1. Introduction

Bone resorption process is critical to define bone remodeling and to allow constant renewal of the bone matrix. Osteoclasts are unique cells that are responsible for bone resorption by removing mineralized bone matrix and by breaking up the organic bone. Osteoclasts are differentiated from hematopoietic stem cells like macrophages. However, unlike macrophage, they are maturated to multinucleated cells by fusion of mononuclear precursors. These multinucleation processes are necessary required for osteoclastic maturation to undergo efficient bone resorption as mononucleated osteoclasts are unable to resorb bone matrix. However, it remains still uncovered how to maintain and how to coordinately regulate genes in the multinuclei.

Osteoclastogenesis is highly regulated processes requiring multiple stages and is governed by diverse factors. Various soluble factors including RANKL, M-CSF, and steroid hormones have been identified to support such osteoclast differentiation and activity. Similarly, various transcriptional factors including NFxB and AP1 are shown to involve in gene regulations required for osteoclastogenesis. Among them, Nfatc1 is a prime transcriptional factor for osteoclastogenesis. During differentiation stages, Nfatc1 interacts with epigenetic factors to regulate a particular set of target genes at each stage, indicating that distinct epigenetic complexes contribute to the spatiotemporal regulation of each gene expression by Nfatc1 during osteoclastogenesis. However, epigenetic complexes during osteoclastogenesis remain to be identified.

Histone modifications have emerged to be prerequisite for gene regulations by transcriptional factors. Among them, histone methylation is considered as more upstream epigenetic modification to direct cellular differentiation, proliferation, and so on. Similarly, according to previous report that chromatin morphology within multinucleated osteoclasts exhibits different shape by electron microscope studies, it is most likely that multinuclear gene expression is regulated through epigenetic mechanisms like chromatin remodeling or histone modifications. Nevertheless, there are still little studies to characterize epigenetic status of multinuclei or to identify epigenetic regulators for maintaining multinuclei.

The present study was undertaken A) to uncover the molecular basis to maintain and to regulate multinuclear genes in single osteoclast cell, and B) to identify novel epigenetic regulatory factors supporting Nfatcl function during the osteoclastogenesis.

2. Results

A. Selective gene expression within multinuclei of mature osteoclasts

Although multinucleation of non-osteoclastic cells are known to cause severe disease, multinucleated osteoclasts are functionally normal and the number of nuclei positively correlates with their bone resorption activity. However, it is minimal understanding about intracellular gene regulation of multinuclei. For that reason, we observed the pattern of multinuclear gene expression.

Firstly, we examined whether all nuclei in a multinucleated osteocaist evenly have transcriptional activity, by expression of several nuclear proteins through immunocytochemistry. Although the tested nuclear receptors were distributed in all nuclei, active transcriptional factors such as RNAPII-Ser2P and Nfatc1 showed selective distribution in only certain nuclei. Similarly, when the primary transcripts of osteoclast-specific genes (Nfatc1 , CtsK, and TRAP) and house-keeping gene (βTubulin) were directly detected by RNA FISH, their distribution was limited among multinuclei. These results suggested that all nuclei in a multinucleated osteoclast are not transcriptionally active. We then examined the possibility of their natural diffuse by shuttling among multinuclei by using two cell lines that stably express ZsGreen-NLS and DsRed-NLS proteins individually. When two individual cells were mixed and differentiated to multinucleated osteoclasts, all nuclei were stained by two colors. This result indicated that in osteoclasts, nuclear proteins are transmitted among multinuclei through shared cytoplasm.

Taken together, each nucleus in a multinucleated osteoclast appears to bear selective transcriptional activity.

B. Identification of epiaenetic reaulators to understand molecular mechanism about selective aene expression within multinuclei.

According to our results, Nfatc1 looks important for gene regulation only in certain nuclei. To explore the molecular basis of such Nfatc1 function during osteoclastogenesis, epigenetic regulators supporting Nfatc1 were searched.

a. Characterization of a novel osteoclastogenic repressor, Jmjd5

Firstly, we tried to identify epigenetic regulators supporting Nfatcl function by comparing expression pattern during osteoclastogenesis induced by RANKL treatment. Particularly we focused on JmjC-domain-containing proteins which are known as histone demethylase family, because their physiological roles in bone are totally unknown.

From the expression of ten kinds of JmjC-domain containing proteins, we observed the reduced expression of Jmjd5 during osteoclastogenesis. To verify its functions in the osteoclast differentiation, a stable cell line constitutively expressing shRNAs against Jmjd5 was established, and then osteoclast formation was assayed. Notably, down-regulation of Jmjd5 expression resulted in promotion of osteoclast formation and upregulated expression of osteoclast-specific genes. These results indicated that Jmjd5 is a negative regulator for osteoclast differentiation.

To define a target histone of Jmjd5, various in vitro histone demethylase assays') were accomplished. However, we could never detect altered methylation levels on any tested lysine residue of histone H3 by Jmjd5 expression. Instead, we could detect direct interaction between Jmjd5 and Nfatcl through an in vitro binding assay, raising an idea that Nfatcl is a Jmjd5 substrate. Additionally, when expression of Jmjd5 was down-regulated, the protein level of Nfatcl, but not transcript level, was increased. These results suggested that Jmjd5 regulates the protein stability of Nfatc1.

Jmjd5 is still believed as a histone demethylase due to presence of its JmjC-domain, however the present study has proved that Jmjd5 regulates the protein stability of Nfatc1 by their transient direct interaction.

b. Characterization of a novel co-activator for the Nfatc1, Kiaa1903

Secondly, we tried to biochemically identify novel regulatory components associated with Nfatc1. To purify complexes in similar condition of in vivo status, an endogenous-antibody column against Nfatc1 was established and applied for nuclear extracts of Raw264 cells stimulated by RANKL. Among several candidates, an uncharacterized protein, Kiaa1903, was identified as an associating component for the known chromatin remodeler SWI/SNF complex. This protein appeared a novel component of SWI/SNF chromatin remodeler complex because of its SANT domain which is also found in some factors of chromatin remodeler and its association with components of SWI/SNF complex. Indeed, expression of Kiaa1903 was osteoclast-specific among the tested tissues and cells.

To test a co-regulator role of Kiaa1903 for Nfatc1, a luciferase assay was performed with known co-regulators against Nfatc1. The overexpression of Kiaa1903 enhanced the transcriptional activity of Nfatcl1 Similarly, when expression of Kiaa1903 was knock-down by shRNA, osteoclast formation was decreased with lowered expression of osteoclast-specific genes. From these results, Kiaa1903 looks to serve as a co-activator for Nfatc1 during osteoclast differentiation. To characterize Kiaa1903 as an epigenetic regulator, a ChIP experiment was performed with histone methylation markers. When expression of Kiaa1903 was down-regulated, the levels of histone H3Lys4 trimethylation on promoters of Nfatcl-target genes were decreased. Also, the complexes of an Nfatc1 containing SWI/SNF complex could not be recruited on their target genes in absence of Kiaa1903.

From these findings, we presume that Kiaa1903 is a novel component of SWI/SNF-type chromatin remodeling complex during osteoclast differentiation.

3. Conclusion

So far, function of various regulators involved in osteoclastogenesis has been characterized by observation of bone phenotype about their knockout mice. However, it is still insufficient to understand osteoclastic differentiation and function at molecular levels. In this study, we detected the intracellular regulation within multinuclei of mature osteoclasts by using various in vitro approaches. Also, through biochemical approaches, we identified novel co-regulatory factors required for fine control of Nfatcl function during osteoclastogenesis.

Our results have revealed that the osteoclasts bear specialized system in order to advantageously maintain their multinuclei. They limit gene expression to only certain nuclei among multinuclei and share necessary nuclear proteins through intracellular shuttling. We presume that by this mechanism osteoclasts reduce the energy required for cellular maintenance and can resorb bone matrix more efficiently.

Although several transcriptional factors such as Nfatc1, NF(B, and AP1 were identified as regulators for normal osteoclast differentiation, epigenetic regulators supporting their function remain to be identified. Present study uncovered an uncharacterized Jmjd5 as a negative regulator for Nfatc1. Any histone demethylase activity was not detectable in Jmjd5 in spite of containing JmjC-domain. Rather, it had nonhistone protein, Nfatc1, as its target protein for protein destabilization. We expect that Jmjd5 functions as a hydroxylase like FIH and Jmjd6 already reported, through which it regulates the stability of Nfatc1 proteins through posttranslational modification like hydroxylation.

Recently, it has been demonstrated that involvement of cell type-specific subunit in polymorphic SWI/SNF complex is important for cellular specification and differentiation. However, this SWI/SNF-type complex was uncharacterized in osteoclast although osteoclasts are very unique differentiated cells. In this study, we purified a novel component of SWI/SNF-type chromatin remodeler, Kiaa1903, which is osteoclast-specifically expressed. Any components of SWI/SNF complex could not be recruited on promoters of Nfatc1-target genes, without Kiaa1903 protein. We expect that Kiaa1903 is a key determiner for osteoclast-specific SWI/SNF complex and is a necessary epigenetic partner of Nfatc1 required throughout osteoclastogenesis from early osteoclastogenesis for chromatin reconfigulation.

From present study, we uncovered new molecular mechanisms required for normal osteoclastogensis. Now, we are trying to confirm that newly identified factors by biochemical skills are significant in osteoclastogenesis and osteoclastic function by preparing the knockout mice.

破骨細胞、骨吸収を行う多核のマクロファージ様巨大細胞である。そのため破骨細胞の分化・成熟には、複数の段階で制御され、実際多くの制御因子が見出されている。しかしながら、クロマチンレベルでの遺伝子発現制御については未だ不明である。

クロマチン構造調節は、ピストンの翻訳後修飾と、ヌクレオソームの再構築とに大別され、それぞれ特異的な転写共役因子群によって担われている。転写共役因子群は、核内で複合体を形成することで協調的に機能し、選択的なクロマチン構造調節を行うと考えられている。即ち、転写共役因子は、複合体を形成することで基質特異性や酵素活性、さらには複合体安定化等の多様な制御機構を獲得しているといえる。そこで本研究では、破骨細胞の分化・成熟に伴うクロマチン構造調節の分子機構解明を目的に、破骨細胞分化制御を担う転写共役因子複合体を生化学的手法で単離・同定しその機能解析を試みている。

第一章の序論に引き続き、第二章では破骨細胞株を用いて、エストロゲンの分化初期段階への影響を解析した。具体的には、破骨細胞分化系にエストロゲンを添加し、分化への影響を調べたところ、骨吸収を担う遺伝子発現がエストロゲンによって抑制されるが分化への直接的な影響はないことを明らかにした。

第三章では、多核化破骨細胞の各々の核について転写活性の違いを検討した。細胞生物学的な解析の結果、多核化破骨細胞では転写活性化の指標であるセリンリン酸化RNAポリメラーゼ∬やNfatc1の局在が均一ではなく、一部の核に限局していることを明らかにした。

第四章では、ピストン脱メチル化酵素であるJmjCファミリーに着目し、中でもJmjd5が破骨細胞分化を抑制していることを見出し、さらにJmjd5がNfatc1と相互作用しNfatc1の安定化を制御していることを明らかにした。

また、本章ではNfatc1の転写共役因子として1(iaa1903を生化学的なアプローチで同定し、この因子が破骨細胞分化を促進する新規のクロマチン再構築因子複合体の構成因子であることを示唆した。

本論文は、これまで不明であった多核化破骨細胞の多核化の生理学的な意味についての新たな知見を得、さらに破骨細胞分化を制御する新たな因子群の同定を行い、クロマチン構造調節の分子機構の一端を明らかにすることができた。本研究は、破骨細胞を材料とした細胞生物学的、および生化学的解析の新たな試みであり、生体内における時期・組織特異的な転写制御におけるエピゲノム調節の理解に繋がるものであると期待される。以上より、審査委員一同は本論文が博士(農学)の学位論文として価値あるものと認めた。