ABSTRACT
Osteoclasts derived from hematopoietic stem cells control bone resorption. Identifying novel molecules that can epigenetically regulate osteoclastogenesis has been an important basic and clinical issue. The polycomb group (PcG) protein enhancer of zeste homologue 2 (Ezh2), a histone lysine methyltransferase is associated with epigenetic regulation of numerous cellular processes, it is not yet clear on its involvement in bone cell development and homeostasis. Here, we crossed LysM-Cre mice with Ezh2flox/flox mice to delete Ezh2 in myeloid cell lineage mature macrophages. Conditional deletion of Ezh2 in macrophages resulted in significant increases in postnatal bone growth in the first 6 months of life, but tibia length and body weight gains were not different in knockout mice compared with their wild type controls. Significantly decreased osteoclastogenesis but increased bone mass without osteopetrosis were found in Ezh2 conditional knockout (CKO) mice. In contrast to female mice, one floxed Ezh2 gene copy recombinant with LysM-Cre+ (Ezh2flox/+LysM-Cre+) produced increased bone mass in young adult male mice compared with control mice (Ezh2flox/flox, LysM-Cre+ and wild type). Inflammatory milieu in bone was significantly lower in both male and female CKO mice compared with their respective controls. Deletion of Ezh2 in macrophages triggered increased gene expression of osteoclast suppressors, IRF8, MafB and Arg1 due to decreased Ezh2-induced trimethylation of H3K27me3. Conversely, NFATc1 and Cathepsin k expression were decreased. These findings suggest that pre-osteoclastic cell differentiation is under epigenetic control of osteoclast suppressive gene expression via an Ezh2-dependent mechanisms.
Introduction
Osteoclasts derived from hematopoietic stem cells and during early development, monitor osteoblasts to control bone resorption and to shape newly formed bone (Siddiqui & Partridge, 2016; Bethel et al, 2011; Zaidi et al, 2018). While osteoclastic bone resorption is an important cellular function in the development and physiology of the skeleton, suppressing osteoclastogenesis is considered an effective therapeutic approach for bone destructive or resorptive diseases, such as osteoporosis and rheumatoid arthritis (Hirayama et al, 2002; Tanaka, 2018). On the other hand, defective osteoclastic bone resorption leads to accumulation of bone with altered architecture, resulting in osteopetrotic bone susceptible to fractures (Stattin et al, 2017). Receptor activator of NF-κB ligand (RANKL) and macrophage colony-stimulating factor (M-CSF) (Feng & Teitelbaum, 2013) have been characterized as two crucial factors that influence the formations of osteoclasts. Moreover, it is also known that many inflammatory factors such as TNF-α mediates osteoclast formation, although the role of those inflammatory factors including TNF-α signaling in osteoclastogenesis remains poorly understood (Cao et al, 2017). Myeloid hematopoietic cells are thought to be precursors for osteoclasts, other cell lineages derived from hematopoietic stem cells share similar differentiation signaling pathways, it is not yet clear how to target specific osteoclastogenic signaling for intervening persistently increased osteoclastic cell differentiation and bone resorption. Therefore, identifying novel molecules that can regulate, especially epigenetically control osteoclastogenesis has been an important basic and clinical issue.
As pathologies of osteoporosis or rheumatoid arthritis are not characterized by common targetable genetic alterations, epigenetic mechanisms and strategies for preventions might be relevant and promising options. Indeed, epigenetic regulation of osteoclastic cell differentiation and activity has been increasingly investigated (Yasui et al, 2011; Kyung-Hyun, 2019). Recent cell culture studies suggested that generalized accumulation of methylation at the CpG island in the mouse RANKL promoter may result in the recruitment of histone deacetylase and chromatin condensation, leading to the stable epigenetic silencing of RANKL (Kitazawa & Kitazawa, 2007). Moreover, DNA methyltransferase 3a (DNMT3a) has been reported an important function in promoting osteoclastogenesis by increasing DNA methylation of anti-osteoclastogenic genes (Nishikawa et al, 2015). DNA methylation in the NFATc1 has profound effects on osteoclast differentiation and activity (Kurotaki et al, 2020; Yasui et al, 2011). These studies are somewhat limited in mainly being performed in vitro, and often focus on a few genes without determining upstream factors (Akimzhanov et al, 2008). Furthermore, DNA methylation is not an isolated event, but also affects and is itself affected by other gene regulatory mechanisms, such as gene silencing via histone modification (methylation or acetylation). While there is a dearth of evidence on epigenetic regulation of bone cell development in association with changes of life styles during early growth, we have previously hypothesized that high fat diet (HFD)-induced obesity may alter enhancer of zeste (Ezh) polycomb repressive complex subunit expression to epigenetically control osteoclastogenesis and therefore bone resorption (Chen et al, 2018). Such regulation may be through activation of Ezh2 to promote osteoclastogenesis by downregulating IRF8, a negative regulator of osteoclastogenesis (Fang et al, 2016).
Ezh2, a histone lysine methyltransferase associated with transcriptional repression (Donaldson-Collier et al, 2019), catalyzes the addition of methyl groups to histone H3 at Lys 27 (H3K27me3) in target genes involved in numerous cellular processes, predominantly leading to gene repression (Lavarone et al, 2019). Loss or inhibition of Ezh2 leads to enhanced osteogenic but inhibited adipogenic differentiation of bone marrow-derived mesenchymal stem cells (BMSCs) both from rodents and humans (Hemming et al, 2017), as well as in murine pre-osteoblastic cell line MC3T3 (Dudakovic et al, 2020). Expression of Ezh2 suppresses the osteogenic genes and ligand-dependent signaling pathways (e.g., WNT, PTH, and BMP2) to favor adipogenic differentiation (Wang et al, 2010). Several studies have shown that inhibition of Ezh2 prevents estrogen deficiency-induced bone loss in animal models (Dudakovic et al, 2016). Despite evidence for a role of Ezh2 in bone cell differentiation, it is not known if this system is involved in maternal HFD-obesity-associated offspring programming of bone development. Depend on different Cre mouse models were used, deletion of Ezh2 in osteoblastic cell lineage resulted in unclear bone developmental patterns (Chen et al, 2020). We recently reported that, using Osterix Cre mouse model, although deletion of Ezh2 in osteoblast precursors resulted in clear osteogenic gene higher expression in bone, bone mass evaluated by peripheral quantitative CT scan showed obscure phenotype when conditional Ezh2 deletion mice compared with Osterix Cre or Ezhflox/flox mice as controls (Levaot et al, 2015). Bone development and remodeling are not separated osteoblastic events, two other types of bone cells, osteoclasts and osteocytes are also involved. Our interests in determining the role of Ezh2 during osteoclast differentiation and bone resorption derived from the evidence that Ezh2 might be involved in bone osteoblastic cell activity and differentiation.
In the present report, we set to examine if deletion of Ezh2 gene in osteoclast precursors would result in changes of bone mass phenotype in young and adult mice, and what is the pattern of post-natal skeletal development after deletion of Ezh2 gene in osteoclast precursors. In addition, we present if Ezh2 gene expression in pre-osteoclasts plays different roles in bone homeostasis in different sex of mice, and inform if Ezh2 gene controls or modifies specific osteoclast gene expression to epigenetically regulate bone resorption.
Results
Changes of Ezh2 and osteoclast suppressive gene expression during bone cell and post-natal bone development
We determined Ezh2 and osteoclast suppressive gene expressions during bone development of wild type C57Bl mice post weaning (30 days old) to 6 months old, and during osteoclast and osteoblast differentiation. We found that, although Ezh2 mRNA was abundantly expressed in bone (total RNA was isolated from spine L4), there were no significant differences among those male wild type mice of 30 days old, 56 days old and 6 months old groups (Fig 1A). Expression of RIF8, a key osteoclast suppressive gene, was highest in bone from 56 days old mouse group compared with two other age groups (Fig 1B). However, the osteoblast bone formation marker osteocalcin (Fig 1C) and osteoclast bone resorption marker Cathepsin K (Fig 1D) mRNA expression were both significantly lower in 6 months old group compared with either 30 or 56 days old mouse group, indicating bone remodeling in 6 months old young adult mice is lower than in younger mice. The protein expression of Ezh2 (Fig 1E) in bone was lower in mice 30 days old compared with those of 56 days and 6 months old mice, but IRF8 and Arg1 (another osteoclast suppressive factor) expression were significantly lower in 6 month old mice compared with those of either 30 days or 56 days old mouse groups (Fig 1E). On the other hand, NFATc1 (a positive osteoclastogenic regulator) protein expression was significantly higher in 6 month old mice compared with two other age groups, indicating osteoclastogenesis increase with age (regardless of lower bone remodeling). Since RNA and protein isolated from L4 bone were mixed with other cell types, we took murine osteoclastic and osteoblastic cell lines to investigate those gene expressions during bone cell differentiation. In the presence of RANKL, it took about 5 days to differentiate osteoclast precursors to mature multi-nuclear mature osteoclasts marine cell line RAW264.7 osteoclastic cell cultures (Fig 1F). During osteoclast differentiation, Ezh2 gene expression significantly dropped from day 3 to day 5 (Fig 1G) compared with its expression in osteoclastic cells on day 1. On the other hand, mRNA expression of IRF8 and MafB (another negative osteoclast regulator), two osteoclast inhibitory genes, were significantly increased at day 3 and day 5 compared with their expression in day 1 (Fig 1G), indicating Ezh2 and osteoclast inhibitory gene expression has inverse association. In osteoblast stromal cell line ST2 cell cultures, it took 7 days for ST2 cells to differentiate into ALP positive mature osteoblast in the presence of osteoblast culture medium (Fig 1H). During this period, Ezh2 mRNA expression was significantly lower on day 3 and 7 compared to day 1 (Fig 1I). Bone specific ALP mRNA expression was significantly increased on day 3 and 7 compared with on day 1, indicating Ezh2 and ALP expression has an inverse relationship during osteoblast differentiation. However, osteocalcin mRNA expression was significantly lower on day 3 and 7 compared with its expression on day 1 (Fig 1I). These data indicated that Ezh2 expression is higher during both osteoclast and osteoblast early differentiation compared with their later differentiation stages, and Ezh2 is inversely associated with either osteoclast suppressive gene or osteoblastic marker gene expression.
Ezh2 gene expression in osteoclastic cells is not required for maintaining better bone mass in 6 months old young adult mice
The inverse association between Ezh2 and osteoclast inhibitory gene expression during osteoclast differentiation triggered us to investigate the role of Ezh2 gene on controlling osteoclastogensis in vivo and postnatal bone development. We have used a large amount of mouse numbers and created substantial amount of data to determine postnatal bone mass in osteoclastic cell specific Ezh2 gene knockout mice. Bone structures were assessed using micro-CT, we first show those results in 6 months old male and female mice. As shown in representative scans of tibia of male mice in Fig 2A, among parameters which were analyzed on trabecular bone, we present percentage of trabecular bone volume BV/TV (bone volume / total tissue volume), actual bone volume (BV), bone surface (BS), BS/TV, trabecular number (Tb.N), trabecular separation (Tb.Sp) from Ezh2flox/flox/LysM-Cre+ (CKO, conditional knockout), Ezh2flox/flox (f/f, flox control), LysM-Cre+ (Cre+, Cre+ control), Ezh2flox/+/LysM-Cre+ (f/+ Cre+, Ezh2 one copy floxed with LysM-Cre positive), Ezh2flox/+/LysM-Cre- (f/+ Cre-, Ezh2 one copy floxed with LysM-Cre negative) and wild type (Wt) control mice in Fig 1B. One-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison analysis was performed, there was no significant difference of BV/TV between CKO and f/+ Cre+ mouse groups, but it was significantly higher in those two groups compared with any other four different genotypic control groups of mice (f/f, Cre+, f/+ Cre- and Wt) (Fig 2B). There were no significant differences among those four different genotype control groups of mice on BV/TV. Like BV/TV, other parameters of BV, BS, BS/TV and Tb.N showed similar patterns among CKO and other four groups (Fig 2B). In contrast, Tb.Sp was significantly lower in CKO and f/+ Cre+ mouse groups compared with all other genotypes (Fig 2B). In females, as shown in representative scans of tibia from all genotypic mouse groups in Fig 2C, we found BV/TV was significantly higher in CKO group compared to other genotypic and wild type mouse groups (Fig 2D). It was not like we observed in males that there were no differences between CKO and f/+ Cre+ mouse groups on such as BV/TV, we found that BV, BS, BS/TV and Tb.N were all significantly higher, but Tb.Sp was significantly lower only in CKO mice compared with those from any other genotypic and wild type mouse groups (Fig 2D). Micro-CT was also performed on vertebrae (L5), as shown in representative scans of L5 of male mice in Fig 3A, BV/TV was significantly higher in CKO and f/+ Cre+ mouse groups compared with any other four genotypic mouse groups (Fig 3B), but BV/TV was also significantly higher in CKO mice compared with f/+ Cre+ mice (Fig 3B). Significant difference of BV was only found between CKO and f/f mice (Fig 3B). BS and BS/TV results were similar to those of the tibia (Fig 3B). Tb.N was significantly higher in CKO compared with all other groups (Fig 3B). Tb.Sp in CKO mice was significantly lower compared with f/f, Cre+, f/+ Cre- and Wt mice, but was not different when compared to f/+ Cre+ mice (Fig 3B). In females, as shown in micro-CT representative L5 scans in Fig 3C, BV/TV in CKO mice had significantly higher than in any other genotypic mouse groups (Fig 3D). BV in CKO and f/+ Cre+ mice were significantly different than in f/f and f/+ Cre-mice (Fig 3D). BS/TV and Tb.N were significantly higher in CKO mice compared with any other groups (Fig 3D), but BS and Tb.Sp were not significantly different across groups (Fig 3D). We found no differences of mouse body weights and tibia lengths among groups and additional micro-CT data of tibia trabecular of 6 months old male mice in Supplemental Table 1 and female mice in Supplemental Table 2. Moreover, micro-CT tibia cortical parameters were also analyzed, those significantly changed parameters in CKO mice compared with other genotypic mouse groups can be found in Supplemental Table 3 for males and in Supplemental Table 4 for females.
Postnatal bone growth is faster in Ezh2flox/flox/LysM-Cre+ CKO mice
We performed and analyzed micro-CT on 30 days old (right after weaning) tibia trabecular bone of all those genotypic male and female mice. As shown in representative tiabia scans of 30 –day old male mice in Supplemental Fig 1A, among parameters which were analyzed on trabecular bone, we found BV/TV, BS/TV and Tb.N were significantly higher in CKO mice compared with f/+ Cre- and Wt control mice in Supplemental Fig 1B. BV/TV showed significantly higher in f/+ Cre+ mice compared with f/+ Cre- and Wt control mice in Supplemental Fig 1B. Tb.Sp showed significantly lower in CKO mice compared with f/+ Cre+ and Wt control mice in Supplemental Fig 1B. We did not observe differences of any other parameters in tibia trabecular (Supplemental Table 5), and tibia cortical parameters (Supplemental Table 6). In females, none of those micro-CT parameters were different from the tibia trabecular compartment among groups in Supplemental Fig 1C and 1D and Supplemental Table 7, and there were not statistically different to micro-CT tibia cortical parameters among groups in Supplemental Table 8. We also analyzed micro-CT on 56 days old tibia trabecular bone. In males, BV/TV showed significantly higher in CKO mice compared with f/f, Cre+ and Wt control mice (Supplemental Fig 2A and 2B). BS/TV was significantly higher in CKO mice compared with Wt mice (Supplemental Fig 2B). Tb,N was significantly higher in CKO mice compared with Cre+, f/+ Cre+, f/+ Cre- and Wt mice, but Tb.Sp was significantly lower in CKO mice compared with Cre+ and Wt control mice (Supplemental Fig 2B). Other trabecular micro-CT parameters and cortical parameters were presented in Supplemental Table 9 and 10. In females, BV/TV was significantly lower in CKO mice compared with Cre+, f/+ Cre+, f/+ Cre- and Wt control mice (Supplemental Fig 2C and 2D). In this age of female mice, we found more micro-CT parameters in both trabecular (Supplemental Table 11) and cortical (Supplemental Table 12) in CKO mice to be significantly different compared with other genotypic control mouse groups. From 30 days old young to 6 months old young adults, we did not observe differences on body weight gains between CKO mice and two other control groups (f/f and Cre+) in either males or females (Fig 4A and 4B). The growth of tibia length was also tightly matched in those three groups of mice in both males and females (Fig 4A and 4B). However, in male trabecular bone, BV/TV in CKO mice steadily increased from 30-days-old to 6-months-old, but it went down after 56 days- in f/f and Cre+ control mice (Fig 4C). The changes of TV with age were no differences between CKO and two control mouse groups (Fig 4C). BV and Tb.N were found increased from 30 days old to 56 days old in all three groups, they significantly declined from 56 days old to 6 months old, but such significant declines did not happen in CKO mice (Fig 4C). With regards to cortical bone, the change in BV/TV, BV, B.Ar and Cs.Th from 30 days old to 6 months old were similar across all three groups (Fig 4C). In females, BV/TV gradually increased from 30 days old to 6 months old in CKO mice, however, it increased from 30 days old to 56 days old, and significantly declined from 56 days old to 6 months old in f/f and Cre+ two control mouse groups (Fig 4D). While the differences at 6 months old were described above, the pattern of changes of BV, TV and Tb.N from 30 days old to 56 days to 6 months old were not significantly different among three mouse groups (Fig 4D). On cortical side, the changes of BV/TV, BV, B.Ar and Cs.Th from 30 days old to 6 months old were found similar to we observed in male mice with no differences among all three groups (Fig 4D).
Deceased inflammatory milieu and osteoclast numbers in bone of Ezh2flox/flox/LysM-Cre+ CKO mice at 6 months old
We isolated total proteins from the spine (L4) of CKO, and f/f and Cre+ two control mouse groups at 6 months old, and inflammatory antibody array analysis was performed. In males, almost half of the inflammatory factors measure were significantly lower in CKO mice compared with those from f/f mice (Fig 5A). Among those, M-CSF was significantly lower in CKO mice compared with f/f mice (Fig 5A), notice that M-CSF is a key positive regulator for osteoclast activation and differentiation. When CKO was compared with Cre+ mice, we found more than two third (60%) of those inflammatory factors were significantly lower in CKO mice (Fig 5A). Among those factors, CSFs were also significantly lower in CKO mice compared with Cre+ control mice (Fig 5A). In females, there were 14 out of 40 of those inflammatory factors were significantly lower in CKO mice compared with f/f mice (Fig 5B), and 24 out of 40 (60%) of those inflammatory factors were significantly lower in CKO mice compared with Cre+ control mice (Figure 5B). Notably, in addition to CSF TNFα was also significantly lower in CKO mice compared either f/f or Cre+ control mice (Fig 5B). While the significant differences of expression of some inflammatory factors between CKO and control mice were not profound, these data strongly indicated that inflammatory milieu in bone of both male and female CKO mice was lower compared with their respective control mice, and suggested that octeoclast activity or differentiation is affected by deletion of Ezh2 in osteoclastic cells. We next performed bone histomorphometric analysis to evaluate cellular changes in bone of CKO mice compared with their controls. As shown in Fig 6A, trabecular number and bone area (Fig 6B) were significantly higher in CKO mice compared with either f/f or Cre+ control mice. Bone from CKO mice had less TRAPase positive staining compared with samples from f/f and Cre+ control mice (Fig 6A). Osteoclast number (Fig 6C) and osteoclast surface (Fig 6D) were significantly lower in CKO mice compared with f/f and Cre+ control mice, consistent with findings of the inflammatory array data. Unexpectedly, osteoblast number in bone of Ezh2 osteoclastic cell specific knockout CKO mice significantly higher than it in bone from f/f and Cre+ mice (Fig 6E), suggesting deletion of Ezh2 in osteoclastic cells indirectly affecting on osteoblastogenesis. Consistent to the results we found in males, data from female mice were presented in Supplemental Fig 3. Indeed, when we used Ezh2 and IRF8 antibody double staining on proximal tibia section from CKO and Cre+ mice, we found inverse association between Ezh2 and IRF8 expression on the bone surface of osteoclastic cells (Fig 6F).
Ezh2 controls osteoclast suppressive gene expression regulating osteoclastogenesis
Western blots were performed using total protein isolated from L4 of CKO and f/f and Cre+ mice at 6-monthof age. Two osteoclast bone resorptive markers Cathepsin K and MMP9 weresignificantly lower of their protein expression in CKO mice compared with f/f or Cre+ control mice (Fig 7A). On the other hand, protein levels of osteoclast suppression markers including IRF8, MafB and Arg1 were significantly higher in bone from CKO mice compared with f/f or Cre+ mice (Fig 7A). Unexpectedly, the protein levels of the osteoblastic bone formation markers osteocalcin and Col1 were significantly higher in bone from CKO mice compared with f/f and Cre+ mice (Fig 7A). Serum bone remodeling marker measurements were performed using ELISA. Serum CTX1 and TRAP-5b levels were significantly lower in CKO mice compared with f/f and Cre+ mice (Fig 7B). On the other hand, serum P1NP and osteocalcin levels were significantly higher in CKO mice compared with f/f and Cre+ control mice (Fig 7B). These Western blot and serum bone remodeling marker measurement results found in males were consistent with results found in female mice (Supplemental Fig 4). Indeed, when we isolated total RNA from bone, real-time PCR showed that NFATc1 and Cathepsin K mRNA expression were significantly lower in CKO mice compared with f/f and Cre+ mice (Fig 7C). On the other hand, IRF8, MafB and Arg1 those osteoclast inhibitory factor mRNA and osteocalcin mRNA expression were significantly higher in CKO mice compared with f/f and Cre+ mice (Fig 7C). Moreover, using genomic DNA and chromatins isolated andprecipitated from bone of CKO, f/f and Cre+ mice, we performed H3K27me3 and Ezh2 ChIP analysis. We found that IRF8, MafB and Arg1 those osteoclast inhibitory factors were significantly associated with H3K27me3 in CKO mice compared with f/f and Cre+ mice, but less associated with Ezh2 (Fig 7D). Notice that CKO mice were Ezh2 gene deleted only in osteoclastic cell population, chromatins were precipitated from multi mixed cell types.
We used pharmacologic inhibitors of Ezh2 to investigate osteoclastogenesis and inverse association of Ezh2 and IRF8 expression using in ex vivo bone marrow non-adherent hematopoietic cell and RAW246.7 cell cultures. We used commercially available Ezh2 inhibitors GSK126, GSK343 and DZNep which inhibited osteoclastogenesis in ex vivo non-adherent bone marrow hematopoietic cell cultures in a dose dependent manner (Fig 8A and 8B). GSK343 at a concentration of 0.5 µM did inhibit Ezh2 expression (Fig 8C and 8D), when we looked at a single osteoclastic cell which Ezh2 was inhibited (Figure 8C), but IRF8 mRNA expression was significantly higher in RAW246.7 cell cultures (Fig 8D). These data once again indicated that Ezh2 and osteoclast inhibitory factors has inverse relationships of their expression and osteoclast inhibitory factor expression is under epigenetic control through Ezh2 to regulate osteoclastogenesis and bone resorption.
Discussion
In our current report, we have determined that deletion of Ezh2 gene in osteoclast precursors increases bone mass in both male and female mice. Densitometric and structural analysis on cortical compartment showed bone was not osteopetrotic in Ezh2 CKO mice, indicating that Ezh2 is not essential for late stage of osteoclast differentiation, but it significantly contributes to controlling osteoclstogenesis and bone resorption in young adult mice. Postnatal bone accretion was accelerated in mice lacking Ezh2. In addition, Ezh2 deletion differential influences bone mass acquisition and bone homeostasis in weanling and adult mice. Moreover, we provide evidence of sxual dimorphism with regards to Ezh2 function in bone development and turnover. In males, not only CKO mice but also Ezh2flox/+/LysM-Cre+ genotypic mice showed increased bone mass, while in females only CKO mice had increased bone mass phenotype compared with other genotypic young adult control mice.
Our study provide new mechanistic data concerning the role of Ezh2 in osteoclast function and bone mass accretion. Specifically, in the absence or inactivated Ezh2 in pre-osteoclasts, osteoclastic suppressive genes are activated leading to suppression of osteoclastogenesis and osteoclast activity. This suggests that osteoclastogenesis and bone resorption are under epigenetic regulation through Ezh2. On the other hand, deletion of Ezh2 in osteoclast precursors also inhibited genes of those known as osteoclastogenic or required for osteoclast differentiation and activity, but indirectly activated osteoblast markers. These data suggest that Ezh2 gene expression plays a key role in governing osteoclast signaling, thereby altering the balance between bone resorption and bone formation. An altered profile of inflammatory markers in pre-osteoclastic cells from mice lacking Ezh2 may provide an environmental signal for osteoclast and osteoblast differentiation and activity. Collectively, our data identify Ezh2 as a potential target for strategies aimed at treating both osteoclastic bone resorptive disorders such as osteoporosis and rheumatoid arthritis.
Modification of osteoclast gene expression through histone modification during osteoclastogenesis might be a good approach to control osteoclast differentiation or activity and thus alter osteoclastic bone resorption (Yi et al, 2019). Osteoclasts are large multinucleated cells with ruffled membranes, and they are differentiated from mononuclear cells of the hematopoietic stem cell lineage by fusing those monocytes to form tartrate-resistant acid phosphatase (TRAP)-positive cells (Levaot et al, 2015). It has been known for many years that two factors, macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor kappa-B ligand (RANKL) stimulate osteoclast differentiation, and RANKL is essential for the formation, fusion, activation and survival of osteoclasts (Asagiri & Takayanagi, 2007). Now, there is a growing body of evidence suggested that histone methyltransferase is involved in bone cell differentiation (Cakouros & Gronthos, 2019). Ezh2 has been reported to suppress osteoblastogenesis via H3K27me3 methylation on the promoters of osteoblast-related genes (Wei et al, 2011). On the other hand, in vitro studies have shown that Ezh2 promotes osteoclastogenesis by downregulating IRF8, a negative regulator of osteoclastogenesis (Adamik et al, 2020). Studies have also shown that H3 monomethylation at lysine 27 by G9a is essential for the enzymatic activity of MMP-9 and facilitates osteoclast differentiation (Shin et al, 2019), and DOT1L a histone methyltransferase of H3K79me suppressed osteoclast differentiation (Gao & Ge, 2018). In our current report, we have successfully generated a pre-osteoclast, macrophage specific Ezh2 gene deletion mouse model. Before mice reach to 56 days old, bone development of those CKO mice seem to be no significantly different compared with their control littermates. Compared with all other genotypic littermates generated by in-bred breeding, Ezh2 deletion in pre-osteoclastic LysM-Cre positive cells promotes bone mass accretion. Indeed, we provide clear evidence in vivo that deletion of Ezh2 in pre-osteoclastic cells inhibits osteoclastogenesis in young adult mice (Chen et al, 2020).
Ezh2 is known as a histone lysine methyltransferase associated with predominantly transcriptional or gene repression (Nutt et al, 2020), it catalyzes the addition of methyl groups to H3K27me3 in target genes involved in numerous cellular processes. We have originally showed evidence that Ezh2 might be involved in bone cell differentiation in the context of maternal or early life diet-induced obesity (Chen et al, 2016). Although our current studies focused on determining the role of Ezh2 in osteoclastogenesis and bone resorption, we have previously investigated the role of Ezh2 in the function of osteoblastic cells. These data suggest that maternal HFD-induced obesity increases Ezh2 expression in fetal osteoprogenitors, while increased Ezh2 activity suppressed expression of osteoblastic marker genes such as SATB2. Ezh2 facilities association of H3K27me3 with SATB2 in fetal cells from both obese dams in rodents and obese mothers in human (Chen et al, 2020). We described that deletion of Ezh2 gene in osteoblastic specific cell lineage increased bone density, especially trabecular bone density and bone mineral content, and increased SATB2 expression in bone (Chen et al, 2020). Together with data presented in our current report, these results clearly demonstrats that Ezh2 expression controls both osteoblast and osteoclast differentiation and activity. Repressive H3K27me3 and active H3K27ac are epigenetic marks that control gene expressions associated with them, some of them are essential in controlling osteoclast and osteoblast differentiation (Yi et al, 2019), however, there are only few studies offer little and inconsistent insight on how epigenetic remodeling of bone-specific chromatin maintains bone cell development and bone mass in vivo. Genetic inactivation of Ezh2 in calvarial bones using Ezh2flox/flox and Prrx1-Cre model enhanced expression of osteogenic genes (Dudakovic et al, 2020). Evidence from several other studies also support a role for Ezh2inhibition in stimulating bone accretion (Ferguson et al, 2018). It has been shown that loss or inhibition of Ezh2 not only resulted in enhanced osteogenesis, but also inhibited adipogenic differentiation of mesenchymal stem cells from both human and rodents (Hemming et al, 2014; Lui et al, 2016). These were interesting evidence describing the role of Ezh2 on bone cell development, however, bone development and modeling are controlled by both osteoblastic and osteoclastic cells, therefore, to determine how Ezh2 controls osteoclast development in vivo is important. However, we have only seen few evidence from in vitro or in ex vivo studies demonstrated the role of Ezh2 during osteoclastogenesis (Adamik et al, 2020). We have shown that deletion of Ezh2 in pre-osteoclastic cells resulted in reduced trimethylation in those genes that play suppressive functions on osteoclast differentiation. Interestingly, for those genes that are responsible for stimulating osteoclast were down-regulated.
Osteoclastic suppressive genes have been identified such as IRF8 (interferon regulatory factor 8), MafB (V-mad musculoaponeurotic fibrosarcoma oncogene homologue B) and Arg1 (Arginase 1), IRF8 has received significant attention lately due to its role in inhibiting osteoclast differentiation (Thumbigere-Math et al, 2019). It has recently been reported that both homo and heterozygous of IRF8 gene mutations promote osteoclastogenesis and increased osteoclast activity, which is accompanied by increased NFATc1 expression (Izawa et al, 2019). IRF8 plays critical roles in the regulation of lineage commitment and in myeloid cell maturation including the decision for a common myeloid progenitor to differentiate into a monocyte precursor cell (Das et al, 2020). The LysM-Cre mouse model used in our current study has a nuclear-localized Cre recombinase inserted into the first coding ATG of the lysozyme 2 gene (Lyz2); both abolishing endogenous Lyz2 gene function and placing NLS-Cre expression under the control of the endogenous Lyz2 promoter/enhancer elements (The Jackson Laboratory). These LysM-Cre mice were generated to be used for Cre-lox studies of the myeloid cell lineage (monocytes, mature macrophages and granulocytes) and the innate immune response. Therefore, we expected that Ezh2 gene deletion in the myeloid cell lineage could interfere with IRF8 expression and function. We have showed that Ezh2 gene deletion resulted in increased IRF8 gene expression, therefore osteoclastogenesis and bone resorption were inhibited. Indeed, ChIP assay indicated that Ezh2 and IRF8 are highly associated, and Ezh2 controls IRF8 gene trimethylation are obviously through H3k27me3. Further, overexpression of Ezh2 in osteoclastic cells (murine macrophage cell line RAW264.7) suppresses IRF8 expression, and osteoclastogenesis was increased. These data suggest that Ezh2 and IRF8 axis in the myeloid cell lineage is at least one of mechanisms of how Ezh2 controls osteoclastogenesis and bone resorption.
Deletion of Ezh2 in pre-osteoclasts resulted in a significant reduction in activities of inflammatory factors within bone. We do not have explanation for why Ezh2 deletion alters the bone inflammasome, we only speculate that inflammatory signatures in macrophages in bone can be programmed through Ezh2 expression. Although decreased osteoclastogenesis and increased bone mass in 6 months old Ezh2 lacking male and female mice are clear, the postnatal bone modeling and growth in young CKO mice were more ambiguous. However, we did not observe obvious differences of bone mass in 30 days old CKO mice compared with all other genotypic and wild type mice in both males and females. Since bone mass in young adult CKO mice was significantly greater than control mice, overall bone growth in CKO mice was faster from 30 days old to 6 months old. The reason for why 30 days old CKO mice showed no obvious bone phenotypic changes compared to control may be because, 1) Ezh2 deletion is in pre-osteoclastic cells, 2) 30 days old mice were just after weaning, osteoblastic bone formation is a dominate event for rapid bone growth during this age of mice (Chen et al, 2017). Moreover, the function of osteoclasts during rapidly growing phase are probably to make the bone in a good shape (Chen et al, 2017), but not much on bone mass. This is supported by our data, although overall there were no statistically different among groups on bone density, but trabecular centroid (x) in 30 days old CKO mice was significantly different compared to their control littermates (Cre+ mice) in both sexes. We stated that CKO mice were notosteopetrotic. Osteopetrosis is defined by increased bone density due to a defect in bone resorption by osteoclasts. In mice laking Ezh2, we did not observe clear differences in the cortical bone compartment by micro-CT analysis, and there were still osteoclasts in bone of CKO mice. Therefore, our determination is that in CKO mice, osteoclastic bone resorption was decreased but not completely abolished. Lastly, we have shown that in males, not only CKO mice but also Ezh2flox/+/LysM-Cre+ genotypic mice showed increased bone mass, while in females only CKO mice had increased bone mass phenotype compared with control mice (compared with either flox, Cre and wild type control) in young adult mice. This observation suggests a potential role for sex hormones or Ezh2 interfered with X chromosome in bone accretion and modeling, which warrants further investigation.
In summary, we show that Ezh2 deletion in osteoclast precursors increased postnatal bone development as well as bone mass in young adult mice. Our data demonstrate a role for Ezh2 in osteoclast gene expression, modulating the bone inflammasome, as well as controlling bone resorption. We suggest that manipulation of Ezh2 expression may be a viable strategy to alter bone development or manage bone resorptive disorders such as osteoporosis and rheumatoid arthritis.
Materials and Methods
Animals and diets
C57BL/6J mouse background myeloid cell-specific targeted mutant LysMcre (LysM-Cre+) mice (Stock No: 004781) and loxP sites flanking exons 14-15 of the zeste homolog 2 (Ezh2) gene Ezh2flox/flox mice (stock no. 022616) were purchased from the Jackson Laboratory (Bar Harbor, ME). To delete targeted gene of Ezh2 in the myeloid cell lineage of osteoclast precursors, including monocytes, mature macrophages, and granulocytes, LysM-Cre+ mice were crossed with Ezh2flox/flox mice. Male Ezh2flox/+/LysM-Cre+ and female Ezh2flox/+/LysM-Cre- offspring were intercrossed to generate the LysM-Cre+ (Cre-positive control); Ezh2flox/flox /LysM-Cre+ mice (CKO) and others Ezh2flox/+/LysM-Cre-, Ezh2flox/flox/LysM-Cre- (flox control), Ezh2flox/+/LysM-Cre+, and wild type (Wt) mice. A staggering mouse breeding strategy was used and all mice were fed standard rodent chow diet. Experiments involved 30 days old, 56 days old and 6 months old littermates of male and female CKO and their corresponding genotypic and wild type control male and female mice (8-12 per group) for bone phenotyping studies. Mice were weighed, randomized by their weights, and housed 5 per cage in mouse small shoe box cages. All animal procedures were approved by the Institutional Animal Care and Use Committee at University of Arkansas for Medical Sciences (AUP#3595 UAMS, Little Rock, AR), and housed in animal facility at the Arkansas Children’s Research Institute with constant humidity and lights on from 06:00-18:00 hrs at 22°C. At the end of the studies, mice were anesthetized by injection with 100 mg Nembutal/kg body weight (Avent Laboratories). Blood was collected via cardiac puncture, which was followed by decapitation, femur, tibia and vertebrae bones were collected and stored in -80 °C.
Bone analyses
Micro-computed tomography (micro-CT) measurements of the trabecular and cortical compartments from the left tibial and L5 spine bone were evaluated using SkyScan μCT scanner (recently upgraded SkyScan 1272, Bruker.com) at 8 μm pixel size with X-ray source power of 60 kV and 166 μA and integration time of 950 ms. The trabecular compartment included a 0.9 mm region extending distally 0.03 mm from the physis. The grayscale images were processed by using a filter (=Al, σ=0.5, mm) to remove noise, and a fixed threshold of 125 was used to extract the mineralized bone from the soft tissue and marrow phase. Cancellous bone was separated from the cortical regions by semi-automatically drawn contours. A total of 120 slices starting from about 0.1 mm distal to growth plate, constituting 0.80 mm length, were evaluated for trabecular bone structure, bone volume fraction (BV/TV, %), trabecular thickness (Tb.Th, mm), trabecular separation (Tb.Sp, mm), trabecular number (Tb.N, 1/mm), Degree of anisotropy (DA) were calculated based on description by Bouxsein et al. (Bouxsein et al, 2010), and by using software provided by SkyScan, Bruker. For cortical bone, the cortical compartment was a 0.6 mm region extending distally starting 5 mm proximal to the tibia-fibula junction. Total cross-sectional area (CSA, mm2), medullary area (MA, mm2) and cortical thickness (Ct.Th, mm) were assessed.
Bone histology
Mouse right tibia samples were embedded, cut and TRAPase stained by Histology Special Procedures at the Arkansas Children’s Nutrition Center Histology Core. TRAPase staining kit was utilized according to the manufacturer’s protocols (Sigma-Aldrich, Acid phosphatase leukocyte, procedure No. 386). TRAPase positive pink-purple-stained cells were visualized with a digitizing morphometry system, which consists of an epifluorescent microscope (model BH-2, Olympus), a color video camera, and a digitizing pad (Numonics 2206) coupled to a computer (Sony) (OsteoMetrics, Inc.).
Measurements of bone turnover markers in bone marrow plasma and in serum
Bone marrow plasma were prepared at the time of tissue harvest. Bone marrow was flushed out from femur using 300 µl of PBS, vortexed and spun (1700 g) for collecting supernatant as bone marrow plasma. Serum bone resorption marker C-terminal telopeptides of type I collagen (CTX-1) was measured by Rat-LapsTM ELISA from Nordic Biosciences Diagnostic (Herlev, Denmark). The serum P1NP (Procollagen 1 N-terminal Propeptide) levels were measured by direct immunoassay P1NP assay Kit. The P1NP level measurement ELISA kit was purchased from Mybiosource.com (Catalog No: MBS2500076) and measurement procedure followed the manufacturer’s recommendations. According to the protocol, pre-coated with total-P1NP antibody, total-P1NP present in the sample is added and binds to antibodies coated on the wells, and then biotinylated total-P1NP antibody is added and binds to total-P1NP in the sample. Then, streptavidin-HRP is added and binds to the biotinylated total-P1NP antibody. After incubation unbound streptavidin-HRP is washed away during a washing step, substrate solution is then added and color develops in proportion to the amount of total-P1NP. ELISA for serum osteocalcin and TRAP-5b were performed according to procedure from manufactory. The reaction is terminated by addition of acidic stop solution and absorbance is measured at 450 nm.
Cell cultures
Murine osteoclast cell line RAW264.7 cells (from ACCT) and non-adherent bone marrow cells were cultured in 96-well plates (2×104 cells/well) in the presence or absence of 30 ng/ml of RANKL, in α-MEM supplemented with 10% fetal bovine serum (FBS) (Hyclone, Logan, UT), penicillin (100 Units/ml), streptomycin (100 µg/ml), and glutamine (4 mM). These cell cultures were previously described in our laboratory (Chen et al, 2020). After 5 days for bone marrow cell cultures, the cells were fixed with 4% paraformaldehyde and stained for TRAPase activity using a TRAPase staining kit according to the manufacturer’s protocols (Sigma-Aldrich, Acid phosphatase leukocyte, procedure No. 386). TRAP-positive cells containing >3 nuclei in each well were counted as osteoclasts under an epifluorescent microscope (model BH-2, Olympus, Imaging America Inc.; Center Valley, PA). For RNA and protein expression study, non-adherent bone marrow cells were seeded in triplicate in 6-well collagen-coated plates (BD Biosciences) at a density of 1×105 cells/well, and cells were treated with Ezh2 inhibitors GSK126, GSK343 and DZNep at four different concentrations in the presence of 30 ng/ml RANKL for 5 days. Murine bone marrow mesenchymal stromal cell line ST2 cells were culture in the osteoblast differentiation medium.
RNA isolation, real-time reverse transcription-polymerase chain reaction
Bone and cell RNA isolation was performed using TRI Reagent (MRC Inc., Cincinnati, OH) according to the manufacturer’s recommendations followed by DNase digestion and column cleanup using QIAGEN mini columns (Chen et al, 2018). Reverse transcription was carried out using an iScript cDNA synthesis kit from Bio-Rad (Hercules, CA). All primers for real-time PCR analysis used in this report were designed using Primer Express software 2.0.0 (Applied Biosystems).
Western blotting
Total protein extracts from L4 and culture cells were prepared using radioactive immunoprecipitation assay (RIPA) buffer (Solarbio). Western blots were performed using standard protocols (Chen et al, 2018). The protein lysates were quantified and separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes (Millipore). This was followed by immunoblotting with primary antibodies NFATc1 (SAB2101576, Sigma-Aldrich, St. Louis, MO, USA), MMP9 (sc-6840, Santa Cruz, Dallas, TX, USA), Cathepsin K (sc-48353, Santa Cruz, Dallas, TX, USA), Ezh2 (Cell Signaling, #9346), IRF8 (Invitrogen, #39-8800), Arg1 (Cell Signaling, #93668), MafB (Bethyl, A700-046), Osteocalcin (Millipore, AB10911), Col1 (C2456, Sigma-Aldrich, St. Louis, MO, USA) and β-Actin (A1978, Sigma-Aldrich, St. Louis, MO, USA); 1:1000 dilution and then by the corresponding horseradish peroxidase conjugated secondary antibodies. Blots were developed using chemiluminescence (PIERCE Biotechnology) according to the manufacturer’s recommendations. Bands of interest were visualized and imaged under chemiluminescent detection using a FluorChem E System (ProteinSimple, San Jose, CA). Quantitation of the intensity of the bands in the autoradiograms was performed using a VersaDocTM imaging system (Bio-Rad).
Chromatin-immunoprecipitation (ChIP)
The procedure for standard ChIP assay using Ezh2 (Cell Signaling, #9346) and H3K27me antibody (ChIP grade from Cell Signaling, #07-449) has been described previously (Chen et al, 2018), and information for all primers used for ChIP assay and real-time PCR were designed using Primer Express software 2.0.0 (Applied Biosystems). The procedure for standard ChIP assay using H3K27me3 and Ezh2 (Cell Signaling) antibodies was described previously (Chen et al, 2018).
Inflammatory antibody array
Simultaneous detection of multiple cytokines undoubtedly provides a powerful tool to study cell signaling pathways. RayBio C-Series mouse (#AAM-INF-1-8, RayBiotech, Inc) inflammatory antibody arrays were performed for the semi-quantitative detection of 40 mouse proteins in cell lysate according to protocols provided by the company. L4 vertebra bone tissue proteins were extracted using a cell lysate buffer as described previously (Chen et al, 2016). Total protein 50 µg from each sample which the original lysate concentration ranged 1-5 mg/ml was loaded. According to description provided by the company, this array can detect IL-2 at concentration of 25 to 250,000 pg/ml, a range of 10,000-fold, and as determined by densitometry, as little as 4 pg/ml of MCP-1 can be detected, the inter-array Coefficient of Variation (CV) of spot signal intensities is 5-10%. Quantitation of the intensity of the bands in the autoradiograms was performed using a VersaDocTM imaging system (Bio-Rad).
Statistics
Numerical variables were expressed as means ± SD (Standard Deviation), and n represents the number of samples/group. For in vitro and ex vivo experiments, differences within groups were evaluated using t test or one-way ANOVA followed by Tukey’s post hoc test. P<0.05 was considered significant. Cell culture experiments were conducted at least three independent times, and representative images (including bone scan images) are displayed. Dose or time response was assessed using Cruick’s non-parametric test for trend. For in vivo animal experiments, comparisons among groups were performed with the nonparametric Kruskall-Wallis test followed by a Dunnett’s test comparing each genotype group to the control or CKO group. The nonparametric Wilcoxon rank-sum test was used to compare CKO to individual LysM-Cre positive or Ezh2 flox and other genotypic and wild type mouse groups. Values were considered statistically significant at p<0.05.
Author contributions
J.R.C. designed and performed the study, and wrote the paper; O.P.L. and M.L.B. performed cell, biochemical and molecular experiments, and in vivo sample analysis, helped to perform ChIP analysis; D.G. and C.L. performed in ex vivo and in vitro experiments; and F.Z. contributed in the performance of experiments and study designs.
Acknowledgement
This work was supported by sub-objective to J.R.C. by United States Department of Agriculture (USDA) / Agricultural Research Service (ARS) Project # USDA-ARS Project 6026-51000-012-06S to the Arkansas Children’s Nutrition Center. Authors would like to thank Jim Sikes, Matt Ferguson, Hoy Pittman and Bobby Fay for their technical assistances on animal experiments.
Footnotes
Conflict of interest statement: The authors have declared that no conflict of interest exists.
Funding: Supported by USDA-ARS Project 6026-51000-012-06S