1.Introduction

Sex steroid hormones, such as estrogen and androgen, have various functions in maintenance of homeostasis: development of reproductive organs and mammary gland, potentiation of muscles, and osteoprotection. Osteoprotective actions of sex steroid hormones have been illustrated that deficiency of sex steroid hormones cause the reduction of bone mass, particularly in post-menopausal women and aged men. However, it is largely unknown how sex steroid hormones exert their osteoprotective effects. The effects of sex steroid hormones on bone tissue can be considered as the sum of the direct effects to bone cells and the indirect effect to other tissue. While the various mechanisms of indirect effects have been proved like the effects through cytokines produced by immune cells and the increased levels of pituitary gland hormones, the direct effects are not fully understood.

Estrogen and androgen exert their effects by binding to its own nuclear receptors, such as Estrogen Receptor(ER)α,β and Androgen Receptor (AR), which work as transcription factors. Since it is known that ERβ has limited function for maintenance of bone metabolism among these receptors, we focused on the analyses of functions of ERα and AR in this study. The analyses of conventional ERα or AR knockout mice failed to address the direct functions of these receptors in bone because these mice exhibited endocrine disturbances. From these backgrounds, generations and analyses of cell type specific knockout (KO) mice are required to clarify the functions of ERα and AR in bone.

Osteoclastic ERα KO mice were generated and it was proved that osteoclastic ERα shorten life span of osteoclasts by promoting their apoptosis to date. Also in my master thesis, it was shown that osteoblastic ERα might be important for osteoprotection in male. Unpublished data in our lab indicated that AR also plays important roles for bone remodeling in osteoblasts and osteoclasts. These results indicate that ERα and AR in both osteoblasts and osteoclasts have osteoprotective function. However, the functions of ERα and AR in osteocyte, which is the other cell type in bone, are more ambiguous than in osteoblasts and osteoclasts. Osteocyte is embedded in extracellular matrix of bone and represents more than 90% of cells existing in bone. Recently it is postulated that osteocyte can orchestrate bone metabolisms by regulating number and activity of osteoblasts and osteoclasts through secreting various kinds of proteins. Thus, we presumed that the functions of sex steroid hormone receptors in osteocytes play possible significant role in bone metabolism.

From these backgrounds, in this study, we tried to elucidate the functions of ERα and AR in osteocytes by generating and analyzing the mice lacking ERα or AR in osteocytes.

2.Results

A.Function of Estrogen Receptor α in osteocytes

To investigate the role of ERα in osteocytes, we generated the mice lacking ERα in late-osteoblasts/osteocytes by crossing ERα floxed mice with Dmp1 Cre mice, which express Cre recombinase driven by the Dmp1 promoter. The mice harboring the genotypes of Dmp1(Tg/0); ERα(L2/L2) and ERα(L2/L2) were analyzed as ERα(ΔOcy/ΔOcy) and ERα (flox/flox) mice, respectively. As the results from bone radiological analyses, 12-week-old female ERα(ΔOcy/ΔOcy) mice, not male mice, exhibited significantly decreased BMD at the proximal tibia when compared to ERα (flox/flox). To further explore the cellular basis for the bone loss observed in the ERα(ΔOcy/ΔOcy) , bone histomorphometry was performed. As the results, in ERα(ΔOcy/ΔOcy) , the bone formation parameters, such as the number of osteoblasts, osteoblast surface and bone formation rate as well as the number of osteocytes/area were significantly decreased, when compared to the ERα (flox/flox). On the other hand, the bone resorption parameters such as osteoclast numbers or surfaces were not altered in ERα(ΔOcy/ΔOcy) . Also, the number of TUNEL staining positive osteoblasts/osteocytes was not significantly different between ERα (flox/flox) and ERα(ΔOcy/ΔOcy) . Also, to verify whether osteocytic ERα mediates osteoprotective estrogen actions, the mice were ovariectomized (ovx) at 8 weeks old with or without E2 replacement treatment at 10 weeks old. As expected, the recovery of BMD by E2 treatment in the ovx ERα(ΔOcy/ΔOcy) was significantly less than that in the ovx ERα (flox/flox), suggesting that a part of osteoprotective estrogen action indeed mediated through osteocytic ERα. These data suggest that osteocytic ERα play a role in bone metabolism by positive regulation of osteoblastic bone formation.

Moreover, estrogen/ ERα signaling is known to be involved in mechano-sensing and increasing bone formation under over loading conditions. Also, it is reported that tail suspension induced bone loss is significantly enhanced by ovariectomy. To determine whether ERα in osteocytes plays a role in unloading induced bone loss, tail suspension experiments were performed starting at 8 weeks of age for 4 weeks. As the results, tail suspension induced bone loss in the femoral diaphysis of ERα(ΔOcy/ΔOcy) female mice was significantly greater than that of ERα (flox/flox). This result indicates that ostecytic ERα exerts its function for resistance against unloading induced osteopenia.

B.Trial to explore the molecular basis of ERα function in osteocytes

To explore the molecular basis of bone loss in ERα(ΔOcy/ΔOcy) , osteocyte isolation method was established and gene expression profiles of isolated osteocytes obtained from ERα(ΔOcy/ΔOcy) and ERα (flox/flox) were compared. To isolate highly purified population of osteocytes, the mice with GFP labeled osteocytes by crossing ERα(ΔOcy/ΔOcy) or ERα (flox/flox) line with a GFP expressing mouse line driven by the Dmp1 promoter were generated. GFP+ and GFP- cell population were sorted by FACS (Fluorescence-activating cell sorter) from cells obtained by sequential enzymatic digestion of calvariae of mice at the age of 10 days. It was confirmed that GFP+ population represented osteocytes and GFP- population represented osteoblasts by the expression of marker genes in each population. Then, gene expression profiles of osteocytes were compared between ERα(ΔOcy/ΔOcy) and ERα (flox/flox) by a GeneChip analysis. As the results, 276 genes were found to be significantly at least 2-fold differentially expressed between ERα(ΔOcy/ΔOcy) and ERα (flox/flox) (p<0.01). The differentially expressed genes were further classified according to biological processes and molecular functions using the Functional Annotation Clustering tool of the DAVID Bioinfomatics Resources. In regard to Biological Processes, categories of cell differentiation, developmental process, and cellular developmental process were highly ranked. In addition, the category of "extra-cellular component" in Cellular Component and "binding" in Molecular Function was highly ranked. Furthermore, 'Secreted' was listed top in the Keyword analysis when sorted by p value. In terms of GO analysis, Sostdc1 (Sclerostin domain containing 1), reported as one of the antagonists of Wnt signaling, was raised as a possible candidate gene of ostecytic ERα. The expression level of Sostdc1 was increased more than 3-fold in osteocytes sorted from ERα(ΔOcy/ΔOcy) calvariae. In fact, there is a report implying that ERα is necessary for contribution of Wnt/β-catenin signaling, which plays pivotal roles in bone formation, in bone mechano-sensing.

Taken together, it was suggested that osteocytic ERα regulates gene expression of secreted proteins, which can be involved in the regulation of osteoblastic bone formation.

C.Function of Androgen Receptor in osteocytes

From the data mentioned above, osteocytes seem to transmit positive signals to osteoblasts through ERα in osteocytes in female mice. However, the osteoprotective actions of estrogen through osteocytes were unlikely from the similar type of analyses in male ERα(ΔOcy/ΔOcy) . This fact led us to raise an idea that the androgenic actions are dominant in male skeletal health. To address this idea, mice selectively deleted AR in osteocytes were generated as same as did for ERα(ΔOcy/ΔOcy) . BMD of 12-week-old male AR(ΔOcy/Y) mice, not in female mice, was significantly decreased in both metaphysis and diaphysis of femurs when compared to those of the control mice, showing that male AR(ΔOcy/Y) exhibited greater decrease of bone mass than that of female ERα(ΔOcy/ΔOcy) . To further explore the cellular basis for the bone loss observed in the AR(ΔOcy/Y), bone histomorphometry was performed. As the results, the number of osteoclasts, osteoclast surface, the number of osteoblasts, osteoblast surface, bone formation rate and mineral apposition rate were significantly increased in AR(ΔOcy/Y) compared to ARflox/Y. Thus, AR(ΔOcy/Y) exhibited osteoporotic bone phenotype with high bone turnover. Also, bone strength was evaluated by three-point bending test using femurs, showing that maximum load of AR(ΔOcy/Y) was 37% lower than that of ARflox/Y. These results suggest that AR in osteocytes contributes to maintaining bone strength by repressing bone resorption in male.

3.Discussion

On the basis of the reported evidences on the functions of sex steroid hormone receptors in bone, it is considered that estrogen is osteoprotective by regulating life span of osteoclasts through osteoclastic and osteoblastic ERα. However, little is known about the functions of osteocyte in osteoprotective sex steroid hormone actions for skeletal homeostasis. To decipher the direct functions of ERα and AR in osteocytes, the mice lacking ERα and AR in osteocytes were genetically generated and their bone phenotypes were analyzed in this study. ERα in osteocytes was found to play a significant role for maintenance of bone mass by regulating osteoblastic bone formation only in female. It was further revealed that ERα in osteocytes is supportive for maintaining bone mass not only under normal loading condition but also under tail suspension induced unloading condition, which can be considered as experimental recapitulation of recumbency or space flight. These results are consistent with a previous report, in which bone mass adaptation induced by over mechanical loading was impaired in ERα total KO mice. Furthermore, to investigate possible molecular basis underlying ERα function in osteocytes, we established osteocyte isolation technique from the conditional knockout mice by FACS. The results obtained from the Functional Annotation Clustering of differentially expressed genes suggested that osteocytic ERα might regulate transcription of the genes related to secrete proteins, which may regulate osteoblastic bone formation and contribute to maintenance of bone homeostasis. In fact, Sostdc1, an antagonist for Wnt signal, was raised as a candidate target gene of osteocytic ERα. These observations indicate that osteocytic ERα might play a role in estrogen's osteoprotective action by regulating osteoblastic bone formation through facilitating Wnt signaling pathway.

The results from the analyses of AR(ΔOcy/Y) revealed that AR in osteocytes mediates osteoprotective function through inhibition of osteoclastic bone resorption in male. Since the previous report revealed that male AR total KO exhibited increased bone resorption with the increased expression levels of RANKL, which is a crucial factor in osteoclastogenesis, RANKL might be a possible target gene of AR in osteocytes, which recently reported to be a major source for RANKL production.

Taken together, this study uncovered that sex steroid hormones play essential roles to maintain skeletal homeostasis through their receptors' action in osteocytes.

女性ホルモンであるエストロゲンや男性ホルモンであるアンドロゲンの骨防御作用は、閉経や加齢による血中性ホルモン濃度の低下が骨粗鬆症の原因となることからよく知られている。しかし、その詳細な作用機序については不明な点が多い。エストロゲンやアンドロゲンはエストロゲン受容体ERα、βやアンドロゲン受容体ARを介してその作用を発揮することが知られている。また、エストロゲンやアンドロゲンの作用は、骨組織のERα、β、ARを介する直接作用と、いったん他組織に作用し血中サイトカインの変動が骨組織に影響を及ぼす間接作用の2通りに大別できるが、本研究室では直接作用に着目し、研究を行ってきた。先行研究である破細胞特異的ERα欠損マウスの解析では、エストロゲンが破骨細胞のERαを介して破骨細胞の寿命を調節していることが明らかとなった。また申請者本人の修士論文の研究により、骨芽細胞のERαが雄性における骨量維持作用の一端を担っていることが示唆された。このように骨芽細胞、破骨細胞におけるERα,ARの機能は明らかになりつつある。しかしながら骨細胞におけるERα,ARの機能の大部分は未だ不明である。

本研究が対象としている骨細胞は骨基質中に埋まっており、骨組織の細胞の9割以上を占める細胞である。しかしその機能については、メカニカルストレスの感知、ミネラル代謝調節、骨芽細胞による骨形成・破骨細胞による骨吸収の調節と多様であり、不明な点も多い。

そこで、本研究ではこの骨細胞におけるERα、ARの機能を明らかにすべく、Cre-loxPシステムにより骨細胞ERαおよび骨細胞AR欠損マウスを作出し、その機能解析を行った。

第一章の序論に引き続き、第二章では骨細胞ERα欠損マウスを作出し、骨細胞ERα欠損マウスが骨密度の低下を呈することを明らかにした。また卵巣摘出実験から、この骨細胞のERαを介した骨量維持作用がリガンド依存的であることが明らかになった。更に、本マウスに対し尾部懸垂実験や骨折治癒実験の負荷実験を行い、尾部懸垂の非荷重下でも、骨細胞のERαが骨量維持作用を発揮していること、言い換えると、骨細胞はメカニカルストレスを感知し、ERαを介した骨量維持作用を発揮することが明らかになった。

更に、第三章では骨細胞のERαの骨量維持作用の詳細なメカニズムを明らかにすべく、骨形態計測などの解析を行った。その結果、骨細胞のERαは骨芽細胞数を正に制御することで、骨防御作用を発揮することが明らかになった。更に、骨細胞はタンパクの分泌により骨形成や骨吸収を制御していることが報告されているため、骨細胞ERα欠損マウスと対照群の骨細胞における遺伝子発現プロファイルを比較した。この際、骨細胞は単純な酵素処理のみでは単離が不可能である。そこで、本研究では、Dmp1-GFPマウスと交配させ、骨細胞がGFPでマークされたマウスをまず作出した。これらマウスの頭蓋骨を酵素処理することにより骨芽細胞と骨細胞の混在した細胞集団を取得し、その中からFACSで骨細胞のみを単離する方法を確立した。これにより単離された骨細胞の遺伝子プロファイルを骨細胞ERα欠損マウスと対照群で比較したところ、273遺伝子の発現変動を認めた。これら遺伝子をデータベースDAVIDで解析した結果、分泌タンパクが比較的多数含まれることが判明した。このことから骨細胞のERαは分泌タンパクの転写を制御することにより骨芽細胞の増殖を促し、骨防御作用を発揮していることが示唆された。なお、骨細胞におけるERαの候補遺伝子としてはWntシグナル関連因子がいくつか取得された。

また、第四章では、雄性における骨細胞を介した骨量維持機構を明らかにする目的で、雄で主要な性ホルモンであるアンドロゲンの受容体ARを骨細胞特異的に欠損させたマウスを作出した。その結果、骨細胞AR欠損マウスでも骨量減少が認められ、それが骨吸収の亢進に起因することが示唆された。

本論文では、これまで大部分が不明であった骨細胞におけるERα、ARの生体内機能を明らかにすることができた。更に、遺伝子欠損マウスからの骨細胞単離法を確立し、骨芽細胞の増殖因子の発現を骨細胞のERαが制御することを明らかにすることで、エストロゲンの骨防御作用の一端を解明した。以上、本研究は、エストロゲンが骨細胞のERαを介して骨細胞-骨芽細胞間の機能的インタラクションを調節することを提唱するものであり、骨組織における性ホルモン作用の新たな制御機構を明らかにした。以上の理由により、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。