Estrogen receptor-related receptor gamma (ERR) is a regulator of skeletogenesis
by
Marco Cardelli
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Medical Biophysics University of Toronto
© Copyright by Marco Cardelli 2015
Estrogen receptor-related receptor gamma (ERR) is a regulator of skeletogenesis
Marco Cardelli
Doctor of Philosophy
Medical Biophysics University of Toronto
2015 Abstract
Sex steroids, such as Estrogen, play important roles in physiological and pathological processes, including bone growth and bone homeostasis. Estrogen's effects are mediated through its receptors, Estrogen receptor alpha and beta (ER and ER). Despite its importance, loss of expression of ERs does not result in severe skeletal effects, suggesting that other factors are involved in the Estrogen pathway(s). Estrogen receptor related receptors (ERRs) are a family of orphan nuclear receptors, i.e., with no known natural ligands, comprising ERRα, ERRβ, and
ERRγ. While ERR has been shown to be involved in various aspects of skeletogenesis, there is very little know about the potential role of ERR. To address this question, I used gain-of- function and loss-of-function approaches in mice. In the gain-of-function model with cartilage- specific ERR-overexpression, the size of the craniofacial, axial and appendicular skeletons was reduced compared to that in wild type (WT) mice. Histological analysis revealed a reduction in the length of the transgenic versus WT growth plate, attributable to a reduced proliferative zone accompanied by a decrease in the number of Ki67-positive proliferating cells, with no significant change in apoptosis. Quantification of expression of putative target genes suggested that ERRγ negatively regulates chondrocyte proliferation and positively regulates matrix synthesis to
ii coordinate growth plate height and organization. In an ERRγ global knockout mouse line, I observed no gross morphological anomalies or difference in skeletal length in newborn mice, but by 8 weeks of age ERRγ +/- male but not female mice exhibited increased trabecular bone, which was further increased by 14 weeks. Histomorphometric and serum biochemistry data indicated that the trabecular bone increase was due to an increase in bone formation, with no apparent change in bone resorption. Analysis of ERRγ +/- versus ERRγ +/+ bone marrow stromal cell cultures indicated that ERRγ negatively regulates osteoblast differentiation and matrix mineralization but not mesenchymal precursor number. These data, together with results from co-immunoprecipitation and Runx2 antisense oligonucleotide treatment experiments in vitro, indicated that ERRγ is not required for skeletal development but is a sex-dependent negative regulator of postnatal bone formation, acting in a RUNX2- and apparently differentiation stage-dependent manner. I conclude that ERRγ plays negative regulatory roles in both cartilage and bone in postnatal mice and may be an important new therapeutic target in skeletal diseases and conditions.
iii
Acknowledgments
I would like to thank my supervisor, Dr. Jane Aubin. Dr. Aubin is not only an expert in the fields of bone and cartilage biology as well as mesenchymal stem cells, she is an inspirational mentor and leader. It was a great privilege to have had the opportunity to study in the Aubin lab. I would like to thank the members of the Aubin lab: Tanya Zappitelli, Ralph A. Zirngbl, Ruolin Guo, Frieda Chen, Jonathan F. Boetto and Kristen P. McKenzie for constructive discussion and critique. Special gratitude to Tanya for her many hours of counsel and support, and to Ruolin for her assistance in experimental procedures.
Very many thanks to my committee members, Dr. Armen Manoukian and Dr. Eldad Zacksenhaus for their guidance and support throughout my PhD training.
iv
Table of Contents
Acknowledgments ...... iv
Table of Contents ...... v
List of Tables ...... ix
List of Figures ...... x
Chapter 1 ...... 1
1 Introduction ...... 1
1.1 Bone Development ...... 1
1.1.1 Endochondral Ossification ...... 1
1.1.2 Intramembranous Ossification ...... 4
1.2 Chondrogenesis/chondrocyte differentiation ...... 7
1.2.1 Growth factors ...... 7
1.2.2 Cyclins ...... 10
1.2.3 Extracellular matrix proteins ...... 10
1.2.4 Transcription factors ...... 11
1.3 Osteogenesis/Osteoblast Differentiation ...... 13
1.3.1 Growth factors ...... 13
1.3.2 Transcription factors ...... 16
1.4 Estrogen in skeletal development ...... 17
1.4.1 Effect of Estrogen and its receptors ...... 17
1.4.1.1 Longitudinal growth ...... 17
1.4.1.2 Bone remodeling ...... 19
1.5 Estrogen receptor-related receptors ...... 23
1.5.1 ERR ...... 27
1.5.1.1 ERR and cartilage ...... 27
v
1.5.1.2 ERR and bone ...... 28
1.5.2 ERR ...... 29
1.5.3 ERR ...... 30
1.5.3.1 ERRand energy metabolism ...... 30
1.5.3.2 ERRand skeletogenesis ...... 31
1.6 Objectives ...... 34
Chapter 2 ...... 35
2 Cartilage-specific overexpression of ERR results in chondrodysplasia and reduced chondrocyte proliferation ...... 35
2.1 Abstract ...... 36
2.2 Introduction ...... 37
2.3 Material and Methods ...... 40
2.3.1 Ethics Statement ...... 40
2.3.2 Construction of pCol2a1mERRγ2 and generation of transgenic mice ...... 40
2.3.3 LacZ stain for detection of transgene ...... 41
2.3.4 Gene expression analysis ...... 41
2.3.5 Western Blotting ...... 41
2.3.6 Whole mount skeletal staining ...... 42
2.3.7 Histological and immunofluorescence analysis ...... 42
2.3.8 TUNEL assay ...... 44
2.3.9 Expression analysis of putative target genes ...... 44
2.3.10 Statistical Analysis ...... 45
2.3.11 Mouse gene nomenclature ...... 45
2.4 Results ...... 48
2.4.1 Overexpression of ERRγ2 results in dwarfism ...... 48
2.4.2 Overexpression of ERRγ impairs chondrocyte proliferation, differentiation- maturation, cartilage matrix production and growth plate organization ...... 52 vi
2.5 Discussion ...... 58
Chapter 3 ...... 63
3 ERR is not required for skeletal development but is a RUNX2-dependent negative regulator of postnatal bone formation in male mice ...... 63
3.1 Abstract ...... 64
3.2 Introduction ...... 65
3.3 Materials and Methods ...... 67
3.3.1 Ethics Statement ...... 67
3.3.2 Animals ...... 67
3.3.3 Whole mount skeletal staining ...... 67
3.3.4 Microcomputed tomography (CT) ...... 68
3.3.5 Histology and histomorphometry ...... 68
3.3.6 Immunohistochemical detection of Ki67 and TUNEL assay ...... 69
3.3.7 Serum biochemistry ...... 69
3.3.8 Gene expression analysis ...... 70
3.3.9 Isolation of Bone Marrow Cells, Runx2 antisense assay, and CFU assay ...... 70
3.3.10 Stable cell line constructs ...... 71
3.3.11 Co-immunoprecipitation (Co-IP) Analysis ...... 72
3.3.12 Western blotting ...... 72
3.3.13 Statistical Analysis ...... 73
3.3.14 Mouse gene nomenclature ...... 73
3.4 Results ...... 75
3.4.1 ERR -/- mice die perinatally, but display no skeletal abnormalities ...... 75
3.4.2 Adult ERR +/- male mice have increased trabecular bone, increased osteoblast number and surface, but no change in osteoclast number or surface ... 77
3.4.3 Increased osteoblast differentiation in ERR +/- mice ...... 82
vii
3.4.4 ERR interacts with RUNX2, and knockdown of RUNX2 leads to at least partial rescue of the ERR +/- skeletal phenotype in vitro ...... 84
3.5 Discussion ...... 87
Chapter 4 ...... 92
4 Conclusions and Future Directions ...... 92
References ...... 99
viii
List of Tables
Table 1.1 Osteoblast-specific ERKO mouse models ...... 22
Table 2.1 Primer sequences used in gene expression analysis ...... 46
Table 2.2 Results of in silico search for putative ERRE sites in known regulators of growth plate chondrocytes ...... 47
Table 3.1 Primer sequences used in gene expression analysis ...... 74
ix
List of Figures
Figure 1.1 Endochondral ossification process ...... 3
Figure 1.2 Intramembranous bone formation...... 6
Figure 1.3 Modular structures of nuclear receptors...... 25
Figure 1.4 ERRs bind DNA ...... 26
Figure 1.5 ERR is expressed in murine bone and cartilage...... 33
Figure 2.1 Col2::ERRγ2 transgenic mouse generation and protein expression analysis...... 50
Figure 2.2 Skeletal analysis of P7 animals...... 51
Figure 2.3 Growth plate analysis of P7 pups...... 54
Figure 2.4 Decreased proliferation, but not apoptosis, is responsible for reduction of growth plate height...... 55
Figure 2.5 Expression profile of putative target genes...... 57
Figure 3.1 Neither ERR ablation nor haploinsufficiency has a detectable effect on embryonic bone development and growth...... 76
Figure 3.2 Trabecular bone formation is increased in 14-week old male distal femurs...... 78
Figure 3.3 ERR regulates osteoblast number and differentiation, but does not affect osteoclast number or activity...... 80
Figure 3.4 ERR regulates markers of osteoblast differentiation and mineralization ...... 81
Figure 3.5 Osteoblast differentiation is increased in ERR +/- stromal cell cultures...... 83
Figure 3.6 ERR interacts with RUNX2 through the DBD and modulates osteoblast differentiation...... 86
x 1
Chapter 1
1 Introduction 1.1 Bone Development
There are two main modes of bone development: endochondral ossification, whereby the ultimate bone structure is preceded by cartilage anlagen, and intramembranous ossification, where mesenchymal condensations differentiate into osteoblasts, without a cartilage scaffold. In both cases, the process is complex and involves many signaling and transcriptional pathways that influence proliferation and cell survival, as well as differentiation and maturation events. I'll first briefly summarize the two developmental processes, then review major growth factor and transcriptional pathways involved, ending with hormonal (Estrogen) regulation.
1.1.1 Endochondral Ossification
Endochondral bone formation is responsible for the development of the vertebrate axial and appendicular skeletons, as well as the cranial base. The process begins with condensation of undifferentiated mesenchyme, characterized morphologically by the close packing of cells as compared to non-condensing cells (1) (Figure 1.1A). The initiation of cell condensations takes place in an avascular environment, and chondroblast differentiation takes place at the centre of these condensed regions. Chondroblasts express a cartilage-associated gene profile including the matrix genes collagen type 2 (Col2a1) and aggrecan (Agg) and mature into chondrocytes completely encased in matrix (2). SRY-box containing gene 9 (Sox9), a member of the Sox family of transcription factors, and major regulator of Col2a1 expression, also promotes chondrocyte proliferation (3). Surrounding the condensed cell population is a layer of stacked
2 cells called the perichondrium, separating the differentiated chondroblasts from the mesenchyme (Figure 1.1B) (2).
The growth of the cartilage element depends on both rapid proliferation and further maturation of chondrocytes. These two processes generate a histologically dynamic structure called the growth plate (Figure 1.1C-F). Within the growth plate, zones of proliferation, maturation, hypertrophy, and bone formation can be clearly identified. Resting chondrocytes at the epiphysis rapidly proliferate and as they mature, begin to secrete cartilage matrix rich in
AGG and COL2A1. At the centre, or diaphysis, the chondrocytes mature further, become flattened and organize into longitudinal columns. These cells then differentiate into hypertrophic (enlarged) chondrocytes, cease expression of Col2a1 as well as proliferative markers, and now secrete a matrix that is rich in collagen type 10 (COL10A1) (1). At the same time, perichondrial cells differentiate into osteoblasts expressing Runt-related transcription factor 2 (Runx2, a transcriptional activator of osteoblast differentiation), to form a mineralized tubular structure surrounding the cartilage centre, termed a bone collar (2).
The terminal hypertrophic chondrocytes become enclosed in a calcified extracellular matrix and die via apoptosis. The mineralized extracellular matrix (ECM) is then vascularized through a vascular endothelial growth factor (VEGF)-dependent pathway (4). Degradation of the
ECM follows, and the infiltrating osteoblasts use the remnants of calcified cartilage as a scaffold to secrete bone matrix and this site is known as the primary ossification centre (5). The first bone that is deposited is known as trabecular bone, which is a woven, or cancellous type. Osteoclasts, multinucleated cells of hematopoietic origin, subsequently invade the ossification centre and resorb the trabecular bone, which will be replaced by more mature bone of a lamellar type (6)
(Figure 1.1F).
3
Figure 1.1 Endochondral ossification process. (A-B) Condensations form and cells at the centre differentiate into chondrocytes. (C-E) Chondrocytes at the centre hypertrophy, their matrix becomes mineralized, and the cells die via apoptosis. (F) The ECM is degraded and infiltrating osteoblasts secrete their own matrix on the mineralized remnants, and subsequently this matrix is mineralized and is primary (trabecular) bone. The primary bone will then be resorbed and replaced by mature, lamellar bone
4
1.1.2 Intramembranous Ossification
Intramembranous ossification is characteristic of the bones of the skull vault, mandible, and majority of the face as well as part of the clavicle. The initial condensation of undifferentiated mesenchyme is similar to that of endochondral ossification (Figure 1.2). Although the condensation event occurs in an avascular condition, there is an immediate onset of blood vessel invasion as soon as the cells begin to differentiate into osteoblasts. These cells begin to secrete a matrix rich in collagen type I (COL1A1), the major collagenous component of bone. The matrix is subsequently mineralized at areas referred to as ossification centres (1). The initial bone formed is irregular in shape and is surrounded by osteoblasts, which, in turn become enclosed in the matrix they secrete and become known as osteocytes. These cells project canaliculi through the bone and maintain contact with other osteocytes. It is thought that this network maintains the integrity of the bone in response to mechanical strain (7).
At the periphery of the bone lie preosteoblasts that rapidly proliferate, mature to osteoblasts and secrete matrix. As in endochondral bone formation, this incremental addition of bone creates a trabecular beam, as more trabeculae form the ossification centre expands.
Trabeculae interlace to form the same type of cancellous bone as found in the developing growth plate. The pattern of growth is dependent on the type of bone being formed. The most widely recognized intramembranous bones are the flat bones of the skull. The centres commence at the lateral edges of the skull rudiment and grow medially by generating a radial network of bone. Trabeculae also form perpendicular to the skull surface, adding to the overall thickness of the bone. Most of the prenatal woven bone that is formed will eventually be resorbed and replaced by mature, lamellar bone, characterized by its organization of collagen fibres.
5
RUNX2 is a master regulator of osteogenesis, necessary for the commitment of mesenchymal cells to the osteoblast lineage, differentiation, and regulation of genes responsible for matrix synthesis.
6
Figure 1.2 Intramembranous bone formation. The osteogenic front (OF) is a region of highly proliferating osteoprogenitors. The osteogenic layer (OL) is made up of osteoblasts and bone lining cells along the bone surface (labeled (B)). The suture matrix (SM) is made up of mesenchymal cells as well as osteoprogenitors.
7
1.2 Chondrogenesis/chondrocyte differentiation
1.2.1 Growth factors
Growth hormone (GH) has been established as an important regulator of longitudinal bone growth. It exerts its effects directly by binding its receptor (GHR) (8), and indirectly by stimulating production of insulin like growth factor-1 (IGF-1) in the liver (9) as well as in other tissues (10). Although the majority of IGF-1 is produced by the liver, a liver specific Igf-1 knockout mouse had little effect on skeletal growth, despite reducing circulating IGF-1 levels by
75% (11). It has also been shown that in a transgenic mouse expressing only liver-derived Igf-1 mice grew to only 30% of wild type (WT) adult body size (12), suggesting that locally produced
IGF-1 is a major regulator of skeletal growth. This was further evidenced through osteoblast and chondrocyte specific Igf null mice, with both displaying reduced body and bone growth (13-
15)
Indian hedgehog (IHH) is a master regulator of chondrocyte proliferation and differentiation, and essential to bone formation (16, 17). Ihh -/- mice display shortened limbs, reduced proliferation and abnormal chondrocyte hypertrophy, as well as a complete absence of osteoblasts within the perichondrium (16). IHH is synthesized by prehypertrophic chondrocytes and binds to its receptor patched (PTC-1), which leads to the activation of the membrane-bound protein smoothened (SMO) which, in turn, triggers a cascade of IHH target gene activation (18,
19). Together with parathyroid related peptide (PTHRP) they form a negative feedback loop to maintain proliferation and delay chondrocyte hypertrophy (17). PTHRP is secreted by perichondrial cells at the articular ends of the growth plate as well as by early proliferating chondrocytes. It diffuses through the growth plate and binds to its receptor, PTH1R which is expressed at low levels in the proliferating chondrocytes, and at much higher levels in
8 prehypertrophic chondrocytes, keeping the chondrocytes in the proliferative state, and preventing chondrocyte hypertrophy (20). Chondrocytes that are at a sufficient distance from
PTHRP action differentiate into prehypertrophic chondrocytes, stimulating IHH expression (21,
22). This, in turn stimulates PTHRP expression by periarticular chondrocytes. This feedback loop regulates the site of hypertrophic differentiation and proliferative zone size (23).
The bone morphogenic proteins (BMP's) are a set of growth factors belonging to the
Transforming growth factor (TGF superfamily (24), which exert their effects on prechondrogenic condensations as well as later in chondrocyte development. It was shown that
BMP inhibits the expansion of the early limb bud, and promotes the expression of chondrocyte markers (25). At later stages of chondrocyte development, BMP's have multiple functions and are expressed throughout the growth plate. Cartilage-specific Bmpr1a/Bmpr1b double knockout mice showed that chondrocyte differentiation is impaired, with a reduction in proliferation, growth plate disorganization and a reduction in cartilage matrix protein expression (26).
Moreover, Col2a1-Smad1/Smad5 double mutant mice (key regulators of canonical BMP signaling) displayed severe chondrodysplasia (27). Taken together, this suggests that it is canonical BMP signaling that is essential for chondrocyte differentiation. More recently, it has been shown that BMP induces chondrocyte hypertrophy in human mesenchymal stem cells in vitro (28).
Investigation of fibroblast growth factor (FGF) signaling in endochondral bone growth is complex due to the presence of multiple ligands and receptors. It has been shown that FGFR3 has an important role in the regulation of chondrocyte proliferation and maturation. Activating mutations of FGFR3 in humans cause the severe dwarfing chondrodysplasias, achondroplasia
(29, 30) , hypochondroplasia(31), and thanatophoric dysplasia(32, 33). In contrast, Fgfr3 -/- mice
9 display skeletal overgrowth (34, 35). While a number of FGFs can activate FGFR3, FGF18 has been shown to be the most important (36). The suppressive effects of FGFR3 are mediated via signal transducer and activator of transcription 1 (STAT1) (36).
To investigate a link between the FGF and IHH pathways, Minina et al. treated limb explants with FGF2, which resulted in a decrease in Ihh expression and acceleration of chondrocyte hypertrophy, an effect inhibited by IHH, suggesting that FGF acts upstream of IHH
(37). In contrast, IHH could not rescue the suppressive effects of FGF2 on proliferation, indicating FGF signaling regulates chondrocyte proliferation in an IHH-independent manner (37).
It has also been shown that the actions of FGF signaling on chondrocyte proliferation and hypertrophy are inhibited by BMP signaling (37, 38).
The Wingless-related integration site (WNT) family of secreted proteins are responsible for several aspects of skeletal development. WNT activation occurs via the canonical pathway, through the binding of the frizzled/low-density lipoprotein receptor-related protein 5/6 (FRZ
/LRP5/6) co-receptor complex, which, in turn, inhibits the phosphorylation of -catenin, its subsequent accumulation and translocation to the nucleus where it regulates target gene expression. WNTs can also be activated through the non-canonical pathway by binding FRZ alone, which results in the degradation of -catenin, thus preventing its translocation to the nucleus (39). Multiple members of the WNT family are expressed throughout the growth plate, with highest expression levels observed in the proliferative and hypertrophic zones. The data indicate that activation of both pathways is essential to proper chondrocyte proliferation and hypertrophy (40, 41), and that they act in a chondrocyte maturational stage-dependent way. For example, constitutive -catenin expression in immature chondrocytes suppresses hypertrophy while it promotes terminal hypertrophy in mature chondrocytes (42).
10
1.2.2 Cyclins
Chondrocyte proliferation is mediated by cyclin/cyclin-dependant kinase (CDK) complex, activating cell cycle progression (43). By analysis of CcnD1 -/- mice, Cyclin D1 (CcnD1) has been shown to be an important regulator of chondrocyte proliferation. These mice displayed reduced growth plate height due to a decrease in proliferation (44), but more importantly it was shown that the actions of TGF and PTHRP were dependent on CCND1 activation (44).
Inhibition of chondrocyte proliferation, principally by FGFR3, caused the induction of CDK inhibitors, such as p21, leading to cell cycle arrest (45).
1.2.3 Extracellular matrix proteins
The major ECM proteins present in the growth plate are the collagenous proteins COL2A1 and
COL10A1, as well as the proteoglycan AGG. COL2A1 and AGG are abundantly expressed in the resting and proliferative zones and are downregulated as chondrocytes mature to hypertrophy, where they begin to synthesize COL10A1 (46). While the ECM is important for the cartilage structure, the protein components are also responsible for activating signaling pathways for chondrocyte proliferation and maturation. For example, it has been shown in chondrocytes from mouse ribcage that lack of β1 integrin results in reduced chondrocyte motility and COL2A1 adhesion, as well as reduced CcnD1, and increased expression of the cyclin dependent kinase inhibitor Cdkn1a, resulting in decreased chondrocyte proliferation (47).
Transgenic mice with targeted inactivation of Col2a1 exhibit increased chondrocyte apoptosis with a marked decrease in the anti-apoptotic protein BCL-2, providing evidence that COL2A1 is important to chondrocyte survival (48).
AGG is sulphated by the enzyme chondroitin 4-sulphotransferase-1, encoded by the
C4st1 gene. Mutation of the C4st1 gene and thus failure of proteoglycan sulphation results in
11 severe dwarfism and neonatal death. The growth plate displays accelerated proliferation and maturation, leading to premature cell death (49). Although not fully elucidated it was hypothesized that the altered AGG in the growth plates of these mice leads to improper sequestration of TGF- in the ECM resulting in constitutive activation of the TGF- receptor
(49).
COL10A1 mutations in humans result in Schmid type of metaphyseal chondrodysplasia
(SMCDS), characterized by short stature, bowing of the long bones, and widened growth plates
(50-54). A Col10a1 -/- mouse was generated and featured some similarities to human SMCDS, including growth plate alterations and coxa vara, the reduction in the angle between the femoral neck and shaft and shortening of femoral neck (55). Electron microscopy revealed drastic changes in growth plate distribution of matrix components, including proteoglycans (55). Trabecular bone in null mice was also affected, with abnormal pattern of mineral deposition (55). In an attempt to recapitulate the SMCDS phenotype in mice, a transgenic mouse model harbouring a frame-shift mutation was generated. These mice also displayed short stature and coxa vara, a hip deformity characterized by reduction in angle between the femoral head and shaft. Further, these mice also displayed a significantly wider hypertrophic zone, a feature not present in the Col10a1 -/- mouse.
Additional data suggests that COL10A1 through its interaction with discoidin domain receptor-2
(DDR2) may affect the intracellular signaling pathways in chondrocytes (56, 57). Taken together the data suggests that COL10A1 is an important factor to proper growth plate development.
1.2.4 Transcription factors
SOX9 expression is required to maintain chondrogenesis. The importance of SOX9 to chondrocyte development was first demonstrated by the failure of Sox9 -/- cells in mouse chimeras to differentiate into chondrocytes, instead remaining undifferentiated mesenchyme (58).
12
Analyses of a knockout mouse model showed that happloinsufficiency of Sox9 resulted in perinatal death and malformation of cartilage primordia and premature skeletal mineralization, suggesting SOX9 was required for 2 processes: initial chondrocyte fate and maturation of chondrocytes (59). To elucidate these processes, conditional Sox9 mutants, where Sox9 was deleted in mesenchymal condensations, were generated and found to completely lack cartilage and bone development in long bones, along with an increase in mesenchymal cell apoptosis (3), indicating that Sox9 is essential to chondrogenesis, and endochondral bone development. When
Sox9 was deleted at later stages of chondrocyte development via generation of Col2a1:Sox9 knockout mice, it was found that SOX9 is required for chondrocyte proliferation and regulation of chondrocyte hypertrophy, but not osteogenesis (3), confirming the observations made in the initial knockout mouse model (59). In vitro assays have shown that SOX9 directly regulates a multitude of cartilage matrix genes, including Agg (60) and Col2a1 (61).
Two related transcription factors, Sox5 and Sox6, are also upregulated by SOX9, and form protein-protein interactions with SOX9 to assemble a transcriptional activation complex (3).
This complex is required for the overt differentiation of condensed mesenchymal cells to prechondrocytes and then to proliferative chondrocytes. The SOX9/SOX5/SOX6 complex also acts to inhibit hypertrophy in the prehypertrophic zone (62).
While prechondrogenic Runx2 expression is directly repressed by bagpipe homeobox homolog 1 (BAPX1 or NKX3-2)(63), allowing Sox9 expression and thus chondrogenic initiation, chondrocyte hypertrophy depends on the downregulation of Sox9 expression, with an increase in Runx2 and myocyte enhancer factor-2 (Mef2c), a transcription factor originally found to regulate muscle and cardiovascular hypertrophy, but was shown to be required for chondrocyte hypertrophy and ossification (64). Runx2 is expressed throughout the
13 hypertrophic zone of the growth plate and is considered a master regulator of chondrocyte progression to hypertrophy and regulates the expression of the major hypertrophic matrix gene
Col10a1. RUNX2 impinges on the IHH/PTHRP negative feedback loop by activating Ihh expression through interaction with its promoter (65), thus maintaining a proper balance between chondrocyte proliferation and hypertrophy. Runx2 -/- mice die at birth, are completely devoid of osteoblasts (66) and have disturbed chondrocyte differentiation, with a decrease in hypertrophic chondrocytes and aberrant Col10a1 expression (67). Conversely, transgenic mice expressing full length Runx2 under the control of the Col2a1 promoter, resulted in mice with severe skeletal malformations characterized by accelerated chondrocyte differentiation and osteogenesis (68).
1.3 Osteogenesis/Osteoblast Differentiation
1.3.1 Growth factors
In addition to their role in the growth plate, the GH/IGF-1 axis also plays an important role in osteoblastogenesis. IGF-1 mediates the osteoblastic effects of GH through the IGF1R, as GH did not increase osteoblast numbers in osteoblast-specific Igf1r null mice (69). Further, it has been shown that bone formation is decreased (70) and increased (71) in conditional Igf1r null and
Igf1r overexpressing transgenic mice, respectively. Although the majority of IGF-1 is produced by the liver, recent evidence suggests that local IGF-1 production is essential to osteoblast differentiation. A Col12-Igf-1 null mouse had reduced mineralization at embryonic stages as well as reduced bone formation in adult mice (14). Further, it has been shown that bone growth and bone turnover were reduced in osteocyte-specific Dmp1-Igf-1 knockout mice (15).
IHH is essential for endochondral osteoblast differentiation, as the Ihh -/- mouse completely lacked osteoblasts (16). Subsequently, it was shown that IHH signaling is directly required for perichondrial cells to initiate osteoblast differentiation and expression of RUNX2
14
(72). The role of IHH in postnatal osteoblast function is still not fully clear. An inducible Col2a1-
Ihh d/d knockout mouse resulted in a loss of a postnatal growth plate as well as decreased trabecular bone accompanied with a decrease in osteoblast markers, suggesting postnatal expression of IHH is required for proper osteoblast differentiation (73). A more recent paper examined osteoblast-specific HH signaling and found that activation of HH in mature osteoblasts resulted in increased bone formation and accelerated bone turnover with a significant loss of trabecular bone (74). Inactivation of HH signaling in mature osteoblasts resulted in an increase in bone mass in older mice due to suppression of osteoclast differentiation (74), suggesting that HH signaling is required for proper bone turnover. In contrast, pharmacological
HH inhibition in mice led to decreased bone mass, highlighted by decreased osteoblast number and function and decreased osteoclast number (75). What is clear is that HH signaling is complex and may exert differential effects in a cell- and stage-specific manner.
The WNT family plays an important role in osteoblast differentiation. Members of the canonical (-catenin-mediated) WNT signaling pathway have been shown to regulate postnatal bone accrual. For example, Lrp5 -/- mice have reduced bone mass and osteoblast proliferation
(76). Conversely, a transgenic mouse harbouring a human LRP5 high bone mass mutation also displays increased bone mass (77). There is evidence that suggests the actions of LRP5 may be cell- and stage-dependent. Yadav et al. (2009) showed that loss of Lrp5 in the gut, and not in mature osteoblasts led to a decrease in trabecular bone formation and osteoblast proliferation
(78). However, Cui et al. (2011) used a Dmp1:Lrp5 -/- knockout mouse to show that Lrp5 was necessary for proper bone mass (77). Deletion of another key molecule in the WNT pathway, - catenin, has revealed its importance in osteoblast differentiation. Multiple groups have shown that -catenin deletion in mesenchymal cells abolishes osteoblast maturation (79-82). In addition
15 to its role in the developing skeleton, -catenin has also been shown to be important for postnatal bone homeostasis by regulating osteoclast bone resorption (83-85).
The BMP pathway has been extensively studied and shown to be important in many cellular processes. Global Bmp knockout mice have not been very useful in elucidating the role of BMPs in osteogenesis, due to either early embryonic lethality (Bmp2 and Bmp4) or functional redundancy (Bmp7). The generation and characterization of conditional knockout mice has elucidated the importance of BMP signaling in osteoblast differentiation. The limb-specific knockout of Bmp2 resulted in normal embryonic limb development but subsequent post natal defects in bone formation and impaired fracture healing (86). Bmp2-/-;Bmp4-/- limb-specific double knockout mice resulted in the disruption of osteoblast differentiation (87), demonstrating the importance of these growth factors in osteogenesis. Osteoblast-specific deletion of the BMP receptor Bmp1ra has revealed a regulatory role for BMP signaling in osteoblast differentiation
(88), function of mature osteoblasts (88), and osteoclastogenesis (89). Taken together, the data indicate that BMP signaling is crucial to multiple aspects of bone homeostasis and osteoblast differentiation.
Genetic studies of the FGF pathway have revealed diverse roles in osteoblast proliferation (90, 91) and function (92-94). Although not fully elucidated, the evidence indicates that
FGF signaling induces BMP expression and activates Runx2 (95-97). It has also been shown that
RUNX2 forms a transcriptional complex with T-cell factor/lymphoid enhancing factor
(TCF/LEF), binding the Fgf18 promoter region, inducing its expression (98). It is clear, from the above, and other studies, that FGF ligands and receptors are required for proper osteoblast proliferation and differentiation, but the precise stages at which the signals act is still yet to be fully elucidated.
16
1.3.2 Transcription factors
A variety of transcription factors have been implicated in formation of the skeleton and osteogenesis; review of all these factors is beyond the scope of this background and I focus here on two that are key. In addition to its importance in chondrogenesis, RUNX2 is also a major regulator of osteogenesis, as evidenced through Runx2-/- mice, where a complete lack of ossification (intramembranous and endochondral) was observed and phenotypic characteristics of cleidocranial dysplasia (CCD) were observed (66). RUNX2 regulates the expression of several osteoblast matrix genes including Col1a1, osteopontin (Opn), bone sialoprotein (Bsp) and osteocalcin (Ocn) (99). Interestingly, Col1a1:Runx2 transgenic mice exhibit severe osteopenia with fractures, with an increase in less mature osteoblasts at the expense of fully functioning, mature osteoblasts and osteocytes and lack of expression of the above-mentioned osteoblast- associated genes (100). Concomitant with the phenotype was the failure to induce. Taken together, the in vivo data coupled with in vitro data (99) indicate that RUNX2 is necessary for the early differentiation of osteoblasts and prevents terminal differentiation of late stage osteoblasts.
Osterix (Osx) is another transcription factor downstream of Runx2 (101), regulating osteogenesis in a Runx2-dependent and -independent manner. Osx-/- mice have a similar phenotype to Runx2-/- mice, indicating its requirement for osteogenesis (101), albeit by promoting proliferation of progenitors, as opposed to Runx2, which has anti-proliferative effects, but promotes differentiation.
17
1.4 Estrogen in skeletal development
1.4.1 Effect of Estrogen and its receptors
1.4.1.1 Longitudinal growth
The role of Estrogen in longitudinal bone growth is complex, due to its seemingly opposite functions, that is, both enhancing and ceasing longitudinal bone growth. Although not fully elucidated, it has been shown that during early sexual maturation, low Estrogen levels enhance growth, while high Estrogen levels (at sexual maturity) lead to cessation of growth (102) . While often considered a “female” hormone, loss of Estrogen in an aromatase-deficient male resulted in tall stature, osteopenia, and delayed bone age (i.e., continued growth post-puberty) (103).
Further, a 28 year old Estrogen receptor alpha (ER)-deficient male had severe osteoporosis, increased height, a bone age of 14 years, and continual growth (104). Conversely, males and females that undergo sexual precocity (premature Estrogen secretion) experience an increase in growth velocity and decreased final height, due to accelerated growth plate fusion (105).
The effects of Estrogen are mediated through its 2 receptors, termed Estrogen receptor alpha (ER) and beta (ER). The ERs are transcription factors that recognize a palindromic
DNA consensus sequence, termed the Estrogen response element (ERE), in the regulatory region of target genes (106).
To better understand their roles, a number of groups have generated and analyzed ER,
ER, and double knockout mice have been generated and analyzed. Two ER knockout lines
(ERKO) have been generated. The first was generated by the Smithies’ lab, by insertion of a neo cassette into exon 2 (107). Phenotypic analysis revealed smaller bone length in male and female mice (108, 109). However, it was also found that a short ER isoform was expressed, containing a
DBD and LBD and capable of activating ERE dependent target genes, and that this isoform can
18 act in a dominant negative fashion (110). The second ERKO, generated by the Chambon group, is a deletion of exon 3, and has been shown to be a complete ER null (111). Skeletal analysis of these mice revealed no change in femur length of female mice at any age assayed (112), and either decreased femur length (113) or no change (112) in male mice. Old female ERKO mice (>1 year) were found to have fused growth plates, a phenotype not seen in WT mice, or ERKO mice with remaining ER activity (114).
Deletion of ER (BERKO) did not result in any femur length difference in male mice
(108), but resulted in increased bone length in female mice at sexual maturity (peak Estrogen levels) (114). Several groups have analyzed double knockout mice. In most analyses, femur length in DERKO male and ERKO male mice was equivalent, suggesting that ER is sufficient for bone growth. However, the Sims group observed a decrease in femur length in sexually mature DERKO males, as compared to ERKO males (112). It is important to note that the targeting of ER in the latter studies did not result in complete deletion of the gene, as alternatively spliced transcript variants were expressed, although no ER protein was observed(111, 112), thus characterized as a complete ER and ER null mouse line. Female
DERKO mice at sexual maturity displayed either no change in femur length to ERKO or
BERKO females, or an increase in femur length as compared to ERKO females. Female
DERKO mice at old age (>1 year) exhibited an increase femur length compared to ERKO females, and no growth plate fusion although, like ERKO females, these mice had an increase in serum Estrogen levels (114).
Taken together, the data suggest that ER is a positive regulator of bone growth, while
ER is a negative regulator of this process. At sexual maturity, when Estrogen levels are
19 peaking, ER and ER are in balance to keep bone growth stable, and a loss of either results in either an increase (BERKO) or a decrease (ERKO) in bone length.
To examine the role of ER more specifically in bone growth, conditional knockout and knock-in mice have been generated. It has been shown in a Col21-ER-/- mouse model, where
ER has been deleted in a cartilage-specific manner, that both male and female mice exhibited normal long bone growth to sexual maturation (4m). Further, there was no observed increase in serum Estrogen, or decrease in serum IGF-1 (115), which contrasts with the global ERKO mice.
However, both ERKO and Col21-ER-/- mice at 12 months of age had longer bones than WT controls, indicating continued growth past sexual maturation. Because Estrogen is known to reduce growth plate height, sexually mature (4m) Col21-ER-/- and WT mice were gonadectomized and treated with high-dose Estrogen. While WT mice displayed reduced growth plate height, Col21-ER-/- mice failed to respond to treatment, further supporting the importance of ER in the growth plate of sexually mature mice (115). So, while ER has indirect effects on long bone growth in young mice (presumptively through effects on GH/IGF-1 axis), it is necessary for normal decline in growth at sexual maturity. Taken together this suggests that
ER is required both systemically and locally for normal long bone growth. Further supporting this hypothesis is the discovery that a caERColII mouse model (in which constitutively active human ER is expressed in a cartilage specific manner via the Col2a1 promoter) had reduced femur lengths, due to a reduction in chondrocyte proliferation and differentiation (116).
1.4.1.2 Bone remodeling
It is widely known that Estrogen is required to maintain bone mass, as a reduction in Estrogen levels in post-menopausal women results in accelerated bone loss, and contributes to
20 osteoporosis (117). With Estrogen deficiency, the bone formation-resorption balance is skewed towards resorption, with an increase in pro-resorption cytokines and factors, such as interleukin-
1 (IL-1) and tumor necrosis factor (TNF), and a decrease in bone protecting factors, such as
TGF-
ER knockout mouse models have revealed that the bone protective effects of Estrogen are mediated principally through ER The Sims group utilized ERKO, BERKO, and DERKO mice and concluded that the loss of ER resulted in decreased cortical bone mass, increased trabecular BMD, and decreased bone turnover in male and female mice. Further, ER deletion had no effect on male mice, but resulted in decreased bone resorption, and increased trabecular bone volume (112). Further studies on gonadectomized mice showed that ER is crucial to restoring bone mass by E2 treatment in male and female mice (119, 120).
In addition to the global knockout models that have been studied, there are a number of conditional knockout and transgenic mice that have been used to elucidate the principle role of
ER in bone remodeling. A caERColI transgenic mouse was generated, in which a constitutively active human ER construct was expressed in an osteoblast-specific manner via the Col1a1 promoter. It was shown that female caERColI mice had increased BMD and increased trabecular bone volume fraction. Interestingly, this was due to a decrease in osteoclast number, without any observed changes in bone formation parameters. However, it was also found that osteoclast cultures supplemented with medium from caERColI osteoblast conditioned medium had a decrease in osteoclast number, compared to WT, suggesting that the osteoclast defect was secondary to altered production of osteoclastogenic factors by the caERColI osteoblasts (121).
21
A number of osteoblast-specific ERKO mice have been generated recently that support the conclusion that in both genders ER is necessary for proper bone mass and remodeling.
Table 1 summarizes the phenotypes of the conditional knockout mice that have been recently been published.
To investigate the cell autonomous effects of ER in the osteoclast, mice with osteoclast-specific deletions have been studied. ER deletion in mature osteoclasts (EROc/Oc) resulted in trabecular bone loss in female mice, with an increase in bone formation and resorption, with a net increase in resorption. Estrogen treatment did not ameliorate the bone phenotype, suggesting that Estrogen action is mediated by osteoclastic ER (122). The authors showed that while osteoclastogenesis itself was not affected, the proapoptotic effects of
Estrogen on osteoclasts through the FAS/FASL system was impaired. In agreement with the above study, ERdeletion in osteoclast progenitors (ERLysM-/-, specific to macrophages, including osteoclast progenitors) resulted in a significant reduction in cancellous bone in female mice (123). However, whereas Estrogen deprivation (OVX) led to no further bone loss in
EROc/Oc mice, there was a significant reduction in cortical bone of OVX ERLysM-/- mice, although it must be mentioned that EROc/Oc mice were deprived of Estrogen for only 2 weeks, while ERLysM-/- mice were deprived of Estrogen for 6 weeks. Interestingly, in the
ERLysM-/- model, the authors concluded that the proapoptotic effects were most likely mediated by extranuclear effects of ER, and not through binding of ER to its ERE. These apparent differences may be due in part by experimental design and/or stage-specific targeting of osteoclasts.
22
Table 1.1 Osteoblast-specific ERKO mouse models
ERKO Cell affected Sex Phenotype Conclusion Prx1-ERf/f uncommitted F/M -↓Ct. bone formation due to -ER in mesenchymal (124) progenitor reduced periosteal BFR cells is required for Ct. -↑CFU-O/AD/F but ↓ Ob. bone formation by differentiation/mineralization potentiating canonical capacity WNT-mediated -blocked osteoblastogenic osteoblast proliferation actions of canonical Wnt and differentiation pathway Osx1-ERf/f osteoprogenitor F/M -↓Ct. bone formation due to -ER in (124) reduced periosteal BFR osteoprogenitor cells is -↑CFU-AD but ↓ Ob. required for Ct. bone differentiation/mineralization formation potentiating capacity canonical WNT- -blocked osteoblastogenic mediated osteoblast actions of canonical Wnt proliferation and pathway differentiation Col1a1-ERf/f osteoblast F/M -no effect on bone mass -ER is not required in (124) committed Ob ERKOOB/OB late osteoblast F -↓Tb. and Ct. bone formation -late osteoblastic ER (125) due to reduced bone turnover is necessary for bone (Ocn -resistance to trabecular bone formation and proper promoter) loss in age and OVX mice bone turnover pOC-ERKO late osteoblast F -↓Tb. and Ct. bone formation -late osteoblastic ER (126) due to reduced osteoblast is necessary for bone (Ocn number formation and bone promoter) -↓bone strength strength Dmp1-ER -/- osteocyte M -↓Tb. bone formation due to -osteocytic expression (127) reduced osteoblast number of ER positively and activity regulates osteoblastic -n/d in Ct. parameters in bone formation in male response to mechanical mice loading F -n/d in unchallenged mice -osteocytic ERis -attenuated E2 response in dispensable for trabecular bone of OVX physiological females regulation of trabecular bone Dmp1- osteocyte F -↓Tb. bone formation due to -osteocytic ER EROcy/Ocy reduced N.Ob, Ob.S and BFR positively regulates (128) osteoblastic bone formation
23
1.5 Estrogen receptor-related receptors
The ERRs are transcription factors that are part of the nuclear receptor superfamily. The superfamily consists of ligand-dependent receptors (such as the ERs discussed above, vitamin D receptor, and retinoic acid receptors), as well as orphan receptors, so called because they lack known ligands. Included in the group of orphan receptors is the Estrogen receptor-related receptor (ERR) subfamily (NR3B), namely ERR NR3B1), ERR(NR3B2)and
ERR(NR3B3). ERR and ERR were discovered by low stringency screening of cDNA libraries using human ER as a probe (129), and ERR was identified via yeast-2 hybrid screening with glucocorticoid receptor 1 (GR-1) as bait (130).
ERRs, as most nuclear receptors, share a common structure composed of a modulator domain (A/B), DNA binding domain [DBD] (C), hinge region (D), and ligand binding domain
[LBD] (E/F). Transcriptional modulation of their target genes is accomplished through interaction with co-factors (131). Figure 1.3 depicts the modular structure of nuclear receptors.
The A/B, or modulator domain is the most variable domain among nuclear receptors. The domain possesses cell and promoter-specific activities, and interacts with cell-specific co-factors
(132). The A/B domain has been shown to be phosphorylated by MAPK (133) as well as cyclin- dependent protein kinases (134) and pp90rsk1 (135). This is of importance to the orphan nuclear receptors because it may explain how their transcriptional activity is controlled in the absence of a ligand.
The DNA binding domain is responsible for the binding of the receptor to specific DNA response elements, as well as for protein-protein interaction. This region contains two zinc finger motifs that are necessary for binding to the specific response elements and is well conserved among nuclear receptors. Specifically, the ERRs share approximately 68% similarity
24 in the DBD with ERs (136). The DNA binding sequence for ERRs was first identified as
TCAAGGTCA (137-139), and has since been confirmed by bioinformatic and ChIP analysis (140).
ERRs can bind this sequence, known as the Estrogen related response element (ERRE), as well as the Estrogen response elements (ERE) as monomers, homodimers, or heterodimers, with either other ERRs or with ERs (137, 140-146) (Figure 1.4).
The hinge region gives the receptor its flexibility, allowing the interaction between the
DBD and LBD. The hinge allows the protein to rotate up to 180o, allowing the proteins to dimerize on inverted and direct repeat sequences (147).
The LBD has multiple functions: it mediates ligand binding, is a binding surface for cofactors, and contains the ligand dependent transactivation domain (AF2). Functional studies have shown that the LBD is also important for homodimerization for some nuclear receptors
(144). In contrast to the DBD, the LBD is not well conserved between ERRs and ERs, displaying only 36% similarity (136). This relatively low amino acid sequence similarity in the LBD may explain why ERRs do not bind Estrogen, which is consistent with structural studies showing that the ligand-binding pocket of ERR is smaller than that of ER and ER. The crystal structure of ERR ligand binding domain reveals a tightly packed binding pocket filled with side chains, meaning that a major conformational change must occur to allow ligand binding
(148).
25
Figure 1.3 Modular structures of nuclear receptors. Most nuclear receptors share a common E structure composed of a modulator domain (A/B), DNA binding domain [DBD] (C), hinge region (D), and ligand binding domain [LBD] (E).
26
Figure 1.4 ERRs bind DNA. ERRs can bind DNA as monomers, homodimers, and heterodimers, and recognize Estrogen response elements (ERE), as well as Estrogen related response elements (ERRE).
27
1.5.1 ERR ERR was the first ERR identified via a reduced stringency screening of cDNA libraries with ER DBD as probe (129). Although ERR expression is nearly ubiquitous in adult tissues, it is predominant in tissues with high energy demands. The first ERR -/- mouse, although fertile and viable, had reduced fat mass and resistance to obesity. Further, the absence of ERR resulted in changes in gene expression of important regulators of lipid metabolism, in both white and brown adipose tissue (149). Subsequent analyses revealed an important role for ERR in adaptive thermogenesis (150), and adaptation to cardiac pressure overload (151). These studies stress the importance of ERR in regulation of oxidative metabolism and regulation of key mitochondrial regulatory genes.
1.5.1.1 ERR and cartilage
ERR protein has been shown to be expressed throughout the rat growth plate, except for the hypertrophic zone, as well as in articular chondrocytes, and throughout the differentiation time course of the chondrogenic rat cell-line C5.18, matching the expression profile of Sox9 (152). It was also shown in C5.18 cells that ERR activated a Sox9 reporter construct (152), which suggests that ERR may regulate early chondrogenic factors. To further elucidate the role of
ERR in chondrogenesis, cultured C5.18 cells were treated with antisense oligonucleotides during either proliferation or differentiation stages. Knockdown during proliferation resulted in decreased expression of early chondrocyte markers of proliferation and differentiation, including Sox9, Ihh, and CcnD1 (152), and knockdown of ERR during differentiation resulted in decreased expression of Sox9, Col2a1, Ihh and Agg, along with an increase in the hypertrophic chondrocyte marker Col10a1, and increased apoptosis (152). Together, the data suggest that
ERR regulates Sox9 to initiate chondrocyte commitment and maturation.
28
In addition to cartilage development, there is also data suggesting ERR is important in cartilage maintenance. ERR was shown to be decreased in a mouse model of collagen-induced inflammatory arthritis (RA) (153), as well as in human osteoarthritis (OA) patient samples (154).
Further analyses suggested that ERR may be part of the early response to cartilage
(154) degradation, specifically a target of the PGE2-COX2-cAMP-PKA pathway .
1.5.1.2 ERR and bone
Our group first showed that ERR is expressed throughout osteoblast differentiation and regulates bone formation in models in vitro (155). We also showed that ERR protein is expressed in fetal and adult rat tibiae (bone and growth plate) as well as in the fetal rat calvaria
(155). Further, antisense knockdown of ERR resulted in a reduction of mineralized bone nodules in primary osteoblast cultures, while overexpression of an ERR enhanced osteoblast differentiation (155), suggesting ERR has a positive regulatory role in osteogenesis. ERR is also regulated, directly or indirectly, by Estrogen in osteoblasts and bone (156). There have been conflicting data published on ERR activity on bone in vivo, with data supporting both a negative regulatory (157, 158), and a positive regulatory role (159). These latter differences may reflect inherent differences in different knockout mouse strains, differences in activities in female (157, 158) versus male (159) mice, and age-related cell autonomous versus non-autonomous roles in bone cells (osteoblasts, osteoclasts) versus other lineages in the bone environment (e.g., adipocytes) (160).
In addition to its role in osteoblast differentiation, ERR has also been implicated in osteoclast differentiation and function. Wei et al. showed that ERR-/- mice exhibited decreased osteoclast numbers and increased trabecular bone, resulting in an osteopetrotic
29 phenotype (161). They went on to show that the absence of ERR downregulated mitochondrial genes involved in osteoclastogenesis in vitro (161). ERR has also been shown to be an important factor in osteoclast mobility and adhesion. Specifically, loss of ERR resulted in the disruption of the actin ring structure (podosome belt) that is essential for proper bone resorption
(162).
1.5.2 ERR
ERR was first identified in a cardiac cDNA library screen with ERR as a probe (129). It is expressed in tissues with high energy demand for example heart and kidney), albeit at lower levels compared to ERR and ERR. It is also expressed at high levels in extra-embryonic ectoderm in the developing placenta and undifferentiated trophoblast stem cell lines (163), which suggested that ERR plays a role in placental development. The latter was confirmed in a mouse lacking ERR expression; by 8.5dpc, ERR -/- embryos already displayed either decreased or complete absence of a placenta, and no ERR -/- mice were observed beyond 10.5dpc (163).
Histological and in situ hybridization analyses revealed an impairment of trophoblast proliferation and differentiation (163).
In a Sox2-Cre:ERRlox/lox conditional knockout line with ERR expression deleted in embryonic tissue, mice lived to adulthood and displayed defects in inner-ear development and rod photoreceptor survival (164, 165). Interestingly, these mice also showed a decrease in expression of genes regulating carbohydrate and fatty acid metabolism. Recent evidence confirms the important role played by ERR in metabolism. For example, it was shown in the
Sox2-Cre:ERRlox/loxmouse line that ERR regulates multiple aspects of metabolism, with reduced food intake rate and decreased fat mass (166). A Nes-Cre:ERR lox/lox mouse (deletion of
30
ERR in the developing nervous system) was used to further elucidate the mechanism by which
ERR regulates metabolism. Thus, Nes-Cre:ERR lox/lox mice exhibited decreased expression of neuropeptide-Y (Npy), a hindbrain neuropeptide known to modulate energy balance and enhanced insulin sensitivity (166). Finally, it was shown that ERR expression and function were enhanced in both Sox2-Cre:ERRlox/loxandNes-Cre:ERR lox/lox mice (166).
To date there are no published data implicating ERR in either chondrogenesis or osteogenesis, although preliminary data from antisense knockdown experiments in our lab did show an increase in mineralized nodule numbers in osteogenic cultures (Cardelli and Aubin, unpublished findings).
1.5.3 ERR
ERR was the third member of the ERR subfamily to be discovered. It was identified by a yeast two-hybrid system using GRIP-1 as bait to screen a 17.5dpc mouse cDNA library, producing a
1.5kb clone (130). A subsequent amino acid sequence search indicated highest sequence homology to ERR and ERR. ERR shares an almost identical DBD to ERR and ERR, and a LBD that shares 57% similarity to ERR and 73% to ERR.
1.5.3.1 ERRand energy metabolism
As with ERR and ERR, ERR expression is nearly ubiquitous in adult tissues, but is predominant in tissues with high energy demands. Due to its similarity to ERR and its ability to interact with PGC1 and PGC1cofactors that are important to mitochondrial biogenesis (168), it was postulated that ERR also played a role in energy homeostasis. Using
ChIP-on-Chip analysis, Dufour et al. identified 231 promoters occupied by ERR in adult mouse heart (140), with significant overlap with ERR. The putative target genes identified were
31 responsible for a myriad of functions including oxidative phosphorylation, TCA cycle, and cholesterol, glucose and fatty acid metabolism (140), which further supported the idea that ERR played an important role in energy metabolism. It was further shown that although ERR and
ERR shared the same recognition motif and functioned as heterodimers, ERR could also bind the motif in the absence of ERR, suggesting a possible compensatory mechanism.
To further elucidate the function of ERR, the Evans group generated a global ERR -/- mouse line (169). ERR null mice were perinatal lethal due to a defect in heart function, with an observed defect in the ability of the ventricular cardiomyocytes to repolarize. Absence of ERR resulted in a failure to switch from fetal carbohydrate to postnatal fatty acid oxidation (169).
Further investigation revealed that ERR is a master regulator of energy metabolism, responsible for not only mitochondrial genome number, but also regulation of key mitochondrial genes.
1.5.3.2 ERRand skeletogenesis
While ERRhas been extensively investigated for its role in bone and cartilage, the role of
ERR by contrast is poorly understood. There are no data in the literature describing a role for
ERR in chondrogenesis, and only very little evidence of its role in osteogenesis. It has been reported that an ESRRpolymorphism is associated with bone mineral density in females of
European ancestry (170). In vitro reporter assays performed in our lab have revealed that ERR is a better transactivator of the osteopontin promoter than ERR, in both HeLa and ROS (rat osteosarcoma) cell lines (Zirngibl et al, unpublished findings). We have also observed that a truncated form of ERR, where the last 2 helices of the LBD are deleted, is functionally dead in
HeLa cells, and exhibits a repressive action in ROS cells.
32
In pilot studies, I showed that ERR is expressed in bone and cartilage tissues (Figure 1.5A), as well as throughout the differentiation time course of primary osteoblasts cultured from neonatal mouse calvaria (Figure 1.5B). Further, I showed that modulating ERR activity using antisense oligonucleotides led to an increase in bone nodule number (Figure 1.5C), which suggested that
ERR may be a negative regulator of osteogenesis. Nuclear receptor gene expression profiling has also shown ERR to be expressed early in the differentiation time course of mouse calvaria cells (171) and mouse bone marrow derived stromal cells (172). ERR expression was reported to be increased in cultured mouse calvaria cells, as well as the murine myoblast cell line C2C12 upon stimulation BMP2 (173). Further, an ERR reporter construct was transactivated by BMP2 in a dose responsive manner, suggesting that BMP2 regulates ERR activity. The latter authors went on to show that overexpression and knockdown of ERR decreased and increased, respectively, BMP2 stimulated osteoblast differentiation in vitro (173). They further showed that
ERR, through protein interaction with RUNX2, prevented normal cofactor interaction with
RUNX2, resulting in repression of transactivation of its target genes Bsp and Ocn (173). ERR was also recently been reported to regulate BMP-stimulated osteogenesis in vitro through upregulation of the microRNA miR-433, which targeted the 3'-UTR region of Runx2, decreasing its protein expression (174). Taken together, the data suggest that ERR is an important regulator of skeletogenesis.
33
Figure 1.5 ERR is expressed in murine bone and cartilage. (A) ERR gene expression was observed in adult mouse bone and cartilage tissue (B) ERR expression over the proliferation- differentiation time course of rat calvarial cells in culture (C) Modulating ERR activity using antisense oligonucleotides led to an increase in bone nodule number. All values are expressed as mean ± SD
34
1.6 Objectives
As summarized above, the data suggest that ERR is a negative regulator of osteogenesis in vitro, but the proof-of-concept that ERR plays an endogenous role in osteogenesis or skeletogenesis more generally, i.e., in chondrogenesis as well, has not yet been characterized. I undertook both a gain-of-function (Chapter 2) and a loss-of-function (Chapter 3) approach to address this gap.
35
Chapter 2 2 Cartilage-specific overexpression of ERR results in chondrodysplasia and reduced chondrocyte proliferation
Reprinted from: Marco Cardelli 1*, Ralph A. Zirngibl 2*, Jonathan F. Boetto 2*, Kristen P.
McKenzie 2, Tammy-Claire Troy 3, Kursad Turksen 3, and Jane E. Aubin 1,2. Cartilage-specific overexpression of ERR results in chondrodysplasia and reduced chondrocyte proliferation.
PLOS ONE (09 Dec 2013)
1 Department of Medical Biophysics and 2 Department of Molecular Genetics, University of
Toronto, Toronto, Ontario, Canada, M5S 1A8
3 Regenerative Medicine Program, Sprott Centre for Stem Cell Research, Ottawa Hospital
Research Institute, Ottawa, Canada, K1Y 8L6
* Co-First Authors
Acknowledgements
We are grateful to Benoit de Crombrugghe for vectors p3000i3020 and pJ251, Peter Greer for the polio IRES, and Lily Morikawa in the Toronto Centre for Phenogenomics for performing paraffin embedding and histological staining. We also thank other members of the Aubin lab for helpful discussions. Authors’ roles: Study design and conduct: MC, RAZ, JB, KPM, T-C T, KT and JEA. Data collection: MC, RAZ, JB, KPM, T-C T. Data analysis: MC, RAZ, JB, KPM, T-C T, KT and JEA. Data interpretation: MC, RAZ, JB, KPM, T-C T, KT and JEA. Drafting manuscript: MC, RAZ, KPM and JB. Revising manuscript content and approving final version of manuscript: MC, RAZ, KPM, T-C T, KT and JEA. MC, RAZ, JB, KPM, T-C T, KT and JEA take responsibility for the integrity of the data analysis.
36
2.1 Abstract
While the role of Estrogen receptor-related receptor alpha (ERRα) in chondrogenesis has been investigated, the involvement of ERR gamma (ERRγ) has not been determined. To assess the effect of increased ERRγ activity on cartilage development in vivo, we generated two transgenic
(Tg) lines overexpressing ERRγ2 via a chondrocyte-specific promoter; the two lines exhibited
~3 and ~5 fold increased ERRγ2 protein expression respectively in E14.5 Tg versus wild type
(WT) limbs. On postnatal day seven (P7), we observed a 4-10% reduction in the size of the craniofacial, axial and appendicular skeletons in Tg versus WT mice. The reduction in bone length was already present at birth and did not appear to involve bones that are derived via intramembranous bone formation as the bones of the calvaria, clavicle, and the mandible developed normally. Histological analysis of P7 growth plates revealed a reduction in the length of the Tg versus WT growth plate, the majority of which was attributable to a reduced proliferative zone. The reduced proliferative zone paralleled a decrease in the number of Ki67- positive proliferating cells, with no significant change in apoptosis, and was accompanied by large cell-free swaths of cartilage matrix, which extended through multiple zones of the growth plate. Using a bioinformatics approach, we identified known chondrogenesis-associated genes with at least one predicted ERR binding site in their proximal promoters, as well as cell cycle regulators known to be regulated by ERRγ. Of the genes identified, Col2al, Agg, Pth1r, and
Cdkn1b (p27) were significantly upregulated, suggesting that ERRγ2 negatively regulates chondrocyte proliferation and positively regulates matrix synthesis to coordinate growth plate height and organization.
37
2.2 Introduction
The bones of the axial and appendicular skeleton arise from condensations of chondrogenic cells that lay down a cartilaginous scaffold, which is then remodeled to give rise to the ossified bone.
The longitudinal growth of endochondral bones is driven by continued chondrogenesis in the growth plate, which can be divided along the longitudinal axis of the bone into distinct zones comprising resting, proliferating, and post-mitotic chondrocytes, respectively, from the articular surface. The process of chondrogenesis is a highly orchestrated proliferation-differentiation sequence that is regulated by a number of signaling pathways and feedback loops. For example, the homeobox transcription factor, SRY-related high-mobility-group box 9 (SOX9), is the primary determinant of chondrogenesis and is required for the initial commitment of mesenchymal stem cells to the chondrogenic lineage (175). The hedgehog protein family member,
Indian hedgehog (IHH), is also crucial to chondrocyte proliferation and differentiation. It is expressed by prehypertrophic cells and binds to its receptor Patched-1 (PTC-1), which, in turn, activates signaling pathways to promote chondrocyte proliferation (176). IHH also forms a negative regulatory feedback loop with parathyroid hormone-related protein (PTHRP) to delay chondrocyte hypertrophy and increase the pool of proliferating chondrocytes (16). On the other hand, the transcription factor RUNX2 promotes the differentiation of chondrocytes from proliferation to hypertrophy (177).
Certain transcription factors belonging to the nuclear hormone receptor family are also involved in chondrocyte differentiation. These include the two Estrogen-binding receptors,
Estrogen receptor alpha and beta (ERα (NR3A1) and ERβ (NR3A2) respectively), and recent reports highlight a role for ERα in the fusion or slowing down of growth plate chondrogenesis at puberty in humans and mice. For example, in a cartilage-specific ERα-deleted mouse,
38 appendicular bones developed normally, but exposure to high levels of Estrogen failed to reduce bone length as it did in wild type (WT) mice, indicating that ERα was required for the natural deceleration of bone growth that occurs in mice upon sexual maturity (115). Conversely, a mouse line that expressed a constitutively active form of ERα in cartilage exhibited fewer proliferating cells in the growth plate and reduced bone length (116).
Three orphan nuclear receptor genes related to the ERs comprise the Estrogen receptor- related receptor (ERR) family: alpha, beta and gamma (NR3B1, NR3B2, and NR3B3, respectively) (129). These genes share a high degree of similarity with the ERs, including 67% identity in the DNA-binding domain (DBD), but are unable to bind Estrogen (131). With their similarity in their DBD, it is not surprising that there is considerable cross-talk at the level of gene regulation between the ERs and the ERRs. However, X-ray crystallography studies have clearly shown that, unlike the ERs, the ERRs assume an active state without a ligand bound to the ligand binding domain (LBD) (148, 178). Consistent with the hypothesis that the ERRs are constitutive transcriptional activators, in vitro transcription assays demonstrated that ERRα and
ERRγ induce expression of target genes without addition of potential ligand to the media (130,
179). ERRα-/- mice display a significant decrease in body mass and ERRγ-/- mice are perinatal lethal due to cardiac failure (149, 169), phenotypes connected to the roles that both of these isoforms play in energy metabolism.
The role of ERRs in bone and cartilage are also beginning to be investigated, with most data published on ERRα (160). ERRα is expressed in proliferating chondrocytes in vivo and throughout chondrocyte differentiation in vitro (152). In addition, it has been shown that ERRα is dysregulated in murine models of inflammatory arthritis (153), as well as in human osteoarthritis
(154). There is no data in the literature describing ERRγ in cartilage, and only very little described
39 on its role in bone. Results of an epidemiological study of ERRγ polymorphisms in humans indicated a correlation between a subset of ERRγ variants and elevated bone mass (180). In vitro overexpression of ERRγ causes a decrease in the expression of bone sialoprotein (Bsp) and osteocalcin (Ocn), markers of mature osteoblasts, in the MC3T3 pre-osteoblast cell line (173).
Taken together, these results suggest that ERRγ is a negative regulator of osteogenesis. To determine whether ERRγ also has a biologically relevant function in cartilage, we have generated transgenic (Tg) mice with a collagen α1 (II) (Col2a1) promoter driving expression of a full length ERRγ2 (long isoform) transcript (Col2::ERRγ2FL). We report here that overexpression of ERRγ2 in a cartilage-specific manner leads to abnormalities in the axial and appendicular skeletons.
40
2.3 Material and Methods
2.3.1 Ethics Statement All experimental procedures were performed in accordance with protocols approved by the
Canadian Council on Animal Care and the University of Toronto Faculty of Medicine and
Pharmacy Animal Care Committee.
2.3.2 Construction of pCol2a1mERRγ2 and generation of transgenic mice We made a generic transgene vector containing the mouse Col2a1 promoter (gift from B. deCrombrugghe) and flanking intron (181) fused to a splice acceptor site (182) with stop codons in all reading frames, an extensive multiple cloning site (MCS), a Polio internal ribosome entry site
(gift from P.A. Greer), a nuclear localized β-galactosidase (183) and the protamine minigene pA
(184). During the construction of this vector, silent mutations were introduced to abolish restriction sites within the transgene so that the MCS could be expanded for future cloning of cDNAs. The transgene is flanked by NotI restriction sites. The longer 458 amino acid open reading frame for ERRγ2 (NM_001243792) was PCR amplified from mouse muscle cDNA with primers containing restriction enzymes to facilitate cloning. All vectors were verified by sequencing. Transgenic lines carrying the pCol2a1mERRγ2 construct were generated by pronuclear injection, as described previously (185). Hemizygous founders were screened for transmission of the transgene by performing PCR on DNA isolated from tail clips taken from the F1 generation of progeny using primers in the IRES (TGC TCC TTT GAA ATC TTG TGC
A) and LacZ (AAG TTG GGT AAC GCC AGG GT) portion. To determine sex of the P0 or P7 pups, we performed PCR targeting the sex-determining region Y (Sry) (Table 2.1).
41
2.3.3 LacZ stain for detection of transgene Embryos of 14.5 dpc were dissected, and processed for LacZ detection as previously described
(186).
2.3.4 Gene expression analysis Embryonic tissues and adult articular knee cartilage were homogenized using an Ultra Turrax
T25 homogenizer, while postnatal tissues were manually ground with a mortar and pestle under liquid nitrogen. The RNA was extracted using TRIzol (Invitrogen), precipitated in isopropanol, and resuspended in 50-200μL of DEPC dH2O. To remove potential contaminating DNA, RNA samples were subjected to DNase treatment, using an Ambion Turbo DNA-Free kit (Invitrogen), as per the manufacturer’s directions. Three μg of DNased RNA were reverse transcribed using
Superscript II Reverse Transcriptase (Invitrogen), according to the manufacturer’s directions.
All primers were designed with intron inclusion in corresponding genomic DNA, and are common to all potential transcript variants (Table 2.1).
2.3.5 Western Blotting Dissected limbs in PBS were homogenized (Ultra Turrax T25), followed by lysis in RIPA buffer
(50 mM Tris HCl pH 8, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate and 0.1% sodium dodecyl sulphate) with added protease inhibitors. Protein samples were quantified using the Bio-Rad DC Protein Assay kit, following the manufacturer’s instructions. Thirty μg of each sample was run in a 10% SDS-PAGE gel, transferred to polyvinylidene fluoride (PVDF) membrane, followed by blocking in 5% milk-TBS-T for 30 minutes at room temperature.
Immunodetection was carried out using a rabbit polyclonal anti-ERRγ antibody (H38x, Santa
Cruz Biotechnology Inc.) diluted 1:5000 in blocking buffer, or rabbit anti-β-ACTIN antibody diluted to 1:2000 (Sigma). This was followed by a one hour incubation with a goat anti-rabbit
IgG, conjugated to horse radish peroxidase (HRP; Santa Cruz Biotechnology), diluted 1:5000-
42
1:8000 in blocking buffer. The HRP was visualized using the Amersham ECL Western Blotting
Detection kit (GE Healthcare), as per the manufacturer’s instructions. The autoradiographic films from both ERRγ and β-ACTIN detection were scanned and the density and size of the bands were quantified using Image Lab software (Bio-Rad). The ERRγ band was normalized to the β-ACTIN band to assess proportionate protein levels. The 29 amino acid difference between the ERRγ1 and ERRγ2 protein isoforms is not clearly obvious in the gels that we run and we attribute the increased expression we see to the ERRγ2 isoform.
2.3.6 Whole mount skeletal staining P0 and P7 animals were dissected, eviscerated, and fixed in 95% ethanol overnight or up to two weeks, and then processed for whole mount skeletal staining as previously described (187). When samples were fully cleared, skeletons were dissected and photographed in a Petri dish containing
100% glycerol, using a Nikon Coolpix P5100 digital camera affixed to a dissecting scope. The images were then quantified in Image J by taking linear measurements of individual skeletal components.
2.3.7 Histological and immunofluorescence analysis The left limbs taken from each skeleton analyzed by whole mount staining were fixed in 4% paraformaldehyde for 24 hours, transferred to PBS for 2 – 3 days and then decalcified (10%
EDTA, 0.1 M Tris pH 7.4) for 15 days. Once decalcified, limbs were serially dehydrated with one 20 minute wash in 30% ethanol, three 20 minute washes in 50% ethanol and a final wash and storage in 70% ethanol. Samples were paraffin-embedded, sectioned (5μm), and stained
(hematoxylin and eosin, and Safranin-O) at the Toronto Centre for Phenogenomics (TCP). The
Safranin-O-stained slides were photographed under a Nikon Eclipse TS100 fluorescence microscope, at magnifications of 40X and 100X. In the 40X micrographs, measurements were taken from the absolute edge of Safranin-O staining before the articular surface, down to the last
43 hypertrophic chondrocyte before the primary spongiosa. Three such measurements were taken at separate axes along the bone, with the average representing the total growth plate height. Using the 100X micrographs, measurements were taken from the first flattened chondrocyte, which appeared in a distinct column of three or more cells, down to the last hypertrophic chondrocyte, and then the subset of that distance that contained only hypertrophic chondrocytes was taken.
This was repeated at five points along the width of each growth plate, and the averages were used to calculate the specific heights of the resting, proliferative and hypertrophic zones. For each animal, growth plates were assessed from the proximal humerus, distal femur and proximal tibia.
To immunodetect Ki67, a common proliferation marker (184), sections were deparaffinized and rehydrated in ethanol washes, followed by antigen retrieval in boiling citrate buffer (10 mM citric acid, 0.05% Tween-20, pH 6) for up to 20 minutes. The slides were blocked in normal goat serum (Invitrogen) for 30 minutes at room temperature, washed, incubated with rabbit polyclonal anti-Ki67 antibody (diluted 1:25 in blocking buffer) for 1 hour at room temperature, washed, then incubated with goat anti-rabbit Alexa-594-conjugated secondary antibody (Invitrogen, diluted 1:50 in blocking buffer) for 30 minutes at room temperature. The samples were counterstained with Hoechst nuclear stain.
Growth plate sections of the proximal humerus were viewed and imaged using a
Bioquant Osteo imager with Photofluor II fluorescence excitation light, a triple Chroma filter, and Bioquant Osteo 2012. Ki67 positive and negative cells within the proliferative and resting zones were counted in Image J. By calculating the percentage of total cells (defined by the
Hoechst stain) that stained positively for Ki67, the mitotic activity in the proliferative zone and resting zone was measured.
44
2.3.8 TUNEL assay Growth plate sections of the proximal humerus were processed as described above, before using the FragEL DNA fragmentation detection kit (Calbiochem), as per the manufacturer’s instructions, and counterstained with methyl green. Sections were imaged at 16X magnification, using a Nikon Eclipse TS100 fluorescence microscope. Images were analyzed using ImageJ.
Cells within the proliferative and hypertrophic zones were counted for TUNEL analysis and normalized to methyl green stained nuclei.
2.3.9 Expression analysis of putative target genes A list of putative target genes was constructed by in silico analysis of the regulatory regions of factors known to be involved in the chondrocyte differentiation process and that contained at least one ERR binding site (ERRE) in the region from -10Kb to +5Kb of their annotated transcription start site. The search used the ERRE consensus sequence TCAAGGTCA and 20 additional sequence variants that had been identified in the literature (140). The search was performed using the freely available Transcriptional Regulatory Element Database (TRED)
(http://rulai.cshl.edu/cgi-bin/TRED/tred.cgi?process=home), which retrieves genomic sequences from the current Ensembl build of the mouse genome. The results from this search are shown in
Table 2.2, along with the primers used to look for differences in gene expression in Table 2.1.
The primers were chosen to pick up all of the known transcript variants and include at least one intron in the corresponding genomic DNA.
The reactions were performed in triplicate on a 96-well plate in a BioRad MyIQ iCycler, for 50 cycles with an annealing temperature of 59oC. The amplification data was uploaded into the PCR miner program (http://www.ewindup.info/miner/version2/) to obtain the Ct and reaction efficiency values. The relative expression levels of the target gene were normalized to the L32 internal control expression.
45
2.3.10 Statistical Analysis All data were analyzed using Graphpad Prism 4.0 software. When three data sets were analysed,
ANOVA was used first to determine significance, followed by Student’s t-Test. All the graphs are plotted as the mean ± standard deviation and the p values listed are for the comparison to the
WT values. Graphs were constructed using Microsoft Excel 2003 software.
2.3.11 Mouse gene nomenclature We followed the mouse nomenclature guide as stated on the Mouse Genome Informatics web page (http://www.informatics.jax.org/mgihome/nomen/short_gene.shtml). Thus, mouse genes are written with first letter capitalized followed by small letters, all italicized, while proteins are written all capitalized and without italics.
46
Table 2.1 Primer sequences used in gene expression analysis
Gene Upstream Sequence Downstream Sequence
L32 CACAATGTCAAGGAGCTGGAAGT TCTACAATGGCTTTTCGGTTCT
Bmp7 GCACTCAGGCAGGGAGTCGG ACCCAGTGGTTGCTGGTGGC
Ihh TGCTGGCGCGCTTAGCAGTG GCAGCGGCCGAATGCTCAGA
Pthrp ATTCCTACACAAGTCCCAGAG ACTTGCCCTTGTCATGCAGTA
Pth1r GACGTGGGCCAACTACAGCG GTGCAGTGCAGCCGCCTAAA
Fgfr3 TAGCGGCCGCCAGTCTCCAC ACGCAGGCCGGGACTACCAT
Sox9 AATGCTATCTTCAAGGCGCTG GGACCCTGAGATTGCCCAG
Sox6 ACAACCACAGACAGATTGAGCAGC TGCCCCTGCCGAGTTTGGTG
Atf4 ATGGCGTATTAGAGGCAGCA GATTTCGTGAAGAGCGCCAT
Runx2 TGTTCTCTGATCGCCTCAGTG CCTGGGATCTGTAATCTGACTCT
Col2a1 ACTGGTGGAGCAGCAAGAGC TCTGGACGTTAGCGGTGTTG
Col10a1 AACGGTACCAAACGCCCAC CTTTGTTCTCCTCTTACTGGAATCCC
Agg GCGTGAGCATCCCTCAACCATC GGCAGTGGTCACAGGATGCATG
CcD1 CCTGTGCGCCCTCCGTATCT TCATGGCCAGCGGGAAGACC
Cdkn1a CAGACCAGCCTGACAGATTTCTA GAGGGCTAAGGCCGAAGATG
Cdkn1b GTTTCAGACGGTTCCCCGAA TCTTAATTCGGAGCTGTTTACGTC
ERRα TCGAGAGATAGTGGTCACCATCAG CTTCCATCCACACACTCTGCAG
ERRβ TGAGATCACCAAACGGAGGC GAACTCGGTCAAGGCGCA
ERRγ TGTGACTTGGCTGACCGAGA TGGAGGAGGCTCATCTGGTCT
ERα GGCTGCGCAAGTGTTACGAA CATTTCGGCCTTCCAAGTCAT
ERβ TTGGTGTGAAGCAAGATCACTAGAA GACTAGTAACAGGGCTGGCACAA
Sry GAGAGCATGGAGGGCCAT CCACTCCTCTGTGACACT
47
Table 2.2 Results of in silico search for putative ERRE sites in known regulators of growth plate chondrocytes Gene Expression Function Putative ERRE Position (bp) Atf4 PZ & PHZ Proliferation ACAAGGACA -7522 TTAAGGTCA -6669 ACAAGGACA -3214 CAAAGGTCA -606 Bmp7 PHZ Proliferation & TAAAGGTCA -4858 Differentiation TCTAGGTCA -4235 TAAAGGTCA +1692 CAAAGTTCA +3245 CTAAGGTCA +4796 CcnD1 RZ, PZ, PHZ, HZ Proliferation TCGAGGTCA -4175 TCTAGGTCA +1117 Fgfr3 PZ, PHZ, HZ Proliferation & GAGAGGTCA -6661 Differentiation TGAAGGACA +4463 Ihh PHZ Proliferation & TAAAGGTCA -5988 Differentiation TCAAGGTCA -2716 TCAAGGACA +1353 CAAAGGTCA +4046 Pth1r PZ, PHZ Proliferation & CAAAGGTCA -9433 Differentiation CCAAGGTCA -2698 Runx2 PHZ, HZ Differentiation CAAAGTTCA -6895 CAAAGGTCA -3156 CAAAGTTCA +4145 Sox6 RZ, PZ, PHZ Proliferation & TCAAGGACA -6264 Differentiation ATGAGGTCA -2802 TCAAGGTCA -978 Sox9 RZ, PZ, PHZ Proliferation & CCAAGGTCA -8797 Differentiation CCAAGGTCA -2966 Pthrp RZ, PZ, PHZ Proliferation & TTCAGGTCA -2728 Differentiation Col2a1 RZ, PZ, PHZ Cell-matrix TAGAGGTCA -8714 communication CAGAGGTCA -7002 GTAAGGTCA -2325 RZ = resting zone; PZ = proliferative zone; PHZ = pre-hypertrophic zone; HZ = hypertrophic zone. Position is relative to the start site.
48
2.4 Results
2.4.1 Overexpression of ERRγ2 results in dwarfism
To begin to investigate the putative role of ERRγ in chondrogenesis, we first asked whether
ERRγ is expressed in cartilage. We found that ERRγ is expressed in mouse cartilage at levels similar to ERRβ and ERβ, but approximately 50 fold less than ERRα, and 85 fold less than ERα
(Figure 2.1A). To evaluate the consequences of cartilage-specific overexpression of ERRγ, we generated two independent Tg mouse lines, Line 1 (#6486) and Line 2 (#4094), that express the longer protein isoform of ERRγ2 under the control of the collagen α1 (II) promoter
(Col2::ERRγ2FL) (Figure 2.1B). LacZ staining of 14.5dpc embryos revealed distinct staining in the developing craniofacial skeleton and both the axial and appendicular skeletons of Tg mice but not wild type (WT) embryos (Figure 2.1B). Western blot analysis of 14.5 dpc limbs demonstrated increased expression of ERRγ2 protein relative to the β-ACTIN control and quantification demonstrated approximately 3-fold overexpression of ERRγ2 in Line 1, and 5- fold overexpression in Line 2 (Figure 2.1C). Mice were born in the expected Mendelian ratios and up to at least 8 months of age appeared healthy, suggesting that there is no overt detriment due to the integration or expression of the transgene.
Significant anomalies in all endochondral bones studied and measured were already apparent at birth in Tg versus WT mice (data not shown), but data are reported here only for postnatal day 7 (P7). Measurement of P7 pup weight showed no significant difference in Line 1, but a significant reduction in Line 2 compared to WT mice (Figure 2.2A). Quantification revealed a small but significant decrease in total body length (data not shown) as well as in the crown-rump length (Figure 2.2B) in pups of both lines. In P7 whole mount skeletal preparations,
Tg mice also exhibited a shortened snout and domed head (Figure 2.2C) and a significant
49 decrease in skull length, but not width compared to WT mice (Figure 2.2C). Analysis of specific long bones showed a reduction in the length of the femur and the tibia (Figure 2.2D) in both lines, as well as the humerus, ulna, radius and scapula (length but not width was affected) (data not shown). No significant differences were observed in Tg versus WT bones that form by intramembranous ossification, e.g., the mandible and the clavicle (data not shown). To rule out the possibility that the phenotypes we observed were due to the sex of the pups, we genotyped them for gender, and found that the phenotypes we describe occur in both male and female mice. Taken together, the data indicate that targeted ERRγ overexpression in cartilage results in mild dwarfism.
50
Figure 2.1 Col2::ERRγ2 transgenic mouse generation and protein expression analysis. (A) ERR and ER gene expression (relative to L32) in adult mouse cartilage. Graphs represent mean ± SD from a minimum of 3 independent samples. (B) Schematic of the transgene used to express the long protein isoform, ERRγ2, in cartilage. Whole mount LacZ staining of a WT and transgenic 14.5 dpc embryo with a close-up of the forelimbs showing robust staining in the developing cartilage. Some nonspecific staining along the neural tube is evident. (C) Western blots and resulting quantification showing the levels of ERRγ and β-ACTIN proteins in 14.5 dpc limbs of the two transgenic lines compared to WT. Graphs represent the mean ± SD from at least three independent Western blots. ** p≤0.01.
51
Figure 2.2 Skeletal analysis of P7 animals. (A) Body weight measured in P7 pups from WT (n= 19), Line 1 (n=11) and Line 2 (n=9) animals. (B) Crown-rump length measurement showed reduced axial skeleton. Measurements were taken from the snout to the base of the tail from WT (n= 24), Line1 (n=16) and Line 2 (n=13) pups. (C) Close-up photographs of the skulls from WT and Tg animals, and resulting quantification showing reduced length, but not width, in Tg compared to WT animals, WT (n=6) and Line 1 (n=11) pups. (D) Alcian blue/Alizarin red double stain of WT and Line 2 hindlimbs. Quantification shows reduced bone length and mineralized component of the axial skeleton in Tg compared to WT animals. Total femur and tibia length were measured as well as the length of the mineralized portion of the bone as demarcated by the Alizarin red staining for WT (n=22), Line 1 (n=14), Line 2 (n=14). Graphs represent the mean ± SD * p≤0.05, ** p≤0.01, *** p≤0.001.
52
2.4.2 Overexpression of ERRγ impairs chondrocyte proliferation, differentiation-maturation, cartilage matrix production and growth plate organization
Quantification of P7 proximal humerus, distal femur, and proximal tibia revealed trends or significant decreases in total growth plate height as well as the proliferative and hypertrophic zones in Tg versus WT mice (Figure 2.3A and B). The most dramatic decrease was in the proliferative zone, which displayed a 22% height reduction in the transgenic versus WT mice, whereas hypertrophic zone changes were less pronounced and not detectable in all bones
(Figure 2.3B). The quantification of zone heights in the Tg mice was complicated by 2 factors: a smaller and disorganized proliferative zone, and the presence of acellular swaths of cartilage matrix that often spanned the resting, proliferative and hypertrophic zones (Figure 2.3A).
Although we observed 2 WT samples containing acellular swaths, such areas were much smaller and less pronounced than those in Tg samples.
To determine the basis of the reduction in growth plate zone heights, we quantified the number of proliferating cells by immunostaining for the proliferation marker, Ki67 (Figure
2.4A). No significant difference was observed in resting zone chondrocytes, but a 30% decrease in Ki67-positive cells was seen in the Tg compared to WT proliferative zone (Figure 2.4B).
Since acellular masses of matrix have been reported previously in other genetically-modified mice, including ones with unusually wide or generally disorganized growth plates in which hypoxic cell death occurs (175), we next performed a TUNEL assay. No significant difference in
TUNEL staining was observed in either the hypertrophic or proliferative zones of Tg versus WT growth plates (quantified for Line 2; Figure 2.4C). RT-qPCR also revealed no differences in expression of the apoptosis-associated markers, Bax and Bcl2, in RNA isolated from the growth plate of WT versus Tg mice (Figure 2.4D). The data suggest that the reduced length of the
53 proliferative zone is a consequence of decreased chondrocyte proliferation, whereas that of the hypertrophic zone may be due to a disruption in chondrocyte differentiation, a delay in chondrocyte maturation, or secondary effects to the reduction in proliferation.
54
Figure 2.3 Growth plate analysis of P7 pups. (A) Growth plates from WT and Line 2 femurs showing the reduced proliferative zone and the acellular swath observed in the transgenic animals. (B) Analysis of total, resting, proliferating, and hypertrophic growth plate heights from proximal humerus (PH), distal femur (DF) or proximal tibia (PT). The largest decrease in growth plate height was observed in the proliferating zones of Tg animals compared to WT.* p≤0.05, ** p≤0.01, *** p≤0.001. RZ, Resting Zone; PZ, Proliferative Zone; HZ, Hypertrophic Zone; AS, Acellular swath.
55
Figure 2.4 Decreased proliferation is responsible for reduction of growth plate height. (A) Hoechst labeling and immunofluorescence of proximal humerus growth plate from WT and Line 2 samples, showing Ki67 positive cells in the proliferating zone (PZ) (red boxes), and resting zone (RZ) (white box). 2º Ab is a section stained with secondary antibody only and Hoechst. Lower panel shows the proliferative zone (red box) and part of the resting zone (blue box) under higher magnification, highlighting the difference in Ki67 positive cells within the proliferative zone of WT and Tg mice (B) Quantification of Ki67 positive cells from Line 2 proximal humerus growth plates demonstrated a clear reduction in the number of proliferating cells in the proliferating zone compared to the resting zone. WT (n=6), Line 2 (n=6); *** p≤0.001 (C) Quantification of TUNEL positive cells shows no difference in apoptotic cells in either the proliferating or hypertrophic zones. (D) Gene expression for apoptotic markers, Bax and Bcl2, from 14.5 dpc limbs confirms lack of increased apoptosis.
56
To investigate the molecular basis of changes observed in the growth plate as a consequence of overexpression of ERRγ2, we next used an in silico database search to screen for putative ERR binding sites in the region spanning 10 Kb upstream to 5 Kb downstream of the transcriptional start site in a variety of genes known to be involved in chondrogenesis. The genes that met both criteria are listed in Table 1 and included transcription factors (Sox9, Sox6,
Atf4, and Runx2), extracellular signaling molecules (Ihh, Bmp7, and Pthrp), hormone and growth factor receptors (Pth1r and Fgfr3), and cartilage matrix proteins (Col2a1, Col10a1, and
Agg). Because we observed a reduction in proliferation within the proliferative zone of transgenic animals, and because ERRγ has been shown to suppress proliferation (188), we also assessed the expression of cell cycle regulators cyclin D1 (CcnD1), and cyclin dependant kinase inhibitor 1a and 1b (Cdkn1a and Cdkn1b). RT-qPCR revealed a significant increase in Pth1r
(Figure 2.5A), but no differences in the transcription factors tested (Figure 2.5B). We also observed an increase in cartilage matrix proteins Col2a1 and Agg (Figure 2.5C) and the cell cycle regulator Cdkn1b (p27) (Figure 2.5D) in Line 2 (higher overexpresser). This suggests that overexpression of ERRγ decreases chondrocyte proliferation through regulation of Cdkn1b, impacts growth plate organization, matrix synthesis as evidenced by upregulation of Col2a1 and
Agg, and affects chondrocyte maturation through upregulation of Pth1r.
57
Figure 2.5 Expression profile of putative target genes. Genes are grouped as extracellular growth factors (A), transcription factors (B), ECM proteins (C), and cell cycle regulators (D). Of the genes tested, we observed increased Pth1r, Col2a1, Agg, and Cdkn1b expression in 14.5 dpc limbs of Tg compared to WT animals. Expression by RT-qPCR of putative target genes was normalized to the expression of L32. WT (n=7), Line 2 (n=9), * p≤0.05
58
2.5 Discussion
We report here that overexpression of ERRγ2 in chondrocytes results in decreased axial and appendicular skeleton size and disruption of growth plate height and organization. This phenotype was manifested already in newborn mice, indicating an effect on embryonic skeletal development that persisted postnatally, and included changes in chondrocyte proliferation, differentiation, and maturation, as well as matrix production, suggesting a role for ERRγ in chondrogenesis.
In the two independent Col2::ERRγ2FL mouse lines analyzed, ERRγ2 is overexpressed at moderate levels (3 to 5 fold higher levels than endogenous ERRγ expression in WT littermates) and the resultant skeletal phenotype and concomitant changes in gene expression observed are relatively subtle and ERRγ dose-dependent. It therefore seems likely that the phenotype seen in Col2::ERRγ2FL mice is a direct consequence of modulation of ERRγ transcriptional activity rather than from disruption of the transcriptional machinery.
Nevertheless, it remains possible that the effects of ERRγ2 overexpression are indirect, e.g., resulting from an imbalance in the ratio of the ERRγ1/ERRγ2 protein isoforms and sequestration of required transcription cofactors (see also below). Thus, further studies in ERRγ knockout mice, in particular chondrocyte-specific ERRγ knockout mice to circumvent the perinatal lethality seen in global ERRγ knockout mice (169) (and Cardelli and Aubin, unpublished data), and in Col2::ERRγ1 overexpressing mouse lines, are of interest.
The most pronounced effect observed in the Tg growth plate was the reduction in height of the proliferative zone, with a 22% reduction in proliferative zone length and a 30% decrease in the percentage of proliferative Tg compared to WT cells. Despite the fact that CcnD1 contains 2 putative ERREs in its regulatory region (Table 1), we did not observe any difference
59 in CcnD1 expression in WT versus Tg 14.5 dpc limbs. However, we observed increased expression of the cyclin-dependent kinase inhibitor Cdkn1b in Tg limbs, suggesting regulation of chondrocyte proliferation through regulation of this kinase inhibitor, a possibility consistent with data showing that ERRγ suppresses S-phase progression in an in vitro model of prostate cancer through transactivation of Cdkn1a and Cdkn1b (188). A scan of the Cdkn1b regulatory region reveals 3 putative ERRE binding motifs at positions -5077, -2625, and -2618, suggesting potential direct regulation, however we cannot rule out a novel, indirect protein interaction.
In addition to the changes in cell cycle regulators and proliferation, we observed slightly reduced hypertrophic zone length in some bones, marked disorganization of the proliferative and hypertrophic zones, and acellular swaths that spanned a large portion of the Tg growth plate, which were not due to an increase in apoptosis. Moreover, Col2a1, Agg, and Pth1r expression were increased in 14.5dpc Col2::ERRγ2FL mice. Taken together, the data suggest that ERRγ plays a role in coordinating chondrocyte proliferation-differentiation-matrix synthesis, but whether the changes are directly and causally related remain to be determined.
Nevertheless, increased COL2A1, a cartilage extracellular matrix (ECM) protein that binds to integrins and activates signaling pathways essential to chondrocyte proliferation, may contribute to both acellular swaths of cartilage and decreased chondrocyte proliferation, via a negative feedback loop between abundant matrix production and proliferation. For example, it has been shown in chondrocytes from mouse ribcage that lack of β1 integrin results in reduced chondrocyte motility and COL2A1 adhesion, as well as reduced CcnD1, and increased Cdkn1a expression, resulting in decreased chondrocyte proliferation (47). Further, it has been shown that
ERRα is involved in osteoclast migration and adhesion, in part through regulation of β3 integrin
(162). It is necessary to determine whether integrin expression and/or chondrocyte adhesion and motility are affected in the ERRγ-overexpressing mice.
60
Whether Col2a1 is regulated directly by ERRγ or indirectly through interaction with a factor upstream of Col2a1 remains to be determined. It has been shown that overexpression of
NKX3.2 in vitro increases Col2a1 expression in a SOX9-independent manner, by directly binding to a 48bp chondrocyte-specific enhancer in the Col2a1 regulatory region (189). Scanning the regulatory region using a core 'AGGTCA' sequence reveals 3 putative ERREs through which
ERRγ could directly control Col2a1 expression. Alternatively, ERRγ may regulate Col2a1 expression through the recruitment of cofactors, such as PGC1-α and CBP/P300, as has been shown in SOX9-dependent regulation of Col2a1 (190, 191). Interestingly, it has been shown that
PGC1-α is a cofactor of ERRγ (168), and it is possible that ERRγ and SOX9, together with PGC1-
α, are part of a larger transcriptional complex to regulate Col2a1 expression. Intriguingly, one of the ERREs in the Col2a1 regulatory region is near the SOX9 binding site, supporting the hypothesis of a larger transcriptional regulatory complex.
Analyses of several genetically-engineered mouse models have revealed the importance of FGF (192, 193) and IHH (16, 73, 194) signaling in chondrocyte proliferation and differentiation.
Recently, it has been reported that several transcription factors regulate the transcription of Ihh, including RUNX2 (65), ATF4 (195), and MSX2 (196). The phenotype we observed in the ERRγ- overexpressing mice did not result in transcriptional changes in Ihh or Fgfr3, suggesting either a regulatory mechanism involving a transcriptional complex with one of the above mentioned transcription factors or regulation independent of IHH or FGFR3.
It will also be important to further analyze how ERRγ regulates chondrocyte maturation and hypertrophy, as no significant changes were detected in the hypertrophy markers assessed, such as Runx2, or its target gene Col10a1. While this may reflect simply the small content of hypertrophic cells in the 14.5 dpc samples utilized, it is worth noting that we also detected no
61 differences in Col10a1 expression in 17.5 dpc or P7 bone samples (data not shown). This suggests that ERRγ may have only a small or secondary role in chondrocyte hypertrophy.
Alternatively, the modest difference we observe in the Tg hypertrophic zone height may be too small to quantify significant differences in hypertrophic marker gene expression. As mentioned, the presence of the acellular region in the Tg animals, made measurement of the growth plate zones difficult, and may have influenced the observation of significant differences where observed. On the other hand, the increase in Pth1r expression seen in 14.5 dpc Tg animals may contribute to a reduction in the hypertrophic zone through delay in the differentiation process.
Pth1r knockout mice exhibit shortened limbs characterized by decreased proliferating chondrocytes. However, they also display premature hypertrophy at early embryonic stages (197).
By contrast, transgenic mice that constitutively express Pth1r display a severe delay in the endochondral process, including a reduced zone of hypertrophy (198). Although we did not observe any differences in hypertrophic gene expression in our mouse model, additional studies on the growth plate of early embryonic mice are needed to elucidate the basis of hypertrophic zone anomalies.
In addition to the above factors, Estrogen and expression of soluble and membrane- bound Estrogen receptors (ERs and GPR30, respectively) have also been implicated in regulating growth plate chondrogenesis. Mice expressing Col2a1 promoter-driven ERα have reduced proliferation and differentiation, and subsequent dwarfism (116), while mice with cartilage-specific inactivation of ERα exhibit prolonged longitudinal bone growth (115). Taken together, the evidence suggests that not only ERα, but also ERRγ are negative regulators of chondrocyte proliferation and differentiation, which appear to be opposite to the function of
ERRα (152). Further, ERRα is able to form heterodimers with ERRγ (144) and ERα (199), suggesting that regulation of chondrocyte proliferation and differentiation may require a carefully
62 controlled balance of the nuclear receptors. It remains to be elucidated if any interaction occurs between ERRγ and ERRα during chondrogenesis. It is also possible that the high ERRγ2 expression in our model may turn ERR transcriptional activation into repression. It was demonstrated that ERRα and ERRγ can independently activate an ERRE-driven promoter reporter, but when heterodimers of ERRα/γ were formed, the same reporter was suppressed (144).
In summary, overexpression of ERRγ2 in a cartilage-specific manner leads to dose- dependent abnormalities in the axial and appendicular skeletons due to alterations in Cdkn1b expression and chondrocyte proliferation as well as differentiation-maturation- matrix synthesis.
Work is ongoing to characterize further the mechanism by which ERRγ exerts its actions in the developing growth plate.
63
Chapter 3 3 ERR is not required for skeletal development but is a RUNX2-dependent negative regulator of postnatal bone formation in male mice
Reprinted from: Marco Cardelli 1 and Jane E. Aubin 1,2. ERR is not required for skeletal development but is a RUNX2-dependent negative regulator of postnatal bone formation in male mice. PLOS ONE (14 Oct 2014)
1 Department of Medical Biophysics and 2 Department of Molecular Genetics, University of
Toronto, Toronto, Ontario, Canada, M5S 1A8
Acknowledgements
We are grateful to Ruolin Guo for the CT data collection, Ralph Zirngibl for the ERR2FL,
ERRAF2, ERRC148G stable MC3T3-E1 cell lines, and Renée Bernatchez in the Centre for Bone and Periodontal Research for performing methylmethacrylate embedding and histological staining. We also thank other members of the Aubin lab for helpful discussions. This work was supported by a Canadian Institutes of Health Research (CIHR) operating grant (FRN 88104, JEA) and a Doctoral Research Award (MC). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Authors’ roles: Study design and conduct: MC and JEA. Data collection: MC. Data analysis: MC and JEA. Data interpretation: MC and JEA. Drafting manuscript: MC. Revising manuscript content and approving final version of manuscript: MC and JEA. MC and JEA take responsibility for the integrity of the data analysis.
64
3.1 Abstract
To assess the effects of the orphan nuclear Estrogen receptor-related receptor gamma (ERRγ) deficiency on skeletal development and bone turnover, we utilized an ERR global knockout mouse line. While we observed no gross morphological anomalies or difference in skeletal length in newborn mice, by 8 weeks of age ERR +/- males but not females exhibited increased trabecular bone, which was further increased by 14 weeks. The increase in trabecular bone was due to an increase in active osteoblasts on the bone surface, without detectable alterations in osteoclast number or activity. Consistent with the histomorphometric results, we observed an increase in gene expression of the bone formation markers alkaline phosphatase (Alp) and bone sialoprotein (Bsp) in bone and increase in serum ALP, but no change in the osteoclast regulators receptor activator of NF-κB ligand (RANKL) and osteoprotegerin (OPG) or the resorption marker carboxy-terminal collagen crosslinks (CTX). More colony forming units-alkaline phosphatase and -osteoblast (CFU-ALP, CFU-O respectively) but not CFU-fibroblast (CFU-F) formed in ERR +/- versus ERR +/+ stromal cell cultures, suggesting that ERR negatively regulates osteoblast differentiation and matrix mineralization but not mesenchymal precursor number. By co-immunoprecipitation experiments, we found that ERR and RUNX2 interact in an ERR DNA binding domain (DBD)-dependent manner. Treatment of post-confluent differentiating bone marrow stromal cell cultures with Runx2 antisense oligonucleotides resulted in a reduction of CFU-ALP/CFU-O in ERR +/- but not ERR +/+ mice compared to their corresponding sense controls. Our data indicate that ERR is not required for skeletal development but is a sex-dependent negative regulator of postnatal bone formation, acting in a
RUNX2- and apparently differentiation stage-dependent manner.
65
3.2 Introduction
The Estrogen receptor-related receptors (ERR) are orphan nuclear receptors comprising three family members: ERR, ERR and ERR (NR3B1, NR3B2 and NR3B3 respectively) (200).
They are similar in structure to the classic Estrogen receptors, ER and ER, with high amino acid identity in their DNA binding domain (DBD; e.g., over 60% in human ERR and ER), but lower (e.g., less than 35%) identity in the ligand binding domain (LBD); the low sequence identity in the LBD is consistent with the observation that the ERRs do not bind Estrogen (129,
199). Mouse knockout studies have revealed that ERR and ERR are important regulators of energy metabolism (140, 151, 160, 169). ERR in particular is a key regulator of mitochondrial genes, and its absence results in perinatal lethality, as a consequence of a failure to transition from carbohydrate dependence to fatty acid oxidation (169).
The role of ERRs in bone formation and turnover is also being investigated. ERR is expressed in osteoblasts throughout the skeleton and was shown to be a positive regulator of osteoblast proliferation and differentiation in vitro (155). There have been conflicting data published on ERR activity on bone in vivo, with data supporting both a negative regulatory (157,
158) and a positive regulatory (159) role. These latter differences may reflect inherent differences in different knockout mouse strains, differences in activities in female (157, 158) versus male (159) mice, and age-related cell autonomous versus non-autonomous roles in bone cells (osteoblasts, osteoclasts) versus other lineages in the bone environment (e.g., adipocytes) (160).
There are fewer studies addressing the role of ERR in bone. It has been reported that an
ESRRpolymorphism is associated with bone mineral density in females of European ancestry
(170). Nuclear receptor gene expression profiling has shown ERR to be expressed early in the
66 differentiation time course of mouse calvaria cells (171) and bone marrow-derived stromal cells
(172) in vitro, with a steady decrease in expression as differentiation progresses. ERR expression was reported to be increased in mouse calvaria cells in culture upon stimulation by bone morphogenetic protein 2 (BMP2) and, through protein interaction with RUNX2, to prevent normal cofactor interaction, resulting in repression of transactivation of its target genes, bone sialoprotein (Bsp) and osteocalcin (Ocn) (173). ERR was also recently reported to regulate BMP- stimulated osteogenesis in vitro through up-regulation of the microRNA miR-433, which targeted the 3'-UTR region of Runx2, decreasing its protein expression (174). However, the endogenous role of ERR during early bone development and postnatal bone turnover has not yet been characterized.
We report here that ERR is not required for skeletal development but is a sex-dependent negative regulator of postnatal bone formation. Further, we show that ERR acts in a RUNX2- and apparently differentiation stage-dependant manner in vitro.
67
3.3 Materials and Methods
3.3.1 Ethics Statement
All experimental procedures were performed in accordance with protocols approved by the Canadian Council on Animal Care and the University of Toronto Faculty of Medicine and Pharmacy Animal Care Committee
3.3.2 Animals
The ERR knockout mouse line was obtained from the Mutant Mouse Regional Resource
Center (MMRRC). The ERR null mouse strain originated from Deltagen (San Carlos, CA) by insertion of a LacZ-neomycin cassette into ERR such that the endogenous gene promoter drives expression of beta-galactosidase and nucleotides from base 586 to 610 of exon 2 were deleted.
DNA was isolated from either yolk sacs (embryonic mice) or tail clips (postnatal mice) and were genotyped using the following primers:
ERR forward: TGGTTTCCATGGGAATGGTCTTGAG
ERR reverse: TCTCTTTCCCTCAAACTCTCTGCTG
Neo: GGGCCAGCTCATTCCTCCCACTCAT
Breeding was performed by crossing heterozygous male with female C57BL/6J mice.
For all experiments, littermates were used as controls. Animals were sacrificed by cervical dislocation.
3.3.3 Whole mount skeletal staining E15 - P0 animals were dissected, eviscerated and fixed in 95% ethanol overnight or up to two weeks, and then processed for whole mount skeletal staining as previously described (187). When samples were fully cleared, skeletons were dissected and photographed in a Petri dish containing
68
100% glycerol, using a Nikon Coolpix P5100 digital camera affixed to a dissecting scope. The images were then quantified in Image J by taking linear measurements of individual skeletal components.
3.3.4 Microcomputed tomography (CT)
8, 14, and 52-week old mouse femurs were dissected and stored in 70% ethanol. CT imaging was performed on a GE Xplore Locus SP imager. A manual trace beginning from just below to
2 mm below the growth plate of the distal femur was used to analyze the cancellous bone. The cortical bone region of interest was defined as a 2mm long region beginning 2mm below the growth plate. For P0 mice the whole femur was analyzed. Quantification was performed by an observer blinded to genotype.
3.3.5 Histology and histomorphometry
14-week old male femurs were dissected and fixed in 4% paraformaldehyde (PFA) overnight at
4oC, dehydrated and stored in 70% ethanol prior to methylmethacrylate (MMA) embedding and sectioning. The distal femur was used to evaluate all histomorphometric parameters. 5m sections were double stained for Von Kossa/toluidine blue to evaluate osteoblast properties, or tartrate-resistant acid phosphatase (TRAP)-stained according to manufacturer’s instructions
(387A TRAP kit, Sigma-Alderich) to evaluate osteoclast properties. These measurements were performed on 3 separate sections from each animal, and the average number or surface was calculated. To evaluate dynamic properties, mice were injected intraperitoneally with 30g/g body weight of calcein (C0875, Sigma-Alderich) 10 and 3 days prior to dissection. Mineral apposition rate (MAR) was calculated by measuring the distance between 2 calcein labels and dividing by 7 days. Four different regions within the trabecular area were used for calculations.
BFR was calculated by multiplication of MAR by the ratio of mineralizing surface (MS)
69
(calcein positive) to the bone surface (BS). Five images were taken at different regions within a section, and 3 different sections were used per animal, and the average was calculated. All quantification was performed by observers blinded to genotype.
3.3.6 Immunohistochemical detection of Ki67 and TUNEL assay To immunodetect Ki67, a common proliferation marker (184), sections were de-plasticized, and rehydrated in ethanol washes, followed by antigen retrieval in citrate buffer (10 mM citric acid,
0.05% Tween-20, pH 6) in a 65oC water bath overnight. The slides were blocked in normal goat serum (Invitrogen) for 30 minutes at room temperature, washed, incubated with rabbit polyclonal anti-Ki67 antibody (diluted 1:25 in blocking buffer) for 1 hour at room temperature, washed, then incubated with biotinylated goat anti-rabbit secondary antibody, and visualized using the Vectastain Elite ABC kit (Vector Labs, Burlingame, California), and counterstained with methylene blue.
To perform TUNEL assay, femoral sections were processed as described above, before using the FragEL DNA fragmentation detection kit (Calbiochem), as per the manufacturer’s instructions, and counterstained with methyl green. In each case, femoral sections were imaged, and Ki67 (or TUNEL) positive and negative cells on the trabecular bone surface were quantified using Bioquant Osteo 2012; only cells with the morphological characteristics of osteoblasts were quantified. All quantification was performed by observers blinded to genotype.
3.3.7 Serum biochemistry
Whole blood was collected through the saphenous vein, and the plasma was separated from whole blood by centrifugation and stored at -80° C until biochemical analysis (Vita-Tech,
Ontario, Canada). OPG and RANKL in serum were assayed using a Quantikine M Murine OPG
ELISA kit and a Quantikine M Murine TRANCE/RANKL ELISA kit (No.MOP00 and
70
No.MTR00, R&D Systems, Minneapolis, MN), respectively following the manufacturer’s directions. CTX-1 levels in serum was determined from fasted mice using Serum CrossLaps®
ELISA (RatLaps EIA No. AC-06F1, Immunodiagnostic Systems, Fountain Hills, AZ). Serum testosterone levels were quantified by Cornell University Animal Health Diagnostic Center
(Ithaca, NY).
3.3.8 Gene expression analysis The trabecular bone from 14-week old male femur and tibiae were dissected, had marrow removed, and were manually ground with a mortar and pestle under liquid nitrogen. Stromal cell cultures were rinsed twice with phosphate buffered saline. In either case, total RNA was extracted using TRIzol (Invitrogen), and reverse transcribed using Superscript II Reverse
Transcriptase (Invitrogen), according to the manufacturer’s directions. All primers were designed with intron inclusion in corresponding genomic DNA, and are common to all potential transcript variants (Table 1). The reactions were performed in triplicate on a 96-well plate in a
BioRad MyIQ iCycler, for 50 cycles with an annealing temperature of 59oC. The amplification data was uploaded into the PCR miner program (http://www.ewindup.info/miner/version2/) to obtain the Ct and reaction efficiency values. The relative expression levels of the target gene were normalized to expression of ribosomal protein L32 as internal control; we confirmed that the expression of L32 remained constant throughout cultures under the conditions used, including in knockdown experiments.
3.3.9 Isolation of Bone Marrow Cells, Runx2 antisense assay, and CFU assay Bone marrow cells were isolated from dissected tibiae and femora, using a modification of a previously published method (201). Cells were plated in α-MEM supplemented with 10% heat- inactivated fetal bovine serum (FBS) and antibiotics (1 IU penicillin, 1μg/mL streptomycin,
71
50μg/mL gentamicin, 250ng/mL fungizone) (standard medium) at 3x106 nucleated cells/35-mm dish. After 4 days, the medium was changed to differentiation medium (standard medium with
50 µg/mL ascorbic acid and 10 mM β-glycerophosphate). For Runx2 antisense assay, differentiation medium was supplemented with 10M Runx2 oligonucleotides from d5 - d19 of culture. Runx2 oligonucleotide sequences were as follows, as previously described (202):
Runx2A* G*T*G* TGG TAG TGA GTG GT*G* G*C; Runx2S G*C*C* ACC ACT CAC
TAC CA*C* A*C (* denotes phosphorothioate modification) (Integrated DNA Technologies,
Coralville, Iowa). At day 19, cultures were stained for ALP activity and mineralization (Von
Kossa), counted, then re-stained with methylene blue.
3.3.10 Stable cell line constructs The plasmids pcDINmERR2, pcDINmERR2AF2 and pcDINmERR2C148G have been described (203). For the generation of stable overexpressing MC3T3-E1 clone 26 cell lines, 20g of plasmid was linearized with SspI and precipitated with ethanol and then air dried. The plasmid pellet was resuspended in 200l of MEM and then combined with 200l MEM with
40l Lipofectamine 2000 (Invitrogen) before being added onto the cells at 60% confluence in a
10cm dish. The following day medium was changed to allow the cells to recover. On day two after transfection, regular medium was supplemented with 180g/ml (final) of active G418. One non-transfected plate of cells was used as a control. After 2-3 weeks of selection more than 10 colonies could easily be identified on the plate for any of the transfected plasmids and the G418 was reduced to 120g/ml until the plate grew confluent. Cells were split for expansion and freezing. When thawing out the vials for experiments, the cells were plated in 180g/ml of
G418 for initial expansion to ensure high levels of expression, but subsequent passaging was in
72 medium containing 120g/ml of G418. High levels of expression were verified by RT-qPCR and Western blot.
3.3.11 Co-immunoprecipitation (Co-IP) Analysis
Co-IP assay was performed essentially as previously reported (173). Briefly, stable overexpressing MC3T3-E1 cell lines were lysed in Co-IP buffer, and lysates were precleared with 50l of pansorbin cells (Calbiochem) for 2 h, which were removed by centrifugation. A total of 2g of rabbit polyclonal antibody against RUNX2 (Santa Cruz Biotechnology), or normal rabbit IgG (negative control) were added to the precleared lysates, and incubated at 4oC overnight with rotation. After washing cells in Co-IP buffer, the samples were centrifuged, and
SDS sample buffer was added, and boiled. The immunoprecipitated complexes were separated by SDS-PAGE, transferred to polyvinylidene difluoride, and immunoblotted with a rabbit polyclonal antibody against ERR (Santa Cruz Biotechnology).
3.3.12 Western blotting
Whole cell extracts were lysed in RIPA buffer (50 mM Tris HCl pH 8, 150 mM NaCl, 1% NP-
40, 0.5% sodium deoxycholate and 0.1% sodium dodecyl sulphate) with added protease inhibitors. Protein samples were quantified using the Bio-Rad DC Protein Assay kit, following the manufacturer’s instructions. Thirty μg of each sample was run in a 10% SDS-PAGE gel, transferred to polyvinylidene fluoride (PVDF) membrane, followed by blocking in 5% milk-
TBS-T for 30 minutes at room temperature. Immunodetection was carried out using a rabbit polyclonal anti-ERR antibody (H38x, Santa Cruz Biotechnology Inc.) diluted 1:5000 in blocking buffer, or rabbit polyclonal anti-RUNX2 antibody (Santa Cruz Biotechnology) diluted to 1:1000, or rabbit anti--ACTIN antibody diluted to 1:2000 (Sigma). This was followed by a one hour incubation with a goat anti-rabbit IgG, conjugated to horse radish peroxidase (HRP;
73
Santa Cruz Biotechnology), diluted 1:5000-1:8000 in blocking buffer. The HRP was visualized using WEST-one Western Blot Detection, as per the manufacturer’s instructions. Densitometry was performed using Image Lab software (Bio-Rad). The proteins of interest were normalized to the -ACTIN band to assess proportionate protein levels.
3.3.13 Statistical Analysis All data were analyzed using Graphpad Prism 4.0 software, or Microsoft Excel 2003 software.
In most cases, datasets were compared using student's t-Test. When three datasets were analysed, ANOVA was used to determine significance. All the graphs are plotted as the mean ± standard deviation and the p values listed are for the comparison to the WT values. Graphs were constructed using Microsoft Excel 2003 software.
3.3.14 Mouse gene nomenclature We followed the mouse nomenclature guide as stated on the Mouse Genome Informatics web page (http://www.informatics.jax.org/mgihome/nomen/short_gene.shtml). This guide states that genes should be written with first letter capitalized followed by small letters, all italicized, while proteins are written all capitalized and non-italics.
74
Table 3.1 Primer sequences used in gene expression analysis
Gene Upstream sequence Downstream sequence
L32 CACAATGTCAAGGAGCTGGAAGT TCTACAATGGCTTTTCGGTTCT
Bmp2 CTCAGCGAATTTGAGTTGAGGC GGCTTCTAGTTGATGGAACGTG
Runx2 TGTTCTCTGATCGCCTCAGTG CCTGGGATCTGTAATCTGACTCT
Osx ATGGCGTCCTCTCTGCTTG TGAAAGGTCAGCGTATGGCTT
Alp CCAACTCTTTTGTGCCAGAGA GGCTACATTGGTGTTGAGCTTTT
Bsp CAGGGAGGCAGTGACTCTTC AGTGTGGAAAGTGTGGCGTT
Ocn CTGACCTCACAGATCCCAAGC TGGTCTGATAGCTCGTCACAAG
Nfatc1 CAGCTGTTCCTTCAGCCAAT GGAGGTGATCTCGATTCTCG
Ctsk ACCCATATGTGGGCCAGGATGA GAGATGGGTCCTACCCGCGC
75
3.4 Results
3.4.1 ERR -/- mice die perinatally, but display no skeletal abnormalities
ERR null mice (ERR -/-) died perinatally, with no pups observed beyond P1. This is consistent with the observation that mice of a different ERR -/- mouse strain, made using a similar gene targeting strategy, also die perinatally due to mitochondrial abnormalities and impaired oxidative metabolism (169). ERR +/- mice were healthy and viable, and were used for subsequent breeding.
No significant differences in forelimb length were observed between E15-P0 in
ERR+/+, ERR +/- and ERR -/- whole mount skeletal preparations (Figure 3.1A, B, P0 shown). There was also no significant difference observed in any of the growth plate zones in sections of P0 humeri (data not shown). MicroCT (CT) analyses confirmed no significant difference in bone mineral density, content, or bone volume fraction between genotypes in P0 mice (data not shown). Taken together, the data indicate that ERR, similarly to ERs and
ERR(108, 149, 157, 158), either plays no apparent role or plays a redundant role in embryonic/early postnatal endochondral bone growth and mineralization.
76
Figure 3.1 Neither ERR ablation nor haploinsufficiency has a detectable effect on embryonic bone development and growth. (A) Whole mount alizarin red/alcian blue staining revealed no observable morphological differences in newborn mouse ERR +/+, ERR +/- and ERR -/- skeletons (upper panel) or in the craniofacial regions (lower panel). (B) Skeletal elements that make up the mouse forelimb of each genotype were measured and found to have no significant differences. (N = 5 ERR +/+; N = 3 ERR -/-). Values are expressed as mean ± SD
77
3.4.2 Adult ERR +/- male mice have increased trabecular bone, increased osteoblast number and surface, but no change in osteoclast number or surface
To address whether ERR plays a role in adult bone, we first quantified ERR by immunoblotting of whole cell lysates of femoral trabecular bone of 14-week old mice; ERR protein was reduced approximately 67% in ERR +/- mice compared to ERR +/+ (Figure
3.2A). CT analyses of 8- and 14-week old male mice (representative image shown in Figure
3.2B) was performed. Trabecular bone volume (BV/TV) and thickness (Tb.Th) were increased
(21% and 10.8%, respectively) at 14 weeks, while trabecular separation (Tb.S) showed a trend
(p = 0.069) towards a decrease at 8 weeks, and a significant decrease (16.5%) at 14 weeks in
ERR +/- versus ERR +/+ mice (Figure 2C). Trabecular number (Tb.N) was significantly increased (12.4%) at 8, but not 14 weeks of age. The differences observed were specific to trabecular bone, as there were no changes detectable in cortical bone parameters, including cortical area (Ct.Ar) and cortical BMD (Ct.BMD) (Figure 3.2C), or when periosteal and endosteal perimeters were quantified (data not shown). Further, no significant differences were observed in any of these parameters in female ERR +/- mice (Figure 3.2D).
78
Figure 3.2 Trabecular bone formation is increased in 14-week old male distal femurs. (A) Quantification of Western blots revealed that ERR protein expression was reduced in whole cell lysates of trabecular bones of 14-week old ERR +/- mice compared to ERR +/+ mice. (B) Representative CT images of 14-week old ERR +/+ and ERR +/- male distal femurs. (C) Quantitative analysis revealed a significant increase in trabecular bone volume fraction (BV/TV) and thickness (Tb.Th), and a decrease in separation (Tb.S) at 14 weeks. Trabecular number (Tb.N) was significantly increased at 8 weeks, but not significantly at 14 weeks (upper panels). There were no significant differences in cortical bone area (Ct.Ar) or bone mineral density (BMD) (lower panel). (D) Analysis on 8, 14 and 52-week old female mice revealed no differences in any bone parameters assessed. Values are expressed as mean ± SD (Male: 8w: N = 7 ERR +/+; N = 5 ERR +/-; 14w: N = 10 ERR +/+; N = 14 ERR +/-; Female: 8w: N = 3 ERR +/+; N = 5 ERR +/-; 14w: N = 10 ERR +/+; N = 7 ERR +/-; 52w: N = 6 ERR +/+; N = 5 ERR +/- ) * = p<0.05; ** = p<0.01; # = 0.069
79
Histomorphometric analyses indicated that osteoblast number per bone surface
(N.Ob/BS) and osteoblast surface per bone surface (Ob.S/BS) were increased in 14-week ERR
+/- compared to ERR +/+ male femurs (Figure 3.3A). Ki67, an established marker of proliferating cells, and TUNEL staining indicated that neither changes in proliferation nor in apoptosis could account for this increase (Figure 3.3B). Consistent with the increased N.Ob/BS and Ob.S/BS, bone formation rate (BFR) and mineralizing surface per bone surface (MS/BS) were significantly increased, with no significant difference in mineral apposition rate (MAR)
(Figure 3.3C). No significant difference was seen in either osteoclast number per bone surface
(N.Oc/BS) or osteoclast surface per bone surface (Oc.S/BS) in ERR +/- versus ERR +/+ trabecular bone (Figure 3.3D).
Consistent with the morphometric analyses, a statistically significant increase in serum alkaline phosphatise (ALP) levels, a marker of bone formation, was observed in ERR +/- compared to ERR +/+ mice, while no significant differences were seen between genotypes in serum RANKL, OPG or CTX (Figure 3.4A). Because adult male ER knockout mice also exhibit high trabecular bone volume and high serum levels of testosterone, which account for the bone phenotype (112, 204), we assessed serum testosterone levels; no significant difference in testosterone levels between genotypes were detectable (Figure 3.4A).
Taken together, the data indicate that the high trabecular bone phenotype observed in
ERR +/- male mice was due to an increased number of active osteoblasts but no change in osteoclast number, activity (Figure 3.3D) or osteoclast marker (Nfatc1, Ctsk) gene expression
(Figure 3.4B), suggesting that ERR regulates bone formation but not bone resorption in male but not female mice.
80
Figure 3.3 ERR regulates osteoblast number and differentiation. (A) von Kossa/toluidine blue double stain was used to quantify osteoblast number and osteoblast surface per bone surface; both are significantly increased in 14 week ERR +/- compared to ERR +/+ male femurs. Scale = 50m (N = 3 ERR +/+; N = 3 ERR +/-) (B) Adjacent femoral sections were immunostained for proliferating (Ki67) and apoptotic (TUNEL) osteoblasts; no significant differences were seen between genotypes. (N = 5 ERR +/+; N = 4 ERR +/-) (C) Double calcein labels were used to quantify bone formation rate (BFR), mineralizing surface per bone surface (MS/BS) and mineral apposition rate (MAR); both BFR and MS/BS were significantly increased, while MAR was not significantly increased in ERR +/- compared to ERR +/+ mice. (N = 3 ERR +/+; N = 3 ERR +/-) (D) There was no significant difference observed between genotypes in either TRAP-positive osteoclast number (Oc.N/BS) or osteoclast surface per bone surface (Oc.S/BS) (N = 5 ERR +/+; N = 4 ERR +/-). All values are expressed as mean ± SD. * = p<0.05; ** = p<0.01
81
Figure 3.4 ERR regulates markers of osteoblast differentiation and mineralization. (A) Serum analysis of bone turnover markers revealed a significant increase in ALP, an indication of increased bone formation, with no changes in the bone resorption marker CTX or osteoclastogenic factors (OPG, RANKL) or serum testosterone. (N = 4 ERR +/+; N = 4 ERR +/-) (B) Gene expression analysis of trabecular bone from 14-week old mice revealed an increase in formation markers and BMP target genes Alp and Bsp, but no difference in Bmp2, Runx2, Osx, Ocn, and no difference in the osteoclast markers Nfatc1 and Ctsk. (N = 8 ERR +/+; N = 6 ERR +/-). All values are expressed as mean ± SD. * = p<0.05
82
3.4.3 Increased osteoblast differentiation in ERR +/- mice
The data thus far suggest that ERR negatively regulates bone formation through regulation of osteoblast number. We therefore next assessed osteoblast differentiation in primary cultures of bone marrow stromal cells from ERR +/- and ERR +/+ mouse hind limbs. Colony forming unit-osteoblast (CFU-O) and CFU-alkaline phosphatase (CFU-ALP) numbers were increased in
ERR +/- versus ERR +/+ stromal cell cultures, without an increase in CFU-fibroblast (CFU-
F) number (Figure 3.5A). Consistent with the in vivo analyses (Figure 3.3B), no difference was detectable in proliferation of stromal cells between genotypes (Figure 3.5A).
Because ERR has previously been implicated as a negative regulator of BMP- stimulated osteogenesis in vitro (173), we next assessed expression of Bmp and its downstream targets in vivo and in vitro. Runx2 expression was unchanged in stromal cell cultures (Figure
5B), consistent with the results of Jeong et al. (173), or in trabecular bone (Figure 3.4B) of 14- week old ERR +/- versus ERR +/+ male mice. Bmp2 expression was also either unchanged
(Figure 3.4B) or exhibited a downward trend (p = 0.06) (Figure 3.5B) in ERR +/- versus ERR
+/+ trabecular bone and differentiated stromal cells respectively. Expression of Alp and Bsp but not Osx or Ocn was increased in both trabecular bone (Figure 3.4B) and stromal cells (Figure
3.5B) of ERR +/- versus ERR +/+ mice. There were no significant differences in expression of osteocyte markers (Dmp1 and Sost) or in ER, ER, ERR or ERR in either stromal cell cultures or in trabecular bone samples (data not shown). Taken together, the data suggest that
ERR negatively regulates osteoblast differentiation and matrix mineralization at a stage after osteoblast commitment (i.e., after upregulation of Runx2 and Osx expression) but before osteoblast maturation or terminal differentiation to osteocytes (i.e., prior to acquisition of Ocn expression).
83
Figure 3.5 Osteoblast differentiation is increased in ERR +/- stromal cell cultures. (A) Bone marrow stromal cells cultured from ERR +/- mice had significant increases in CFU-O and CFU-ALP, with no change in CFU-F, compared to cells from ERR +/+ mice. Cell counts performed during the proliferation phase revealed no differences between genotypes (N = 8 ERR +/+; N = 8 ERR +/-). (B) Gene expression analysis of ERR +/- versus ERR +/+ cultured stromal cells revealed a trend to decreased expression of Bmp, and a significant increase in Alp and Bsp (N = 6 ERR +/+; N = 6 ERR +/-). All values are expressed as mean ± SD. # = p=0.055; * = p<0.05; ** = p<0.01; *** p<0.001
84
3.4.4 ERR interacts with RUNX2, and knockdown of RUNX2 leads to at least partial rescue of the ERR +/- skeletal phenotype in vitro
ERR has been reported to outcompete P300 as a co-factor for RUNX2, resulting in repression of RUNX2 transactivity (173). However, it has not been reported which region(s) of ERRis/are required for this interaction, but both the DNA binding domain (DBD) and ligand binding domain (LBD) contain dimerization interfaces that modulate the regulation of target genes (205).
We therefore next explored the potential interaction between ERR and RUNX2 proteins by doing co-immunoprecipitation (Co-IP) of ERR and RUNX2 in MC3T3-E1 preosteoblast cell lines, transfected with either a full length ERR (ERR2), a LBD mutant (ERRAF2), or a
DBD mutant (ERRC148G). All three ERR proteins were expressed, but only ERR2 and
ERRAF2 and not ERRC148G co-immunoprecipitated with RUNX2 (Figure 6A), confirming the physical interaction of ERR and RUNX2 and revealing that the DBD is required for this interaction.
To determine whether RUNX2 is required for the ERR +/- osteoblast phenotype, we treated bone marrow stromal cells with antisense Runx2 oligonucleotides post-confluence throughout the differentiation/mineralization period. Runx2 AS but not S oligonucleotides significantly decreased RUNX2 expression (Figure 6B). Consistent with the observation that abrogation of Runx2 expression has no overt effect on cells at certain differentiation stages, i.e., in cells already committed to the osteoblast lineage (206), we observed no decrease in CFU-O or
CFU-ALP in Runx2 antisense-treated ERR +/+ cultures treated in the time window tested
(Figure 6C). In contrast, CFU-O and CFU-ALP were significantly decreased in ERR +/- cultures (Figure 6C). Taken together, the data suggest that the enhanced osteoblast
85 differentiation seen in ERR +/- mice was both RUNX2-dependent and differentiation stage- dependent.
86
Figure 3.6 ERR interacts with RUNX2 through the DBD and modulates osteoblast differentiation. (A) The MC3T3-E1 preosteoblastic cell line was stably transfected with the indicated plasmids. Whole cell extracts (WCE) from these cell lines were probed with anti- ERR, confirming protein expression in the stable constructs. A co-immunoprecipitation assay revealed interaction between RUNX2 and full length ERR (ERR2), as well as with the LBD mutant form of ERR (ERRAF2), but not the ERR DBD mutant (ERRC148G). (B) Decreased protein levels of RUNX2 in Runx2 antisense (Runx2AS), compared to Runx2 sense (Runx2S) samples. (C) CFU-O and CFU-ALP in ERR +/- bone marrow stromal cells treated with Runx2 antisense oligonucleotides (Runx2AS) were significantly reduced while no changes in either CFU-O or CFU-ALP were observed in ERR +/+ cultures. CFU-F numbers were unchanged regardless of treatment and genotype. All values are expressed as mean ± SD. Graphs shown in (C) are of one representative sample, done in triplicate. In all cases, statistical analyses were performed on a minimum of 3 biological samples. ** = p<0.01
87
3.5 Discussion
We report that whereas ERR-deficient mice manifested no detectable defects in embryonic or early postnatal development, ERR-deficient male mice exhibited increased trabecular, but not cortical bone volume in 14-week animals due to RUNX2-dependent negative regulation of the progression of osteoblast differentiation. Female mice were unaffected by ERRdeficiency
The increase in trabecular bone volume seen in ERR +/- male mice was due to increased trabecular number and thickness, as a consequence of an increased number of osteoblasts on the bone surface, increased BFR and increased MS/BS ratio, indicating an overall increase in active osteoblasts occupying the bone surface. We found no significant change in osteoclast number or surface, suggesting that ERR regulates bone formation and osteoblast differentiation, but not bone resorption. This was further supported by the observation of increased expression of certain osteoblast markers in the trabecular bone, with no change in markers of osteoclast formation or activity. Our findings contrast with ER deficiency in male mice, which also leads to increased trabecular bone volume, but due to decreased bone turnover, with decreases in both bone formation and resorption (112). On the other hand, ERR deficiency also results in increased trabecular bone volume, due to increased bone formation, resulting from increased osteoblast differentiation; this was variously reported to be accompanied by no differences in CFU number or proliferation in female mice only (157) or with an increase in osteoblast proliferation, with no increase in CFU number in both male and female mice; these discrepancies may reflect differences in mouse strains and/or the targeting strategies used (158).
In any case, we detected no significant changes by CT analysis in ERR +/- female mice at the same ages at which differences were seen in ERR +/- versus ERR +/+ male mice.
88
The fact that a bone consequence of ERR haploinsufficiency is seen only in adult male mice, and not in developing or neonatal mice or in female mice at any age tested is of interest.
This is likely not due to age-related changes in expression levels of ERR, since we have found
ERR to be expressed at relatively low levels in male and female bone (similar to ER levels) at all ages tested, and expression is flat/constitutively low in cultures undergoing osteoblast differentiation (data not shown). It is possible that ERR has a redundant role in endochondral bone or growth plate development and bone growth, but we also observed no change in BV/TV in newborn mice, which suggests that ERR is not necessary for osteoblast differentiation or precursor fate at developmental or neonatal stages, but rather plays a regulatory role in mice only at sexual maturity. This suggests that negative regulation of osteoblast differentiation by
ERRis sex hormone-dependent. Regulation of sex hormone receptor-dependent bone phenotypic outcomes is complex and multifaceted, but it is notable that sex-dependent adult bone phenotypes have also been seen in ER knockout mice. For example, ER ablation decreased bone turnover and increased trabecular bone mass in both male and female mice, whereas ER ablation decreased bone resorption and increased trabecular bone mass only in female mice (112). The elevated BV/TV observed in the male gonad-intact ER knockouts was found to be due to the high serum testosterone levels and androgen receptor (AR) function (112).
We therefore tested but found no difference in serum testosterone levels between ERR +/- and
ERR +/+ mice. While complete elucidation of the underlying direct and indirect mechanisms for sex-dependent effects on bone of ERs or AR in a background of altered ERR expression is ongoing, interactions between ER and ERR signaling pathways and transcriptional activity would not be unexpected to play roles in mice with altered expression of ERRs. Specifically with respect to ERREstrogen via ER has been shown to regulate ERR expression in breast
89 cancer cells (207). Transcriptional coactivators such as peroxisome proliferator-activated receptor
(PPAR) coactivator-1 (PGC-1) and PGC-1amongst others, play important roles in both
ER- and ERR-mediated transcription (167, 208-210), suggesting that transcriptional cofactors are partially shared between ERs and ERRs, such that altered levels of ERR may alter cofactor stoichiometry for ERs and other ERRs. ERRs can bind to both Estrogen response elements
(EREs) and ERR response elements (ERREs) (137, 141, 142, 199), suggesting that ERRs can also affect ER-mediated signaling. Consistent with this, ERR has been shown to modulate ER responsiveness in prostate, breast, and uterine endometrial cancers (188, 207, 211). Thus, we postulate that ERR haploinsufficiency may alter the actions of ER in male mouse bone, an effect potentially abrogated by ER activities and Estrogen signaling in females. Of course, since ERRs can form not only homo- but heterodimers with each other and ERs (141, 143, 144, 179,
199, 212, 213), we cannot rule out ER-independent effects of ERR on bone target genes and bone or a role for ERR or potentially ERR in the bone phenotype seen.
Our observations on bone marrow stromal cells in vitro indicate that ERR deficiency leads to an increase in osteoblast differentiation (increase in CFU-ALP and CFU-O), without a change in the number of mesenchymal precursors (CFU-F) or proliferation. Further, we observed an increase in Alp and Bsp, markers of osteoblast differentiation and activity. Although we cannot rule out the possibility that ERR directly regulates these genes, previously, Jeong et al. showed that ERR repressed the transcriptional activity of RUNX2 on Bsp and Ocn (173); they also showed that ERR physically interacted with RUNX2, an interaction that repressed
RUNX2 transactivity on a 6x OSE-luc promoter construct (173). We now show that ERR interacts with RUNX2 in vitro, and that the DBD is necessary for this protein interaction, indicating that the DBD is responsible not only for binding to response elements, but also for
90 proper protein-protein interaction. It has been reported previously that both the LBD and the
DBD contain regions for dimerization (144). Interestingly, it has been shown that ER interacts with RUNX2 in MCF7 breast cancer cell line, and the MC3T3-E1 pre-osteoblast cell line, and that this interaction was Estrogen dependant, and resulted in the inhibition of RUNX2 activity
(214). Further, these authors reported that the DBD harboured the interaction motif that directly binds RUNX2. Thus, one can hypothesize that a larger transcription complex, including ERR and ER may form to modulate RUNX2 activity, and regulate osteoblast differentiation/mineralization. Our data suggest that ERR may bind to DNA while interacting with RUNX2. There is no reported putative ERRE within the 6x OSE response element, and it has not yet been shown that ERR can bind this region directly.
Alp and Bsp, but not Runx2, Osx or Ocn are upregulated in ERR +/- compared to ERR
+/+ trabecular bone samples and in differentiating stromal cell cultures, suggesting that ERR negatively regulates osteoblast differentiation and matrix mineralization at a stage after osteoblast commitment (i.e., after upregulation of Runx2 and Osx expression) but before osteoblast maturation or terminal differentiation to osteocytes (i.e., prior to acquisition of Ocn expression, and expression of osteocyte markers Dmp1 and Sost [data not shown]). This, together with the observation that Runx2 antisense treatment of post-confluent differentiating stromal cell cultures significantly reduced CFU-O and CFU-ALP in ERR +/- cultures compared to sense controls, but not in ERR +/+ cultures, supports the conclusion not only that
RUNX2 is required for the ERR +/- osteoblast phenotype but that ERR's role is differentiation stage-dependent. This is interesting in view of recent data that Runx2 also has differentiation stage-dependent effects, with no effect of knockdown or knockout at certain developmental times. Thus, although RUNX2 was originally thought to be necessary at all stages of osteoblast
91 differentiation, it has now been shown via analysis of the 1(I)-collagen-Cre;Runx2flox/flox
(2.3kb 1(I)-collagen promoter) conditional knockout mouse that Runx2 does not have an overt effect in cells already committed to the osteoblast lineage (206). On the other hand, several kinds of data, including results from the global Runx2 knockout and an OG2-Cbfa1 transgenic mouse indicate that RUNX2 is required in uncommitted osteoprogenitors (66) and in mature
OCN-expressing osteoblasts (215). The lack of effect of Runx2 antisense knockdown in ERR +/+ cells is consistent with the fact that the majority of osteoblast precursors in the mouse cultures under the experimental conditions used are already committed and differentiating at the time of antisense treatment (216). Further analysis of additional genetically-modified mouse models will be useful for extending the evidence that ERR is a differentiation stage-dependent negative regulator of osteoblast differentiation.
In summary, this is the first report of the consequences of ERR ablation and haploinsufficiency on bone metabolism in vivo and extends previous reports of ERR function in osteoblasts to the whole animal level. Our data indicate that ERR is a sex-dependent and
RUNX2-dependent negative regulator of osteoblast differentiation, and that its regulatory activity may also be differentiation stage-dependent.
92
Chapter 4 4 Conclusions and Future Directions
As outlined in the Objectives to the work undertaken in this thesis, my goal was to provide proof-of-concept that ERR plays a role in skeletogenesis, by using mouse gain-of-function and loss-of-function models. Cartilage-specific overexpression of ERRγ2 led to dose-dependent abnormalities in mouse axial and appendicular skeletons; histomorphometric and cellular analyses led me to conclude that ERRγ negatively regulates chondrocyte proliferation and positively regulates matrix synthesis to coordinate growth plate height and organization
(Chapter 2). On the other hand, the bone abnormalities seen with ERR ablation and haploinsufficiency in mice and additional studies completed to dissect the underlying mechanisms, prompted me to conclude that ERR is a sex-dependent and RUNX2-dependent negative regulator of osteoblast differentiation, and that its regulatory activity may also be differentiation stage-dependent (Chapter 3). We observed newborn skeletal anomalies which worsened with age in our overexpressing mouse model while early developmental events in skeletogenesis were not detectably affected but manifested in mature ERR null mice. Taken together, the data support an important negative regulatory role for ERR in the postnatal skeleton, and open the door for additional studies to further address the mechanism(s) by which
ERR acts in the skeleton.
Overexpression of ERR in a cartilage-specific manner led to a decrease in the size of the postnatal axial and appendicular skeletons and disruption of growth plate height and organization. The number of proliferating cells as well as the length of the proliferation zone were decreased, and large acellular swaths and general disorganizaton were evident in our
93
Col2::ERR2FL mouse transgenic model compared to WT mice. An important question remains whether ERR directly regulates the genes whose expression was also affected in
Col2::ERR2FL mice, and which of these was functionally responsible for the skeletal changes observed. For example, Cdkn1b (p27) expression was upregulated. Given that Cdnk1b is expressed throughout the growth plate, that its deletion leads to increased proliferative cells within the proliferative zone of the growth plate (217), and that ERR suppresses proliferation of an in vitro model of prostate cancer by binding to response elements within the Cdnk1b promoter region (188), we speculate that Cdkn1b (p27) expression is directly regulated by ERR and is at least partly responsible for the growth plate anomalies in the Col2::ERR2FL mouse model. Although I have not shown whether corresponding protein levels were affected, expression of the matrix genes Col2 and Agg was increased in Col2::ERR2FL mice. It is possible that increases in COL2 and AGG contributed to the presence of acellular swaths and chondrodysplasia, as the former is important to chondrocyte motility and proliferation (47), and the latter important for chondrocyte-chondrocyte and chondrocyte-matrix interaction (218). All 3 of these genes contain multiple putative ERREs within their regulatory regions so an important next step would be to perform ChIP assays to test whether the genes are direct targets of ERR
I have also shown that ERR negatively regulates osteoblast differentiation and mineralization in male mice, and my data suggest this is in a RUNX2-dependent manner, through protein-protein interaction. Scanning the regulatory region of Runx2 does not reveal putative ERREs, but it has been shown that ERR can regulate the transcription of Runx2 through induction of the microRNA-433 (miR-433) in the mouse embryonic mesenchymal stem cell line, C3H10T1/2, inhibiting BMP-stimulated osteoblast differentiation (174). Taken together, the data suggest that ERR may regulate osteoblast differentiation through one or more
94 mechanisms (e.g., direct transcriptional regulation of Runx2 alone or as part of a larger complex, and/or via regulation of other osteoblast-associated genes and/or other genes involved in bone formation) which I discuss further below.
Although my data suggest that ERR regulation of osteoblast differentiation is RUNX2- dependent, whether solely through regulation of RUNX2 or through regulation of RUNX2 and/or direct regulation of target genes within the osteogenic pathway remains unknown. For example, examination of the regulatory regions of Bsp and Ocn, two osteoblast-associated genes with known roles in osteogenesis, reveals multiple putative ERREs to which ERR may bind, either alone or as part of a larger complex with RUNX2, and directly repress Bsp and Ocn expression. I have shown that ERR interacts with RUNX2 and that this interaction requires the
DBD, but I do not yet know whether the interaction between ERR and RUNX2 is direct. Data from Jeong et al. (2009), by way of GST assay, suggest that this interaction is direct (173), and it would be beneficial to confirm this in our system. It is possible that ERR forms a protein complex with RUNX2 on a putative ERRE within the Bsp or Ocn regulatory region.
Alternatively, ERR may interact with RUNX2 on the RUNX2 response elements on the Bsp and Ocn regulatory regions, as was shown in vitro (173). Thus, while it is known that the DBD is also responsible for protein-protein interaction, we cannot rule out the possibility that ERR is binding an ERRE to perform its actions. In fact, I have noted putative ERREs proximal to putative RUNX2 binding sequences (TCCCACA for Bsp; AACCACA for Ocn) in the Bsp and
Ocn promoters, which further supports the suggestion of a larger transcriptional regulatory complex. A next step would be to perform a ChIP assay targeting the putative ERREs within the mentioned target genes. To further elucidate the roles of the consensus sequences, I would
95 construct clones of said regulatory regions with mutated or deleted ERRE/RUNX2 response elements.
To further elucidate the functional interaction between ERR and RUNX2, it would be interesting to extend our findings to the whole animal level by crossing the ERR +/- mouse with the Runx2 +/- mouse line to generate the compound heterozygous ERR +/-;Runx2 +/- mouse. I hypothesize, based on my in vitro results, that the osteopetrotic phenotype observed in
ERR+/- male mice, and perhaps amelioration of the cleidocranial dysplasia phenotype CCD), a hallmark feature of Runx2 +/- mice (219), would be at least partially rescued in ERR +/-;Runx2
+/- mice.
Whether other genes not analyzed in these studies play roles in the phenotype observed in the ERRdeficientmice is also of interest. In this regard, the global ERR knockout mouse that was first reported by the Evans group was used to show, through ChIP-on-chip analysis, that ERR is a master regulator of metabolism and is necessary for transition from carbohydrate to fatty acid oxidation (169). In a more recent study with a slow-twitch muscle-specific ERR- overexpressing mouse (ERRGO) line, it was shown that ERR regulates the angiogenic factor
Vegfa (220), which is known to be essential for normal endochondral ossification. While this result is not entirely surprising, considering the positive effects that ERR has on muscle metabolism, what is interesting is that Vegfa is a transcriptional target of RUNX2. This suggests that ERR could form a complex with RUNX2 to activate Vegfa, whose consequences one would expect to be opposite to what we observed in our mouse model. However, as noted above, the ability for regulation of Vegfa depends on other putative transcriptional partners available and involved in the complex. For instance, it was shown in the ERRGO mouse that
96
ERR was able to activate the Vegfa promoter in the absence of its transcriptional co-factor
PGC1, which has been otherwise shown to enhance the transactivity of ERR. Further, it has been shown that ERR can suppress proliferation in a prostate cancer model (188), while it was also shown to promote proliferation in an ER positive breast cancer cell line (207). From the data, it appears that ERR affects proliferation in a cell and cofactor-specific manner. To this point, I have not examined the expression levels of Vegfa or Pgc1 (or the presence of PGC1 protein) in the mouse models, but it would be interesting to investigate Vegfa regulation by
ERR with and without its classic cofactors in an osteoblast cell system.
The mouse models that I used were important to extend in vitro data and provide proof of concept for a role for ERR in skeletogenesis, but to further investigate the role of ERR and to circumvent the perinatal lethality that I encountered in the global ERR -/- mouse, it will be necessary to generate tissue-specific ERR knockout mice. Generation of a Col2::ERR-/- mouse may very well elucidate the role of ERR in cartilage, in light of our findings discussed in Chapter 2, but the lack of phenotype observed in the ERR -/- embryonic/newborn skeleton would suggest either a compensatory mechanism in development, perhaps involving ERRor
ERRor that the skeletal effects of ERR are modulated by sex hormones and their receptors, as suggested by my data on the mouse models. However, our current mouse models do not allow us to decipher what role estrogen and its receptors may play. An important next step could include crossing either of our models to an ER null mouse, which would disturb the estrogen affect on the skeleton and allow us to better understand any possible interaction between ERR and ER. To examine the local effects on the skeleton without disturbing systemic
97 estrogen/ER activity, we could use the conditional models that I been discussed in the
Introduction (Table 1.1).
By deleting ERR in not only an osteoblast-specific manner, but a stage-specific manner
I would be able to further our understanding of how ERR regulates osteoblast differentiation/mineralization. A number of osteoblast specific Cre mouse lines have been established, including Col1a1-Cre (221), Ocn-Cre (70) and Dmp1-Cre (222), which drive the expression of Cre recombinase under the control of the above promoters, which are expressed at different developmental stages of the osteoblast. By crossing these mice with a mouse carrying a floxed ERR allele (223), I would be able to delete ERR expression at discreet stages of osteoblast differentiation. While no osteoblast-specific models have been generated for aberrant expression of ERRs, there are multiple ER null models reported (summary in Table 1.1).
While it was shown that ER plays a role in many of the osteoblast stages examined, there was no phenotype observed in the Col1a1-ERf/f (124), which suggests ER is not overtly required in committed osteoblasts. Interestingly, it has also been shown via analysis of the 1(I)-collagen-
Cre;Runx2flox/flox (2.3kb 1(I)-collagen promoter) conditional knockout mouse that Runx2 does not have an overt effect in cells already committed to the osteoblast lineage (206). This conflicts with data from a different 2.3kb 1(I)-collagen Runx2 knockout model (Runx2 E8/E8, deletion of exon 8), which resulted in decreased bone formation and resorption parameters, with a net loss in bone (224). However, it must be noted that exon 8 is not common to all Runx2 transcript variants, and the deletion construct was still able to transactivate the Ocn and Sost promoters, whereas the 1(I)-collagen-Cre;Runx2flox/flox targeted exon 4, common to all Runx2 transcript variants, and did not induce target promoters (206). In light of these reports, it makes the generation of a Col1a1-ERR-/- mouse ever more crucial, and could potentially shed light on
98 the stage-specific effects of ERR. It would then be possible to perform similar in vitro knockdown experiments we reported in chapter 3, this time with the conditional ERR model.
In summary, the work presented in this thesis has begun to bridge the gap of our understanding of the regulatory role of ERR in skeletogenesis. Using both gain of function and loss of function models, I have shown that ERR is a negative regulator of chondrocyte proliferation and osteoblast differentiation. Future work will concentrate on investigating the transcriptional complex and how ERR interacts within it as well as how ERR itself is regulated in the chondrocyte and osteoblast. In this way, we can begin to better complete the picture of the regulatory network involved in skeletogenesis. At the same time, this foundational work opens new door to future clinical application. My data suggest that ERR would be an ideal candidate for small molecule screening as a path towards a therapeutic target for such pathologic conditions as osteoarthritis and osteoporosis. Taken further, based on our data indicating interaction between ERR and RUNX2, a small molecule modulating ERR activity could have a therapeutic effect on individuals suffering from CCD.
99
References
1. St-Jacques B, Helms JA. Prenatal Bone Development. In: Glorieux FH, editor. Pediatric Bone: biology and diseases. Amsterdam ; Boston: Academic Press; 2003. p. 77-115.
2. Karsenty G, Wagner EF. Reaching a genetic and molecular understanding of skeletal development. Dev Cell. 2002;2(4):389-406.
3. Akiyama H, Chaboissier MC, Martin JF, Schedl A, de Crombrugghe B. The transcription factor Sox9 has essential roles in successive steps of the chondrocyte differentiation pathway and is required for expression of Sox5 and Sox6. Genes & development. 2002;16(21):2813-28.
4. Zelzer E, McLean W, Ng YS, et al. Skeletal defects in VEGF(120/120) mice reveal multiple roles for VEGF in skeletogenesis. Development. 2002;129(8):1893-904.
5. Kronenberg HM. Developmental regulation of the growth plate. Nature. 2003;423(6937):332-6.
6. Cormack DaH, A. Ham's Histology. Philadelphia: Lippincott Williams & Wilkins; 1987.
7. Stains JP, Civitelli R. Cell-to-cell interactions in bone. Biochem Biophys Res Commun. 2005;328(3):721-7.
8. Gevers EF, van der Eerden BC, Karperien M, Raap AK, Robinson IC, Wit JM. Localization and regulation of the growth hormone receptor and growth hormone- binding protein in the rat growth plate. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2002;17(8):1408-19.
9. Melmed S. Insulin-like growth factor I--a prototypic peripheral-paracrine hormone? Endocrinology. 1999;140(9):3879-80.
10. Ohlsson C, Mohan S, Sjogren K, et al. The role of liver-derived insulin-like growth factor-I. Endocr Rev. 2009;30(5):494-535.
11. Sjogren K, Liu JL, Blad K, et al. Liver-derived insulin-like growth factor I (IGF-I) is the principal source of IGF-I in blood but is not required for postnatal body growth in mice. Proceedings of the National Academy of Sciences of the United States of America. 1999;96(12):7088-92.
12. Stratikopoulos E, Szabolcs M, Dragatsis I, Klinakis A, Efstratiadis A. The hormonal action of IGF1 in postnatal mouse growth. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(49):19378-83.
13. Govoni KE, Lee SK, Chung YS, et al. Disruption of insulin-like growth factor-I expression in type IIalphaI collagen-expressing cells reduces bone length and width in mice. Physiological genomics. 2007;30(3):354-62.
100
14. Govoni KE, Wergedal JE, Florin L, Angel P, Baylink DJ, Mohan S. Conditional deletion of insulin-like growth factor-I in collagen type 1alpha2-expressing cells results in postnatal lethality and a dramatic reduction in bone accretion. Endocrinology. 2007;148(12):5706-15.
15. Sheng MH, Zhou XD, Bonewald LF, Baylink DJ, Lau KH. Disruption of the insulin-like growth factor-1 gene in osteocytes impairs developmental bone growth in mice. Bone. 2013;52(1):133-44.
16. St-Jacques B, Hammerschmidt M, McMahon AP. Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes & development. 1999;13(16):2072-86.
17. Vortkamp A, Lee K, Lanske B, Segre GV, Kronenberg HM, Tabin CJ. Regulation of rate of cartilage differentiation by Indian hedgehog and PTH-related protein. Science. 1996;273(5275):613-22.
18. Ingham PW, McMahon AP. Hedgehog signaling in animal development: paradigms and principles. Genes & development. 2001;15(23):3059-87.
19. McMahon AP, Ingham PW, Tabin CJ. Developmental roles and clinical significance of hedgehog signaling. Current topics in developmental biology. 2003;53:1-114.
20. Lee K, Lanske B, Karaplis AC, et al. Parathyroid hormone-related peptide delays terminal differentiation of chondrocytes during endochondral bone development. Endocrinology. 1996;137(11):5109-18.
21. Bitgood MJ, McMahon AP. Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Developmental biology. 1995;172(1):126-38.
22. Chung UI, Schipani E, McMahon AP, Kronenberg HM. Indian hedgehog couples chondrogenesis to osteogenesis in endochondral bone development. The Journal of clinical investigation. 2001;107(3):295-304.
23. Amizuka N, Warshawsky H, Henderson JE, Goltzman D, Karaplis AC. Parathyroid hormone-related peptide-depleted mice show abnormal epiphyseal cartilage development and altered endochondral bone formation. The Journal of cell biology. 1994;126(6):1611-23.
24. Wozney JM, Rosen V, Celeste AJ, et al. Novel regulators of bone formation: molecular clones and activities. Science. 1988;242(4885):1528-34.
25. Niswander L, Martin GR. FGF-4 regulates expression of Evx-1 in the developing mouse limb. Development. 1993;119(1):287-94.
26. Yoon BS, Ovchinnikov DA, Yoshii I, Mishina Y, Behringer RR, Lyons KM. Bmpr1a and Bmpr1b have overlapping functions and are essential for chondrogenesis in vivo.
101
Proceedings of the National Academy of Sciences of the United States of America. 2005;102(14):5062-7.
27. Retting KN, Song B, Yoon BS, Lyons KM. BMP canonical Smad signaling through Smad1 and Smad5 is required for endochondral bone formation. Development. 2009;136(7):1093-104.
28. Steinert AF, Proffen B, Kunz M, et al. Hypertrophy is induced during the in vitro chondrogenic differentiation of human mesenchymal stem cells by bone morphogenetic protein-2 and bone morphogenetic protein-4 gene transfer. Arthritis research & therapy. 2009;11(5):R148.
29. Rousseau F, Bonaventure J, Legeai-Mallet L, et al. Mutations in the gene encoding fibroblast growth factor receptor-3 in achondroplasia. Nature. 1994;371(6494):252-4.
30. Shiang R, Thompson LM, Zhu YZ, et al. Mutations in the transmembrane domain of FGFR3 cause the most common genetic form of dwarfism, achondroplasia. Cell. 1994;78(2):335-42.
31. Bellus GA, McIntosh I, Smith EA, et al. A recurrent mutation in the tyrosine kinase domain of fibroblast growth factor receptor 3 causes hypochondroplasia. Nature genetics. 1995;10(3):357-9.
32. Tavormina PL, Rimoin DL, Cohn DH, Zhu YZ, Shiang R, Wasmuth JJ. Another mutation that results in the substitution of an unpaired cysteine residue in the extracellular domain of FGFR3 in thanatophoric dysplasia type I. Human molecular genetics. 1995;4(11):2175-7.
33. Tavormina PL, Shiang R, Thompson LM, et al. Thanatophoric dysplasia (types I and II) caused by distinct mutations in fibroblast growth factor receptor 3. Nature genetics. 1995;9(3):321-8.
34. Colvin JS, Bohne BA, Harding GW, McEwen DG, Ornitz DM. Skeletal overgrowth and deafness in mice lacking fibroblast growth factor receptor 3. Nature genetics. 1996;12(4):390-7.
35. Deng C, Wynshaw-Boris A, Zhou F, Kuo A, Leder P. Fibroblast growth factor receptor 3 is a negative regulator of bone growth. Cell. 1996;84(6):911-21.
36. Ornitz DM. FGF signaling in the developing endochondral skeleton. Cytokine & growth factor reviews. 2005;16(2):205-13.
37. Minina E, Kreschel C, Naski MC, Ornitz DM, Vortkamp A. Interaction of FGF, Ihh/Pthlh, and BMP signaling integrates chondrocyte proliferation and hypertrophic differentiation. Developmental cell. 2002;3(3):439-49.
38. Yoon BS, Pogue R, Ovchinnikov DA, et al. BMPs regulate multiple aspects of growth- plate chondrogenesis through opposing actions on FGF pathways. Development. 2006;133(23):4667-78.
102
39. Topol L, Jiang X, Choi H, Garrett-Beal L, Carolan PJ, Yang Y. Wnt-5a inhibits the canonical Wnt pathway by promoting GSK-3-independent beta-catenin degradation. The Journal of cell biology. 2003;162(5):899-908.
40. Akiyama H, Lyons JP, Mori-Akiyama Y, et al. Interactions between Sox9 and beta- catenin control chondrocyte differentiation. Genes & development. 2004;18(9):1072-87.
41. Mak KK, Chen MH, Day TF, Chuang PT, Yang Y. Wnt/beta-catenin signaling interacts differentially with Ihh signaling in controlling endochondral bone and synovial joint formation. Development. 2006;133(18):3695-707.
42. Tamamura Y, Otani T, Kanatani N, et al. Developmental regulation of Wnt/beta-catenin signals is required for growth plate assembly, cartilage integrity, and endochondral ossification. The Journal of biological chemistry. 2005;280(19):19185-95.
43. Wuelling M, Vortkamp A. Transcriptional networks controlling chondrocyte proliferation and differentiation during endochondral ossification. Pediatric nephrology. 2010;25(4):625-31.
44. Beier F, Ali Z, Mok D, et al. TGFbeta and PTHrP control chondrocyte proliferation by activating cyclin D1 expression. Molecular biology of the cell. 2001;12(12):3852-63.
45. Dailey L, Laplantine E, Priore R, Basilico C. A network of transcriptional and signaling events is activated by FGF to induce chondrocyte growth arrest and differentiation. The Journal of cell biology. 2003;161(6):1053-66.
46. van der Eerden BC, Karperien M, Wit JM. Systemic and local regulation of the growth plate. Endocr Rev. 2003;24(6):782-801.
47. Aszodi A, Hunziker EB, Brakebusch C, Fassler R. Beta1 integrins regulate chondrocyte rotation, G1 progression, and cytokinesis. Genes & development. 2003;17(19):2465-79.
48. Yang C, Li SW, Helminen HJ, Khillan JS, Bao Y, Prockop DJ. Apoptosis of chondrocytes in transgenic mice lacking collagen II. Experimental cell research. 1997;235(2):370-3.
49. Kluppel M, Wight TN, Chan C, Hinek A, Wrana JL. Maintenance of chondroitin sulfation balance by chondroitin-4-sulfotransferase 1 is required for chondrocyte development and growth factor signaling during cartilage morphogenesis. Development. 2005;132(17):3989-4003.
50. Makitie O, Susic M, Ward L, Barclay C, Glorieux FH, Cole WG. Schmid type of metaphyseal chondrodysplasia and COL10A1 mutations--findings in 10 patients. American journal of medical genetics Part A. 2005;137A(3):241-8.
51. McIntosh I, Abbott MH, Francomano CA. Concentration of mutations causing Schmid metaphyseal chondrodysplasia in the C-terminal noncollagenous domain of type X collagen. Human mutation. 1995;5(2):121-5.
103
52. Warman ML, Abbott M, Apte SS, et al. A type X collagen mutation causes Schmid metaphyseal chondrodysplasia. Nature genetics. 1993;5(1):79-82.
53. Wallis GA. Here today, bone tomorrow. Current biology : CB. 1993;3(10):687-9.
54. Wallis GA, Rash B, Sweetman WA, et al. Amino acid substitutions of conserved residues in the carboxyl-terminal domain of the alpha 1(X) chain of type X collagen occur in two unrelated families with metaphyseal chondrodysplasia type Schmid. American journal of human genetics. 1994;54(2):169-78.
55. Kwan KM, Pang MK, Zhou S, et al. Abnormal compartmentalization of cartilage matrix components in mice lacking collagen X: implications for function. The Journal of cell biology. 1997;136(2):459-71.
56. Leitinger B, Kwan AP. The discoidin domain receptor DDR2 is a receptor for type X collagen. Matrix Biol. 2006;25(6):355-64.
57. Luckman SP, Rees E, Kwan AP. Partial characterization of cell-type X collagen interactions. The Biochemical journal. 2003;372(Pt 2):485-93.
58. Bi W, Deng JM, Zhang Z, Behringer RR, de Crombrugghe B. Sox9 is required for cartilage formation. Nature genetics. 1999;22(1):85-9.
59. Bi W, Huang W, Whitworth DJ, et al. Haploinsufficiency of Sox9 results in defective cartilage primordia and premature skeletal mineralization. Proceedings of the National Academy of Sciences of the United States of America. 2001;98(12):6698-703.
60. Sekiya I, Tsuji K, Koopman P, et al. SOX9 enhances aggrecan gene promoter/enhancer activity and is up-regulated by retinoic acid in a cartilage-derived cell line, TC6. The Journal of biological chemistry. 2000;275(15):10738-44.
61. Lefebvre V, Huang W, Harley VR, Goodfellow PN, de Crombrugghe B. SOX9 is a potent activator of the chondrocyte-specific enhancer of the pro alpha1(II) collagen gene. Molecular and cellular biology. 1997;17(4):2336-46.
62. Smits P, Dy P, Mitra S, Lefebvre V. Sox5 and Sox6 are needed to develop and maintain source, columnar, and hypertrophic chondrocytes in the cartilage growth plate. The Journal of cell biology. 2004;164(5):747-58.
63. Lengner CJ, Hassan MQ, Serra RW, et al. Nkx3.2-mediated repression of Runx2 promotes chondrogenic differentiation. J Biol Chem. 2005;280(16):15872-9.
64. Arnold MA, Kim Y, Czubryt MP, et al. MEF2C transcription factor controls chondrocyte hypertrophy and bone development. Dev Cell. 2007;12(3):377-89.
65. Yoshida CA, Yamamoto H, Fujita T, et al. Runx2 and Runx3 are essential for chondrocyte maturation, and Runx2 regulates limb growth through induction of Indian hedgehog. Genes & development. 2004;18(8):952-63.
104
66. Komori T, Yagi H, Nomura S, et al. Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell. 1997;89(5):755- 64.
67. Inada M, Yasui T, Nomura S, et al. Maturational disturbance of chondrocytes in Cbfa1- deficient mice. Developmental dynamics : an official publication of the American Association of Anatomists. 1999;214(4):279-90.
68. Ueta C, Iwamoto M, Kanatani N, et al. Skeletal malformations caused by overexpression of Cbfa1 or its dominant negative form in chondrocytes. The Journal of cell biology. 2001;153(1):87-100.
69. DiGirolamo DJ, Mukherjee A, Fulzele K, et al. Mode of growth hormone action in osteoblasts. The Journal of biological chemistry. 2007;282(43):31666-74.
70. Zhang M, Xuan S, Bouxsein ML, et al. Osteoblast-specific knockout of the insulin-like growth factor (IGF) receptor gene reveals an essential role of IGF signaling in bone matrix mineralization. The Journal of biological chemistry. 2002;277(46):44005-12.
71. Zhao G, Monier-Faugere MC, Langub MC, et al. Targeted overexpression of insulin-like growth factor I to osteoblasts of transgenic mice: increased trabecular bone volume without increased osteoblast proliferation. Endocrinology. 2000;141(7):2674-82.
72. Long F, Chung UI, Ohba S, McMahon J, Kronenberg HM, McMahon AP. Ihh signaling is directly required for the osteoblast lineage in the endochondral skeleton. Development. 2004;131(6):1309-18.
73. Maeda Y, Nakamura E, Nguyen MT, et al. Indian Hedgehog produced by postnatal chondrocytes is essential for maintaining a growth plate and trabecular bone. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(15):6382-7.
74. Mak KK, Bi Y, Wan C, et al. Hedgehog signaling in mature osteoblasts regulates bone formation and resorption by controlling PTHrP and RANKL expression. Developmental cell. 2008;14(5):674-88.
75. Ohba S, Kawaguchi H, Kugimiya F, et al. Patched1 haploinsufficiency increases adult bone mass and modulates Gli3 repressor activity. Developmental cell. 2008;14(5):689- 99.
76. Kato M, Patel MS, Levasseur R, et al. Cbfa1-independent decrease in osteoblast proliferation, osteopenia, and persistent embryonic eye vascularization in mice deficient in Lrp5, a Wnt coreceptor. The Journal of cell biology. 2002;157(2):303-14.
77. Cui Y, Niziolek PJ, MacDonald BT, et al. Lrp5 functions in bone to regulate bone mass. Nature medicine. 2011;17(6):684-91.
78. Yadav VK, Oury F, Suda N, et al. A serotonin-dependent mechanism explains the leptin regulation of bone mass, appetite, and energy expenditure. Cell. 2009;138(5):976-89.
105
79. Day TF, Guo X, Garrett-Beal L, Yang Y. Wnt/beta-catenin signaling in mesenchymal progenitors controls osteoblast and chondrocyte differentiation during vertebrate skeletogenesis. Developmental cell. 2005;8(5):739-50.
80. Hill TP, Spater D, Taketo MM, Birchmeier W, Hartmann C. Canonical Wnt/beta-catenin signaling prevents osteoblasts from differentiating into chondrocytes. Developmental cell. 2005;8(5):727-38.
81. Hu H, Hilton MJ, Tu X, Yu K, Ornitz DM, Long F. Sequential roles of Hedgehog and Wnt signaling in osteoblast development. Development. 2005;132(1):49-60.
82. Rodda SJ, McMahon AP. Distinct roles for Hedgehog and canonical Wnt signaling in specification, differentiation and maintenance of osteoblast progenitors. Development. 2006;133(16):3231-44.
83. Glass DA, 2nd, Bialek P, Ahn JD, et al. Canonical Wnt signaling in differentiated osteoblasts controls osteoclast differentiation. Developmental cell. 2005;8(5):751-64.
84. Holmen SL, Zylstra CR, Mukherjee A, et al. Essential role of beta-catenin in postnatal bone acquisition. The Journal of biological chemistry. 2005;280(22):21162-8.
85. Kramer I, Halleux C, Keller H, et al. Osteocyte Wnt/beta-catenin signaling is required for normal bone homeostasis. Molecular and cellular biology. 2010;30(12):3071-85.
86. Tsuji K, Bandyopadhyay A, Harfe BD, et al. BMP2 activity, although dispensable for bone formation, is required for the initiation of fracture healing. Nature genetics. 2006;38(12):1424-9.
87. Bandyopadhyay A, Tsuji K, Cox K, Harfe BD, Rosen V, Tabin CJ. Genetic analysis of the roles of BMP2, BMP4, and BMP7 in limb patterning and skeletogenesis. PLoS genetics. 2006;2(12):e216.
88. Mishina Y, Starbuck MW, Gentile MA, et al. Bone morphogenetic protein type IA receptor signaling regulates postnatal osteoblast function and bone remodeling. The Journal of biological chemistry. 2004;279(26):27560-6.
89. Kamiya N, Ye L, Kobayashi T, et al. Disruption of BMP signaling in osteoblasts through type IA receptor (BMPRIA) increases bone mass. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2008;23(12):2007-17.
90. Liu Z, Xu J, Colvin JS, Ornitz DM. Coordination of chondrogenesis and osteogenesis by fibroblast growth factor 18. Genes & development. 2002;16(7):859-69.
91. Ohbayashi N, Shibayama M, Kurotaki Y, et al. FGF18 is required for normal cell proliferation and differentiation during osteogenesis and chondrogenesis. Genes & development. 2002;16(7):870-9.
106
92. Eswarakumar VP, Horowitz MC, Locklin R, Morriss-Kay GM, Lonai P. A gain-of- function mutation of Fgfr2c demonstrates the roles of this receptor variant in osteogenesis. Proceedings of the National Academy of Sciences of the United States of America. 2004;101(34):12555-60.
93. Jacob AL, Smith C, Partanen J, Ornitz DM. Fibroblast growth factor receptor 1 signaling in the osteo-chondrogenic cell lineage regulates sequential steps of osteoblast maturation. Developmental biology. 2006;296(2):315-28.
94. Yu K, Xu J, Liu Z, et al. Conditional inactivation of FGF receptor 2 reveals an essential role for FGF signaling in the regulation of osteoblast function and bone growth. Development. 2003;130(13):3063-74.
95. Choi KY, Kim HJ, Lee MH, et al. Runx2 regulates FGF2-induced Bmp2 expression during cranial bone development. Developmental dynamics : an official publication of the American Association of Anatomists. 2005;233(1):115-21.
96. Kim BG, Kim HJ, Park HJ, et al. Runx2 phosphorylation induced by fibroblast growth factor-2/protein kinase C pathways. Proteomics. 2006;6(4):1166-74.
97. Naganawa T, Xiao L, Coffin JD, et al. Reduced expression and function of bone morphogenetic protein-2 in bones of Fgf2 null mice. Journal of cellular biochemistry. 2008;103(6):1975-88.
98. Reinhold MI, Naski MC. Direct interactions of Runx2 and canonical Wnt signaling induce FGF18. The Journal of biological chemistry. 2007;282(6):3653-63.
99. Komori T. Regulation of skeletal development by the Runx family of transcription factors. Journal of cellular biochemistry. 2005;95(3):445-53.
100. Liu W, Toyosawa S, Furuichi T, et al. Overexpression of Cbfa1 in osteoblasts inhibits osteoblast maturation and causes osteopenia with multiple fractures. The Journal of cell biology. 2001;155(1):157-66.
101. Nakashima K, Zhou X, Kunkel G, et al. The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation. Cell. 2002;108(1):17-29.
102. Juul A. The effects of oestrogens on linear bone growth. Human reproduction update. 2001;7(3):303-13.
103. Morishima A, Grumbach MM, Simpson ER, Fisher C, Qin K. Aromatase deficiency in male and female siblings caused by a novel mutation and the physiological role of estrogens. J Clin Endocrinol Metab. 1995;80(12):3689-98.
104. Smith EP, Boyd J, Frank GR, et al. Estrogen resistance caused by a mutation in the estrogen-receptor gene in a man. N Engl J Med. 1994;331(16):1056-61.
107
105. Kaplan SL, Grumbach MM. Clinical review 14: Pathophysiology and treatment of sexual precocity. J Clin Endocrinol Metab. 1990;71(4):785-9.
106. Driscoll MD, Sathya G, Muyan M, Klinge CM, Hilf R, Bambara RA. Sequence requirements for estrogen receptor binding to estrogen response elements. J Biol Chem. 1998;273(45):29321-30.
107. Lubahn DB, Moyer JS, Golding TS, Couse JF, Korach KS, Smithies O. Alteration of reproductive function but not prenatal sexual development after insertional disruption of the mouse estrogen receptor gene. Proc Natl Acad Sci U S A. 1993;90(23):11162-6.
108. Vidal O, Lindberg MK, Hollberg K, et al. Estrogen receptor specificity in the regulation of skeletal growth and maturation in male mice. Proceedings of the National Academy of Sciences of the United States of America. 2000;97(10):5474-9.
109. Parikka V, Peng Z, Hentunen T, et al. Estrogen responsiveness of bone formation in vitro and altered bone phenotype in aged estrogen receptor-alpha-deficient male and female mice. Eur J Endocrinol. 2005;152(2):301-14.
110. Flouriot G, Brand H, Denger S, et al. Identification of a new isoform of the human estrogen receptor-alpha (hER-alpha) that is encoded by distinct transcripts and that is able to repress hER-alpha activation function 1. EMBO J. 2000;19(17):4688-700.
111. Dupont S, Krust A, Gansmuller A, Dierich A, Chambon P, Mark M. Effect of single and compound knockouts of estrogen receptors alpha (ERalpha) and beta (ERbeta) on mouse reproductive phenotypes. Development. 2000;127(19):4277-91.
112. Sims NA, Dupont S, Krust A, et al. Deletion of estrogen receptors reveals a regulatory role for estrogen receptors-beta in bone remodeling in females but not in males. Bone. 2002;30(1):18-25.
113. Tozum TF, Oppenlander ME, Koh-Paige AJ, Robins DM, McCauley LK. Effects of sex steroid receptor specificity in the regulation of skeletal metabolism. Calcif Tissue Int. 2004;75(1):60-70.
114. Chagin AS, Lindberg MK, Andersson N, et al. Estrogen receptor-beta inhibits skeletal growth and has the capacity to mediate growth plate fusion in female mice. J Bone Miner Res. 2004;19(1):72-7.
115. Borjesson AE, Lagerquist MK, Liu C, et al. The role of estrogen receptor alpha in growth plate cartilage for longitudinal bone growth. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2010;25(12):2690-700.
116. Ikeda K, Tsukui T, Imazawa Y, Horie-Inoue K, Inoue S. Conditional expression of constitutively active estrogen receptor alpha in chondrocytes impairs longitudinal bone growth in mice. Biochemical and biophysical research communications. 2012;425(4):912-7.
108
117. Richelson LS, Wahner HW, Melton LJ, 3rd, Riggs BL. Relative contributions of aging and estrogen deficiency to postmenopausal bone loss. N Engl J Med. 1984;311(20):1273-5.
118. Oursler MJ, Cortese C, Keeting P, et al. Modulation of transforming growth factor-beta production in normal human osteoblast-like cells by 17 beta-estradiol and parathyroid hormone. Endocrinology. 1991;129(6):3313-20.
119. Borjesson AE, Windahl SH, Lagerquist MK, et al. Roles of transactivating functions 1 and 2 of estrogen receptor-alpha in bone. Proceedings of the National Academy of Sciences of the United States of America. 2011;108(15):6288-93.
120. Borjesson AE, Farman HH, Engdahl C, et al. The role of activation functions 1 and 2 of estrogen receptor-alpha for the effects of estradiol and selective estrogen receptor modulators in male mice. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2013;28(5):1117-26.
121. Ikeda K, Tsukui T, Horie-Inoue K, Inoue S. Conditional expression of constitutively active estrogen receptor alpha in osteoblasts increases bone mineral density in mice. FEBS letters. 2011;585(9):1303-9.
122. Nakamura T, Imai Y, Matsumoto T, et al. Estrogen prevents bone loss via estrogen receptor alpha and induction of Fas ligand in osteoclasts. Cell. 2007;130(5):811-23.
123. Martin-Millan M, Almeida M, Ambrogini E, et al. The estrogen receptor-alpha in osteoclasts mediates the protective effects of estrogens on cancellous but not cortical bone. Molecular endocrinology. 2010;24(2):323-34.
124. Almeida M, Iyer S, Martin-Millan M, et al. Estrogen receptor-alpha signaling in osteoblast progenitors stimulates cortical bone accrual. The Journal of clinical investigation. 2013;123(1):394-404.
125. Maatta JA, Buki KG, Gu G, et al. Inactivation of estrogen receptor alpha in bone- forming cells induces bone loss in female mice. FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 2013;27(2):478-88.
126. Melville KM, Kelly NH, Khan SA, et al. Female mice lacking estrogen receptor-alpha in osteoblasts have compromised bone mass and strength. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2014;29(2):370-9.
127. Windahl SH, Borjesson AE, Farman HH, et al. Estrogen receptor-alpha in osteocytes is important for trabecular bone formation in male mice. Proceedings of the National Academy of Sciences of the United States of America. 2013;110(6):2294-9.
128. Kondoh S, Inoue K, Igarashi K, et al. Estrogen receptor alpha in osteocytes regulates trabecular bone formation in female mice. Bone. 2014;60:68-77.
109
129. Giguere V, Yang N, Segui P, Evans RM. Identification of a new class of steroid hormone receptors. Nature. 1988;331(6151):91-4.
130. Hong H, Yang L, Stallcup MR. Hormone-independent transcriptional activation and coactivator binding by novel orphan nuclear receptor ERR3. The Journal of biological chemistry. 1999;274(32):22618-26.
131. Giguere V. Orphan nuclear receptors: from gene to function. Endocr Rev. 1999;20(5):689-725.
132. Tora L, Gronemeyer H, Turcotte B, Gaub MP, Chambon P. The N-terminal region of the chicken progesterone receptor specifies target gene activation. Nature. 1988;333(6169):185-8.
133. Kato S, Endoh H, Masuhiro Y, et al. Activation of the estrogen receptor through phosphorylation by mitogen-activated protein kinase. Science. 1995;270(5241):1491-4.
134. Krstic MD, Rogatsky I, Yamamoto KR, Garabedian MJ. Mitogen-activated and cyclin- dependent protein kinases selectively and differentially modulate transcriptional enhancement by the glucocorticoid receptor. Mol Cell Biol. 1997;17(7):3947-54.
135. Joel PB, Smith J, Sturgill TW, Fisher TL, Blenis J, Lannigan DA. pp90rsk1 regulates estrogen receptor-mediated transcription through phosphorylation of Ser-167. Mol Cell Biol. 1998;18(4):1978-84.
136. Bonnelye E, Aubin JE. Estrogen receptor-related receptor alpha: a mediator of estrogen response in bone. J Clin Endocrinol Metab. 2005;90(5):3115-21.
137. Johnston SD, Liu X, Zuo F, et al. Estrogen-related receptor alpha 1 functionally binds as a monomer to extended half-site sequences including ones contained within estrogen- response elements. Molecular endocrinology. 1997;11(3):342-52.
138. Sladek R, Bader JA, Giguere V. The orphan nuclear receptor estrogen-related receptor alpha is a transcriptional regulator of the human medium-chain acyl coenzyme A dehydrogenase gene. Molecular and cellular biology. 1997;17(9):5400-9.
139. Vega RB, Kelly DP. A role for estrogen-related receptor alpha in the control of mitochondrial fatty acid beta-oxidation during brown adipocyte differentiation. The Journal of biological chemistry. 1997;272(50):31693-9.
140. Dufour CR, Wilson BJ, Huss JM, et al. Genome-wide orchestration of cardiac functions by the orphan nuclear receptors ERRalpha and gamma. Cell metabolism. 2007;5(5):345- 56.
141. Vanacker JM, Bonnelye E, Chopin-Delannoy S, Delmarre C, Cavailles V, Laudet V. Transcriptional activities of the orphan nuclear receptor ERR alpha (estrogen receptor- related receptor-alpha). Molecular endocrinology. 1999;13(5):764-73.
110
142. Vanacker JM, Pettersson K, Gustafsson JA, Laudet V. Transcriptional targets shared by estrogen receptor- related receptors (ERRs) and estrogen receptor (ER) alpha, but not by ERbeta. The EMBO journal. 1999;18(15):4270-9.
143. Hentschke M, Susens U, Borgmeyer U. Domains of ERRgamma that mediate homodimerization and interaction with factors stimulating DNA binding. European journal of biochemistry / FEBS. 2002;269(16):4086-97.
144. Huppunen J, Aarnisalo P. Dimerization modulates the activity of the orphan nuclear receptor ERRgamma. Biochemical and biophysical research communications. 2004;314(4):964-70.
145. Gearhart MD, Holmbeck SM, Evans RM, Dyson HJ, Wright PE. Monomeric complex of human orphan estrogen related receptor-2 with DNA: a pseudo-dimer interface mediates extended half-site recognition. Journal of molecular biology. 2003;327(4):819-32.
146. Barry JB, Laganiere J, Giguere V. A single nucleotide in an estrogen-related receptor alpha site can dictate mode of binding and peroxisome proliferator-activated receptor gamma coactivator 1alpha activation of target promoters. Molecular endocrinology. 2006;20(2):302-10.
147. Glass CK. Differential recognition of target genes by nuclear receptor monomers, dimers, and heterodimers. Endocr Rev. 1994;15(3):391-407.
148. Kallen J, Schlaeppi JM, Bitsch F, et al. Evidence for ligand-independent transcriptional activation of the human estrogen-related receptor alpha (ERRalpha): crystal structure of ERRalpha ligand binding domain in complex with peroxisome proliferator-activated receptor coactivator-1alpha. The Journal of biological chemistry. 2004;279(47):49330-7.
149. Luo J, Sladek R, Carrier J, Bader JA, Richard D, Giguere V. Reduced fat mass in mice lacking orphan nuclear receptor estrogen-related receptor alpha. Molecular and cellular biology. 2003;23(22):7947-56.
150. Villena JA, Hock MB, Chang WY, Barcas JE, Giguere V, Kralli A. Orphan nuclear receptor estrogen-related receptor alpha is essential for adaptive thermogenesis. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(4):1418-23.
151. Huss JM, Imahashi K, Dufour CR, et al. The nuclear receptor ERRalpha is required for the bioenergetic and functional adaptation to cardiac pressure overload. Cell metabolism. 2007;6(1):25-37.
152. Bonnelye E, Zirngibl RA, Jurdic P, Aubin JE. The orphan nuclear estrogen receptor- related receptor-alpha regulates cartilage formation in vitro: implication of Sox9. Endocrinology. 2007;148(3):1195-205.
153. Bonnelye E, Laurin N, Jurdic P, Hart DA, Aubin JE. Estrogen receptor-related receptor- alpha (ERR-alpha) is dysregulated in inflammatory arthritis. Rheumatology. 2008;47(12):1785-91.
111
154. Bonnelye E, Reboul P, Duval N, Cardelli M, Aubin JE. Estrogen receptor-related receptor alpha regulation by interleukin-1beta in prostaglandin E(2)- and cAMP- dependent pathways in osteoarthritic chondrocytes. Arthritis and rheumatism. 2011;63(8):2374-84.
155. Bonnelye E, Merdad L, Kung V, Aubin JE. The orphan nuclear estrogen receptor-related receptor alpha (ERRalpha) is expressed throughout osteoblast differentiation and regulates bone formation in vitro. The Journal of cell biology. 2001;153(5):971-84.
156. Bonnelye E, Kung V, Laplace C, Galson DL, Aubin JE. Estrogen receptor-related receptor alpha impinges on the estrogen axis in bone: potential function in osteoporosis. Endocrinology. 2002;143(9):3658-70.
157. Delhon I, Gutzwiller S, Morvan F, et al. Absence of estrogen receptor-related-alpha increases osteoblastic differentiation and cancellous bone mineral density. Endocrinology. 2009;150(10):4463-72.
158. Teyssier C, Gallet M, Rabier B, et al. Absence of ERRalpha in female mice confers resistance to bone loss induced by age or estrogen-deficiency. PloS one. 2009;4(11):e7942.
159. Rajalin AM, Pollock H, Aarnisalo P. ERRalpha regulates osteoblastic and adipogenic differentiation of mouse bone marrow mesenchymal stem cells. Biochemical and biophysical research communications. 2010;396(2):477-82.
160. Bonnelye E, Aubin JE. An energetic orphan in an endocrine tissue: a revised perspective of the function of estrogen receptor-related receptor alpha in bone and cartilage. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2013;28(2):225-33.
161. Wei W, Wang X, Yang M, et al. PGC1beta mediates PPARgamma activation of osteoclastogenesis and rosiglitazone-induced bone loss. Cell metabolism. 2010;11(6):503-16.
162. Bonnelye E, Saltel F, Chabadel A, Zirngibl RA, Aubin JE, Jurdic P. Involvement of the orphan nuclear estrogen receptor-related receptor alpha in osteoclast adhesion and transmigration. Journal of molecular endocrinology. 2010;45(6):365-77.
163. Luo J, Sladek R, Bader JA, Matthyssen A, Rossant J, Giguere V. Placental abnormalities in mouse embryos lacking the orphan nuclear receptor ERR-beta. Nature. 1997;388(6644):778-82.
164. Chen J, Nathans J. Estrogen-related receptor beta/NR3B2 controls epithelial cell fate and endolymph production by the stria vascularis. Developmental cell. 2007;13(3):325-37.
165. Onishi A, Peng GH, Poth EM, et al. The orphan nuclear hormone receptor ERRbeta controls rod photoreceptor survival. Proceedings of the National Academy of Sciences of the United States of America. 2010;107(25):11579-84.
112
166. Byerly MS, Al Salayta M, Swanson RD, et al. Estrogen-related receptor beta deletion modulates whole-body energy balance via estrogen-related receptor gamma and attenuates neuropeptide Y gene expression. The European journal of neuroscience. 2013;37(7):1033-47.
167. Hentschke M, Susens U, Borgmeyer U. PGC-1 and PERC, coactivators of the estrogen receptor-related receptor gamma. Biochemical and biophysical research communications. 2002;299(5):872-9.
168. Huss JM, Kopp RP, Kelly DP. Peroxisome proliferator-activated receptor coactivator- 1alpha (PGC-1alpha) coactivates the cardiac-enriched nuclear receptors estrogen-related receptor-alpha and -gamma. Identification of novel leucine-rich interaction motif within PGC-1alpha. The Journal of biological chemistry. 2002;277(43):40265-74.
169. Alaynick WA, Kondo RP, Xie W, et al. ERRgamma directs and maintains the transition to oxidative metabolism in the postnatal heart. Cell metabolism. 2007;6(1):13-24.
170. Elfassihi L, Giroux S, Bureau A, Laflamme N, Cole DE, Rousseau F. Association with replication between estrogen-related receptor gamma (ESRRgamma) polymorphisms and bone phenotypes in women of European ancestry. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2010;25(4):901-11.
171. Pirih FQ, Abayahoudian R, Elashoff D, Parhami F, Nervina JM, Tetradis S. Nuclear receptor profile in calvarial bone cells undergoing osteogenic versus adipogenic differentiation. Journal of cellular biochemistry. 2008;105(5):1316-26.
172. Roforth MM, Liu G, Khosla S, Monroe DG. Examination of nuclear receptor expression in osteoblasts reveals Rorbeta as an important regulator of osteogenesis. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2012;27(4):891-901.
173. Jeong BC, Lee YS, Park YY, et al. The orphan nuclear receptor estrogen receptor-related receptor gamma negatively regulates BMP2-induced osteoblast differentiation and bone formation. The Journal of biological chemistry. 2009;284(21):14211-8.
174. Kim EJ, Kang IH, Lee JW, Jang WG, Koh JT. MiR-433 mediates ERRgamma- suppressed osteoblast differentiation via direct targeting to Runx2 mRNA in C3H10T1/2 cells. Life sciences. 2013;92(10):562-8.
175. Blumbach K, Niehoff A, Paulsson M, Zaucke F. Ablation of collagen IX and COMP disrupts epiphyseal cartilage architecture. Matrix Biol. 2008;27(4):306-18.
176. Day TF, Yang Y. Wnt and hedgehog signaling pathways in bone development. The Journal of bone and joint surgery American volume. 2008;90 Suppl 1:19-24.
177. Takeda S, Bonnamy JP, Owen MJ, Ducy P, Karsenty G. Continuous expression of Cbfa1 in nonhypertrophic chondrocytes uncovers its ability to induce hypertrophic
113
chondrocyte differentiation and partially rescues Cbfa1-deficient mice. Genes & development. 2001;15(4):467-81.
178. Wang L, Zuercher WJ, Consler TG, et al. X-ray crystal structures of the estrogen-related receptor-gamma ligand binding domain in three functional states reveal the molecular basis of small molecule regulation. The Journal of biological chemistry. 2006;281(49):37773-81.
179. Xie W, Hong H, Yang NN, et al. Constitutive activation of transcription and binding of coactivator by estrogen-related receptors 1 and 2. Molecular endocrinology. 1999;13(12):2151-62.
180. Elfassihi L, Giroux S, Bureau A, Laflamme N, Cole DE, Rousseau F. Association with replication between estrogen-related receptor gamma (ESRRgamma) polymorphisms and bone phenotypes in women of European ancestry. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research.25(4):901-11.
181. Zhou G, Garofalo S, Mukhopadhyay K, et al. A 182 bp fragment of the mouse pro alpha 1(II) collagen gene is sufficient to direct chondrocyte expression in transgenic mice. Journal of cell science. 1995;108 ( Pt 12):3677-84.
182. Friedrich G, Soriano P. Promoter traps in embryonic stem cells: a genetic screen to identify and mutate developmental genes in mice. Genes & development. 1991;5(9):1513-23.
183. Mirski SE, Gerlach JH, Cummings HJ, Zirngibl R, Greer PA, Cole SP. Bipartite nuclear localization signals in the C terminus of human topoisomerase II alpha. Experimental cell research. 1997;237(2):452-5.
184. Rossert J, Eberspaecher H, de Crombrugghe B. Separate cis-acting DNA elements of the mouse pro-alpha 1(I) collagen promoter direct expression of reporter genes to different type I collagen-producing cells in transgenic mice. The Journal of cell biology. 1995;129(5):1421-32.
185. Turksen K, Troy TC. Permeability barrier dysfunction in transgenic mice overexpressing claudin 6. Development. 2002;129(7):1775-84.
186. Gossler A, Joyner AL, Rossant J, Skarnes WC. Mouse embryonic stem cells and reporter constructs to detect developmentally regulated genes. Science. 1989;244(4903):463-5.
187. Parr BA, McMahon AP. Dorsalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb. Nature. 1995;374(6520):350-3.
188. Yu S, Wang X, Ng CF, Chen S, Chan FL. ERRgamma suppresses cell proliferation and tumor growth of androgen-sensitive and androgen-insensitive prostate cancer cells and its implication as a therapeutic target for prostate cancer. Cancer research. 2007;67(10):4904-14.
114
189. Kawato Y, Hirao M, Ebina K, et al. Nkx3.2 promotes primary chondrogenic differentiation by upregulating Col2a1 transcription. PloS one. 2012;7(4):e34703.
190. Kawakami Y, Tsuda M, Takahashi S, et al. Transcriptional coactivator PGC-1alpha regulates chondrogenesis via association with Sox9. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(7):2414-9.
191. Tsuda M, Takahashi S, Takahashi Y, Asahara H. Transcriptional co-activators CREB- binding protein and p300 regulate chondrocyte-specific gene expression via association with Sox9. The Journal of biological chemistry. 2003;278(29):27224-9.
192. Naski MC, Colvin JS, Coffin JD, Ornitz DM. Repression of hedgehog signaling and BMP4 expression in growth plate cartilage by fibroblast growth factor receptor 3. Development. 1998;125(24):4977-88.
193. Naski MC, Ornitz DM. FGF signaling in skeletal development. Frontiers in bioscience : a journal and virtual library. 1998;3:d781-94.
194. Razzaque MS, Soegiarto DW, Chang D, Long F, Lanske B. Conditional deletion of Indian hedgehog from collagen type 2alpha1-expressing cells results in abnormal endochondral bone formation. The Journal of pathology. 2005;207(4):453-61.
195. Wang W, Lian N, Li L, et al. Atf4 regulates chondrocyte proliferation and differentiation during endochondral ossification by activating Ihh transcription. Development. 2009;136(24):4143-53.
196. Amano K, Ichida F, Sugita A, et al. MSX2 stimulates chondrocyte maturation by controlling Ihh expression. The Journal of biological chemistry. 2008;283(43):29513-21.
197. Lanske B, Karaplis AC, Lee K, et al. PTH/PTHrP receptor in early development and Indian hedgehog-regulated bone growth. Science. 1996;273(5275):663-6.
198. Schipani E, Lanske B, Hunzelman J, et al. Targeted expression of constitutively active receptors for parathyroid hormone and parathyroid hormone-related peptide delays endochondral bone formation and rescues mice that lack parathyroid hormone-related peptide. Proceedings of the National Academy of Sciences of the United States of America. 1997;94(25):13689-94.
199. Yang N, Shigeta H, Shi H, Teng CT. Estrogen-related receptor, hERR1, modulates estrogen receptor-mediated response of human lactoferrin gene promoter. The Journal of biological chemistry. 1996;271(10):5795-804.
200. Giguere V. To ERR in the estrogen pathway. Trends in endocrinology and metabolism: TEM. 2002;13(5):220-5.
201. Scutt A, Beier N, Fittschen C. EMD273316 & EMD95833, type 4 phosphodiesterase inhibitors, stimulate fibroblastic-colony formation by bone marrow cells via direct inhibition of PDE4 and the induction of endogenous prostaglandin synthesis. BMC pharmacology. 2004;4:10.
115
202. Wai PY, Mi Z, Gao C, Guo H, Marroquin C, Kuo PC. Ets-1 and runx2 regulate transcription of a metastatic gene, osteopontin, in murine colorectal cancer cells. The Journal of biological chemistry. 2006;281(28):18973-82.
203. Zirngibl RA, Chan JS, Aubin JE. Divergent regulation of the Osteopontin promoter by the estrogen receptor-related receptors is isoform- and cell context-dependent. Journal of cellular biochemistry. 2013.
204. Sims NA, Clement-Lacroix P, Minet D, et al. A functional androgen receptor is not sufficient to allow estradiol to protect bone after gonadectomy in estradiol receptor- deficient mice. The Journal of clinical investigation. 2003;111(9):1319-27.
205. Perlmann T, Umesono K, Rangarajan PN, Forman BM, Evans RM. Two distinct dimerization interfaces differentially modulate target gene specificity of nuclear hormone receptors. Molecular endocrinology. 1996;10(8):958-66.
206. Takarada T, Hinoi E, Nakazato R, et al. An analysis of skeletal development in osteoblast-specific and chondrocyte-specific runt-related transcription factor-2 (Runx2) knockout mice. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2013;28(10):2064-9.
207. Ijichi N, Shigekawa T, Ikeda K, et al. Estrogen-related receptor gamma modulates cell proliferation and estrogen signaling in breast cancer. The Journal of steroid biochemistry and molecular biology. 2011;123(1-2):1-7.
208. Kamei Y, Ohizumi H, Fujitani Y, et al. PPARgamma coactivator 1beta/ERR ligand 1 is an ERR protein ligand, whose expression induces a high-energy expenditure and antagonizes obesity. Proceedings of the National Academy of Sciences of the United States of America. 2003;100(21):12378-83.
209. Schreiber SN, Knutti D, Brogli K, Uhlmann T, Kralli A. The transcriptional coactivator PGC-1 regulates the expression and activity of the orphan nuclear receptor estrogen- related receptor alpha (ERRalpha). The Journal of biological chemistry. 2003;278(11):9013-8.
210. Tcherepanova I, Puigserver P, Norris JD, Spiegelman BM, McDonnell DP. Modulation of estrogen receptor-alpha transcriptional activity by the coactivator PGC-1. The Journal of biological chemistry. 2000;275(21):16302-8.
211. Yamamoto T, Mori T, Sawada M, et al. Estrogen-related receptor-gamma regulates estrogen receptor-alpha responsiveness in uterine endometrial cancer. International journal of gynecological cancer : official journal of the International Gynecological Cancer Society. 2012;22(9):1509-16.
212. Bonnelye E, Vanacker JM, Dittmar T, et al. The ERR-1 orphan receptor is a transcriptional activator expressed during bone development. Molecular endocrinology. 1997;11(7):905-16.
116
213. Greschik H, Wurtz JM, Sanglier S, et al. Structural and functional evidence for ligand- independent transcriptional activation by the estrogen-related receptor 3. Molecular cell. 2002;9(2):303-13.
214. Khalid O, Baniwal SK, Purcell DJ, et al. Modulation of Runx2 activity by estrogen receptor-alpha: implications for osteoporosis and breast cancer. Endocrinology. 2008;149(12):5984-95.
215. Ducy P, Starbuck M, Priemel M, et al. A Cbfa1-dependent genetic pathway controls bone formation beyond embryonic development. Genes & development. 1999;13(8):1025-36.
216. Malaval L, Modrowski D, Gupta AK, Aubin JE. Cellular expression of bone-related proteins during in vitro osteogenesis in rat bone marrow stromal cell cultures. Journal of cellular physiology. 1994;158(3):555-72.
217. Emons JA, Marino R, Nilsson O, et al. The role of p27Kip1 in the regulation of growth plate chondrocyte proliferation in mice. Pediatric research. 2006;60(3):288-93.
218. Yang BB, Zhang Y, Cao L, Yang BL. Aggrecan and link protein affect cell adhesion to culture plates and to type II collagen. Matrix Biol. 1998;16(9):541-61.
219. Otto F, Thornell AP, Crompton T, et al. Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development. Cell. 1997;89(5):765-71.
220. Narkar VA, Fan W, Downes M, et al. Exercise and PGC-1alpha-independent synchronization of type I muscle metabolism and vasculature by ERRgamma. Cell metabolism. 2011;13(3):283-93.
221. Dacquin R, Starbuck M, Schinke T, Karsenty G. Mouse alpha1(I)-collagen promoter is the best known promoter to drive efficient Cre recombinase expression in osteoblast. Developmental dynamics : an official publication of the American Association of Anatomists. 2002;224(2):245-51.
222. Lu Y, Xie Y, Zhang S, Dusevich V, Bonewald LF, Feng JQ. DMP1-targeted Cre expression in odontoblasts and osteocytes. Journal of dental research. 2007;86(4):320-5.
223. Murray J, Auwerx J, Huss JM. Impaired myogenesis in estrogen-related receptor gamma (ERRgamma)-deficient skeletal myocytes due to oxidative stress. FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 2013;27(1):135-50.
224. Adhami MD, Rashid H, Chen H, Clarke JC, Yang Y, Javed A. Loss of Runx2 in Committed Osteoblasts Impairs Postnatal Skeletogenesis. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2014.