggest, based on previous reports that this effect could result from a blockade of osteoclastogenesis and osteoclast reabsorbing activity mediated by CSF-1/CSF-1R pathway. First, clinical and preclinical studies have shown that estrogen-deficiency bone loss results mainly from increased osteoclast number and bone reabsorbing activity [35, 36]. Second, in vitro studies have shown increased levels of both CSF-1 mRNA and protein by stromal cells isolated from mice with OVX and a direct relationship between CSF-1 levels and osteoclasts production [37, 38]. Third, in vivo studies have shown that genetic deletion of CSF-1 or neutralizing antibody to CSF-1 prevents the OVX-induced bone loss in mice [19, 23, 39]. Finally, women with late menopause (≥ 30 years after menopause) have significant increased levels of CSF-1 as compared to women in the premenopausal stage [40] and protein levels of CSF-1 in the femoral head are significantly higher in patients with osteoporotic-related fractures than in patients with osteoarthritis [41].

Osteoporosis is not only a major public health problem in postmenopausal women or aged people but also in premenopausal women as a consequence of treatment and/or chronic disease such as diabetes. Both T1D and T2D have been reported to increase risk of fracture; however, T1D has a stronger association with hip fracture (summary relative risk of 6.3) than T2D (summary relative risk of 1.7) [12]. In our study, we showed that mice with T1D have a significant cortical bone loss at the distal femur and neck of the femur (an anatomical site where hip fractures occur) and that blockade of CSF-1R reverts partially the loss of cortical bone in the distal femur and femoral neck, suggesting that an increased activation of the CSF-1/CSF-1R pathway is also involved in the bone loss by T1D. In support of this hypothesis, both the levels of mRNA and protein for CSF-1 and its receptor, CSF-1R are significantly higher in retina from rats with T1D as compared to age-matched naïve rats [38]. However, it should be recognized that we did not determine the levels or expression of CSF-1 or its receptor, in bone tissue. Future studies are needed to determine serum and/or bone levels of CSF-1 and its receptor as well as the pattern of expression of these proteins in bone tissue under T1D conditions. The observation that mice with STZ-induced T1D did not show trabecular bone loss in the femoral metaphysis was somewhat surprising given the evidence that bone loss in these mice occurred even at femoral cortical bone. The reasons behind this observation are unknown; however, the strain of mice could explain why these mice are ¨resistant¨ to trabecular bone loss at the distal femur induced by STZ. In support of this, it has been reported that the diabetogenic effects of STZ in mice are strain dependent [42] and that magnitude of trabecular bone loss at the femoral metaphysis in a model of disuse osteoporosis is influenced by the mouse strain [43]. Our results show that the model of T1D results in a greater loss of cortical bone compared to OVX model. These results are in agreement with previous studies which have reported that OVX performed in C57BL6 and BALB/c mice did not induce a loss of the tibial cortical bone even at prolonged times (24 weeks) after OVX [44]. In contrast, several studies have shown that mice with STZ-induced T1D display a significant loss of both trabecular and cortical bone [14, 45]. The mechanisms underlying these model-related differences are not fully known. However, it has been suggested that T1D-induced bone loss is more severe compared to OVX-induced bone loss due to, at least in part, that T1D negatively affects both bone formation and bone reabsorption, in contrast, the OVX-induced bone loss results mainly from an increased bone reabsorption [46].

On the other hand, the microCT analysis revealed that treatment with 2G2 antibody in naïve mice tended to increase the trabecular bone at the proximal metaphysis tibia and distal metaphysis femur but not in the cortical bone of the tibia, femur and trabecular femoral neck. According to this, a significant increase in the density of trabecular bone in the femur has been previously reported in mice lacking this receptor [47]. The reasons behind the bone regional differences in terms of response to this antibody are unknown, but possibly relates to the femoral neck having a relatively lower rate of bone turnover than trabecular bone in the distal femur and proximal tibia. Another explanation could relate to periosteal modeling that is less pronounced at the femoral neck compared to the femoral or tibial midshafts [48].

Our study has several limitations. First, the mechanisms behind the anti-osteoporotic effect of the 2G2 antibody were not fully explored. Thus, future studies evaluating the effect of this drug on osteoclastogenesis and reabsorbing activity as well as on osteoblastogenesis and/or bone formation are needed. Second, we did not validate the antibody effect on osteoclast activity or density. While a previous study demonstrated that a different anti-CSF-1R antibody decreases the number of osteoclasts [49], we do not have direct evidence for the effect of the 2G2 antibody on osteoclast number and/or activity. Our rationale for using this antibody was based on the study by Ries C et al., 2014 (who kindly provided us the antibody) that reported a high affinity (KD = 0.2 nM) of the 2G2 antibody binding to mouse CSF-1R, which is essential for the development and maturation of osteoclasts, and reduces the survival of CSF-1-dependent murine M-NFS-60 cells (an in vitro model of bone marrow precursors) [29]. Third, while our results show that blocking the CSF-1/CSF-1R axis reverts the loss of cortical and trabecular bone in two murine models of osteoporosis, it is unknown if pharmacological blockade of this pathway will also improve the mechanical properties of bone and thus increase the bone’s resistance to fracture. Finally, a possible effect of the anti-CSF-1R antibody on disease progression in the T1D was not explored. However, our results show that T1D mice treated with anti-CSF-1R have similar values of body weight and serum glucose compared to T1D mice treated with vehicle. Additionally, it is important to mention that we did not measure potential side effects of anti-CSF-1R treatment, although normal histology in various tissues has been reported after chronic administration of this antibody in mice [50].

Conclusions

In summary, chronic treatment with an anti-CSF-1R antibody (2G2) significantly reversed the bone loss observed in mouse models of primary and secondary osteoporosis. Our results suggest that CSF-1/CSF-1R may play a key role in the bone loss induced by estrogen deficiency and T1D. Further studies elucidating the cellular mechanisms behind the anti-osteoporotic effects of CSF-1R blockade will increase our understanding of the pathogenesis of osteoporosis.