Skip to main content
NCBI home page
Oxford University Press logoLink to Oxford University Press
. 2021 Feb 10;36(6):1104–1116. doi: 10.1002/jbmr.4270

Sexual Dimorphism in Differentiating Osteoclast Precursors Demonstrates Enhanced Inflammatory Pathway Activation in Female Cells

Se Hwan Mun 1, Sandra Jastrzebski 2, Judy Kalinowski 3, Steven Zeng 4, Brian Oh 5, Seyeon Bae 6,7, Giannopoulou Eugenia 8,9, Nazir M Khan 10, Hicham Drissi 11, Ping Zhou 12, Bongjin Shin 13, Sun‐Kyeong Lee 14, Joseph Lorenzo 15,16,*,✉, Kyung‐Hyun Park‐Min 17,18,19,*,✉
PMCID: PMC11140852  PMID: 33567098

ABSTRACT

Sexual dimorphism of the skeleton is well documented. At maturity, the male skeleton is typically larger and has a higher bone density than the female skeleton. However, the underlying mechanisms for these differences are not completely understood. In this study, we examined sexual dimorphism in the formation of osteoclasts between cells from female and male mice. We found that the number of osteoclasts in bones was greater in females. Similarly, in vitro osteoclast differentiation was accelerated in female osteoclast precursor (OCP) cells. To further characterize sex differences between female and male osteoclasts, we performed gene expression profiling of cultured, highly purified, murine bone marrow OCPs that had been treated for 3 days with macrophage colony‐stimulating factor (M‐CSF) and receptor activator of NF‐κB ligand (RANKL). We found that 125 genes were differentially regulated in a sex‐dependent manner. In addition to genes that are contained on sex chromosomes, transcriptional sexual dimorphism was found to be mediated by genes involved in innate immune and inflammatory response pathways. Furthermore, the NF‐κB‐NFATc1 axis was activated earlier in female differentiating OCPs, which partially explains the differences in transcriptomic sexual dimorphism in these cells. Collectively, these findings identify multigenic sex‐dependent intrinsic difference in differentiating OCPs, which results from an altered response to osteoclastogenic stimulation. In humans, these differences could contribute to the lower peak bone mass and increased risk of osteoporosis that females demonstrate relative to males. © 2021 American Society for Bone and Mineral Research (ASBMR).

Keywords: SEXUAL DIMORPHISM, OSTEOCLASTS, OSTEOCLASTOGENESIS, INFLAMMATORY RESPONSES, SIGNALING PATHWAYS, TRANSCRIPTOME

Introduction

Sexual dimorphism of the skeleton is well established.(1) Women, on average, achieve a lower peak bone mass than men, which is involved in their two‐ to fourfold greater incidence of vertebral fractures and osteoporosis.(2) Differences in the type and rate of production of sex steroids contribute to this sexual dimorphism, but this is not the only reason for the differences between bones from females and males.(3) Other variables include genetic,(4) behavioral, and environmental factors.(5) A number of studies in cell systems other than bone have shown that the genetic basis of sexual dimorphism is related both to differential expression of genes on sex chromosomes and divergent epigenetic regulation of autosomal gene transcription.(6) Differences in gene expression can affect the function of bone cells (osteoblasts, osteocytes, and osteoclasts), which regulate skeletal growth, the maintenance of bone mass, and bone quality.(7) Despite the well‐documented observations of a sex‐dependent dimorphism in the skeletal phenotype, the mechanisms that underlie these differences are poorly understood.(8)

Osteoclasts are the only cells that can efficiently resorb bone.(9,10) They differentiate from multipotential, myeloid‐lineage precursor cells under the influence of a variety of cytokines and local factors.(11,12) Chief among these are macrophage colony‐stimulating factor (M‐CSF) and receptor activator of NF‐κB ligand (RANKL),(9,13) which initiate osteoclast differentiation from myeloid precursors. RANKL binds to its receptor, RANK, and, in turn, activates downstream signaling mechanisms like the MAPK and NF‐κB signaling pathways and also induces co‐stimulatory signals via ITAM‐containing immunoreceptors.(14) Overall, these pathways result in the activation of NFATc1, the master regulator of the osteoclastogenesis.

It was previously documented that there is a difference in osteoclastogenesis between male and female osteoclast precursor cells (OCPs).(15) In this study, we examined the skeletal phenotype of male and female C57BL/6J mice and found that bone mass is lower and that osteoclast number is greater in females, without an appreciable difference in osteoblast number. In addition, in vitro osteoclastogenesis occurs with faster kinetics in female cells relative to male cells. Using RNA‐sequencing (RNA‐seq), we analyzed the transcriptomic sexual dimorphisms of OCPs that were differentiating into mature osteoclasts. This revealed a set of 125 genes that are differentially regulated between female and male cultures. This analysis demonstrated that genes related to inflammatory responses are enriched in sexual‐dimorphic genes and their mRNA levels were higher in female differentiating OCPs compared with male differentiating OCPs. Female OCPs also activate NF‐κB/NFATc1 with faster kinetics relative to male OCPs, and a subset of sexual‐dimorphic genes in these cultures is regulated in an NFATc1‐dependent manner during osteoclast differentiation. Taken together, these findings suggest that multigenic, sex‐dependent expression patterns contribute to sexual dimorphism in osteoclastogenesis.

Materials and Methods

Mice

C57BL/6J female and male mice were obtained from either Charles River Laboratories (Wilmington, MA, USA) or Jackson Laboratory (Bar Harbor, ME, USA) and were randomly allocated for experiments at the age of 8 weeks. Both NFATc1fl/fl mice and Mx1‐Cre transgenic mice were purchased from Jackson Laboratory. NFATc1fl/fl mice were crossed to Mx1‐Cre transgenic mice to generate NFATc1 fl/fl Mx1 Cre mice. Six‐week‐old NFATc1fl/fl Mx1 Cre mice (named NFATc1ΔMX) and their respective controls were intraperitoneally administrated with poly I:C for three times every other day. TLR4 KO mice were obtained from Dr Ping Zhou (Weill Cornell Medicine). RUNX1fl/+ mice have been previously described.(16) Animals were housed in a specific pathogen‐free environment either in the Weill Cornell Medicine vivarium or the UConn Health Center for Comparative Medicine. All experiments conformed to the ethical principles and guidelines approved by the Institutional Animal Care and Use Committee of either the Hospital for Special Surgery and Weill Cornell Medical College or UConn Health.

RNA‐sequencing (RNA‐seq) analysis

The analysis was performed as previously described.(17) Briefly, read quality was assessed with FastQC v0.11.6, and adapters were trimmed using Cutadapt v1.15. Reads were then mapped to the mouse genome (mm10), and reads in exons were counted against Gencode v27 with STAR Aligner.(18) Differential gene expression analysis was performed in R using edgeR.(19) Counts per million reads (CPMs) were generated in edgeR, and genes with low expression levels (<3 CPM in at least one group) were filtered from all downstream analyses. Benjamini–Hochberg false discovery rate (FDR) procedure was used to correct for multiple testing. Significantly expressed genes were defined as differentially expressed genes between male and female with p < .05 and fold‐change of at least 1.5. The RNA‐seq experiments analyzed were performed using three biological replicates.

The ingenuity pathway analysis (IPA)

IPA was used to analyze the functional pathways, upstream regulators, and molecular network for differently expressed genes.

Pathway analysis using gene set enrichment analysis (GSEA)

GSEA is used for analyzing the differentially expressed genes.(20) Hallmark gene sets in MSigDB (molecular signature data base) were used for the analysis. Pathways were ranked based on p values.

Reagents

Human and mouse M‐CSF and sRANKL were purchased from Peprotech (Rocky Hill, NJ, USA) or ConStem (Cheshire, CT, USA). The antibodies used for immunoblotting were as follows: NFATc1 (#sc‐7294) and TBP (sc‐204) from Santa Cruz Biotechnology (Dallas, TX, USA); p65 (#4764), HRP‐conjugated anti‐rabbit IgG (#7074), and HRP‐conjugated anti‐mouse IgG (#7076) from Cell Signaling Technology (Danvers, MA, USA); α‐tubulin (Sigma‐Aldrich #T9026, St. Louis, MO, USA).

Mouse osteoclast differentiation from bone marrow macrophage (BMM) cultures

Bone marrow cells were flushed from the femurs, followed by lysis of red blood cells using ammonium chloride potassium lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA).(21) Cells were cultured in M‐CSF conditioning medium (CM), α‐MEM supplemented with 10% FBS, 1% penicillin–streptomycin (Thermo Fisher Scientific), and 10% L929 cell supernatant, which served as a source of M‐CSF.(22) The non‐adherent cell population was recovered the next day and cultured with CM for 3 additional days. Then cells were plated at a seeding density of 1.5 × 105/mL and incubated with CM and RANKL (50 ng/mL). Cells were replenished with CM and RANKL every 2 days. When multinucleated cells were observed, cells were fixed and stained with the TRAP staining kit (Sigma‐Aldrich).

Mouse osteoclast differentiation from osteoclast precursor cell cultures

OCPs were isolated by flow cytometry activated cell sorting (FACS). The osteoclast precursor population (CD45R− CD3− CD11b−/lo CD115+) was sorted from bone marrow cells as described.(12,23) Sorted cells were cultured with M‐CSF (30 ng/mL) with or without RANKL (30 ng/mL) for up to 6 days.

Flow cytometric staining and sorting (FACS)

OCP was isolated as described.(12) Antibodies were purchased from eBiosciences (San Diego, CA, USA). Briefly, bone marrow cells were stained with anti‐CD45R/B220 (RA3‐6B2) for B‐cell lineage cells; anti‐CD3 (145‐2C11) for T‐cell lineage cells; anti‐CD11b/Mac‐1 (M1/70) for myeloid lineage cells; and anti‐c‐fms/CD115 (AFS98). Stained cells were first gated by B220 and CD3 and B220− CD3− population was further gated for CD11b−/low population. B220−CD3−CD11b−/low population was further gated for the absence or presence of CD115 (the receptor for M‐CSF, CSF1R). The osteoclast precursor population (CD45R− CD3− CD11b−/low CD115+) was sorted in a BD‐FACS Aria II (BD Biosciences, San Jose, CA, USA). For determining RANK expression, all antibodies were from BioLegend (San Diego, CA, USA) and Novus Biologicals (Centennial, CO, USA). Bone marrow cells were cultured with M‐CSF‐conditioning medium for 4 days. Culture cells were incubated with anti CD16/CD32 antibodies (Fc block) for 15 minutes to prevent nonspecific binding of antibodies used for staining and were subsequently stained with CD45 (7.1), RANK (9A725, Novus), c‐FMS (AFS98), F4/80 (BM8), and CD11b (M1/70) antibodies (1:100). DAPI staining (Thermo Fisher Scientific) to deplete dead cells from gating was added before the FACS analysis. Stained cells were analyzed on a BD canto instrument (BD Biosciences), and data analysis was done using FlowJo software (Tree Star, Ashland, OR, USA).

Bone pit assay

Bone pit assay was performed as described.(12) Sorted osteoclast precursor cells from male and female C57BL/6 mice were seeded onto UV‐irradiated bovine cortical bone slices to determine whether osteoclast resorptive activity is affected by sex. Cells were treated with M‐CSF and RANKL (both at 30 ng/mL) for 8 days. The bone slices were stained for TRAP as described above, photographed by microscopy, and then cells were removed from the bone slices by sonication. Bone slices were subsequently stained with toluidine blue to visualize the pit area and osteoclast path. Total pit area resorbed per chip (n = 3 replicates per group) was measured using a light microscope (BX53, Olympus Scientific, Waltham, MA, USA) and image analysis software (Olympus CellSens).

RNA preparation and real‐time PCR

Total RNA was obtained with the RNeasy‐Plus Mini Kit from QIAGEN (Valencia, CA, USA) and was reverse transcribed using a First Strand cDNA Synthesis Kit (Fermentas, Thermo Fisher Scientific). Real‐time PCR was performed with Fast SYBR Green Master Mix and mRNA expression was detected in triplicate using a 7500 Fast Real‐Time PCR system (Applied Biosystems, Carlsbad, CA, USA) following the manufacturer's protocols. Primer sequences are provided in Supplemental Table S2.

Immunoblotting analysis

For immunoblotting, whole cell lysates or nuclear lysates were fractionated on 7.5% polyacrylamide gels using SDS‐PAGE and transferred to polyvinylidene difluoride membranes for probing with antibodies.

Micro‐CT and histomorphometry analysis

Micro‐CT (μCT) analysis was performed as described previously,(22,24) and all samples were included in the analysis conducted in a blinded manner. For μCT analysis, before decalcification, femurs were scanned by μCT, with an isotropic voxel resolution of 12 μm (μCT40, Scanco, Bruttisellen, Switzerland; 55 kVp, 145 μA, 600 ms integration time) to evaluate morphological changes in bone. Bone morphology in the femur was examined in two regions: the diaphysis and the metaphysis. For cortical bone, the volume of interest (VOI) encompassed cortical bone within a 50‐slice section in the diaphysis. For trabecular bone, the VOI encompassed an 80‐slice section in the metaphysis, proximal to the growth plate. To ensure exclusion of primary spongiosa in the growth plate, VOIs began 50 slices proximal to the median of the growth plate. The bone marrow/bone cut‐offs used for segmentation are the following: Gauss sigma: 0.5; Gauss support: 1; lower threshold: 506.8; upper threshold: 2127.6; cortical bone: 738.4 mg HA/ccm; trabecular bone: 499.4 mg HA/ccm. Outcome parameters for trabecular bone included trabecular bone parameters including bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp), and trabecular number (Tb.N.). 3D reconstructions were generated by stacking thresholded 2D images from the contoured region.

Histomorphometry experiment was performed with femur from male and female C57BL/6 mice. Bone histomorphometric analysis was performed in a blinded, unbiased manner using OsteoMeasure (OsteoMetrics, Decatur, GA, USA) with light microscopy. Femurs were fixed in 4% paraformaldehyde for 2 days, were decalcified with 10% neutral‐buffered EDTA (Sigma‐Aldrich), and were embedded in paraffin. The quantification of osteoclasts was performed in paraffin‐embedded tissues that were stained for TRAP and methyl green (Vector Laboratories, Burlingame, CA, USA). Multinucleated TRAP‐positive cells adjacent to bone were identified as osteoclasts. The measurement terminology and units from the Nomenclature Committee of the American Society for Bone and Mineral Research(25) were used for histomorphometric analysis.

RNA interference

Bone marrow cells were cultured with M‐CSF for 4 days and cells were seeded at a density of 0.1 × 105 cells per well in 96‐well plates and transfected with 50 nmoL of siRNA oligonucleotides (listed in Supplemental Table S3) using TransIT‐TKO transfection reagent (Mirus Bio, Madison, WI, USA) according to the manufacturer's instructions. Cells were used for experiments after 24 hours of incubation with siRNA oligonucleotides.

Statistical analysis

All statistical analyses were performed with Prism 7.0 software (GraphPad Software, La Jolla, CA, USA) or R (ver. 3.5.0) using the two‐tailed, unpaired t test (two conditions) and one‐way or two‐way ANOVA for multiple comparisons (more than two conditions) with post hoc Tukey's correction for multiple comparisons. Shapiro–Wilk test was used to check normality.

Data availability

The RNAseq data sets that were generated by the authors as part of this study have been deposited in the Gene Expression Omnibus database with the accession code GEO: GSE153299.

Results

Female mice have decreased bone mass with an increased number of osteoclasts

To investigate the differences in bone mass between male and female mice, we first performed micro‐CT analysis. We found that femurs from 8‐week‐old C57BL/6 female mice exhibited decreased bone mass relative to male femurs. The ratio between bone volume and total volume (BV/TV) as well as the trabecular number (Tb.N) were significantly decreased, and the trabecular spacing (Th.Sp) was significantly increased in female femurs compared with male femurs (Fig. 1A). To test for sexual dimorphism in osteoclasts in vivo, we performed histomorphometric analysis of trabecular bone. This revealed that the number of osteoclasts, osteoclast surfaces, and eroded surfaces (a measure of osteoclast activity) were significantly higher in female mice than in male mice (Fig. 1B). In contrast, there were no significant differences in osteoblast surfaces between males and females (Fig. 1B). Although cortical bone parameters, including cortical bone area and cortical porosity, were comparable in female and male endocortical bone, the medullary areas (Ma.Ar) and total subperiosteal areas (Tt.Ar) of cortical bone were lower in female bone than in male bone, implying that female bone had a smaller cross‐sectional diameter (Fig. 1C). In addition, histomorphometry analysis revealed that endocortical osteoclast surface per bone surface (a measure of osteoclast size) and osteoclast eroded surface per bone surface were significantly higher in females compared with males (Fig. 1D). Collectively, our findings demonstrate that female C57BL/6 mice had decreased bone mass with increased osteoclast number and activity on trabecular bone and increased osteoclast size and activity on endocortical bone compared with male mice.

Fig 1.

Fig 1

Open in a new tab

Female mice show decreased bone mass with increased osteoclast numbers compared with male mice. (A, C) μCT analysis of femurs from 8‐week‐old male and female C57BL/6J mice (male [n = 6] and female [n = 7]). (A) μCT analysis of trabecular bone. Left panels show the representative images. Right panels show the indicated parameters in distal femurs. Bone volume/tissue volume ratio (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular spacing (Tb.Sp) were determined by μCT analysis. Scale bars = 100 μm. (C) μCT analysis of cortical bone. Left panels show the representative images. Right panels show the indicated parameters in cortical bone: cortical area/tissue area (Ct. Ar/Tt.Ar), cortical porosity, medullary area (Ma.Ar), and total subperiosteal area (Tt.Ar). Scale bars = 1 mm. (B, D) Histomorphometric analysis of the distal femur of 8‐week‐old male and female C57BL/6J mice (male [n = 6] and female [n = 7]). Representative images showing TRAP‐positive, multinucleated osteoclasts (red box highlights the trabecular bone: B; blue box highlights the endosteal area: D). The red scale bar = 200 μm, and black scale bar = 100 μm. Right panels show number of osteoclasts per bone perimeter (N.Oc/B.Pm), osteoclast surface area per bone surface (Oc.S/BS), eroded surface per bone surface (ES/BS), and osteoblast surface area per bone surface (Ob.S/BS). Data are shown as mean ± SEM. p values by unpaired t test.

Female osteoclasts formed faster than male osteoclasts

To determine if sex affects in vitro osteoclast differentiation, bone marrow cells from male and female mice were differentiated into osteoclasts in vitro as described previously.(22) Similar to the in vivo bone mass data, female bone marrow macrophage (BMM) cultures showed a significantly enhanced rate of osteoclast differentiation relative to male cells on days 2 and 3 after RANKL stimulation (Fig. 2A). Bone marrow cells are able to differentiate into osteoclasts after stimulation with M‐CSF and RANKL. However, these cells are not homogenous and are composed of multiple cell lineages. We previously showed that CD45R− CD3− CD11b−/low CD115high cells (herein named purified osteoclast precursor cells [pOCPs]) are highly enriched in their osteoclastogenic potential.(23) The total number of bone marrow cells was comparable between females and males (Supplemental Fig. S1A). Using our established flow cytometry paradigm,(23) we first measured the frequency and number of pOCPs, fraction IV (CD45R− CD3− CD11b−/low CD115high CD117high cells) in bone marrow. This is the population that is the most efficient at producing osteoclasts. We found that the number and frequency of pOCPs were comparable between male mice and female mice (Supplemental Fig. S1B, C). We then isolated multipotential pOCPs and stimulated them to differentiate toward mature osteoclasts by treatment with M‐CSF and RANKL for up to 6 days. Consistent with the BMM culture results, female pOCPs formed osteoclasts faster than male pOCPs (Fig. 2B) and produced a larger total pit area on cortical bone slices than did cultures from male mice (Fig. 2C). Taken together, these results argue that female C57BL/6 mice have more osteoclasts and resorbing activity in their bones than males and that in vitro osteoclastogenesis is accelerated in female OCPs compared with male OCPs.

Fig 2.

Fig 2

Open in a new tab

Osteoclast differentiation shows faster kinetics in female cells relative to male cells. (A, B) Osteoclastogenesis assay. Upper panel shows representative images of TRAP‐stained cells on the indicated days. Lower panel shows the number of TRAP‐positive multinuclear cells (MNCs = more than three nuclei) per well from three independent experiments. Scale bar = 200 μm. (A) Bone marrow cells from male and female mice were cultured with M‐CSF for 4 days and then were cultured with RANKL for up to 3 days. (B) Osteoclast precursor cells (OCPs) were sorted from bone marrow cells by FACS. OCPs were cultured with M‐CSF (30 ng/mL) and RANKL (30 ng/mL) for the indicated times. (C) Bone resorption pit assay showing total pit area resorbed per bone chip. (D) RT‐qPCR analysis of Itgb3 and Ctsk mRNA after 72 hours of culture of OCPs with or without RANKL (30 ng/mL) normalized relative to Hprt mRNA. All data are shown as mean ± SEM. p values by unpaired two‐way ANOVA with a post hoc Tukey test (A, B), Student's t test (C), or one‐way ANOVA with a post hoc Tukey test (D).

Osteoclast‐specific sex‐signature dimorphisms in the transcriptome

We next measured the expression of genes in male and female differentiating OCPs. It has been shown that RANKL stimulation increases osteoclast‐specific genes, such as integrin beta 3 (Itgb3) and cathepsin K (Ctsk), which play important roles in osteoclast activity. As expected, these osteoclast‐specific genes were induced by RANKL in both male and female osteoclasts. However, the difference in gene expression between female and male cultures was minimal, except for a few genes like Ctsk (Fig. 2D). To identify genes that are differentially regulated by sex in differentiating pOCPs, we performed RNA‐sequencing (RNA‐seq) and bioinformatic analysis. FACS‐purified OCPs were cultured with M‐CSF and RANKL for 3 days to differentiate them toward osteoclasts (hereafter differentiating OCPs). This time point was chosen because it represents the time in the cultures when mononuclear osteoclast fusion begins, and there is limited mature osteoclast apoptosis. Three biological replicates of differentiating OCPs from male and female RUNX1fl/+ mice were used for the RNA‐seq experiments. Cells from RUNX1fl/+ mice had been previously used as our study controls.(26,27) Because of the limited number of differentially expressed genes (DEGs), PCA analysis separated samples by treatment but not by sex (Supplemental Fig. S2A). The expression of 125 genes was significantly (p < 0.05, fold changes >1.5) changed between male and female differentiating OCPs (Fig. 3A), while the expression of 93 genes was differentially expressed between male and female cells treated only with M‐CSF (called M‐CSF conditions) (Supplemental Fig. S2B). Among 125 DEGs in differentiating OCPs, 116 of these genes were upregulated, whereas 9 were downregulated in female cells compared with male cells (Fig. 3A). Twenty‐eight sexual dimorphic genes in M‐CSF conditions were commonly detected in differentiating OCPs (Supplemental Fig. S2C), listed in Supplemental Table S2. We focused on these 125 DEGs in a pairwise comparison between female (F) differentiating OCPs (M‐CSF and RANKL) and male (M) differentiating OCPs (M‐CSF and RANKL). Of the 125 genes, only 9 were located on sex chromosomes, and 5 genes including Xist, Kdm5d, Uty, Ddx3y, and Eif2s3y were also detected in common sexual dimorphic genes between M‐CSF conditions and differentiating OCPs (Supplemental Fig. S2B, D). A total of 125 sexual dimorphic genes were classified into four clusters by k‐means clustering (Fig. 3B, C). However, these clusters did not capture known osteoclast‐specific genes (Supplemental Fig. S2E). Although class I genes were higher in female differentiating OCPs compared with male differentiating OCPs, class II contained only two genes and were lower in female cells relative to male cells (Fig. 3C and Supplemental Fig. S2F). Class III genes included genes on the Y chromosome and were more highly expressed in males (Supplemental Fig. S2F). Class IV genes were RANKL‐inducible genes and were higher in female cells compared with male cells (Fig. 3B, D).

Fig 3.

Fig 3

Open in a new tab

RNA‐sequencing analysis of differentiating OCPs. CD11b‐/lowCD45R− CD3− CD115+ cells (named osteoclast precursors cells [OPCs]) were sorted from bone marrow cells by flow cytometry. Sorted cells were treated with or without RANKL in the presence of M‐CSF for 3 days. (A) Volcano plot of RNA‐sequencing analysis of differentially expressed genes between female and male osteoclasts. Red dots show genes with significant (p < 0.05) and >1.5‐fold changes. (B) Heat map showing relative expression (Z‐score) of 125 genes from (A). The genes are listed in Supplemental Table S1. (C, D) Box plots showing the log2(CPM) of genes in class I (C) and IV (D). Of note, class II has only two genes and class II and III genes are displayed in Supplemental Fig. S2. All class genes are listed in Supplemental Table S1.

Inflammatory responses are higher in female differentiating OCPs relative to male differentiating OCPs

To gain insight into which functions and downstream molecules/pathways are important for accelerating female osteoclastogenesis, we performed a bioinformatic analysis of the 125 sex dimorphic genes in differentiating OCPs. Ingenuity pathway analysis (IPA) identified granulocyte adhesion and diapedesis, toll‐like receptor signaling, altered T‐cell and B‐cell signaling in rheumatoid arthritis, neuroinflammation signaling pathway, and LXR/RXR activation (Fig. 4A). In line with these results, gene set enrichment analysis (GSEA) showed that the pathways related to inflammatory response, allograft rejection, IL6_JAK_STAT3 signaling, UV‐response UP, and interferon gamma response were enriched in sex‐dimorphic genes in differentiating OCPs (Supplemental Fig. S2G). Moreover, IPA analysis of 93 sexual dimorphic genes in M‐CSF conditions revealed that sexual dimorphic genes in M‐CSF conditions were enriched in pyruvate fermentation to lactate, airway pathology in chronic obstructive pulmonary disease, role of IL‐17A in psoriasis, and the white adipose tissue browning pathway (Supplemental Fig. S2F), suggesting that the different sexual dimorphic pathways were turned on in M‐CSF conditions compared with differentiating OCPs. Upstream regulator analysis of IPA also showed that genes that can be regulated by inflammatory mediators, including lipopolysaccharide (a TLR4 ligand), poly IC (a TLR3 ligand), IFN gamma, and tetradecanoylphorbol acetate, were higher in female differentiating OCPs compared with male differentiating OCPs (Fig. 4B). To further elucidate the impact of inflammatory mediators on sexual dimorphism, we tested whether inflammatory mediators such as TNFα, IL1β, IL6, or IFNγ can contribute to sexual dimorphism in osteoclasts. Bone marrow cells from male and female mice were differentiated into osteoclasts in vitro, and the levels of inflammatory mediators including TNFα, IL1β, IFNγ, or IL‐6 were measured by qPCR. RANKL regulated the expression of inflammatory mediators, although the expression of TNFα, IL1β, IFNγ, or IL‐6 in differentiating osteoclast precursors were very low (Supplemental Fig. S3A). Only IL‐6 expression at the later phase of differentiation was significantly different between male and female cells. We next tested whether inflammatory mediators influence sexual dimorphism. We inhibited TNFα, IL1β, IFNγ, or IL‐6 using inhibitors or neutralizing antibodies. Blocking IL6 had no effects on sexual dimorphism in RANKL‐induced osteoclastogenesis. When TNFα, IFNγ, or IL1β were blocked, the difference in the number of osteoclasts between male osteoclasts and female osteoclasts was diminished (Supplemental Fig. S3B–D). However, female osteoclasts were still bigger and more abundant than male osteoclasts (Supplemental Fig. S3B–D). These data suggested that some inflammatory mediators affect sexual dimorphism in osteoclastogenesis.

Fig 4.

Fig 4

Open in a new tab

Class I genes are enriched in inflammatory pathways. (A) Ingenuity pathway analysis (IPA) showed the top five canonical pathways enriched in 125 genes from Fig. 3A. (B) Upstream regulators from IPA were shown. (C) Cumulative values for TLR2 and CD14, selected class I genes, from RNA‐sequencing. (D) CD11b‐/lowCD45R− CD3− CD115+ cells from male and female C57BL/6J mice were cultured with M‐CSF and RANKL for 3 days. RT‐qPCR analysis of CD14 and TLR2 normalized relative to Hprt mRNA. (E) Osteoclastogenesis assay. Bone marrow cells were cultured with M‐CSF for 4 days and then were primed by RANKL for 1 day. RANKL‐primed cells were stimulated with LPS (0.5 μg/mL) and were cultured with RANKL for an additional 2 days. The upper panel shows representative images of TRAP‐stained cells. The lower panel shows the number of TRAP‐positive multinuclear cells (MNCs = more than three nuclei) per well from three independent experiments. Scale bar = 100 μm. (F) Bone marrow cells were cultured with M‐CSF for 4 days, and then cells were cultured with M‐CSF and RANKL for 1 day. Cells were stimulated with LPS for 2 hours. Immunoblot for nuclear lysates with antibodies against p65, phospho‐STAT1, and lamin B1 antibodies (upper panel). Representative images from three independent experiments. Lower panels show the densitometric quantitation of p65 and phospho‐STAT1 band intensity after 24 hours of culture with RANKL from three independent donors. (G) Osteoclastogenesis assay. TLR4‐deficient bone marrow cells were cultured with M‐CSF and RANKL. The upper panel shows representative images of TRAP‐stained cells. The lower panel shows the number of TRAP‐positive multinuclear cells (MNCs) per well from four independent experiments. Scale bar = 200 μm. Data are shown as mean ± SEM. p values by Student's t test (C, D, F, G) or one‐way ANOVA with Sidak's multiple comparison test (E).

The log2(CPM) values of CD14 and TLR2, representative inflammatory genes in class I, were displayed from the data of the RNA‐seq (Fig. 4C). Because our RNA‐seq analysis was performed on OCPs from Runx1fl/+ mice, we also isolated OCPs from wild‐type C57BL/6 mice to test whether our findings were affected by the additional lox P sites of the Runx1 gene. The expression of CD14 and TLR2 was measured by qPCR in FACS‐isolated OCPs from C57BL/6 mice (Fig. 4D). As in our original RNA‐seq analysis, both CD14 and TLR2 were higher in female differentiation OCPs compared with cultures from males. Hence, this effect was independent of any effects of the Runx1‐flox allele in cells that were used for our RNA‐seq analysis (Fig. 4C, D). CD14 is a pattern recognition molecule in the innate immune response and a co‐receptor for TLR4.(28) To further determine if increased expression of CD14 contributes to the enhanced osteoclastogenesis of female cells, we measured the effect of TLR4 activation and TLR4 deficiency on the sexual dimorphic response of differentiating OCPs. We activated TLR4‐CD14 complex‐mediated signals using lipopolysaccharide (LPS), a TLR4 ligand. BMMs were cultured with RANKL for 1 day, and the RANKL‐primed cells were stimulated with LPS. Although pretreatment of cells with LPS before RANKL stimulation has been shown to suppress RANKL‐induced osteoclastogenesis,(29) LPS stimulation of RANKL‐primed cells had the reverse effect and enhanced osteoclast differentiation (Fig. 4E). Intriguingly, LPS‐mediated osteoclastogenesis was higher in female cells compared with male cells, and the difference in osteoclasts between basal conditions and LPS‐treated conditions was also significantly different between male and female cells (Fig. 4E and Supplemental Fig. S3E). Moreover, LPS‐induced NF‐κB and STAT1 activation was higher in female cells than in male cells (Fig. 4F). To corroborate our findings, TLR4‐deficient cells were used. Interestingly, TLR4 deficiency negated the sexual dimorphism of differentiating OCPs despite a slight increase in osteoclasts in female cells compared with male cells (Fig. 4G and data not shown). These results suggest that TLR4 activation enhanced the sexual dimorphic difference, whereas TLR4 deficiency eliminated the sexual dimorphic difference in OCP differentiation. Thus, our study suggests that inflammatory pathway genes that have a higher expression in female differentiating OCPs might be linked to the sexual‐dimorphism of osteoclastogenesis.

Female‐specific enhancement of RANKL responsive genes

Class IV genes were mostly RANKL‐inducible and more highly expressed in female differentiating OCPs compared with male differentiating OCPs (Fig. 3B and Supplemental Table S1). However, the function of RANKL‐inducible class IV genes in OCP differentiation was mainly unknown. The log2(CPM) value of selected class IV genes, including Isoc2b, Slc6a12, Vcan, Clec7a, and ATP6V0c from RNA‐seq, is displayed (Fig. 5A and Supplemental Fig. S4A). We also measured mRNA expression of class IV genes by RT‐qPCR in OCPs from wild‐type C57BL/6 mice (Fig. 5B and Supplemental Fig. S4B). The pattern of selected class IV genes was similar between Runx1fl/+ differentiating OCPs and wild‐type C57BL/6 differentiating OCPs. To test whether class IV genes contribute to sexual dimorphism in differentiating OCPs, we knocked down expression of some of these using short interfering RNAs (siRNAs) in BMMs before cells were cultured with M‐CSF and RANKL for 3 days. Although decreased expression of Isoc2b inhibited osteoclastogenesis, the knock‐down of Clec7a enhanced osteoclastogenesis (Fig. 5C–F). Intriguingly, the decreased expression of Isco2b or Clec7a negated sexual dimorphism in osteoclastogenesis (Fig. 5C–F). In addition, decreased expression of ATP6V0c, another class IV gene, also eliminated difference in the number of osteoclasts between female and male cells while suppressing osteoclastogenesis (Supplemental Fig. S4C, D). Thus, our results suggest that sexual dimorphism in osteoclastogenesis is multigenic in origin.

Fig 5.

Fig 5

Open in a new tab

Comparison of the expression of class IV genes. (A) Cumulative values for Isoc2b, Slc6a12, Vcan, and Clec7a, selected class IV genes, from RNA‐sequencing. (B) CD11b‐/lowCD45R− CD3− CD115+ cells from male and female C57BL/6J mice were cultured with or without RANKL in the presence of M‐CSF for 3 days. RT‐qPCR analysis of Isoc2b, Slc6a12, Vcan, and Clec7a normalized relative to Hprt mRNA. (C–F) Knock‐down of Isoc2b or Clec7a in BMMs from male and female C57BL/6J mice. (C, E) The efficiency of knock‐down. RT‐qPCR analysis of Isoc2b or Clec7a normalized relative to Hprt mRNA. (D,F) Osteoclastogenesis assay. The upper panel shows representative images of TRAP‐stained cells. The lower panel shows the number of TRAP‐positive multinuclear cells (MNCs = more than three nuclei) per well from at least three independent experiments. Data are shown as mean ± SEM. p values by one‐way ANOVA with a post hoc Tukey test.

p65‐NFATc1 axis is enhanced in female differentiating OCPs compared with male differentiating OCPs

To elucidate the underlying mechanism of the enhanced osteoclastogenesis in female cells, we examined the differences in RANKL‐signaling pathways. We first measured the level of RANK by flow cytometry. RANK expression on F4/80+ macrophage lineage cells was comparable between females and males, as measured by % of RANK+ cells (Supplemental Fig. S5A, B) and MFI (Supplemental Fig. S5C). To find a regulator for sex‐dimorphic genes in osteoclasts, we used GSEA to identify transcription factor‐binding motifs that are enriched in the promoters of sex‐dimorphic genes. The binding motifs for Ets2, PU.1, EN1, NF‐κB, and NFATc1 were significantly enriched in the promoters of sex‐dimorphic genes (Fig. 6A). PU.1 and Ets2 are ETS family transcription factors, and PU.1 is a key transcription factor for the development of myeloid cells, especially macrophages. The deletion of PU.1 in hematopoietic cells in mice resulted in severe osteopetrosis, suggesting the positive role of PU.1 in osteoclasts.(30) In contrast, Ets2 is suppressed during osteoclastogenesis, and overexpression of Ets2 protects mice from bisphosphonate‐induced osteoclast apoptosis.(31) EN1 is linked to osteoporosis but is not expressed in osteoclasts.(32) Of these factors, only NF‐κB and NFATc1 were significantly induced by RANKL stimulation (Supplemental Fig. S6). These results led us to investigate differences in the activation of NF‐κB and NFATc1, key signaling components downstream of RANKL, between male and female differentiating OCPs. Cultures were stimulated with RANKL for 1 day, and nuclear NFATc1 and p65 were measured by immunoblot. As expected, RANKL stimulation induced its direct target, NFATc1, as well as inducing nuclear p65 translocation and activation. Strikingly, accumulation of both nuclear p65 and NFATc1 was significantly increased in female cells compared with male cells on day 1 after RANKL stimulation (Fig. 6B). Our results suggest that the early phase of the RANKL response was accelerated in female cells relative to male cells, despite similar levels of RANK expression. To determine if increased NF‐κB p65 activation in female cells results in a difference in the NF‐κB p65 binding on sexual dimorphic genes between male and female cells, we performed p65 ChIP‐qPCR for Isoc2b, Slc6a12, ATP6v03, and Clec7a. p65 binding in both male and female differentiating OCPs was increased by RANKL stimulation at the p65 binding sites in the promoter of Isoc2b, Slc6a12, ATP6v03, and Clec7a(33) compared with control conditions (Supplemental Figs. S7 and S8). However, the p65 binding was comparable between male and female differentiating OCPs, suggesting that the sexually dimorphic genes may be regulated by additional mechanisms.

Fig 6.

Fig 6

Open in a new tab

RANKL‐induced NF‐κB activation increased in female osteoclasts compared with male osteoclasts. (A) Transcription factor binding sites enriched in 125 sexually dimorphic genes from GSEA was shown. (B) Bone marrow cells were cultured with M‐CSF for 4 days and then cells were stimulated with RANKL for 1 day. Immunoblot of nuclear lysates with anti‐NFATc1 and p65 antibodies. TBP and α‐tubulin were used as controls for nuclear and cytoplasmic proteins, respectively. Left two panels showed the densitometric quantitation of p65 (top) and NFATc1 (bottom) band intensity after 24 hours of culture with RANKL from three independent donors. (C) The mRNA expression of Nfatc1, Isoc2b, Slc6a12, Vcan, and Clec7a (relative to the Hprt housekeeping gene) after RANKL stimulation for 1 day (n = 3). Data are shown as mean ± SEM. p values by one‐way ANOVA with a post hoc Tukey test (B, C).

We next wished to determine the effect of enhanced NFATc1 activation in female cells on differentially expressed sexually dimorphic genes in the early phase of osteoclastogenesis. To link the enhanced NFATc1 in females to RANKL‐inducible class IV genes, we measured mRNA expression of selected class IV genes in NFATc1‐deficient BMMs from NFATc1ΔMX mice. NFATC1 was low in NFATc1‐deficient BMMs, and mRNA expression of Isoc2b, Vcan, and Clec7a mRNA were significantly diminished in female NFATc1‐deficient cells relative to control cells (Fig. 6C). However, Slc6a12 was not affected by NFATc1 deficiency (Fig. 6C). Taken together, these results suggest that an early induction of NFATc1 may contribute to the accelerated expression in females of many but not all RANKL‐inducible class IV genes.

Discussion

Bone loss starts at a younger age in women than in men,(34) and the average bone loss in women is significantly higher.(35) Conversely, peak bone mass is higher in males than in females.(36) Thus, net bone mass and skeletal size are lower in females than males. However, the mechanisms that underly sexual dimorphism in bone are not fully determined.(37) We found that sex‐specific changes in bone mass are closely linked to an inherent difference in the response of OCP cultures to treatment with M‐CSF and RANKL. We further investigated sex‐based differential gene expression in differentiating OCPs and found that 125 genes were differentially expressed between male and female cells. Furthermore, female cells showed an enhanced activation of NF‐κB and NFATc1 when they were stimulated with RANKL, and some of these RANKL‐inducible sexually dimorphic genes were dependent on NFATc1. Therefore, our study revealed that an interplay between transcriptomic and signaling networks underlies sexual dimorphism in differentiating OCPs.

We found that female mice have smaller skeletons and a greater number of osteoclasts compared with male mice. In addition, trabecular bone parameters were significantly different between male and female bones. Consistently, osteoclast number in female trabecular bone was higher than in male trabecular bone. However, cortical bone parameters such as cortical bone area and cortical porosity were comparable between male and female bone. Callewaert and colleagues showed that bone strength, measured by three‐point bending experiments, is lower in female bone compared with male bone.(38) Intriguingly, endocortical bone formation was significantly higher in female mice relative to male mice, whereas periosteal bone formation was lower.(38) We found that subendosteal and subperiosteal areas were lower in female mice compared with male mice and endosteal osteoclast number was not significantly different between female and male bone. However, osteoclast surface and erosion surface per unit bone surface were greater on the endocortical surface of female bones. These results imply that the endocortical surface of females had bigger osteoclasts, which resorbed more bone. Hence, it appears that enhanced osteoclast activity contributes to the sexual dimorphism of endocortical bone.

Despite a clear functional sexual dimorphism in various cell types, only a few sexually dimorphic genes have been discovered. For example, other myeloid cells such as microglia and macrophages display widespread sexual dimorphism in their functions. However, they contain very few sexually dimorphic genes.(39–41) Consistently, our results showed that a limited number of genes are differentially expressed between male and female differentiating OCPs. These data suggested that other factors, besides these genes, may be associated with the sex‐based difference in OCP differentiation. There are several possible explanations for the differences in the responses of female and male differentiating OCPs. The effect of sex hormones can be a major reason for sexual dimorphism in osteoclasts because sex hormones greatly contribute to skeletal remodeling and homeostasis. It is also possible that genes on sex chromosomes contribute to osteoclast differentiation and activity in a sex‐dependent manner. Among the 125 sexually dimorphic genes that we identified in differentiating OCPs, we found 9 on sex chromosomes: 4 Y‐linked genes and 5 X‐linked genes. The function of these genes in osteoclasts is not well characterized. Cybb and Xist are linked to bone metabolism. Cybb is a gene required for the activation of oxidative burst. Cybb enhances RANKL‐induced NFATc1, and Cybb‐deficient mice show excessively high bone mass with decreased osteoclasts.(42) Xist is a long‐non‐coding RNA, which controls X‐chromosome inactivation in females.(43) Xist has been shown to be significantly upregulated in plasma and monocytes from patients with osteoporosis compared with healthy controls. Although overexpression of Xist in mesenchyme stem cells suppresses osteogenesis, the role of Xist in osteoclasts is unknown. Differential activation of signaling pathways is also a good candidate to be a mediator of sexual dimorphism in differentiating OCPs. Surprisingly, we found more robust activation of RANKL signaling pathways in female cells, which had increased activation of NF‐κB p65 and NFATc1 compared with male cells. These signals are crucial for the differentiation of OCPs toward mature osteoclasts. In addition, the activation of NF‐κB p65 by LPS in RANKL‐primed differentiating OCPs was higher in female cells compared with male cells. However, we did not detect a difference in p65 binding in the promoter of a few class VI genes or NFATc1. We cannot rule out that increased NF‐κB activation may not correlate with differences in NF‐κB DNA binding between female and male differentiating OCP cultures. However, based on our study, enhanced p65 activation in female cells may influence other pathways to regulate the expression of sexual dimorphic genes. The specific upregulation of RANKL signaling may be directly involved in sexual dimorphism in these cultures. In addition, our results suggested that inflammatory responses have increased activation in female differentiating OCPs compared with male differentiating OCPs. Thus, although the exact mechanisms of increased NF‐κB activation in female OCP by RANKL and LPS was not fully determined, differences in inflammatory signaling likely affected the sexual dimorphisms of osteoclastogenesis that we observed.

Inflammation induces osteoclast‐mediated bone resorption in many pathological conditions such as rheumatoid arthritis (RA) and periodontitis.(44) Although the role of inflammatory genes in osteoclastogenesis is not clear, inflammatory signals typically enhance in vitro osteoclast differentiation of RANKL‐primed cells.(45) However, during osteoclast differentiation, RANKL suppresses genes that are enriched in inflammatory responses. Sex‐based differential activation or sensitivity in inflammatory responses has been well studied.(46,47) It has been shown that females have enhanced innate and adaptive immune responses compared with males, who are more susceptible to infection than females. Autoimmune diseases such as RA and systemic lupus erythematosus (SLE) are more abundant in females.(48) Inflammatory responses are initiated when pathogens are recognized by pattern recognition receptors, like toll‐like receptors (TLRs) and lectin receptors. Some pattern recognition receptors and associated adaptors such as TLR1, TLR2, CD14, and CLEC7A display higher expression in female differentiating OCPs compared with male cells. However, the role of these proteins in the sexual dimorphism of osteoclastogenesis has not been studied previously. We found that TLR4 deficiency attenuated the sexual dimorphism of differentiating OCPs, since we were not able to detect a significant difference in osteoclast number between TLR4KO female cells and TLR4KO male cells. Similarly, knock‐down of CLEC7A negated sexual dimorphic differences in OCP differentiation. However, consistent with the previous reports showing that TLR4 deficiency or CLEC7A deficiency enhanced osteoclastogenesis,(49,50) both TLR4 deficiency and the decreased expression of CLEC7A enhanced osteoclastogenesis compared with control cells. These results suggest that TLR4‐ and CLEC7A‐mediated signals play an important role as both inhibitors of RANKL‐induced osteoclastogenesis and mediators of the sexual dimorphism of osteoclastogenesis. Several studies of sexual dimorphism in response to pathogens have been reported, although the results are contradictory. Administration of group B streptococcus (GBS) recruits fewer neutrophils and induces a less severe immune response in females relative to males.(51) Another study showed that when mice were challenged with LPS, female mice had increased neutrophil infiltration and inflammatory cytokine production compared with male mice. In contrast to our results, the same group demonstrated that LPS‐driven osteoclastogenesis from sorted OCPs is significantly higher in male cells than in female cells.(52) In our system, LPS stimulation enhanced osteoclastogenesis with faster kinetics in female differentiating OCPs relative to male differentiating OCPs, suggesting that TLR4 activation accelerated the sexual dimorphism of osteoclasts. However, without exposure to pathogens, osteoclastogenesis is faster in females relative to males, and genes related to the inflammatory/immune responses are higher in female differentiating OCPs compared with male differentiating OCPs. Therefore, the mechanisms by which increased immune response genes are connected to enhanced osteoclastogenesis as well as the reasons that immune response genes are more highly expressed in female cells need to be further investigated.

Although transcriptomic sexual dimorphism in mammals is being studied to explain phenotypic differences between males and females,(53) the mechanisms regulating these differences as well as the genes that are differentially expressed are only beginning to be defined. Since knock‐down of sexual dimorphic genes resulted in the attenuation of the difference in osteoclastogenesis between male and female differentiating OCPs, these genes play an important role in the sexual dimorphism of this process. Cellular differentiation requires cell‐specific and stimulus‐dependent epigenetic programs.(54) Although few studies have been published on osteoclast biology,(54–56) both RANKL signals and inflammatory signals have been shown to influence epigenetic modification. Sexually dimorphic gene expression can be differentially programmed and may be correlated with imprinted epigenetic modifications. However, it is still possible that enhanced osteoclastogenesis in female OCPs impacts on the difference in the expression of sexually dimorphic genes. To investigate if there are such differences between male and female OCP cultures, studies of the differences in epigenetic programs between male and female differentiating OCPs are needed. The inherent differences in the response of female differentiating OCPs to osteoclastogenic stimuli relative to male cells may contribute to the sexual dimorphism of skeletal mass and osteoclast number in mice. In turn, these differences, if they also occur in humans, may be a factor in the lower peak bone mass and increased rates of osteoporosis that occur in females relative to males.

In summary, our results demonstrated that in vivo osteoclast number is greater while bone mass is less in female mice compared with male mice. In vitro, osteoclastogenesis is also enhanced in females compared with males. This response is, in part, associated with enhanced RANKL signaling and inflammatory pathway activation. Our results suggested that cross‐talk between inflammatory signaling pathways and the RANKL signaling pathway plays an important role in accelerating osteoclast differentiation and activity in female differentiating OCPs relative to male differentiating OCPs. Finally, our results argue that sexually dimorphic differences in OCP differentiation result from a multigenic process, with many genes being critical for the phenomena, since in their absence sexually dimorphic differentiation of OCPs is no longer found.

Disclosures

All authors declare no conflict of interest.

Acknowledgments

This work was supported by the National Institute of Arthritis and Musculoskeletal and Skin diseases (NIAMS) of NIH under award numbers R56 AR074391‐01A1 (to J.L), R01AR063661‐05, and 3R01AR063661‐04S1 (to J.L. and H.D.), R01AR069562 and AR073156 (to K‐HP‐M) and by the Rosensweig Genomics Center from The Tow Foundation (to K‐HP‐M).

Authorsʼ roles: SHM: conceptualization; data curation; investigation; formal analysis; writing—original draft; methodology. SJ: data curation. JK: data curation. SZ: data curation. BO: data curation. SB: data curation. EG: data curation. NMK: conceptualization. HD: conceptualization. PZ: resources. BS: data curation. S‐KL: conceptualization; data curation. JL: conceptualization; supervision; writing—original draft (supporting); writing—review and editing; funding acquisition. K‐HP‐M: conceptualization; supervision; writing—original draft (supporting); writing—review and editing; funding acquisition.

Author contributions: Sehwan Mun: Conceptualization; data curation; formal analysis; investigation; methodology; writing‐original draft. Sandra Jastrzebski: Data curation. Judith Kalinowski: Data curation. Steven Zeng: Data curation. Brian Oh: Data curation. Seyeon Bae: Data curation. Giannopoulou Eugenia: Data curation. Nazir Khan: Conceptualization. Hicham Drissi: Conceptualization. Ping Zhou: Resources. Bongjin Shin: Data curation. Sun‐Kyeong Lee: Conceptualization; data curation. Joseph Lorenzo: Conceptualization; supervision; writing‐original draft; writing‐review & editing. Kyung‐Hyun Park‐Min: Conceptualization; supervision; writing‐original draft; writing‐review & editing.

 

PEER REVIEW

The peer review history for this article is available at https://publons.com/publon/10.1002/jbmr.4270.

Supplementary Material

jbmr4270-sup-0001-Supinfo

Appendix S1. Supporting Information.

jbmr4270-sup-0001-supinfo.pdf (571.3KB, pdf)
jbmr4270-sup-0002-Tables

Appendix S2. Tables.

jbmr4270-sup-0002-tables.pdf (206KB, pdf)

Contributor Information

Se Hwan Mun, Arthritis and Tissue Degeneration Program, David Z. Rosensweig Genomics Research Center Hospital for Special Surgery New York NY USA.

Sandra Jastrzebski, Department of Medicine University of Connecticut Health Farmington CT USA.

Judy Kalinowski, Department of Medicine University of Connecticut Health Farmington CT USA.

Steven Zeng, Arthritis and Tissue Degeneration Program, David Z. Rosensweig Genomics Research Center Hospital for Special Surgery New York NY USA.

Brian Oh, Arthritis and Tissue Degeneration Program, David Z. Rosensweig Genomics Research Center Hospital for Special Surgery New York NY USA.

Seyeon Bae, Department of Medicine Weill Cornell Medical College New York NY USA; Arthritis and Tissue Degeneration Program, David Z. Rosensweig Genomics Research Center Hospital for Special Surgery New York NY USA.

Giannopoulou Eugenia, Biological Sciences Department New York City College of Technology, City University of New York Brooklyn NY USA; Arthritis and Tissue Degeneration Program, David Z. Rosensweig Genomics Research Center Hospital for Special Surgery New York NY USA.

Nazir M Khan, Department of Orthopaedics School of Medicine, Emory University Atlanta GA USA.

Hicham Drissi, Department of Orthopaedics School of Medicine, Emory University Atlanta GA USA.

Ping Zhou, Feil Family Brain & Mind Research Institute (BMRI), Weill Cornell Medical College New York NY USA.

Bongjin Shin, Center on Aging University of Connecticut Health Farmington CT USA.

Sun‐Kyeong Lee, Center on Aging University of Connecticut Health Farmington CT USA.

Joseph Lorenzo, Department of Orthopaedic Surgery University of Connecticut Health Farmington CT USA; Department of Medicine University of Connecticut Health Farmington CT USA.

Kyung‐Hyun Park‐Min, BCMB Allied Program Weill Cornell Graduate School of Medical Sciences New York NY USA; Department of Medicine Weill Cornell Medical College New York NY USA; Arthritis and Tissue Degeneration Program, David Z. Rosensweig Genomics Research Center Hospital for Special Surgery New York NY USA.

References

  1. Gordon  CM, Zemel  BS, Wren  TA, et al.  The determinants of peak bone mass. J Pediatr.  2017;180:261‐269. [DOI] [PubMed] [Google Scholar]
  2. Lofman  O, Larsson  L, Toss  G.  Bone mineral density in diagnosis of osteoporosis: reference population, definition of peak bone mass, and measured site determine prevalence. J Clin Densitom.  2000;3(2):177‐186. [DOI] [PubMed] [Google Scholar]
  3. Gilsanz  V, Wren  TAL, Ponrartana  S, Mora  S, Rosen  CJ.  Sexual dimorphism and the origins of human spinal health. Endocr Rev.  2018;39(2):221‐239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Pocock  NA, Eisman  JA, Hopper  JL, Yeates  MG, Sambrook  PN, Eberl  S.  Genetic determinants of bone mass in adults. A twin study. J Clin Invest.  1987;80(3):706‐710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Krall  EA, Dawson‐Hughes  B.  Heritable and life‐style determinants of bone mineral density. J Bone Miner Res.  1993;8(1):1‐9. [DOI] [PubMed] [Google Scholar]
  6. Mank  JE.  The transcriptional architecture of phenotypic dimorphism. Nat Ecol Evol.  2017;1(1):6. [DOI] [PubMed] [Google Scholar]
  7. Lorenzo  JA, Canalis  E, Raisz  LG.  Metabolic bone disease. In Kronenberg  H, Melmed  S, Polonsky  KS, Larsen  PR, eds. Williams text book of endocrinology. 11th ed. Philadelphia: Saunders‐Elsevier; 2008. pp 1269‐1310. [Google Scholar]
  8. Lorenzo  J.  Sexual dimorphism in osteoclasts. Cells.  2020;9(9):2086. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Teitelbaum  SL.  Bone resorption by osteoclasts. Science.  2000;289(5484):1504‐1508. [DOI] [PubMed] [Google Scholar]
  10. Tsukasaki  M, Takayanagi  H.  Osteoimmunology: evolving concepts in bone‐immune interactions in health and disease. Nat Rev Immunol.  2019;19(10):626‐642. [DOI] [PubMed] [Google Scholar]
  11. Xiao  Y, Zijl  S, Wang  L, et al.  Identification of the common origins of osteoclasts, macrophages, and dendritic cells in human hematopoiesis. Stem Cell Reports.  2015;4(6):984‐994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Jacome‐Galarza  CE, Lee  SK, Lorenzo  JA, Aguila  HL.  Identification, characterization, and isolation of a common progenitor for osteoclasts, macrophages, and dendritic cells from murine bone marrow and periphery. J Bone Miner Res.  2013;28(5):1203‐1213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Park‐Min  KH.  Mechanisms involved in normal and pathological osteoclastogenesis. Cell Mol Life Sci.  2018;75(14):2519‐2528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Humphrey  MB, Nakamura  MC.  A comprehensive review of immunoreceptor regulation of osteoclasts. Clin Rev Allergy Immunol.  2016;51(1):48‐58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Delahunty  KM, Shultz  KL, Gronowicz  GA, et al.  Congenic mice provide in vivo evidence for a genetic locus that modulates serum insulin‐like growth factor‐I and bone acquisition. Endocrinology. 2006;147(8):3915‐3923. [DOI] [PubMed] [Google Scholar]
  16. Growney  JD, Shigematsu  H, Li  Z, et al.  Loss of Runx1 perturbs adult hematopoiesis and is associated with a myeloproliferative phenotype. Blood.  2005;106(2):494‐504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kusnadi  A, Park  SH, Yuan  R, et al.  The cytokine TNF promotes transcription factor SREBP activity and binding to inflammatory genes to activate macrophages and limit tissue repair. Immunity. 2019;51(2):241‐57.e9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dobin  A, Davis  CA, Schlesinger  F, et al.  STAR: ultrafast universal RNA‐seq aligner. Bioinformatics.  2013;29(1):15‐21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Robinson  MD, McCarthy  DJ, Smyth  GK.  edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics.  2010;26(1):139‐140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Subramanian  A, Tamayo  P, Mootha  VK, et al.  Gene set enrichment analysis: a knowledge‐based approach for interpreting genome‐wide expression profiles. Proc Natl Acad Sci U S A.  2005;102(43):15545‐15550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Estrada  K, Styrkarsdottir  U, Evangelou  E, et al.  Genome‐wide meta‐analysis identifies 56 bone mineral density loci and reveals 14 loci associated with risk of fracture. Nat Genet.  2012;44(5):491‐501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Bae  S, Lee  MJ, Mun  SH, et al.  MYC‐dependent oxidative metabolism regulates osteoclastogenesis via nuclear receptor ERRalpha. J Clin Investig.  2017;127(7):2555‐2568. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jacquin  C, Gran  DE, Lee  SK, Lorenzo  JA, Aguila  HL.  Identification of multiple osteoclast precursor populations in murine bone marrow. J Bone Miner Res.  2006;21(1):67‐77. [DOI] [PubMed] [Google Scholar]
  24. Bouxsein  ML, Boyd  SK, Christiansen  BA, Guldberg  RE, Jepsen  KJ, Muller  R.  Guidelines for assessment of bone microstructure in rodents using micro‐computed tomography. J Bone Miner Res.  2010;25(7):1468‐1486. [DOI] [PubMed] [Google Scholar]
  25. Parfitt  A, Drezner  MK, Vlorieux  FH, et al.  Bone histomorphometry: stardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res.  1987;2:595‐610. [DOI] [PubMed] [Google Scholar]
  26. Soung do  Y, Kalinowski  J, Baniwal  SK, et al.  Runx1‐mediated regulation of osteoclast differentiation and function. Mol Endocrinol.  2014;28(4):546‐553. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Paglia  DN, Yang  X, Kalinowski  J, Jastrzebski  S, Drissi  H, Lorenzo  J.  Runx1 regulates myeloid precursor differentiation into osteoclasts without affecting differentiation into antigen presenting or phagocytic cells in both males and females. Endocrinology.  2016;157(8):3058‐3069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Wright  SD, Ramos  RA, Tobias  PS, Ulevitch  RJ, Mathison  JC.  CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science. 1990;249(4975):1431‐1433. [DOI] [PubMed] [Google Scholar]
  29. Ji  JD, Park‐Min  KH, Shen  Z, et al.  Inhibition of RANK expression and osteoclastogenesis by TLRs and IFN‐gamma in human osteoclast precursors. J Immunol.  2009;183(11):7223‐7233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Carey  HA, Hildreth  BE  3rd, Geisler  JA, et al.  Enhancer variants reveal a conserved transcription factor network governed by PU.1 during osteoclast differentiation. Bone Res.  2018;6:8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Zhang  Q, Badell  IR, Schwarz  EM, et al.  Tumor necrosis factor prevents alendronate‐induced osteoclast apoptosis in vivo by stimulating Bcl‐xL expression through Ets‐2. Arthritis Rheum.  2005;52(9):2708‐2718. [DOI] [PubMed] [Google Scholar]
  32. Zheng  HF, Forgetta  V, Hsu  YH, et al.  Whole‐genome sequencing identifies EN1 as a determinant of bone density and fracture. Nature.  2015;526(7571):112‐117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Shen  Z, Hoeksema  MA, Ouyang  Z, Benner  C, Glass  CK.  MAGGIE: leveraging genetic variation to identify DNA sequence motifs mediating transcription factor binding and function. Bioinformatics. 2020;36(Suppl 1):i84‐i92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Jones  G, Nguyen  T, Sambrook  P, Kelly  PJ, Eisman  JA.  Progressive loss of bone in the femoral neck in elderly people: longitudinal findings from the Dubbo osteoporosis epidemiology study. BMJ.  1994;309(6956):691‐695. [PMC free article] [PubMed] [Google Scholar]
  35. Hannan  MT, Felson  DT, Dawson‐Hughes  B, et al.  Risk factors for longitudinal bone loss in elderly men and women: the Framingham Osteoporosis Study. J Bone Miner Res.  2000;15(4):710‐720. [DOI] [PubMed] [Google Scholar]
  36. Nieves  JW, Formica  C, Ruffing  J, et al.  Males have larger skeletal size and bone mass than females, despite comparable body size. J Bone Miner Res.  2005;20(3):529‐535. [DOI] [PubMed] [Google Scholar]
  37. Alswat  KA.  Gender disparities in osteoporosis. J Clin Med Res.  2017;9(5):382‐387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Callewaert  F, Venken  K, Kopchick  JJ, et al.  Sexual dimorphism in cortical bone size and strength but not density is determined by independent and time‐specific actions of sex steroids and IGF‐1: evidence from pubertal mouse models. J Bone Miner Res. 2010;25(3):617‐626. [DOI] [PubMed] [Google Scholar]
  39. Thion  MS, Low  D, Silvin  A, et al.  Microbiome influences prenatal and adult microglia in a sex‐specific manner. Cell.  2018;172(3):500‐16.e16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Gal‐Oz  ST, Maier  B, Yoshida  H, et al.  ImmGen report: sexual dimorphism in the immune system transcriptome. Nat Commun.  2019;10(1):4295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Bhasin  JM, Chakrabarti  E, Peng  DQ, Kulkarni  A, Chen  X, Smith  JD.  Sex specific gene regulation and expression QTLs in mouse macrophages from a strain intercross. PLoS One. 2008;3(1):e1435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Kang  IS, Kim  C.  NADPH oxidase gp91phox contributes to RANKL‐induced osteoclast differentiation by upregulating NFATc1. Sci Rep.  2016;6:38014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Brown  CJ, Hendrich  BD, Rupert  JL, et al.  The human XIST gene: analysis of a 17 kb inactive X‐specific RNA that contains conserved repeats and is highly localized within the nucleus. Cell.  1992;71(3):527‐542. [DOI] [PubMed] [Google Scholar]
  44. Redlich  K, Smolen  JS.  Inflammatory bone loss: pathogenesis and therapeutic intervention. Nat Rev Drug Discov.  2012;11(3):234‐250. [DOI] [PubMed] [Google Scholar]
  45. Zhang  P, Liu  J, Xu  Q, et al.  TLR2‐dependent modulation of osteoclastogenesis by porphyromonas gingivalis through differential induction of NFATc1 and NF‐kappaB. J Biol Chem.  2011;286(27):24159‐24169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Jaillon  S, Berthenet  K, Garlanda  C.  Sexual dimorphism in innate immunity. Clin Rev Allergy Immunol.  2019;56(3):308‐321. [DOI] [PubMed] [Google Scholar]
  47. Fish  EN.  The X‐files in immunity: sex‐based differences predispose immune responses. Nat Rev Immunol.  2008;8(9):737‐744. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Oertelt‐Prigione  S.  The influence of sex and gender on the immune response. Autoimmun Rev.  2012;11(6–7):A479‐A485. [DOI] [PubMed] [Google Scholar]
  49. Zhu  X, Zhao  Y, Jiang  Y, et al.  Dectin‐1 signaling inhibits osteoclastogenesis via IL‐33‐induced inhibition of NFATc1. Oncotarget. 2017;8(32):53366‐53374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wang  D, Gilbert  JR, Taylor  GM, et al.  TLR4 inactivation in myeloid cells accelerates bone healing of a calvarial defect model in mice. Plast Reconstr Surg.  2017;140(2):296‐306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Scotland  RS, Stables  MJ, Madalli  S, Watson  P, Gilroy  DW.  Sex differences in resident immune cell phenotype underlie more efficient acute inflammatory responses in female mice. Blood.  2011;118(22):5918‐5927. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Valerio  MS, Basilakos  DS, Kirkpatrick  JE, et al.  Sex‐based differential regulation of bacterial‐induced bone resorption. J Periodontal Res.  2017;52(3):377‐387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Rinn  JL, Snyder  M.  Sexual dimorphism in mammalian gene expression. Trends Genet.  2005;21(5):298‐305. [DOI] [PubMed] [Google Scholar]
  54. Park‐Min  KH.  Epigenetic regulation of bone cells. Connect Tissue Res.  2017;58(1):76‐89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Carey  HA, Hildreth  BE  3rd, Samuvel  DJ, et al.  Eomes partners with PU1 and MITF to regulate transcription factors critical for osteoclast differentiation. iScience.  2019;11:238‐245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Yasui  T, Hirose  J, Tsutsumi  S, Nakamura  K, Aburatani  H, Tanaka  S.  Epigenetic regulation of osteoclast differentiation: possible involvement of Jmjd3 in the histone demethylation of Nfatc1. J Bone Miner Res.  2011;26(11):2665‐2671. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

jbmr4270-sup-0001-Supinfo

Appendix S1. Supporting Information.

jbmr4270-sup-0001-supinfo.pdf (571.3KB, pdf)
jbmr4270-sup-0002-Tables

Appendix S2. Tables.

jbmr4270-sup-0002-tables.pdf (206KB, pdf)

Data Availability Statement

The RNAseq data sets that were generated by the authors as part of this study have been deposited in the Gene Expression Omnibus database with the accession code GEO: GSE153299.

Articles from Journal of Bone and Mineral Research are provided here courtesy of Oxford University Press

RESOURCES