Abstract
Bone strength is partially determined during cortical bone consolidation, a process comprising coalescence of peripheral trabecular bone and its progressive mineralisation. Mice with genetic deletion of suppressor of cytokine signalling 3 (Socs3), an inhibitor of STAT3 signalling, exhibit delayed cortical bone consolidation, indicated by high cortical porosity, low mineral content, and low bone strength. Since leptin receptor (LepR) is expressed in the osteoblast lineage and is suppressed by SOCS3, we evaluated whether LepR deletion in osteocytes would rectify the Dmp1cre.Socs3fl/fl bone defect. First, we tested LepR deletion in osteocytes by generating Dmp1cre.LepRfl/fl mice and detected no significant bone phenotype. We then generated Dmp1cre.Socs3fl/fl.LepRfl/fl mice and compared them to Dmp1cre.Socs3fl/fl controls. Between 6 and 12 weeks of age, both Dmp1cre.Socs3fl/fl.LepRfl/fl and control (Dmp1cre.Socs3fl/fl) mice showed an increasing proportion of more heavily mineralised bone, indicating some cortical consolidation with time. However, at 12 weeks of age, rather than resolving the phenotype, delayed consolidation was extended in female Dmp1cre.Socs3fl/fl.LepRfl/fl mice. This was indicated in both metaphysis and diaphysis by greater proportions of low-density bone, lower proportions of high-density bone, and greater cortical porosity than Dmp1cre.Socs3fl/fl controls. There was also no change in the proportion of osteocytes staining positive for phospho-STAT3, suggesting the effect of LepR deletion in Dmp1cre.Socs3fl/fl mice is STAT3-independent. This identifies a new role for leptin signalling in bone which opposes our original hypothesis. Although LepR in osteocytes has no irreplaceable physiological role in normal bone maturation, when STAT3 is hyperactive, LepR in Dmp1Cre-expressing cells supports cortical consolidation.
Introduction
Bone strength is partially determined during cortical bone consolidation. This is a three-step process comprising coalescence of peripheral trabecular bone on the endocortical surface, a gradual transformation of the consolidated material from woven to lamellar bone, and progressive mineralisation of bone material (Enlow 1962, Frost 1982, Isojima & Sims 2021).
Recent data have indicated that, in the postnatal period, cortical consolidation is delayed when STAT3 phosphorylation in osteocytes is elevated. STAT3 is a ubiquitous intracellular transcription factor that is phosphorylated by multiple bone-active cytokines, and its signalling receives negative feedback from the intracellular inhibitor suppressor of cytokine signalling 3 (SOCS3) (Sims 2020). Targeted deletion of SOCS3 in osteocytes, in the Dmp1cre.Socs3fl/fl mouse model, led to elevated STAT3 phosphorylation at basal levels and a prolonged and elevated STAT3 phosphorylation response to treatment with interluekin (IL)-6 family cytokines (IL-11, oncostatin M, and leukemia inhibitory factor) (Cho et al. 2017, Walker et al. 2020). This Dmp1cre-targeted hyperactivation of STAT3 signalling was associated with delayed cortical consolidation characterised by high cortical porosity, high proportions of less mineralised and woven bone, and low bone strength (Cho et al. 2017, Walker et al. 2020). This phenotype was exaggerated in two conditions when STAT3 signalling in osteocytes was increased further. First, when the mice were subjected to mechanical loading, a condition that promotes STAT3 signalling in osteocytes (Mantila Roosa et al. 2011, Corry et al. 2019), either by treadmill running or tibial loading, they formed more low-density and woven bone, respectively, than normal (McGregor et al. 2022). Secondly, when the proportion of phospho-STAT3-positive osteocytes were increased even further in Dmp1cre.Socs3fl/fl mice by global deletion of G-CSFR, the defective cortical phenotype was worsened such that a consolidated cortex was still not formed by 26 weeks of age (Isojima et al. 2022). Conversely, the delayed corticalisation phenotype of Dmp1cre.Socs3fl/fl mice was partially rescued when the proportion of phospho-STAT3 positive osteocytes was suppressed by targeted deletion of the common IL-6 family receptor (gp130) (Walker et al. 2020). These studies all indicated that hyperactivation of STAT3 in osteocytes leads to a high level of bone resorption in the cortex that would normally be suppressed for cortical consolidation to occur.
SOCS3 also provides negative feedback to leptin-induced JAK-STAT signalling. It achieves this by binding to tyrosine-985 on the leptin receptor (LepR) and preventing its continued phosphorylation and activation of downstream pathways (Bjørbæk et al. 2000, Eyckerman et al. 2000, Howard & Flier 2006). While increased SOCS3 expression in other tissues (e.g. hypothalamus) can suppress leptin-STAT3 signalling (Münzberg et al. 2004), the interaction between SOCS3 and leptin signalling has not been explored in bone.
There have been several studies examining the role of leptin signalling in osteoblast lineage cells (Wee & Baldock 2014, Reid et al. 2018). Leptin receptor is expressed by mesenchymal stromal cells (Zhou et al. 2014, Tikhonova et al. 2019). Lineage tracing using LepRcre has showed LepR+ cells can form colonies, differentiate into osteoblasts, adipocytes, and chondrocytes, and can support the haematopoietic niche by secreting factors such as stem cell factor (SCF) and CXCL12 (Zhou et al. 2014). Some studies have suggested that osteoblasts do not express the leptin receptor (Ducy et al. 2000, Zhou et al. 2014) and do not respond to leptin stimulation as assessed by STAT3 activation (Ducy et al. 2000). However, there are also several studies that demonstrate LepR expression in osteoblasts (Scheller et al. 2010) and a direct effect of leptin signalling on osteoblasts both in vivo and in vitro (Scheller et al. 2010, Turner et al. 2013). This suggests that detection of leptin signalling within osteoblast lineage cells can be difficult and emphasises the importance of exploring multiple models since the role of leptin signalling may be context-dependent. The presence of leptin signalling in osteocytes has not yet been described, but mRNAs associated with leptin signalling (Lep, Socs3, Stat3) are upregulated in the first 4 h in the bone’s response to mechanical load (Kapur et al. 2010, Mantila Roosa et al. 2011), likely mediated by the osteocyte network and its role in mechano-responsiveness.
SOCS3 depletion in osteocytes would be expected to lead to prolonged and elevated STAT3 signalling downstream of all receptors that rely on it for negative feedback; this would include any STAT3 response to leptin signalling. Our aim in this study was to evaluate whether blocking leptin receptor signalling in SOCS3-deficient cells would rectify the deficit in cortical bone consolidation in Dmp1cre.Socs3fl/fl mice. We find that leptin receptor deletion did not rectify the defect, but worsened it, specifically in female mice, suggesting a role for leptin receptor signalling in Dmp1cre positive cells that promotes cortical bone consolidation.
Materials and methods
Mice
All animal procedures were conducted with approval of the St. Vincent’s Health Melbourne Animal Ethics Committee (Ethics Number: #005/16). The Dmp1cre.Socs3fl/fl line was previously established (Cho et al. 2017, Walker et al. 2020) using mice with a constitutive 10 kb promoter Dmp1cre (Tg(Dmp1-cre)1Jqfe) (Lu et al. 2007) and mice with a Socs3-floxed allele (Socs3tm1Wsa) (Ushiki et al. 2016). To generate mice with specific deletion of leptin signalling, we bred Dmp1cre or Dmp1cre.Socs3fl/fl mice with LepR floxed (Leprtm1Rck/JAusb) mice (Cohen et al. 2001). All mouse strains had previously been crossed to and maintained on a C57BL/6 background. All mice were housed at 22°C. Analyses were performed with observers blinded to the sex and genotype of the animals. Details of the mice used for this study are described below.
Cohort 1: Dmp1cre vs. Dmp1cre.LepRfl/fl
Experimental Dmp1cre.LepRfl/fl and control mice (Dmp1cre) were bred in parallel such that they were cousins. Both sexes were examined. MicroCT analysis was performed ex vivo at 12 weeks of age on a Skyscan 1276. Mice were fasted overnight prior to tissue collection at 12 weeks of age.
Cohort 2: Dmp1cre.Socs3fl/flvs. Dmp1cre.Socs3fl/fl.LepRfl/fl
Experimental Dmp1cre.Socs3fl/fl.LepRfl/fl and control Dmp1cre.Socs3fl/fl mice were bred in parallel such that they were cousins to ensure sufficient mice of the specified genotypes were obtained and evaluated within a single cohort. Both sexes were examined. MicroCT analysis was performed in vivo at 6 weeks of age and ex vivo at 12 weeks of age on a Skyscan 1076. Mice were fasted overnight prior to tissue collection at 12 weeks of age.
Genotyping
DNA was extracted from murine toe tissue. Genotyping was performed using standard methods with specific primers for Cre, Socs3, and LepR alleles. To confirm that Dmp1cre induced recombination of the LepR-floxed allele, frozen femora, that had been flushed to remove the bone marrow, were processed for DNA isolation using the ISOLATE II Genomic DNA Kit (BIO-52066, Bioline) and then assessed by PCR. Primer sequences and product sizes for these alleles are provided in Supplementary Table 1 (see section on supplementary materials given at the end of this article). We have previously confirmed Socs3 recombination in bone using Dmp1cre (Cho et al. 2017).
MicroCT analyses
MicroCT was performed on the right femur using either a SkyScan 1076 system or a Skyscan 1276 system (Bruker-microCT, Kontich, Belgium) (McGregor et al. 2019, Walker et al. 2021). Detailed scan settings are provided in Supplementary Table 2. In vivo scans were performed on mice sedated with ketamine (100 mg/kg)/xylazine (10 mg/kg). Calcium hydroxyapatite phantoms of known densities (0.25 g/cm3 and 0.75 g/cm3, SKY016946, Bruker Skyscan) were scanned to facilitate calibration of the scans to known tissue mineral density. Images were reconstructed and analysed using SkyScan software programs: NRecon (version 1.7.4.2), Dataviewer (version 1.5.4), CT Analyser (version v1.16.4.1) and CtVox (version 2.3.2.0). Femur length was measured from the microCT scans and was used to correct regions of interest (ROI) for any changes in bone length.
Standard trabecular and cortical analyses were performed on Dmp1cre.LepRfl/fl femora compared to Dmp1cre control femora as previously described (McGregor et al. 2019). The trabecular ROI was drawn within the metaphyseal region. This was defined as 15% of the femoral length, commencing at a distance of 7.5% of the total femur length from the distal growth plate extending towards the mid-shaft. A standard binary threshold set at 0.4184 g/cm3 was applied to distinguish bone from non-bone. The diaphyseal ROI was 15% of the femoral length, commencing at a distance of 30% of the total femur length from the distal growth plate and extending towards the proximal end. The same threshold used for the metaphyseal region was applied to the diaphysis.
For multi-level bone thresholding, the metaphyseal ROI (including both trabecular and cortical bone) was used for analysis. We used an unbiased image analysis method (Otsu) to set thresholds for bone volume at four apparent densities (Walker et al. 2020, 2021), based on the local spatial densities obtained from scans of 12-week-old male Dmp1cre.Socs3fl/fl femora. The lowest threshold bins non-bone tissue and is not reported, while the other thresholds stratify bone into low (0.4346–0.7522 g/cm3), medium (0.7523–1.0980 g/cm3), and high density (greater than 1.0980 g/cm3 calcium hydroxyapatite). Total volumetric measurements were determined based on the whole ROI. Slice-by-slice areal analyses commenced at a distance of 7.5% of the total femur length from the distal growth plate and extended for 2 mm towards the mid-shaft as previously described (Walker et al. 2020, 2021).
The diaphyseal ROI was 15% of the femoral length, commencing at a distance of 30% of the total femur length from the distal growth plate and extending towards the proximal end. The same bone thresholds used for the metaphyseal region were applied to the diaphysis.
Mechanical strength testing
Mechanical properties of femora were derived from three-point bending tests using a Bose Biodynamic 5500 Test Instrument (Bose, DE, USA), as described previously (Williamson et al. 2017). Specifically, each femur was loaded cranio-caudally with a span length of 8.5 mm and 0.5 mm/s loading speed. Once whole-bone properties were determined, tissue-level mechanical properties were calculated using microCT analysis of the mid-shaft (Jepsen et al. 2015). Biomechanical measurements were derived as described by Turner and Burr (Turner & Burr 1993), using bone geometry parameters at the mid-diaphysis derived from microCT scans to calculate the material-specific properties. A 0.1% strain offset was applied to determine the yield point (Zhang et al. 2021). One sample was excluded due to an abnormal force/displacement curve.
RNA extraction and real-time quantitative PCR
At the time of collection, epiphyses were cut from the femur and the marrow flushed using phosphate buffered saline. The remaining femoral cortical bone shaft was snap frozen in liquid nitrogen. These samples were homogenised in Qiazol with an LS-10-35 Polytron. The RNA was purified using RNeasy lipid tissue minikits (Qiagen), according to manufacturer’s instructions including on-column DNAse digestion. RNA was quantified using a Nanodrop and yielded 260:280 ratios from 1.84 to 2.03. cDNA synthesis of RNA from each was performed with Tetro cDNA Synthesis kit (BIO-65043, Bioline) as per the manufacturer’s instructions. Quantitative real-time PCR was performed using SYBR Select Master Mix (–4472908, Applied Biosystems) in triplicate and run on a PCR machine either Stratagene MxPro 3000p or Agilent Aria Mx. Specific qPCR primer details are provided in Supplementary Table 1. Results were calculated using the ΔCT method.
Immunohistochemistry for phospho-STAT3
Immunohistochemistry for phospho-STAT3 (pSTAT3) was carried out as previously described (Isojima et al. 2022). Briefly, tibiae were decalcified with EDTA and embedded in paraffin. Five micrometre tibial sections were deparaffinised using histolene, followed by rehydration using a graded series of ethanol solutions (100%, 100%, 90%, and 70%) and deionised water. Endogenous peroxidase was quenched with 3% hydrogen peroxide in methanol. Antigen retrieval used 0.5% Trypsin-EDTA solution (T4049, Sigma) blocked in 10% swine serum (S4000, Vector labs) with 2% Tween-20, pSTAT3 Y705 antibody (1:100, #9131, Cell Signalling) was applied overnight. The next day, slides were washed with Tris-NaCl-Tween (TNT) buffer (0.1 M Tris-HCL, 0.15 M sodium chloride, and 0.05% Tween-20) and incubated with swine anti-rabbit biotinylated secondary antibody (1:250, E0353, Dako) for 45 min and followed by streptavidin horseradish peroxidase (1:500, P0397, Dako). Further amplification of antibody signal was performed using biotin tyramine for 7 min (1:50, TSA Biotin system kit, Perkin Elmer). Detection was completed using a diaminobenzidine colorimetric kit (K3468, Dako) and counterstained with Mayer’s haematoxylin. The number of pSTAT3-positive and pSTAT3-negative osteocytes were counted using Osteomeasure (Osteometrics) as described. The region analysed was 1.2 mm proximal to the growth plate and extended for a length of 1.6 mm. All osteocytes within 400 µm of the periosteal surface were included. Representative images were taken from the lateral side of the tibia, using a Leica DM RB microscope with a DP72 camera and CellSens software (Olympus CellSens Entry3.2 - Build 23706).
Statistical analysis
The sample number was determined based on previous studies using Dmp1cre.Socs3fl/fl mice (Cho et al. 2017, McGregor et al. 2019). Data were plotted with the mean with s.e.m. Statistical analyses were performed in GraphPad Prism (version 9.1.2) and specific test details are included within the figure legends. Body weight and bone mass/structure are known to be different between sexes; thus, the effect of LepR deficiency was compared only to sex-matched control mice. All microCT measurements, histomorphometry, and mechanical testing were conducted with the observers blinded to the sex and genotype of the samples.
Results
Dmp1cre-targeted Leptin receptor deficiency under physiological conditions does not alter bone structure
Dmp1cre-induced recombination of the LepR floxed allele was confirmed in femoral samples by PCR (Fig. 1A). Here, 12-week-old Dmp1cre.LepRfl/fl mice had no difference in body weight in either sex compared to Dmp1cre control mice (Table 1). Male, but not female, Dmp1cre.LepRfl/fl mice had slightly longer femora (2%) compared to Dmp1cre control mice (Table 1). By standard microCT analysis (single threshold), leptin deficiency had no effect on trabecular or cortical bone mass or dimensions in male mice (Table 1). In females, there were also no significant differences apart from a <5% smaller trabecular thickness compared to controls (Table 1).
Body weight and microCT analyses between Dmp1cre and Dmp1cre.LepRfl/fl femora at 12 weeks of age.
Parameter | Dmp1cre (n = 10) | Dmp1cre.LepRfl/fl (n = 10) | P-value |
---|---|---|---|
Males | |||
Body weight (g) | 27.5 ± 0.7 | 26.7 ± 0.8 | 0.46 |
Femur length (mm) | 13.89±0.08 | 14.23±0.10 | 0.02 |
Bone volume/tissue volume ratio (BV/TV) | 10.8 ± 0.9 | 12.8 ± 1.9 | 0.34 |
Trabecular thickness (µm) | 51 ± 1 | 52 ± 1 | 0.28 |
Trabecular number (1/mm) | 2.1 ± 0.2 | 2.4 ± 0.3 | 0.41 |
Cross-sectional area (mm2) | 2.00 ± 0.07 | 2.06 ± 0.03 | 0.58 |
Cortical bone area (mm2) | 0.91 ± 0.04 | 0.93 ± 0.02 | 0.72 |
Marrow area (mm2) | 1.09 ± 0.04 | 1.13 ± 0.02 | 0.49 |
Cortical thickness (mm) | 215 ± 5 | 215 ± 2 | 0.99 |
Periosteal perimeter (mm) | 7.6 ± 0.2 | 7.7 ± 0.1 | 0.70 |
Endocortical perimeter (mm) | 3.7 ± 0.1 | 3.9 ± 0.1 | 0.54 |
Mean polar moment of inertia (mm4) | 0.48 ± 0.04 | 0.50 ± 0.01 | 0.67 |
Females | |||
Body weight (g) | 21.9 ± 0.9 | 22.4 ± 0.50 | 0.65 |
Femur length (mm) | 13.81 ± 0.11 | 13.77 ± 0.05 | 0.67 |
Bone volume/tissue volume ratio (BV/TV) | 5.2 ± 0.5 | 4.33 ± 0.6 | 0.28 |
Trabecular thickness (µm) | 46 ±1 | 43 ±1 | 0.04 |
Trabecular number (1/mm) | 1.1 ± 0.1 | 1.0 ± 0.1 | 0.39 |
Cross-sectional area (mm2) | 1.71 ± 0.05 | 1.73 ± 0.03 | 0.76 |
Cortical bone area (mm2) | 0.82 ± 0.02 | 0.81 ± 0.02 | 0.85 |
Marrow area (mm2) | 0.90 ± 0.03 | 0.92 ± 0.02 | 0.48 |
Cortical thickness (mm) | 214 ± 4 | 210 ± 2 | 0.31 |
Periosteal perimeter (mm) | 6.8 ± 0.1 | 6.9 ± 0.1 | 0.60 |
Endocortical perimeter (mm) | 3.1 ± 0.1 | 3.1 ± 0.1 | 0.43 |
Mean polar moment of inertia (mm4) | 0.35 ± 0.02 | 0.35 ± 0.01 | 0.70 |
Values shown are mean ± s.e.m. P-values were determined by two-tailed t-test. Statistical significance (P < 0.05) are highlighted in italics and bold.
Since Dmp1cre-induced deletion of SOCS3 delayed accrual of high-density bone content (Walker et al. 2020), we performed multi-level thresholding to determine if leptin receptor deficiency alone had any effects on the proportions of low-, mid-, and high-density bone within the femoral metaphysis and diaphysis. When multi-level thresholding is applied, we observe that low- and medium-density thresholds are associated with trabecular bone and the growth plate, while high-density bone is associated with the epiphysis and cortical bone. We neither observed changes to total femoral metaphyseal or diaphyseal bone volumes nor any change in the proportions of metaphyseal or diaphyseal bone at each density threshold in either male or female Dmp1cre.LepRfl/fl femora compared to Dmp1cre controls (Fig. 1C and D). This indicates that leptin receptor deficiency in Dmp1Cre-expressing cells has no effect on bone structure during physiological conditions and that such a basal phenotype would not complicate a study of the effect of LepR deletion in Dmp1Cre.Socs3f/f mice.
Metaphyseal cortical consolidation is further delayed by leptin receptor deletion in female mice
To evaluate metaphyseal bone consolidation in Dmp1cre.Socs3fl/fl.LepRfl/fl mice, we monitored changes in the proportions of low-, medium-, and high-density bone in 6-week-old and 12-week-old femora similar to a previous study (Isojima et al. 2022). Between 6 and 12 weeks of age, we observed no difference in total metaphyseal bone volume in female mice as the skeleton matured (Fig. 2A). However, in male Dmp1cre.Socs3fl/fl control mice, we observed a significant reduction in total metaphyseal bone volume (Fig. 2C) as previously reported (Cho et al. 2017). Despite this difference, both male and female Dmp1Cre.Socs3fl/fl mice accrued higher proportions of high density bone between 6 and 12 weeks, indicating some progression of maturation of the bone cortex with age (Fig. 2).
In female 6-week-old Dmp1cre.Socs3fl/fl.LepRfl/fl mice, there were no differences in volume of bone at any density level compared to Dmp1cre.Socs3fl/fl mice, indicating similar bone phenotypes at this age (Fig. 2A). At 12 weeks of age, however, femora from female Dmp1cre.Socs3fl/fl.LepRfl/fl mice had a higher proportion of low-density bone than Dmp1cre.Socs3fl/fl control femora and lower proportions of both medium- and high-density bone (Fig. 2A, B), indicating a delay in cortical bone maturation. Dmp1cre.Socs3fl/fl.LepRfl/fl mice had accrued half the amount of medium- and high-density bone between the 6 and 12 weeks of age compared to Dmp1cre.Socs3fl/fl mice (Fig. 2A). The presence of low-density bone was also observed to extend further from the metaphyseal region towards the diaphysis in Dmp1cre.Socs3fl/fl.LepRfl/fl femora compared to Dmp1cre.Socs3fl/fl femora (Fig. 2B). In the same area, the amount of high-density bone appeared less in Dmp1cre.Socs3fl/fl.LepRfl/fl femora within the cortical bone (Fig. 2B). This indicates that absence of leptin receptor signalling in female mice with STAT3 hyperactivation further impairs bone consolidation.
No difference in the proportions of low-, medium-, and high-density bone were seen in femora of male Dmp1cre.Socs3fl/fl.LepRfl/fl mice compared to Dmp1cre.Socs3fl/fl mice at either 6 or 12 weeks of age (Fig. 2C and D). This indicates no effect of leptin deficiency on cortical bone maturation in male mice with activated STAT3.
Since we previously reported that the proportions of low-, medium-, and high-density bone change along the length of the bone with increasing bone maturity and distance from the epiphyseal growth plate (Walker et al. 2020), we assessed bone mass at each threshold slice-by-slice along its length in the femora from female mice. Representative cross-sections from female 6-week-old (Fig. 3A) and 12-week-old mice (Fig. 3B) show the distribution of these density thresholds. Low-density bone was observed on bone surfaces and is particularly abundant in regions where coalescence of peripheral trabecular bone on the endocortical surface occurs, with more being detected in the 12-week-old female Dmp1cre.Socs3fl/fl.LepRfl/fl mice than Dmp1cre.Socs3fl/fl (Fig. 3A and B). As previously observed in 12-week-old femora (Walker et al. 2020, McGregor et al. 2022), the proportion of low-density bone was less in the slices taken furthest from the growth plate and there was a concomitant increase in the proportion of high-density bone (Fig. 3D, F and H). At 6 weeks of age (Fig. 3A, C and G), we observed a similar gradient showing cortical maturation along the length of the bone in both genotypes; the younger bone had more low-density bone and accrued less high-density bone than at 12 weeks of age. More low-density bone was observed in Dmp1cre.Socs3fl/fl.LepRfl/fl femora than Dmp1cre.Socs3fl/fl control mice from 1 to 1.5 mm distal from the growth plate end of the ROI (Fig. 3C); however, there were generally no differences in the proportions of medium- and high-density bone (Fig. 3E and G). At 12 weeks of age, a greater proportion of low-density bone and lower proportions of medium- and high-density bone were observed in Dmp1cre.Socs3fl/fl.LepRfl/fl mice compared to Dmp1cre.Socs3fl/fl controls (Fig. 3D, F and H). This indicates that leptin receptor deficiency within Dmp1cre.Socs3fl/fl mice further delays metaphyseal cortical consolidation.
Delay in cortical bone mineralisation is conserved within the diaphysis
We next determined whether the delay in bone consolidation in the metaphysis extended into the most mature bone, found at the diaphysis (mid-shaft). Dmp1cre.Socs3fl/fl.LepRfl/fl mice had greater amounts of low-density bone and lower amounts of high-density bone relative to Dmp1cre.Socs3fl/fl control mice with no change in the total bone volume (Fig. 4A and B). This indicates that the delay in bone mineralisation was conserved within the diaphysis.
We have previously reported that Dmp1cre.Socs3fl/fl femora have a mechanical strength defect characterised by lower work-to-failure and post-yield deformation compared to Dmp1cre controls (Cho et al. 2017). We mechanically tested the femora of Dmp1cre.Socs3fl/fl.LepRfl/fl mice at the midpoint (diaphyseal region) to determine if bone strength was reduced further and in a similar manner. Representative load-displacement (Fig. 4C) and stress–strain curves (Fig. 4D) showed that both control and experimental femora generated similar curves during testing. We observed no differences in the force required to break the femur (Fig. 5E), the stiffness of the bones (Fig. 5F), or the yield strain (Fig. 5G). All other parameters showed no difference (Table 2), indicating that there was no difference in bone strength between Dmp1cre.Socs3fl/fl.LepRfl/fl and control Dmp1cre.Socs3fl/fl femora.
Additional mechanical properties of femora from female Dmp1cre.Socs3fl/fl and Dmp1cre.Socs3fl/fl.LepRfl/fl mice at 12 weeks of age.
Parameter | Dmp1cre.Socs3fl/fl (n = 9) | Dmp1cre.Socs3fl/fl.LepRfl/fl (n = 11) | P-value |
---|---|---|---|
Ultimate stress (MPa) | 65 ± 5 | 61 ± 4 | 0.11 |
Ultimate deformation (mm) | 0.20 ± 0.01 | 0.23 ± 0.02 | 0.35 |
Ultimate strain (%) | 2.0 ± 0.1 | 2.2 ± 0.1 | 0.39 |
Yield strength/stress (MPa) | 58 ± 4 | 57 ± 4 | 0.79 |
Yield force (N) | 13.8 ± 0.5 | 12.8 ± 0.8 | 0.26 |
Yield displacement (mm) | 0.17 ± 0.02 | 0.20 ± 0.02 | 0.28 |
Yield strain (%) | 1.7 ± 0.1 | 1.9 ± 0.2 | 0.32 |
Post-yield displacement (mm) | 0.033 ± 0.009 | 0.029 ± 0.008 | 0.74 |
Post-yield strain (%) | 0.32 ± 0.09 | 0.28 ± 0.08 | 0.74 |
Energy to failure (mJ) | 1.5 ± 0.2 | 1.6 ± 0.3 | 0.84 |
Toughness (mJ/mm3) | 0.64 ± 0.06 | 0.69 ± 0.07 | 0.63 |
Values shown are mean ± S.E.M.P-values were determined by two-tailed t-test. No statistically significant differences were observed between the genotypes.
Quantitative analysis of genes from cortical bone samples was performed to determine if there were changes to osteoblasts or osteoclasts gene expression. No differences in osteoblast lineage genes were identified in Dmp1cre.Socs3fl/fl.LepRfl/fl mice compared to Dmp1cre.Socs3fl/fl control mice (Fig. 5A). When we measured genes associated with osteoclasts, we observed a reduction in DcStamp expression but no significant differences in Tnfsf11, Tnfrsf11b, and Acp5 expression (Fig. 5A). This may suggest a small reduction in osteoclast differentiation/activity in Dmp1cre.Socs3fl/fl.LepRfl/fl mice, which could contribute to the higher levels of low-density bone mass retained in these mice.
Leptin receptor deficiency does not change STAT3 activation in Dmp1cre.Socs3fl/fl mice
Since we previously observed that increased osteocyte STAT3 phosphorylation is associated with delayed cortical consolidation (Cho et al. 2017, Walker et al. 2020), we tested whether leptin receptor deficiency changed the proportion of pSTAT3 positive osteocytes in SOCS3-deficient cortical bone by immunohistochemistry (Fig. 5B). In Dmp1cre.Socs3fl/fl control mice, 30% of osteocytes were pSTAT3-positive (Fig. 5B), consistent with our previous work (Walker et al. 2020). The proportion of osteocytes with STAT3 activation above the threshold of detection was the same in Dmp1cre.Socs3fl/fl.LepRfl/fl and Dmp1cre.Socs3fl/fl control mice (Fig. 5C). This suggests that the further delay of cortical bone consolidation induced by deleting the leptin receptor is occurring through a pSTAT3-independent mechanism.
Discussion
This work identifies a new context-specific role for leptin signalling in bone. We demonstrate that leptin receptor signalling within Dmp1Cre-expressing cells can promote cortical bone consolidation in female mice, and this role is not irreplaceable in physiological circumstances. This is the first evidence for a role of leptin signalling in osteocytes.
The role of leptin signalling in cortical bone maturation in female mice was identified using a technique that measures bone maturation by distinguishing and quantifying low-, medium-, and high-density bone in the metaphysis (Walker et al. 2020, 2021). A lack of leptin receptor signalling in Dmp1cre-positive cells in female mice in the context of STAT3 hyperactivation resulted in a less mature cortical phenotype, indicated by a greater proportion of low-density bone. The greater proportion of low-density bone in vivo can be explained by multiple factors. The first is that there is a delay in the resorption of woven bone that would be replaced by lamellar bone during remodelling. This would be consistent with previous work examining Dmp1Cre.Socs3fl/fl mice which showed that they contained a greater proportion of woven bone in regions with reduced amounts of high-density bone due to a high level of bone resorption in the cortex (Walker et al. 2020). The phenotype we observe here was milder than in that study, and, although histomorphometry was not performed, compared to the already high levels of bone formation and resorption previously described in Dmp1Cre.Socs3fl/fl mice, we observed no significant changes in osteoblast or osteoclast mRNA markers, apart from a reduction in Dcstamp mRNA levels. Another possibility to explain the further delay in consolidation in Dmp1cre.Socs3fl/fl.LepRfl/fl mice is that LepR-deficient osteoblasts and osteocytes may not fully support mineralisation. However, we observed no differences in osteoblast lineage genes such as Col1a1, Bgalp, and Dmp1 that contribute to matrix deposition and mineralisation, which suggests this is less likely.
We observed no change in bone structure in Dmp1cre.LepRfl/fl mice when SOCS3 was at normal levels. This indicates that, under physiologic conditions, LepR expression is not required in late-stage osteoblasts and osteocytes. However, we established that leptin signalling is important when the skeleton is under ‘stress,’ that is, in the context of hyperactivation of STAT3 and/or delayed cortical bone consolidation. STAT3 activation in osteocytes is upregulated during inflammation (Hirano 2021), mechanical loading (Mantila Roosa et al. 2011, Corry et al. 2019), and fracture repair (Mantila Roosa et al. 2011); thus, these may be the contexts where leptin signalling exerts a protective effect. Earlier studies using germline-deficient mice could not distinguish the direct role of leptin signalling on bone cells from other influences; global deletion of leptin (ob/ob) or LepR (db/db) is not only associated with reduced bone mass but also shifts whole energy metabolism toward the accrual of fat mass (Ducy et al. 2000, Hamrick et al. 2004, Williams et al. 2011). Previous works focussed on the role of leptin receptor signalling in mesenchymal stromal cells (Scheller et al. 2010, Yue et al. 2016) identified a mild anabolic effect of deletion of LepR using Prx1cre on trabecular bone, with no changes to cortical bone mass (Yue et al. 2016). Thus, our finding of no skeletal phenotype under physiological conditions is perhaps unsurprising since a broader cre model which would also target Dmp1cre expressing cells (i.e. Prx1cre) had a very small effect. Comparisons to Col1a1(3.6kb) and Col1a1(2.3kb) cre models of leptin receptor deficiency are difficult to interpret since these models also target neuronal cell populations (Scheller et al. 2011) and the Col1a1(3.6kb) cre mice are mildly obese (Scheller et al. 2010). Since Prx1cre has broad expression throughout the mesenchymal lineage, our study has provided a new context-specific role for leptin signalling in more mature cells of the osteoblast lineage, namely those that express Dmp1cre, and showed that removal of this signalling in a SOCS3-deficient model leads to a further delay in consolidation. While Dmp1cre targets late osteoblasts and osteocytes (Lim et al. 2017, Dallas et al. 2018), the use of a single cre driver in our Dmp1cre.Socs3fl/fl.LepRfl/fl mice does not exclude the possibility that SOCS3 and LepR deletion occurs in different Dmp1Cre-expressing cell subsets. It is possible that delayed bone consolidation results from interactions between these two populations.
In our previous studies, in comparison with control mice, Dmp1cre.Socs3fl/fl mice had a greater number of osteocytes with sufficient STAT3 phosphorylation to stain positive by immunohistochemistry, in line with knockdown of the inhibitor SOCS3 (Walker et al. 2020). When G-CSFR was deleted, the number of phospho-STAT3 osteocytes was significantly greater and was associated with a further delay in cortical consolidation (Isojima et al. 2022). In the present study, we anticipated that the absence of Socs3 in Dmp1cre.Socs3fl/fl mice would increase the sensitivity of Dmp1cre-expressing cells to leptin signalling since SOCS3 inhibits JAK activation by binding the phospho-tyrosine-985 site on the leptin receptor (Bjørbæk et al. 2000). We reasoned that the absence of LepR could normalise both the uninhibited phosphorylation and the phenotype, as we had previously observed with gp130 deletion (Walker et al. 2020). However, when LepR was deleted, we observed a worsening of the phenotype. This was not associated with any change in the proportion of osteocytes staining positive for phospho-STAT3. For this reason, we speculate that STAT3-independent pathways downstream of LepR/JAK signalling may promote the positive effect of leptin on bone cortical bone consolidation. This may include activation of phosphoinositide-3 kinase (PI3K) and insulin receptor substrate pathways as described in the CNS and other peripheral tissues (Park & Ahima 2014). Leptin signalling also activates STAT5 in hypothalamic neurons (Gong et al. 2007) and ERK/MAPK signalling in chondrocytes (Liu et al. 2019), suggesting that other downstream signalling pathways may be present within bone. Activation of STAT3-independent pathways may also provide an explanation for why LepR deficiency further impairs bone consolidation and did not rescue the phenotype of Dmp1cre.Socs3fl/fl mice.
Leptin receptor deficiency delayed bone consolidation only in female mice. One explanation for this may arise from the sex-dependent levels of bone formation and resorption in male vs female mice and the threshold of these activities at which cortical consolidation can occur. In all mammals, females have higher levels of bone formation and resorption on endosteal surfaces than males, and the same is true in mice (Isojima & Sims 2021). In Dmp1cre.Socs3fl/fl mice, this higher level of bone cell activity was shown to be responsible for a greater delay in cortical consolidation in female mice than in male mice (Cho et al. 2017). At 12 weeks of age, male mice have a sufficiently low level of osteoblast and osteoclast activity for cortical consolidation to occur and the phenotype of male Dmp1cre.Socs3fl/fl mice begins to recover (Cho et al. 2017). In female mice, they reach the lower level of bone cell activity later in life and resolution of the phenotype occurs later (i.e. between 16 and 26 weeks of age) (Cho et al. 2017). We suggest that LepR deletion may have increased bone cell activity in both male and female mice, but at 12 weeks of age, where the level of activity in female mice is higher than the threshold required for consolidation, the delay induced by LepR deletion is more readily observed. In contrast, in males, only a very dramatic delay would be detectable at this time point. Deletion of IL-6 in the Dmp1cre.Socs3fl/fl model has also given a sex-specific phenotype: Dmp1cre.Socs3fl/fl.IL6−/− mice had a further delay in corticalisation compared with Dmp1cre.Socs3fl/fl mice in male mice but not female mice (Cho et al. 2017). The same study showed that oestrogen administration to orchiectomised mice delayed bone consolidation in male Dmp1cre.Socs3fl/fl mice, while testosterone administration to ovariectomised mice improved bone consolidation in female Dmp1cre.Socs3fl/fl mice. This indicates that androgen levels can also affect the rate of bone consolidation. It is likely that there is an interaction between leptin signalling and oestrogens or androgens that occurs during Socs3 deficiency and may lead to the observed delay in bone consolidation in female mice.
Another possible explanation of the sex difference observed is that fat mass and circulating leptin levels differ between male and female mice (Frederich et al. 1995). Female mice have greater adiposity and higher leptin levels independent of total-body lipid (Frederich et al. 1995), leading to greater levels of circulating leptin. Therefore, by inducing leptin signalling deficiency, it is possible that a greater difference in leptin levels may be observed in female mice. This phenomenon was observed in another model where targeted deletion of leptin receptor signalling in the brain led to a more pronounced metabolic phenotype in female mice compared to male mice (Lee et al. 2020). However, peripheral actions of leptin on JAK-STAT signalling in the skeleton occur at very low concentrations (Philbrick et al. 2017), suggesting this is unlikely to be an explanation for the sexually dimorphic phenotype we observe here. Since leptin regulates energy metabolism, it would be difficult to independently determine whether a sex difference in circulating leptin levels affects bone directly since leptin administration would reduce food intake and fat mass, which each exert effects on bone.
In conclusion, this study indicates that (1) LepR deficiency in female Dmp1cre.Socs3fl/fl mice delays cortical bone consolidation and (2) that these effect of LepR are context-specific and may relate to situations when the skeleton is undergoing a high level of bone remodelling or where signalling within osteocytes is amplified, as in inflammation, mechanical loading, or fracture healing.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/JOE-22-0084.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
N K W was supported by EH Flack Fellowship from the Marion and EH Flack Trust and a St Vincent’s Rising Star Award. T F C d L was supported by a travel grant from the National Council for Scientific and Technological Development, Brazil. N A S is supported by an NHMRC (Australia) Senior Research Fellowship. This work was funded by an NHMRC (Australia) Project Grant and philanthropic support to St. Vincent’s Institute. S V I acknowledges the support of the Victorian State Government Operational Infrastructure Scheme.
Acknowledgements
The authors thank the staff of the St. Vincent’s BioResources Centre for excellent animal care and Dr Paul Baldock, Garvan Institute, Sydney, for providing conditional LepR mice from his colony.
References
Bjørbæk C, Lavery HJ, Bates SH, Olson RK, Davis SM, Flier JS & Myers MG Jr 2000 SOCS3 mediates feedback inhibition of the leptin receptor via Tyr985*. Journal of Biological Chemistry 275 40649–40657. (https://doi.org/10.1074/jbc.M007577200)
Cho DC, Brennan HJ, Johnson RW, Poulton IJ, Gooi JH, Tonkin BA, McGregor NE, Walker EC, Handelsman DJ & Martin TJ et al.2017 Bone corticalization requires local SOCS3 activity and is promoted by androgen action via interleukin-6. Nature Communications 8 806. (https://doi.org/10.1038/s41467-017-00920-x)
Cohen P, Zhao C, Cai X, Montez JM, Rohani SC, Feinstein P, Mombaerts P & Friedman JM 2001 Selective deletion of leptin receptor in neurons leads to obesity. Journal of Clinical Investigation 108 1113–1121. (https://doi.org/10.1172/JCI13914)
Corry KA, Zhou H, Brustovetsky T, Himes ER, Bivi N, Horn MR, Kitase Y, Wallace JM, Bellido T & Brustovetsky N et al.2019 Stat3 in osteocytes mediates osteogenic response to loading. Bone Reports 11 100218. (https://doi.org/10.1016/j.bonr.2019.100218)
Dallas SL, Xie Y, Shiflett LA & Ueki Y 2018 Mouse Cre models for the study of bone diseases. Current Osteoporosis Reports 16 466–477. (https://doi.org/10.1007/s11914-018-0455-7)
Ducy P, Amling M, Takeda S, Priemel M, Schilling AF, Beil FT, Shen J, Vinson C, Rueger JM & Karsenty G 2000 Leptin inhibits bone formation through a hypothalamic relay: a central control of bone mass. Cell 100 197–207. (https://doi.org/10.1016/s0092-8674(0081558-5)
Enlow DH 1962 A study of the post-natal growth and remodeling of bone. American Journal of Anatomy 110 79–101. (https://doi.org/10.1002/aja.1001100202)
Eyckerman S, Broekaert D, Verhee A, Vandekerckhove J & Tavernier J 2000 Identification of the Y985 and Y1077 motifs as SOCS3 recruitment sites in the murine leptin receptor. FEBS Letters 486 33–37. (https://doi.org/10.1016/s0014-5793(0002205-5)
Frederich RC, Hamann A, Anderson S, Löllmann B, Lowell BB & Flier JS 1995 Leptin levels reflect body lipid content in mice: evidence for diet-induced resistance to leptin action. Nature Medicine 1 1311–1314. (https://doi.org/10.1038/nm1295-1311)
Frost HM 1982 Mechanical determinants of bone modeling. Metabolic Bone Disease and Related Research 4 217–229. (https://doi.org/10.1016/0221-8747(8290031-5)
Gong Y, Ishida-Takahashi R, Villanueva EC, Fingar DC, Münzberg H & Myers MG Jr 2007 The long form of the leptin receptor regulates STAT5 and ribosomal protein S6 via alternate mechanisms. Journal of Biological Chemistry 282 31019–31027. (https://doi.org/10.1074/jbc.M702838200)
Hamrick MW, Pennington C, Newton D, Xie D & Isales C 2004 Leptin deficiency produces contrasting phenotypes in bones of the limb and spine. Bone 34 376–383. (https://doi.org/10.1016/j.bone.2003.11.020)
Hirano T 2021 IL-6 in inflammation, autoimmunity and cancer. International Immunology 33 127–148. (https://doi.org/10.1093/intimm/dxaa078)
Howard JK & Flier JS 2006 Attenuation of leptin and insulin signaling by SOCS proteins. Trends in Endocrinology and Metabolism 17 365–371. (https://doi.org/10.1016/j.tem.2006.09.007)
Isojima T & Sims NA 2021 Cortical bone development, maintenance and porosity: genetic alterations in humans and mice influencing chondrocytes, osteoclasts, osteoblasts and osteocytes. Cellular and Molecular Life Sciences 78 5755–5773. (https://doi.org/10.1007/s00018-021-03884-w)
Isojima T, Walker EC, Poulton IJ, McGregor NE, Wicks IP, Gooi JH, Martin TJ & Sims NA 2022 G-CSF receptor deletion amplifies cortical bone dysfunction in mice with STAT3 hyperactivation in osteocytes. Journal of Bone and Mineral Research [epub]. (https://doi.org/10.1002/jbmr.4654)
Jepsen KJ, Silva MJ, Vashishth D, Guo XE & van der Meulen MC 2015 Establishing biomechanical mechanisms in mouse models: practical guidelines for systematically evaluating phenotypic changes in the diaphyses of long bones. Journal of Bone and Mineral Research 30 951–966. (https://doi.org/10.1002/jbmr.2539)
Kapur S, Amoui M, Kesavan C, Wang X, Mohan S, Baylink DJ & Lau KHW 2010 Leptin receptor (Lepr) is a negative modulator of bone mechanosensitivity and genetic variations in Lepr may contribute to the differential osteogenic response to mechanical stimulation in the C57BL/6J and C3H/HeJ pair of mouse strains. Journal of Biological Chemistry 285 37607–37618. (https://doi.org/10.1074/jbc.M110.169714)
Lee NJ, Qi Y, Enriquez RF, Clarke I, Ip CK, Wee N, Baldock PA & Herzog H 2020 Energy partitioning between fat and bone mass is controlled via a hypothalamic leptin/NPY relay. International Journal of Obesity 44 2149–2164. (https://doi.org/10.1038/s41366-020-0550-6)
Lim J, Burclaff J, He G, Mills JC & Long F 2017 Unintended targeting of Dmp1-Cre reveals a critical role for Bmpr1a signaling in the gastrointestinal mesenchyme of adult mice. Bone Research 5 16049. (https://doi.org/10.1038/boneres.2016.49)
Liu X, D’Cruz AA, Hansen J, Croker BA, Lawlor KE, Sims NA & Wicks IP 2019 Deleting suppressor of cytokine signaling-3 in chondrocytes reduces bone growth by disrupting mitogen-activated protein kinase signaling. Osteoarthritis and Cartilage 27 1557–1563. (https://doi.org/10.1016/j.joca.2019.05.018)
Lu Y, Xie Y, Zhang S, Dusevich V, Bonewald LF & Feng JQ 2007 DMP1-targeted Cre expression in odontoblasts and osteocytes. Journal of Dental Research 86 320–325. (https://doi.org/10.1177/154405910708600404)
Mantila Roosa SM, Liu Y & Turner CH 2011 Gene expression patterns in bone following mechanical loading. Journal of Bone and Mineral Research 26 100–112. (https://doi.org/10.1002/jbmr.193)
McGregor NE, Murat M, Elango J, Poulton IJ, Walker EC, Crimeen-Irwin B, Ho PWM, Gooi JH, Martin TJ & Sims NA 2019 IL-6 exhibits both cis- and trans-signaling in osteocytes and osteoblasts, but only trans-signaling promotes bone formation and osteoclastogenesis. Journal of Biological Chemistry 294 7850–7863. (https://doi.org/10.1074/jbc.RA119.008074)
McGregor NE, Walker EC, Chan AS, Poulton IJ, Cho EH, Windahl SH & Sims NA 2022 STAT3 hyperactivation due to SOCS3 deletion in murine osteocytes accentuates responses to exercise- and load-induced bone formation. Journal of Bone and Mineral Research 37 547–558. (https://doi.org/10.1002/jbmr.4484)
Münzberg H, Flier JS & Bjørbaek C 2004 Region-specific leptin resistance within the hypothalamus of diet-induced obese mice. Endocrinology 145 4880–4889. (https://doi.org/10.1210/en.2004-0726)
Park HK & Ahima RS 2014 Leptin signaling. F1000Prime Reports 6 73. (https://doi.org/10.12703/P6-73)
Philbrick KA, Wong CP, Branscum AJ, Turner RT & Iwaniec UT 2017 Leptin stimulates bone formation in ob/ob mice at doses having minimal impact on energy metabolism. Journal of Endocrinology 232 461–474. (https://doi.org/10.1530/JOE-16-0484)
Reid IR, Baldock PA & Cornish J 2018 Effects of leptin on the skeleton. Endocrine Reviews 39 938–959. (https://doi.org/10.1210/er.2017-00226)
Scheller EL, Song J, Dishowitz MI, Soki FN, Hankenson KD & Krebsbach PH 2010 Leptin functions peripherally to regulate differentiation of mesenchymal progenitor cells. Stem Cells 28 1071–1080. (https://doi.org/10.1002/stem.432)
Scheller EL, Leinninger GM, Hankenson KD, Myers MG Jr & Krebsbach PH 2011 Ectopic expression of Col2.3 and Col3.6 promoters in the brain and association with leptin signaling. Cells, Tissues, Organs 194 268–273. (https://doi.org/10.1159/000324745)
Sims NA 2020 The JAK1/STAT3/SOCS3 axis in bone development, physiology, and pathology. Experimental and Molecular Medicine 52 1185–1197. (https://doi.org/10.1038/s12276-020-0445-6)
Tikhonova AN, Dolgalev I, Hu H, Sivaraj KK, Hoxha E, Cuesta-Domínguez Á, Pinho S, Akhmetzyanova I, Gao J & Witkowski M et al.2019 The bone marrow microenvironment at single-cell resolution. Nature 569 222–228. (https://doi.org/10.1038/s41586-019-1104-8)
Turner CH & Burr DB 1993 Basic biomechanical measurements of bone: a tutorial. Bone 14 595–608. (https://doi.org/10.1016/8756-3282(9390081-k)
Turner RT, Kalra SP, Wong CP, Philbrick KA, Lindenmaier LB, Boghossian S & Iwaniec UT 2013 Peripheral leptin regulates bone formation. Journal of Bone and Mineral Research 28 22–34. (https://doi.org/10.1002/jbmr.1734)
Ushiki T, Huntington ND, Glaser SP, Kiu H, Georgiou A, Zhang JG, Metcalf D, Nicola NA, Roberts AW & Alexander WS 2016 Rapid inflammation in mice lacking both SOCS1 and SOCS3 in hematopoietic cells. PLoS ONE 11 e0162111. (https://doi.org/10.1371/journal.pone.0162111)
Walker EC, Truong K, McGregor NE, Poulton IJ, Isojima T, Gooi JH, Martin TJ & Sims NA 2020 Cortical bone maturation in mice requires SOCS3 suppression of gp130/STAT3 signalling in osteocytes. eLife 9 e56666. (https://doi.org/10.7554/eLife.56666)
Walker EC, McGregor NE, Chan ASM & Sims NA 2021 Measuring bone volume at multiple densities by micro-computed tomography. Bio-Protocol 11 e3873. (https://doi.org/10.21769/BioProtoc.3873)
Wee N & Baldock PA 2014 The skeletal effects of leptin. In Leptin: Biosynthesis, Functions and Clinical Significance. Endocrinology Research and Clinical Developments, pp. 129–140. Hauppauge, NY: Nova Science Publishers.
Williams GA, Callon KE, Watson M, Costa JL, Ding Y, Dickinson M, Wang Y, Naot D, Reid IR & Cornish J 2011 Skeletal phenotype of the leptin receptor-deficient db/db mouse. Journal of Bone and Mineral Research 26 1698–1709. (https://doi.org/10.1002/jbmr.367)
Williamson L, Hayes A, Hanson ED, Pivonka P, Sims NA & Gooi JH 2017 High dose dietary vitamin D(3) increases bone mass and strength in mice. Bone Reports 6 44–50. (https://doi.org/10.1016/j.bonr.2017.02.001)
Yue R, Zhou BO, Shimada IS, Zhao Z & Morrison SJ 2016 Leptin receptor promotes adipogenesis and reduces osteogenesis by regulating mesenchymal stromal cells in adult bone marrow. Cell Stem Cell 18 782–796. (https://doi.org/10.1016/j.stem.2016.02.015)
Zhang L, Reed F & Herzog H 2020 Leptin signalling on arcuate NPY neurones controls adiposity independent of energy balance or diet composition. Journal of Neuroendocrinology 32 e12898. (https://doi.org/10.1111/jne.12898)
Zhang G, Luo J, Zheng G, Bai Z, Cao L & Mao H 2021 Is the 0.2%-strain-offset approach appropriate for calculating the yield stress of cortical bone? Annals of Biomedical Engineering 49 1747–1760. (https://doi.org/10.1007/s10439-020-02719-2)
Zhou BO, Yue R, Murphy MM, Peyer JG & Morrison SJ 2014 Leptin-receptor-expressing mesenchymal stromal cells represent the main source of bone formed by adult bone marrow. Cell Stem Cell 15 154–168. (https://doi.org/10.1016/j.stem.2014.06.008)