MicroRNAs involved in bone formation

Abstract

During skeletal development, mesenchymal progenitor cells undergo a multistage differentiation process in which they proliferate and become bone- and cartilage-forming cells. This process is tightly regulated by multiple levels of regulatory systems. The small non-coding RNAs, microRNAs (miRNAs), post-transcriptionally regulate gene expression. Recent studies have demonstrated that miRNAs play significant roles in all stages of bone formation, suggesting the possibility that miRNAs can be novel therapeutic targets for skeletal diseases. Here, we review the role and mechanism of action of miRNAs in bone formation. We discuss roles of specific miRNAs in major types of bone cells, osteoblasts, chondrocytes, osteoclasts, and their progenitors. Except a few, the current knowledge about miRNAs in bone formation has been obtained mainly by in vitro studies; further validation of these findings in vivo is awaited. We also discuss about several miRNAs of particular interest in the light of future therapies of bone diseases.

Introduction

Bone formation requires tightly coordinated proliferation and differentiation of various types of cells. This is achieved through multiple levels of regulatory systems. Cues for proliferation and differentiation of bone cells are usually provided by extracellular signaling molecules, which eventually control gene expression to regulate cellular functions. Gene expression can be regulated by several different mechanisms. Relatively recently, small non-coding RNAs, microRNAs (miRNAs), discovered in diverse organisms, have been shown to play important regulatory roles in gene expression mainly at the post-transcriptional level. The discovery of this novel regulatory system prompted vigorous investigations into the roles of miRNAs in skeletal cells.

Bone- and cartilage-forming cells are originated from mesenchymal cells of the neural crest and lateral and paraxial mesoderm [1]. Bone formation starts with mesenchymal condensation in which mesenchymal cells form clusters at regions where future bones are to develop. In the mesenchymal condensation, cells differentiate to progenitors for chondrocytes that form cartilage and osteoblasts that produce mineralized bone. Mineralized bones are generated through two different processes: membranous and endochondral ossification. In membranous ossification, mesenchymal cells differentiate directly into bone-forming osteoblasts, whereas in endochondral ossification, mesenchymal cells first differentiate to chondrocytes that produce extracellular matrix proteins, such as type II collagen and aggrecan, to form a cartilage template. This cartilage template (the growth plate) is subsequently replaced by mineralized bone.

The growth plate cartilage drives longitudinal bone growth. Proliferation and differentiation of growth plate chondrocytes need to be tightly coordinated for proper growth of long bones. The growth plate consists of three groups of chondrocytes: resting (or periarticular in embryos), proliferating, and hypertrophic chondrocytes. Resting chondrocytes proliferate slowly, and differentiate into proliferating chondrocytes that form orderly columns. Proliferating chondrocytes further differentiate into post-mitotic hypertrophic chondrocytes that express type X collagen. Hypertrophic chondrocytes undergo mineralization and then are replaced by mineralized bone. This continuous proliferation and differentiation of chondrocytes, followed by their replacement by mineralized bone, leads to the longitudinal bone growth.

Signaling molecules play important roles in regulation of chondrocyte differentiation. Indian hedgehog (Ihh), secreted by prehypertrophic and hypertrophic chondrocytes, regulates chondrocyte proliferation and Ihh acts on resting chondrocytes to regulate their proliferation and differentiation, and expression of parathyroid hormone-related peptide (PTHrP, Pthlh), a signaling molecule that inhibits hypertrophic differentiation. Ihh also induces osteoblast differentiation of perichondrial cells, coupling chondrogenesis and osteogenesis [2]. Bone morphogenetic protein (BMP) signaling is essential for chondrocyte differentiation in vivo [3]. BMP signaling in prechondrogenic cells stimulates their differentiation into chondrocytes [4]. BMP signaling also promotes differentiation of proliferating chondrocytes into hypertrophic chondrocytes. BMPs bind to type I and type II BMP receptors and signal through Smads and mitogen-activated protein kinase (MAPK) pathways [5]. Fibroblast growth factors (FGFs) bind to the cell surface tyrosine/kinase receptors, fibroblast growth factor receptors (FGFR1-4 in humans and mice) that activate multiple intracellular signaling pathways, including MAPK, phosphoinositide 3-kinase (PI3K), and JAK/STAT1 (Janus kinase/signal transducer and activator of transcription) to regulate chondrocyte differentiation [6].

Osteoblasts are differentiated from local mesenchymal progenitor cells. Osteoblasts produce mineralized bone matrix whose major component is type I collagen, then die or further differentiate into bone-lining cells or osteocytes that are embedded in the bone matrix. Osteoblast differentiation is regulated by multiple signaling systems. During the endochondral bone formation, Ihh, produced by hypertrophic chondrocytes, induces osteoblast differentiation of mesenchymal progenitor cells in the perichondrium of endochondral bones [7]. Wnt signaling is essential for osteoblast differentiation in both membranous and endochondral bone formation [8–11]. Wnt ligands bind to frizzled receptors and their co-receptor, low-density lipoprotein receptor-related protein 5 (LRP5) or LRP6, and signal through multiple intracellular signaling pathways. In the canonical pathway, Wnt signaling stabilizes β-catenin and promotes its nuclear localization [12]. β-catenin forms a transcriptional complex containing lymphoid enhancer-binding factor 1(Lef1). β-catenin-independent, non-canonical Wnt signaling pathways also regulate functions and differentiation of cells of the osteoblast-lineage [13]. Notch signaling [14], mediated by Notch receptors (Notch 1–4) and the ligands, Jagged 1 and 2, and Delta-like 1, 3, and 4, suppresses differentiation of mesenchymal cells into pre-osteoblastic cells, whereas it promotes differentiation of pre-osteoblasts into osteoblasts.

Mineralized bone is constantly remodeled though coupled bone formation and bone resorption. Bone resorption is mediated by osteoclasts that are differentiated from precursor cells of the monocyte/macrophage lineage. Osteoclast differentiation requires two essential signaling molecules; colony-stimulating factor 1 (CSF1) and ligand to receptor activator of NFκB (RANKL) [15]. These signaling molecules sequentially induce differentiation of osteoclasts that express specific markers including tartrate-resistant acid phosphatase (TRAP; Acp5) and cathepsin K (Ctsk). As bone is resorbed by osteoclasts, growth factors [e.g., transforming growth factor beta (TGF-β), insulin-like growth factor 1 (IGF-1)] are released from the bone matrix; these growth factors, in turn, regulate the function and differentiation of osteoblasts.

Signals from extracellular space are eventually conveyed to the nucleus where transcription factors primarily control gene expression. Sex determining region Y (SRY) box 9 (Sox9) is an essential transcription factor for chondrocyte differentiation. It is expressed in common progenitors for chondrocytes and osteoblasts. Upon differentiation, its expression becomes restricted to chondrocytes. Sox9 is also important for chondrocyte function; it regulates expression of collagen type II, collagen XI, and aggrecan [16, 17]. Sox9-related molecules, Sox 5 and Sox 6, enhance the activity of Sox9 [18]. Myocyte enhancer factor 2 (MEF2) family transcription factors promote hypertrophic differentiation of chondrocytes. Histone deacetylase 4 (HDAC4) is a major negative regulator of MEF2s and thereby suppresses hypertrophic differentiation [19]. Runx2, a member of the Runt transcription factor family, stimulates chondrocyte hypertrophy [20, 21]. Runx2 is essential for osteoblast differentiation [22, 23]. Osterix (Osx, Sp7) that acts at a level genetically downstream of Runx2 is the other transcription factor necessary for osteoblast differentiation [24]. Transcription factors also regulate osteoclast differentiation. The transcription factor, PU.1, is expressed in monocytes and required for differentiation into early osteoclast precursors. In later stages, Fos, Mitf (microphthalmia-associated transcription factor), Nfatc1 (Nuclear factor for activated T cell c1), and NFkB are individually essential for differentiation and function of osteoclasts [25].

Gene expression is also controlled at the post-transcriptional level where miRNAs play significant roles. MiRNAs, ~22 nucleotide-long, non-coding RNAs, directly bind to RNAs in a sequence-complimentary manner to inhibit translation and facilitate degradation of their target transcripts. MiRNAs are encoded in the genome, and are transcribed as long primary transcripts (pri-miRNAs). Pri-miRNAs are subsequently processed by the Drosha/DGCR8 microprocessor complex into small hairpin precursor miRNAs (pre-miRNAs). Pre-miRNAs are transported by exportin-5 to the cytoplasm and are cleaved by Dicer into mature and functional miRNAs [26].

The importance of miRNAs in skeletal development was initially demonstrated by studies deleting Dicer in skeletal cells in vivo. The global reduction in miRNAs caused significant defects in multiple types of bone cells. However, understanding roles of individual miRNAs in bone formation in vivo is still rudimentary; most of our current knowledge about roles of miRNAs is based on in vitro studies using various experimental systems.

In this review, we discuss the role of miRNAs in three major cell types of bone: chondrocytes, osteoblasts, and osteoclasts. For each cell type, we first discuss about in vivo studies and then review major findings from in vitro studies.

The role of miRNAs in chondrocytes

In vivo studies

The effect of a global reduction of miRNAs in cartilage and bone development was shown in mice, in which the Dicer gene was conditionally deleted in the limb bud mesenchymal cells [27] and in chondrocytes [28]. Chondrocyte-specific Dicer deletion resulted in significant growth defects and premature death by the time of weaning. A reduction in proliferating chondrocytes and expansion of the hypertrophic region were observed, suggesting miRNAs regulate both proliferation and differentiation of growth plate chondrocytes.

The RNA, Dnm3os, is transcribed from an intron of the mouse Dynamin 3 (Dnm3) gene and encodes three miRNAs, miR-199a, miR-199a*, and miR-214. It is expressed in multiple types of cells including mesenchymal cells as well as perichondrial cells and periarticular chondrocytes. Its function was studied in mutant mice in which the 5′ region of Dnm3os was replaced by insertion of a lacZ gene [29]. This modification caused a reduction in all three miRNAs. The mutant mice showed growth retardation, craniofacial hypoplasia, dorsal vertebral hypoplasia, and osteopenia, strongly suggesting the essential role of these three miRNAs in normal mouse skeletal development. In human and mouse mesenchymal stem cells (MSCs), miR-199 is upregulated upon chondrocyte differentiation [30, 31].

The role of miR-140, which is expressed in chondrocytes in a relatively specific manner, has been studied in zebrafish [32] and mice [33–35]. In mice, miR-140 expression is mostly found in the cartilage of developing limbs, ribs, vertebrae, sternum, and the skull [36]. MiR-140 is encoded in an intronic sequence of the Wwp2 gene. MiR-140 expression shows an overlapping pattern with those of Wwp2, Sox9, and Col2a1 (encoding type II collagen) [37]. In chondrogenic ATDC5 cells, Sox9 increases miR-140 expression and cell proliferation [37]. In this paper, Sp1 transcription factor, a cell cycle inhibitor, was found upregulated upon miR-140 inhibition in vitro and was suggested as a miR-140 target. Nakamura et al. [35] found a similar regulatory role of Sox9 in miR-140 expression in vivo using Sox9 mutant zebrafish. A study in mice also confirmed this relationship of Sox9 and miR-140 [18]. This study also showed that a proximal upstream region of the pri-miR-140 gene possessed a chondrocyte-specific promoter activity, which was directly regulated by Sox9.

MiR-140-null mice show a mild skeletal growth defect with short endochondral bones and reduced longitudinal growth of the skull [33]. Miyaki et al. [33], found that these mice also showed age-related osteoarthritic changes in the articular cartilage presumably due to aberrant expression of the matrix-degrading enzyme, Adamts-5, a direct target of miR-140. The growth plate phenotype was further analyzed in a study using independently generated miR-140-null mice, [34]. The reduction in endochondral growth was attributed to the premature chondrocyte differentiation toward post-mitotic chondrocytes and the delayed differentiation of resting to proliferating chondrocytes. By comparing profiles of Argonaute2-associated RNA between control and miR-140-null mice, Dnpep that encodes for an aspartyl aminopeptidase was found a direct target of miR-140. In addition to Dnpep, multiple possible miR-140 target genes, including Pdgfra, HDAC4, Smad3, and Rala, have been identified in various experimental systems [32, 36, 38, 39]; however the role of possible regulation of these genes by miR-140 in cartilage and bone in vivo is unclear.

In a recent in vivo study, we have provided evidence that let-7 miRNAs regulate chondrocyte proliferation in the growth plate [40]. Suppression of let-7 miRNAs through overexpression of the let-7 inhibitor, Lin28a, in mouse cartilage caused mild growth impairment because of reduced chondrocyte proliferation. Lin28a-overexpressing chondrocytes showed upregulation of predicted let-7 target genes including the cycle regulators, cell division cycle 34 (Cdc34), and E2F transcription factor 5 (E2f5). Although the phenotype of Lin28 transgenic mice was mild, when the let-7 suppression, via Lin28a overexpression, was combined with miR-140 deficiency, the mice showed a dramatic growth defect. This finding suggests that let-7 miRNAs and miR-140 coordinately regulate skeletal development by regulating chondrocyte proliferation and differentiation, respectively.

In vitro studies

H19 is a non-protein-coding, maternally expressed gene, abundant in embryonic tissues, and known to play a key role in imprinting of the Igf2 gene. In addition to the role of imprinting, H19 also produces miR-675 from its transcripts. MiR-675 is highly expressed in murine and human articular cartilage. Its expression levels in chondrocytes correlate with the differentiation state, and Sox9 and Col2a1 expression. MiR-675 positively regulates Col2a1 expression in human articular chondrocytes [41]. MiR-1247 is another highly expressed miRNA in mouse cartilage. Mir-1247 targets Sox9 whereas Sox9 negatively regulates miR-1247 expression. This reciprocal regulation appears to create a negative feedback loop [42].

Yang et al. [43] found that miR-145 was downregulated during chondrocyte differentiation of murine embryonic mesenchymal cells (C3H10T1/2 cell line). MiR-145 targeted and suppressed expression of Sox9. Overexpression of miR-145 decreased mRNA levels of chondrocyte marker genes, including Col2a1, Acan (encoding aggrecan), Col9a1, Col11a1, and Comp (encoding cartilage oligomeric matrix protein). On the contrary, miR-145 inhibition increased their expression. Thus, miR-145 negatively regulates chondrocyte differentiation, and its downregulation appears to be an important regulatory mechanism for chondrocyte differentiation.

Ohgawara et al. [44] focused on miRNAs that regulated CCN family protein 2/connective tissue growth factor (Ccn2/Ctgf), a molecule that plays an important role in endochondral bone formation. Initially they performed computational analysis and found binding sites for five miRNAs in the 3′-untranslated region of Ccn2 (miR-26, miR-199, miR-375, miR-19, and miR-18). Then, they compared miRNA expression between non-chondrocytic cells (HeLa cells) and chondrocytic cells (HCS-2/8), and found that miR-18a was most differentially expressed in the chondrocytic cells. Subsequently they performed a luciferase assay and showed that miR-18a suppressed Ccn2 expression. Thus, suppression of miR-18a appears to be necessary for Ccn2 expression upon chondrocyte differentiation.

Sumiyoshi et al. [45] found that miR-1 was repressed upon hypertrophic differentiation of chondrocytes. MiR-1 overexpression in a human chondrocytic cell line (HCS-2/8) and chicken primary chondrocytes reduced aggrecan expression, demonstrating the inhibitory role of miR-1 in chondrogenesis.

Zhong et al. [46] compared miRNA expression profiles in rat articular cartilage at post-natal day 0, 21 and 42, and found that miR-146a and miR-195 were upregulated, whereas miR-337 was downregulated as rats grew. MiR-337 showed the greatest change; it was expressed at a high level at birth but quickly downregulated to an undetectable level as rats reached adulthood. Among the potential miR-337 targets predicted by the computational program, TargetScanS, Tgfbr2, encoding type II TGF-β receptor, was confirmed as a target by in vitro reporter experiments. Overexpression of miR-337 in a human chondrocytic cell line, caused downregulation of Tgfbr2 protein, and inhibited chondrocyte differentiation.

Chicken chondrocytes abundantly express miR-181a. Mir-181a represses expression of Ccna2 (encoding for cyclin A2) and Acan. The regulation of these chondrocyte genes by miR-181a may act as a negative feedback for cartilage homeostasis [47]. Another study in chicken cells found that miR-221 was increased upon inhibition of chondrocyte differentiation [48]. Blockade of miR-221 increased proliferation of limb mesenchymal cells. MiR-221 appeared to regulate chondrocyte proliferation by targeting Mdm2 that encodes an inhibitor of the tumor suppressor, p53.

The expression level of miR-375 in chicken limb mesenchymal cells is decreased upon differentiation into chondrocytes. MiR-375 inhibition increased chondrogenesis in cultured chicken limb mesenchymal cells. In wound-healing assay, miR-375 inhibition facilitated formation of cellular condensation by increasing cell migration, and subsequently stimulated chondrocyte differentiation. This study demonstrated that cadherin-7 was a direct miR-375 target that mediated these actions of miR-375 [49].

MiRNAs in the chondrocyte differentiation of stem cells

Using large scale miRNA screening, Guérit et al. [50] identified miR-574-3p as one of the four most upregulated miRNAs during chondrogenesis in MSCs. MiR-574-3p exhibited a similar expression pattern to that of miR-140. MiR-574-3p expression increased at early stages of chondrogenesis and was maintained at an elevated level throughout differentiation. Retinoid X receptor alpha (Rxra), whose expression progressively decreased during the differentiation process, was validated as a direct target of miR-574-3p. It was also shown that Sox9 binds to the promoter region of the miR-574 gene and positively regulates its transcription during chondrogenesis. Since Rxra was previously demonstrated to inhibit Sox9 activity, miR-574-3p may be a constituent of a positive feedback loop that regulates chondrogenesis.

In mouse MSCs, miR-24 and miR-199b were upregulated by more than fivefold upon differentiation into chondrocytes and osteoblasts [30]. MiR-101, miR-124a, and miR-199a were also upregulated, while miR-18 and miR -96 were downregulated. In human MSCs, miR-199a was upregulated by 4 fold after chondrocyte differentiation [31]. These miRNAs may regulate lineage determination during MSC differentiation.

MiR-449 negatively regulates chondrocyte differentiation of MSCs [51]. In human bone marrow MSCs and in human chondrosarcoma cell lines, miR-449 suppresses Lef1, a critical component of the Wnt signaling pathway. This suppression leads to downregulation of chondrocyte phenotypes, including Col2a1 and Sox9 expression, and reduced proteoglycan production.

MiR-335, discussed earlier about its role in osteogenesis, was also studied in chondrocytes [52]. Upon chondrocyte differentiation of murine MSCs, the abundance of miR-335 increased significantly. MiR-335 overexpression in murine MSCs promoted chondrogenesis. Rock1 and Daam1, whose products were reported to negatively regulate Sox9, were found to be miR-335 targets. MiR-335 is co-expressed with its host gene, Mest. Sox9 downregulates miR-29a and miR-29b that suppress Mest expression. Thus, these miRNAs and their target genes may form a positive feedback loop; miR-335 decreases Rock1 and Daam1 to increase Sox9, which in turn increases Mest and miR-335 transcription by suppressing miR-29a and miR-29b.

The role of miRNAs in osteoblasts

In vivo studies

The role of miRNAs in bone was first revealed when Dicer was deleted in an osteoblast-specific manner. Deletion of Dicer in bone marrow stromal cells (BMSCs) caused defective osteoblast differentiation without significant effects on the cell growth or survival. During the growth phase (up to day 19 of culture), Runx2 and Col1a1 were expressed normally but then failed to increase further [53]. Mature osteoblast markers also failed to express. This finding suggests that miRNAs are necessary for normal differentiation of osteoprogenitors into mature osteoblasts. This notion is consistent with an in vivo study in which Dicer was deleted in osteoblast progenitors using Osx-Cre transgenic mice. These mice showed reduced mature osteoblasts expressing osteocalcin [54].

When Dicer was deleted in mature osteoblasts and osteocytes using a transgenic expressing Cre recombinase under the control of an osteocalcin (bglap) promoter, mice showed no obvious skeletal defects at birth except a mild delay in bone formation of some bones [53]. However, at 1 month of age the Dicer-deficient bone showed a greater bone mass with increased Col1a1, Sp7 (Osterix, Osx), and bglap expression compared with control. There were no overt changes in osteoclasts. These findings suggest that miRNAs are required for normal differentiation of pre-osteoblasts into mature osteoblasts, whereas they suppress functions of mature osteoblasts.

Li et al. [55] found that miR-2861 was expressed primarily in mouse osteoblasts. Using a stromal cell line and bone marrow stromal cells, they showed that miR-2861 promoted osteoblast differentiation induced by BMP-2 stimulation, whereas suppression of this miRNA inhibited it. Histone deacetylase 5 (HDAC5) was identified as a miR-2861 target. In addition to inhibiting MEF2 transcription factors through direct physical binding, HDAC5 also deacetylates lysine residues of various proteins, including Runx2, to regulate protein functions. It has been reported that deacetylation of Runx2 facilitates its ubiquitination and degradation [56]. It appears that miR-2861 reduces HDAC5 expression, and thereby increases Runx2 levels to promote osteoblast differentiation. In the same study, in vivo treatment of mice with antisense miR-2861 caused reductions in bone mineral density (BMD) and histological parameters of bone formation. They also found a mutation in the miR-2861 gene in patients with adolescent osteoporosis. This study suggests that a single miRNA can significantly influence bone mass. However, since miR-2861 is located in a promoter region of Cdk9, a gene encodes a cell cycle regulator; further investigation is required to determine the causal role of the miR-2861 dysfunction in adolescent osteoporosis. The primary transcript encoding miR-2861 also generates another miRNA, miR-3960. MiR-3960 also induces osteoblast differentiation, whereas its inhibition reduces osteogenesis [57]. MiR-3960 appears to target and suppress Hoxa2 which inhibits Runx2 expression. Another interesting finding in this study is that Runx2 binds to the common promoter of the miR-3960/miR-2861 gene and increases its expression. These regulations may create a positive feedback loop that maintains Runx2 levels upon osteoblast differentiation.

In a study that aimed at identifying specific miRNAs associated with osteoporosis, miR-214 was found to inhibit bone formation [58]. MiR-214 was expressed at higher levels in elderly individuals with history of bone fractures than those without fractures. The miR-214 level was inversely correlated with bone formation markers. The inhibitory effect of miR-214 in bone formation was confirmed in mouse models and by in vitro experiments. They found that miR-214 suppressed activating transcription factor 4 (Atf4), a transcription factor that promotes the expression of osteoblast-specific genes such as osteocalcin [59]. Transgenic mice overexpressing miR-214 in osteocalcin-expressing cells (mature osteoblasts), showed reduced bone formation, which was rescued by miR-214 inhibition. This miRNA may play a significant role in post-menopausal osteoporosis because its inhibition in ovariectomized mice rescued their osteopenic phenotype. In an in vitro study, miR-214 was found to also target Osx. MiR-214 overexpression in osteoblast cell lines (Saos-2 and U2OS cells) reduced Osx expression, whereas osteoblastic trans-differentiation of C2C12 cells was enhanced upon miR-214 inhibition [60].

MiR-206 is one of the muscle-specific miRNAs. In a study by Inose et al. [61], ectopic expression of miR-206 in mouse osteoblasts in vivo reduced bone mass without any effect in bone resorption. MiR-206 appears to inhibit osteoblast differentiation by downregulating connexin-43 (Gja1), a gap junction protein that plays an important role in osteoblast and osteocyte functions. Knockdown of miR-206 promoted osteoblast differentiation. The biological significance of this finding is not clear because endogenous miR-206 expression in osteoblasts is low.

Another miRNA that was studied in vivo is miR-34c. MiR-34 family miRNAs are direct transcriptional targets of p53 and mediate part of p53′s function [62]. MiR-34c was upregulated during osteoblast differentiation of myoblastic C2C12 cells and BMSCs. Transgenic mice overexpressing miR-34c in osteoblasts displayed age-related osteoporosis. This effect is due to reduced osteoblast proliferation and differentiation as well as increased bone resorption. Transcriptome profiling and reporter assays identified several molecules involved in the Notch signaling pathway as miR-34c targets [63]. MiR-34c likely regulates osteoclastogenesis indirectly via regulation of Notch signaling in osteoblasts. Mir-34c overexpression also caused downregulation of Runx2 and Satb2 (special AT-rich sequence binding protein 2) that facilitate osteoblast differentiation [63].

MiR-182, expressed in osteoblasts, negatively regulates osteoblast proliferation and differentiation [64]. MiR-182 appears to suppress the expression of Foxo1, a transcription factor that regulates the reduction and oxidation balance in osteoblasts. Foxo1 regulates bone mass, as its depletion results in decreased osteoblast proliferation and bone formation [65, 66]. Furthermore, Foxo1 overexpression in vitro rescues the antiproliferative and pro-apoptotic effects of miR-182. In vivo experiments using zebrafish confirmed the inhibitory effect of miR-182 on bone formation [64].

The miR-17-92 cluster is located on chromosome 13 in humans and encodes six individual miRNAs (miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92a). This cluster has two paralogs: the miR-106b-25 and the miR-106a-363 clusters. The miR-106b-25 cluster is located on human chromosome 7 and mouse chromosome 5 and comprises three miRNAs: miR-106b, miR-93, and miR-25, while the miR-106a-363 is located on chromosome X and encodes six miRNAs: miR-106a, miR-18b, miR-20b, miR-19b-2, miR-92a-2, and miR-363. The 15 miRNAs encoded by the miR-17-92 and its two paralogous genes form four functionally distinct “seed” family groups (miR-17, miR-18, miR-19, and miR-92) [67]. The miR-17-92 cluster miRNA gene is the first miRNA gene whose mutations were found to cause skeletal developmental defects in humans. Deletion of these genes causes a brachy-syndactyly syndrome (Feingold syndrome type II) [68, 69]. On the other hand, duplication of this gene causes variable skeletal phenotypes; one family shows skeletal overgrowth with polydactyly [70], whereas the other shows short stature and brachydactyly [71]. Mice hemizygous or homozygous for miR17-92 deletion showed reduced skeletal growth with delayed ossification of digital and skull bones [68]. The miR-17-92 cluster miRNAs are expressed abundantly in osteoprogenitor cells and embryonic stem (ES) cells, but decrease upon osteoblast differentiation [72]. This study also found that haploinsufficiency of the miR-17-92 cluster gene in mice resulted in reduced bone mass likely due to impaired osteoblast proliferation and differentiation.

In vitro studies

In addition to these in vivo studies discussed so far, there are a significant number of in vitro studies that have investigated roles of specific miRNAs in osteoblasts.

As suggested by in vivo models, miR-17-92 miRNAs appear to regulate both proliferation and differentiation of osteoprogenitors. A study using human periodontal ligament tissue-derived mesenchymal stem cells (PDLSCs) found that suppression of miR-17 promoted osteoblast differentiation, whereas its overexpression inhibited it [73]. Another recent study showed that expression levels of miRNA-17 family miRNAs, including miR-17, -106a, and -20a, were decreased in human adipose-derived mesenchymal stem cells (hADSCs) during differentiation toward the osteogenic lineage, whereas they were increased during adipocyte differentiation [74]. Overexpression of miR-17 and miR-106a promoted adipocyte differentiation and inhibited osteogenesis by suppressing Bmp2 expression. Another study using mouse palatal mesenchymal cells showed that miR-17-92 miRNAs targeted multiple molecules involved in the TGF-β pathway to inhibit TGF-β signaling to promote proliferation [75].

MiR-17-92 cluster miRNAs may inhibit apoptosis in osteoblasts by targeting the pro-apoptotic protein, Bim (bcl-2 interacting mediator of cell death) [76]. This study found that miR-17, miR-20a, and miR-92a were significantly reduced in osteoblasts undergoing apoptosis. Overexpression of miR-17-92a miRNAs reduced dexamethasone-induced apoptosis and enhanced the anti-apoptotic effect of estrogen in osteoblasts.

MiR-34a was found to be one of the most significantly upregulated miRNAs upon osteoblast differentiation in human MSCs, whereas miR-34b exhibited modest upregulation and miR-34c, which was discussed earlier, was not detectable [77]. MiR-34a overexpression in vitro seemed to regulate the expression of Jag1 that encodes a Notch ligand. In addition, miR-34a regulated osteoblast proliferation, likely through targeting the cell cycle regulators, Ccnd1 (cyclin D1), Cdk4 (cyclin-dependent kinase 4), Cdk6 (cyclin-dependent kinase 6), E2f3 (E2F transcription factor 3), and Cdc25a (cell division cycle 25A). Pre-miR-34a transfection inhibited heterotopic bone formation of subcutaneously implanted MSCs in immune deficient mice, whereas anti-miR-34a stimulated it [77]. In another study conducted in C2C12 cells, miR-34b as well as miR-34c were upregulated upon activation of Wnt signaling and upon osteoblast differentiation in a mouse osteoblast cell line [78].

MiR-29 regulates osteoblast differentiation by directly and indirectly modulating Wnt signaling [79, 80]. Three miR-29 family miRNAs are transcribed from two separate genomic loci. Mir-29c and miR-29b-2 are generated from the same primary miRNA that is transcribed from a locus located on the chromosome 1; miR-29a and miR-29b1 are produced from another primary miRNA encoded on chromosome 7 in humans [81]. MiR-29 binding sites are found in the RNA encoding osteonectin, the most abundant non-collagen extracellular matrix protein in bone. In primary cultures of mouse osteoblast cells, the osteonectin level decreases upon cell maturation and mineralization. This decrease coincides with an increase in levels of miR-29a and miR-29c, suggesting that these miRNAs might negatively regulate osteonectin expression. Reporter experiments confirmed their interaction. Inhibition of miR-29a/c using antagomirs caused an increase in osteonectin levels. Based on these findings, a model was proposed that during osteoblast differentiation Wnt signaling induces osteoblast differentiation and upregulates miR-29a/c, which in turn, reduces osteonectin protein levels at the mineralization stage [79]. MiR-29a stimulates osteoblast differentiation in primary human osteoblast cultures and suppresses expression of multiple Wnt signaling inhibitors, such as Dkk1 (Dickkopf-related protein 1), to enhance Wnt signaling. This may create a positive feedback loop that facilitates osteoblast differentiation [80]. MiR-29a expression is suppressed by glucocorticoid treatment. Rats treated with glucocorticoids for 2–4 weeks showed reduced BMD and decreased miR-29a expression [82]. In these mice, expression of Col1a1 and osteocalcin was decreased, whereas Dkk1 expression was increased. Overexpression of miR-29a in vivo, via a lentivirus-based gene transfer system, significantly ameliorated these changes caused by glucocorticoid treatment. Furthermore, when rats were treated with a miR-29a inhibitor, they displayed significant reductions in bone mass, mineral apposition, and histomorphometric parameters for osteoblast functions. MiR-29a inhibitor-treated osteoblasts also showed a differentiation defect ex vivo. Additionally miR-29a reduced osteoclast activity possibly by downregulating RANKL expression. This study suggests that miR-29a could be exploited to reduce the negative effects of glucocorticoids on bone. Another in vitro study confirmed the role of miR-29 miRNAs in osteoblast differentiation [83]. This study found that miR-29b targeted multiple negative regulators of osteoblast differentiation, such as HDAC4, Tgfb3 (TGF-β3), Acvr2a (activin receptor IIA), Ctnnb1 (β-catenin), and Dusp2 (dual specificity phosphatase 2). It was also suggested that miR-29b influenced extracellular matrix production by targeting collagen genes, including Col1a1, and might prevent fibrosis of bone tissue.

Li et al. [84] reported that miR-143 suppressed osteogenic differentiation by targeting Osx in osteoblastic MC3T3-E1 cells. MiR-143-mimic transfection upregulated Osx, alkaline phosphatase (ALP) and osteocalcin levels, whereas miR-143 inhibitor reduced their expression. Furthermore, Osx overexpression reversed the negative effect of miR-143 on osteogenic differentiation.

Bhushan et al. [85] provided evidence that miR-181 miRNAs (miR-181a, b, c, and d) promote osteoblast differentiation by downregulating TGF-β signaling. They showed that miR-181a was upregulated upon osteoblast differentiation in C2C12 cells, MC3T3-E1 cells, and primary calvarial osteoblasts. Expression of other miR-181 miRNAs also increased during mouse calvarial and tibial development. Analysis of the mRNA expression profile in MC3T3-E1 cells transfected with miR-181a revealed that Tgfbr1 (TGF-β receptor I) and Tgfbi (TGF-β, induced) that encode molecules in the TGF-β signaling pathway that negatively regulates osteoblastic differentiation were significantly downregulated. Luciferase reporter assay confirmed that these genes were targeted by miR-181.

Gámez et al. [86] reported that miR-322 increased expression of osteogenic genes including Osx, Runx2, Msx2, and Ibsp (integrin binding sialoprotein) in C2C12 cells in response to BMP-2. MiR-322 appears to increase Osx expression by targeting of Tob2. Tob2 facilitates deadenylation and degradation of mRNAs. Tob2 overexpression reduced Osx mRNA, which was inhibited by overexpression of miR-322.

MiR-335-5p was found to be expressed in osteoblasts and hypertrophic chondrocytes in mouse embryos. MiR-335 acts as a suppressor of Dkk1, to enhance Wnt signaling and promote osteogenesis [87].

In expression profiling of miRNAs in MC3T3-E1 cells, miR-141 and miR-200a were found to be downregulated by BMP-2 treatment [88]. Their overexpression in this cell line reduced osteogenesis, as shown by decreased ALP activity. Conversely, inhibition of these miRNAs stimulated osteoblast differentiation. These miRNAs appear to target the osteogenic transcription factor, Dlx5, to inhibit osteoblast differentiation.

MiR-23a, miR-27a, and miR-24-2, transcribed from a miRNA cluster gene, were shown to suppress the osteogenic gene, Satb2, in primary rat osteoblasts [89]. Satb2 facilitates bone formation by enhancing the function of Runx2. Transcription of these miRNAs is negatively regulated by Runx2 through a Runx binding element in the promoter region. The Runx2-mediated suppression of these miRNAs is expected to increase Satb2 expression and enhance osteoblast phenotypes during differentiation. On the other hand, in terminally differentiated osteocytes, miR-23a suppressed Runx2, creating a feedback mechanism to control osteoblast maturation. In another study, miR-23a was overexpressed in a murine pre-osteoblast cell line [90]. MiR-23a overexpression reduced tumor necrosis factor-α (TNF-α)-induced cell apoptosis by inhibiting expression of the apoptosis-inducing gene, Fas. Since TNF-α is one of the inflammatory cytokines that are elevated in estrogen deficiency, miR-23a expression in osteoblasts may be an endogenous counter-mechanism against post-menopausal osteoporosis.

The search for miRNAs targeting Runx2 led to the identification of miR-23a, miR-30c, miR-34c, miR-133a, miR135a, miR137, miR204, miR-205, miR-217, miR218, and miR-338. These miRNAs downregulated Runx2 to inhibit osteoblast and chondrocyte differentiation [91].

Lei et al. [92] investigated the association between genetic polymorphisms in predicted miRNA binding sites in the 3′-UTR of mRNAs and osteoporosis. This study identified three polymorphisms in miRNA target sites of the Fgf2 gene that were significantly associated with femoral neck BMD. These sites were targeted by 9 miRNAs, miR-146a/b, miR-549, miR-563, miR-212, miR-132, miR-505, miR-421, miR-361, miR-582, and miR-302b. Fgf2 expression was decreased in patients of the high BMD group compared with those in the low BMD group.

Lisse et al. [93] investigated regulation of osteoblastic miRNAs by vitamin D. MiR-637 and miR-1228 were upregulated upon vitamin D treatment and were shown to target Col4a1 and Bmp2k (bone morphogenetic protein 2 kinase). MiR-637 was also shown to target Osx to suppress osteoblast differentiation in early differentiation stages [94].

Chen et al. [95] compared miRNAs expressed in calvariae of E18.5 Osx-null embryos with those in wild-type embryos to identify Osx-regulated miRNAs. Ten miRNAs, including miR-133a, and its two paralogous miRNAs, miR-204 and miR-211, were downregulated in the absence of Osx. These miRNAs target Runx2 as discussed earlier. Additionally, miR-204/211 targets Sost that encodes the Wnt antagonist, Sclerostin. Thirty miRNAs were upregulated in Osx deficiency, including miR-141 and miR-220a that target Dlx5.

An et al. [96] performed miRNA and mRNA profiling in bones of ovariectomized mice. MiRNAs -127, -133a, -133*a, -133b, -136, -206, -378, and -378* were upregulated after ovariectomy, whereas miR-204 was downregulated. This study also demonstrated that miR-127 suppressed and miR-136 enhanced osteoblast differentiation, whereas both miRNAs promoted cell death in osteocytic cells.

A recent study aimed to identify specific miRNAs that can be used as biomarkers for osteoporosis [97]. MiR-21, miR-23a, miR-24, miR-100, and miR-125b were upregulated both in the serum and in bone tissues of patients with osteoporosis. This finding is in line with the previous finding that miR-23a, miR-24, and miR-125b suppress osteogenesis in vitro, as discussed earlier. This study presents the possibility of using miRNAs as diagnostic markers in bone diseases.

MiRNAs in mesenchymal stem cells

Mesenchymal stem cells can differentiate into multiple cell lineages including osteoblasts, chondrocytes and adipocytes. Manipulation of MSCs to generate desired types of cells has been intensively studied in fields of tissue engineering and regenerative medicine in the hope of developing cell-based therapies. MiRNAs have drawn attention as a tool to manipulate MSCs to expand and differentiate them into bone cells.

Several miRNAs change their expression during osteoblast differentiation of MSCs. Laine et al. [31] found that MiR-96 and miR-199a were increased by 16- and 2-fold, respectively, in human MSCs during osteoblast differentiation; their exact roles in osteogenesis is currently unclear. MiR-346 is also increased during osteoblast differentiation of MSCs. Mir-346 overexpression promoted osteogenesis via targeting RNA encoding glucogen synthase kinase-3β (Gsk3b) that inhibits canonical Wnt signaling by facilitating β-catenin degradation [98]. Mir-31 was found to be upregulated in MSCs differentiating into mineralizing osteoblasts [99]. In this study, miR-31 was shown to negatively regulate Osx. However, this result is somewhat contradicting to that of a previous study in which miR-31 was downregulated upon hMSC differentiation toward osteoblasts and suppressed osteogenesis by targeting Runx2 and Bmpr2 [100]. Another study also reported downregulation of miR-31 during osteoblast differentiation [101]. MiR-31 inhibition enhanced osteoblast differentiation by de-repressing Satb2 that cooperated with Runx2. It is possible that miR-31 mainly suppresses osteoblast differentiation in early stages, but promotes it during late stages.

Eskildsen et al. [102] found that miR-138 was downregulated upon osteoblast differentiation. MiR-138 overexpression and knockdown in hMSCs decreased and increased bone formation, respectively, after subcutaneous implantation of hMSCs into immuno-compromised mice. MiR-138 was found to target Ptk2, a gene encoding focal adhesion kinase (FAK) that activates the extracellular signal-regulated kinase (ERK) signaling pathway. MiR-138 reduced ERK signaling, inhibited Runx2 phosphorylation, and reduced Osx expression. Zhang et al. [94] showed that miR-637 expression in hMSCs decreased upon osteoblast differentiation and increased during adipogenic differentiation. Since MiR-637 targeted Osx, miR-637 might regulate lineage switching between osteoblasts and adipocytes. In addition, miR-204 and miR-211 appear to be important negative regulators of Runx2. These miRNAs both inhibited osteogenesis, whereas miR-204 promoted adipogenesis in MSCs [103].

MiR-370 was found to be downregulated upon osteoblast differentiation of mesenchymal cells (C3H10T1/2 cells) treated with BMP-2 [104]. MiR-370 overexpression reduced ALP activity and resulted in a mild decrease in mineralization, demonstrating that miR-370 inhibits osteoblast differentiation. Bmp2 and Ets1, a proto-oncogene, were predicted as potential targets of miR-370. MiR-370 overexpression reduced expression of these genes as well as the osteoblast genes including Spp1 that encodes osteopontin and Runx2.

The possible shift of MSC differentiation tendency from the osteoblast to adipocyte lineage has been implicated in osteoporosis. MiRNA profiling in MSCs from ovariectomized mice demonstrated a significant upregulation of miR-705 and miR-3077-5p [105]. After estrogen supplementation, expression of these miRNAs returned to normal levels. In vitro experiments also found that expression levels of these miRNAs and osteoblastic phenotypes were inversely correlated. Furthermore, overexpression of these miRNAs increased adipogenic differentiation, whereas their inhibition decreased it. In this study, it was shown that miR-705 and miR-3077-5p targeted Hoxa10 and Runx2, respectively.

Gong et al. [106] attempted to identify miRNAs regulated by the osteo-promotive gene, Satb2, in murine MSCs. Among ten downregulated miRNAs after Satb2 overexpression, miR-27a was selected for further analysis. MiR-27a was found to target Bmp2, Bmpr1a, and Smad9, and miR-27a overexpression inhibited osteoblast differentiation.

MiR-210 is highly expressed in ST2 mouse mesenchymal cells. Mizuno et al. [107] found that miR-210 was upregulated upon BMP-4-induced osteoblast differentiation. Overexpression of miR-210 in these cells promoted osteoblast differentiation, as assessed by ALP activity and osteocalcin expression. This is likely through targeting activin A receptor type 1B (Acvr1b) that mediates TGF-β/activin signaling. Another miRNA found to be upregulated in ST2 cells during osteoblast differentiation was miR-125b [108]. MiR-125b inhibition promoted osteoblast differentiation. MiR-125b might target ErbB2, which encodes for a protein tyrosine kinase that regulates proliferation of osteoblasts and chondrocytes [109].

MiR-124a and miR-181a were among six downregulated miRNAs upon BMP-4-induced osteoblast differentiation of mouse-induced pluripotent stem (iPS) cells [110] Their downregulation appears to facilitate osteoblast differentiation; suppression of these six miRNAs induced expression of osteoblastic marker genes, such as Runx2 and Spp1, in iPS cells. These miRNAs appear to suppress osteoblast differentiation at early differentiation stages.

MiRNA expression profiling in unrestricted somatic stem cells (USSCs), derived from the human umbilical cord blood, identified miR-135b as the most significantly downregulated miRNA upon osteogenic differentiation [111]. Overexpression of miR-135b reduced expression of bone markers including Ibsp and Osx. Another study using USSCs to define a miRNA signature associated with osteoblast-lineage commitment identified several miRNAs; miR-146a, miR-181a, miR-199b-5p, and miR-30b were upregulated, and miR-1274a, miR-139-3p, miR-221, miR-376a, and miR-376c were downregulated upon osteoblast differentiation [112]. Further study revealed that inhibition of miR-1274a, miR-221, and miR-376a increased osteoblast differentiation.

Human adipose tissue-derived stem cells (hADSCs), isolated from adipose tissues, are capable of differentiating into osteoblastic cells. In hADSCs, miR-26a was found to target Smad1 transcription factor and inhibit osteoblast differentiation [113]. MiR-196a enhanced osteoblast differentiation of hADSCs by targeting Hoxc8 [114]. MiR-218 decreased expression of the Wnt antagonists, Dkk2 and Sfrp2, and thereby promoted osteogenesis in hADSCs. Additionally, Wnt signaling enhanced miR-218 expression, creating a signal-amplification circuit [115]. Like in MSCs discussed earlier, miR-31 knockdown in rat ADSCs enhanced osteogenesis by targeting Satb2. MiR-31 knockdown also promoted repair of “critical-sized calvarial defects” after engraftment of cells treated with a miR-31 inhibitor and synthetic scaffolds [116].

Mesenchymal stem cells from patients with post-menopausal osteoporosis showed decreased osteogenic potential ex vivo [117]. Deregulation of miRNAs in estrogen deficiency may account for the reduced osteogenic potential. Yang et al. [118] found that miR-21 was downregulated in MSCs isolated from ovarectomized mice. MiR-21 was suppressed by TNF-α, an inflammatory cytokine that plays a pathogenic role in post-menopausal osteoporosis. Furthermore, miR-21 overexpression increased the osteoblast differentiation of MSCs. Treatment of ovariectomized mice with anti-TNF-α increased miR-21 expression and bone formation. These data suggest that miR-21 suppression in estrogen deficiency may decrease osteoblast differentiation of MSCs and contribute to the bone loss. Spry1, a gene encoding for a negative regulator of the FGF and ERK-MAPK signaling pathways, was identified as a miR-21 target that mediated the above effects; overexpression of Spry-1 in hMSCs from healthy donors, reduced osteoblast differentiation.

The role of miRNAs in osteoclasts

The significance of miRNAs in osteoclasts was first shown by in vitro experiments in which miRNAs were globally reduced in osteoclast precursors by deleting Dgcr8, Dicer, or Ago2, genes individually necessary for miRNA biogenesis or function. The global reduction in miRNAs inhibited osteoclast differentiation and function. Reduced levels of transcription factors necessary for the osteoclastogenesis, such as PU.1, Mitf, Fos, and Nfatc1, were observed in miRNA-deficient osteoclast precursors [119]. In this study, Dicer deletion in macrophages and osteoclasts in vivo using the CD11b-Cre transgenic mice decreased osteoclast number and increased trabecular bone mass. MiR-223 expressed in cells of the osteoclast lineage targeted NF1A, a negative regulator of macrophage colony-stimulating factor receptor (Csf1r), and facilitated osteoclast differentiation. Osteoclast-specific deletion of the Dicer gene using Cathepsin-Cre transgenic mice confirmed that miRNA deficiency in osteoclasts causes a mild increase in bone mass with a decrease in osteoclast number [120]. Additionally, these mice also showed decreased osteoblast activity and reductions in mineral apposition and bone formation rates. In a recent study, miRNAs were reduced in osteoclasts by conditionally deleting the Dgcr8 gene, which encodes a molecule essential for processing pri-miRNAs into pre-miRNAs, using cathepsin-Cre transgenic mice [121]. Mice missing Dgcr8 in osteoclasts showed increased trabecular and cortical bone with impaired osteoclastogenesis confirming the role of miRNAs in osteoclast differentiation and function.

Chen et al. [122] showed a reduced level of miR-503 in the circulating osteoclast precursor cells, CD14+ peripheral blood mononuclear cells (PBMCs), in patients with post-menopausal osteoporosis. Overexpression of miR-503 decreased TRAP and Nfatc1 expression, and inhibited osteoclastogenesis in CD14+ PBMCs obtained from post-menopausal women. MiR-503 was shown to directly target mRNA encoding receptor activator of NFκB (RANK). Mice treated with miR-503 or antagomir-503 showed an increase or decrease in BMD, respectively. In addition, it was shown that 17β-estradiol (E2) increased miR-503 expression, further supporting the notion that the reduction in miR-503 caused by estrogen deficiency contributes to development of post-menopausal osteoporosis.

A study, in which human CD14+ PBMCs was used to investigate changes in miRNA expression during osteoclastogenesis, found miR-148a as the most dramatically upregulated miRNA [123]. Overexpression of miR-148a in human PBMCs or mouse bone marrow cells increased osteoclast differentiation. V-maf musculoaponeurotic fibrosarcoma oncogene homolog B (Mafb), a negative regulator of osteoclastogenesis, was identified as a target of miR-148a. Ovariectomized mice as well as control mice treated with antagomir-148 showed a significant increase in BMD, accompanied by reduced osteoclast activity and bone resorption. In this study, miR-148 was also found to be increased in patients with systemic lupus erythymatosus (SLE) who had never received corticosteroid therapy. The miR-148 level was correlated with increased osteoclastogenesis, suggesting that miR-148 may play a role in the increased risk for bone loss in these patients.

A miRNA that was significantly downregulated during osteoclastogenesis in human PBMCs was miR-125a [124]. MiR-125a overexpression inhibited osteoclastogenesis, while its inhibition promoted it by upregulating the direct target, TNF receptor-associated factor 6 (Traf6), an osteoclastogenesis-promoting factor. In addition, the osteoclast-promotive transcription factor, Nfatc1, was found to bind to the promoter of the miR-125a gene to inhibit its transcription, potentially creating a positive feedback loop.

MiRNA profiling in bone marrow-derived monocyte/macrophage precursors after RANKL stimulation showed downregulation of 33 miRNAs and upregulation of 38 miRNAs upon osteoclast differentiation [125]. MiR-21 was strongly upregulated by RANKL treatment, and also by transcription factors that regulate differentiation and function of osteoclasts, including Fos. Inhibition of miR-21 reduced osteoclastogenesis in vitro with concomitant upregulation of its target gene Pdcd4 that encodes a negative regulator of Fos. MiR-21 appears to suppress Pdcd4 to augment Fos function, and thereby, facilitate osteoclastogenesis. Another miRNA that was upregulated during osteoclast differentiation of mouse bone marrow cells was miR-31 [126]. Inhibition of miR-31 in mouse bone marrow macrophages using antagomirs compromised the cytoskeleton function and suppressed osteoclast formation. Upregulation of the GTPase, RhoA, a target of miR-31 appears to be responsible for this effect because treatment with a RhoA inhibitor rescued the defect in osteoclastogenesis of miR-31 antagomir-treated cells.

Osteoclast differentiation is inhibited by interferon-β (IFN-β). MiR-155 appears to partly mediate this inhibitory effect of IFN-β on osteoclastogenesis [127]. MiR-155 was strongly induced by IFN-β in bone marrow-derived macrophages in culture. Overexpression of miR-155 suppressed osteoclast differentiation by targeting Socs1 and Mitf, positive regulators of osteoclastogenesis.

The role of miR-29 was studied in human and mouse cells by different groups, and somewhat contradicting results were obtained. In human cells, miR-29b, was found to be decreased during osteoclast differentiation in vitro [128]. MiR-29b overexpression in osteoclast precursors reduced the expression of Nfatc1, Ctsk (Cathepsin K), Mmp9, TRAP, and RANK. It also reduced bone resorption activity, assessed by the lacunae generation (pit formation) assay, and collagen degradation. This effect appeared to be due to the suppression of Fos and Mmp2 expression by miR-29b. In contrast, in mouse cells, all miR-29 family members (a, b, and c) were increased upon osteoclast differentiation in primary bone marrow-derived macrophages as well as in a monocytic cell line [129]. Inhibition of miR-29 significantly reduced osteoclastogenesis in these cells. This study identified several groups of miR-29 target genes that have coherent functions. Cell division control protein 42 (Cdc42) and SLIT-ROBO Rho GTPase-activating protein 2 (Srgap2) are both important for osteoclast function and migration by regulating the cytoskeleton. G protein-coupled receptor 85 (Gpr85) and CD93 potentially regulate macrophage differentiation. Nf1A inhibits macrophage and osteoclast maturation. The reason of the discrepancy between human and mouse cells is currently unclear. It is possible that miR-29 promotes the initial osteoclast lineage commitment while it inhibits osteoclast maturation, since more differentiated cells (peripheral blood osteoclast precursors) were used in the human study whereas bone marrow progenitors were used in the mouse study.

Conclusions

These studies have demonstrated that miRNAs regulate many aspects of bone development, and play significant roles in bone physiology and pathophysiology. The majority of these findings have been obtained from studies using in vitro experimental systems; toward further understanding of physiological roles of individual miRNAs at tissue and organ levels, and toward the ultimate goal of translating these findings into clinical applications, studies using in vivo models are required.

Abbreviations

BMP

Bone morphogenetic protein

MAPK

Mitogen-activated protein kinase

FGF

Fibroblast growth factor

FGFR

Fibroblast growth factor receptor

CSF

Colony-stimulating factor

RANK

Receptor activator of NFκB

RANKL

Ligand to receptor activator of NFκB

TRAP

Tartrate-resistant acid phosphatase

TGF-β

Transforming growth factor beta

IGF

Insulin-like growth factor

HDAC

Histone deacetylase

MEF

Myocyte enhancer factor

MSC

Mesenchymal stem cells

BMSC

Bone marrow stromal cells

ADSC

Adipose-derived stem cells

BMD

Bone mineral density

ALP

Alkaline phosphatase

TNF-α

Tumor necrosis factor alpha

USSC

Unrestricted somatic stem cells