Nuclear receptors are transcription factors comprising both ligand-dependant molecules and a large number of so-called orphan receptors for which natural ligands have not been identified.1 The first orphan nuclear receptors identified were proteins related to estrogen receptor alpha (ERα) that were referred to as estrogen receptor-related receptors (ERRs) and include three members (ERRα, ERRβ, and ERRγ) (ESRRA, NR3B1; ESRRB, NR3B2; and ESRRG, NR3B3, respectively, according to the Nuclear Receptors Nomenclature).2, 3 The sequence alignment of the ERRs and the ERs reveals a high similarity (68%) in the DNA-binding domain; however, the moderate similarity (36%) in other parts of the proteins such as the ligand-binding E domain may explain why ERRα does not bind estrogen.2, 3 Nevertheless, considerable data support the idea that ERRα may impinge on the estrogen pathway.4, 5 Additionally, the ERRs can be activated by nonhormonal signals including interaction with coactivators such as peroxisome proliferator-activated receptor-gamma coactivator-1 (PGC1α and β).3, 6
It is through their interaction with the PGC1 family in highly metabolic tissues that ERRα/γ elicit their main function: the control of mitochondrial biogenesis by the regulation of fatty acid oxidation (FAO) and energy expenditure.7, 8 Less is known about their role in other tissues, including bone and cartilage, although growing data are available for ERRα.
ERRα in Cartilage
The earliest evidence that ERRα might be functionally important in cartilage was the spatial and temporal correlation seen between high ERRα protein expression and chondrogenesis during mouse development and in articular cartilage later in the adult, paralleling expression of the master chondrocyte regulator Sox9.9, 10 ERRα was also found to activate the Sox9 promoter and modulate Sox9 expression in a rat chondrogenic cell line9 (Fig. 1). Sox9 regulation by ERRα has been confirmed in patients with osteoarthritis (OA), a chronic disease characterized by slowly progressive destruction of the articular cartilage,10, 11 and in neonatal mouse calvarial explants overexpressing ERRα.12 Experiments in which ERRα expression was downregulated in chondrogenic cells suggested that modulation of ERRα contributes to the maturation of proliferating chondrocytes to hypertrophy, most likely due to ERRα's direct or indirect regulation of Sox9. The data also implicate ERRα in cartilage maintenance in aging-related diseases such as rheumatoid arthritis and OA, where ERRα is dysregulated.10, 13
A role of ERRα in the Sox9 pathway is supported by the observation that one of its coactivators, PGC1α, is also expressed in chondrogenic sites where it coactivates Sox9 to regulate chondrogenesis14 (Fig. 1). Treatment with the ERRα inverse agonist XCT790, known to disrupt the interaction between ERRα and PGC1, decreased the expression of Sox9 in human OA chondrocytes, consistent with the view that ERRα and PGC1 together regulate Sox9 expression in chondrocytes supporting the hypothesis that they may function together in cartilage formation and maintenance10 (Fig. 1).
ERRs in Bone
ERRα in osteoclasts
Two recent studies have solidified a role for ERRα in osteoclast (OC) differentiation and function in vivo and in vitro. In the first, ERRα knockout mice exhibited osteopetrosis, resulting from an increase in the trabecular bone and a decrease in OC number.15 Concomitantly, osteoclastogenesis was dramatically disturbed in vitro and genes implicated in mitochondrial biogenesis were downregulated (Fig. 2). Interestingly, the authors showed a direct regulation of ERRα expression by peroxisome proliferator-activated receptor gamma (PPARγ) associated with PGC1β on the PPAR response element in the ERRα promoter in the presence of rosiglitazone and RANKL, suggesting that ERRα is a direct PPARγ target gene in the specific context of osteoclastogenesis (Fig. 2). The results are notable because data from clinical trials suggest that long-term use of rosiglitazone increased fracture rates among patients with type 2 diabetes.16 The regulation of ERRα by PPARγ suggests that ERRα is involved in the cellular and molecular mechanisms by which PPARγ and rosiglitazone regulate bone remodeling in a metabolic disease like diabetes. However, it is worth mentioning that the data suggest that the regulation of ERRα in OC may also have a PPARγ/PGC1β independent component15 (Fig. 2). ERRα has been implicated in other aspects of OC function, in particular in regulation of OC mobility and actin cytoskeletal organization. Indeed, blocking ERRα activity or expression reduced expression of osteopontin (OPN) and the integrin β3 chain, altered distribution of activated c-src (phosphorylated at the Tyr416) and vinculin, contributing to the disruption of the podosome belt, a specific actin structure implicated in OC adhesion, migration, and invasion17, 18 (Fig. 2).
Paralleling ERRα expression, PGC1β is upregulated during the transition from bone marrow macrophages to pre-OC, and PGC1β knockout mice exhibited osteopetrosis.19 Although OC number was not affected, PGC1β-deficient OC displayed abnormal morphology and their bone-resorbing activity was significantly impaired due to a reduction in phosphorylation of c-src at Tyr416 and a decrease in actin ring formation.19 Taken together, these data suggest that ERRα and PGC1β may regulate not only genes involved in mitochondrial biogenesis in OC (MCAD, OXPHOS genes; Fig. 2; Table 1) but also genes involved in actin cytoskeletal organization required for OC migration and resorption capacity (Fig. 2, Table 1).
Table 1. ERRα-Regulated Genes That May Impact Bone Physiology
Several studies support a positive regulatory role for ERRα in OB differentiation and function. Inhibiting ERRα blocks differentiation of rat calvarial (RC) cells into mature OBs, whereas overexpression enhances differentiation with associated increased expression of OPN and osteocalcin (OCN) in vitro.20 Bone marrow mesenchymal stem cells (MSCs) isolated from ERRα-deficient mice exhibited reduced proliferation, osteoblastic differentiation, and mineralization.21
In contrast, two separate studies showed increased bone mineral density with deletion of ERRα in mice.22, 23 Species-specific responses, ie, differences between rats and mice in which studies were done, may underlie some of the differences. As likely, however, are the diverse contributions of both cell-autonomous ERRα functions in OBs and nonautonomous functions (impact of ERRα-expressing adipocytes, mesenchymal cells, immune cells, and endothelial cells in the bone environment). It is worth mentioning that PGC1β–/– mice also exhibit reduced bone formation as a result of a decrease in OB number and associated reduced serum osteocalcin concentration.19 RANKL expression was increased in the bone of PGC1β–/– mice, a result paralleling the upregulation of RANKL in primary OBs in which ERRα is decreased, suggesting that ERRα and PGC1β may indirectly decrease OC formation through their activities in OBs.19, 24
ERRγ is also involved in osteoblastogenesis and can regulate the bone morphogenic protein 2 (BMP2) pathway.25 ERRγ heterozygote male mice exhibited a slightly but significantly increased bone volume fraction (BV/TV) at 14 weeks, associated with an increased bone mineral content, and increased expression of alkaline phosphatase (ALP)26 (Cardelli and Aubin, unpublished data).
ERRα Target Genes and Interacting Proteins That May Impact Bone
Many other ERRα target genes and pathways are known regulators of bone formation and function (Tables 1, 2). In a previous review, we summarized genes (c-erbA1, aromatase, endothelial nitric oxide synthase [eNOS], and lactoferrin) known to be directly regulated by ERRα in non-bone tissues that may impact bone physiology.5 More recently, a variety of other ERRα target genes or interacting molecules that also impact bone turnover have been described. In this article we do not review these in detail, but they include: (1) small heterodimer partner (SHP), which physically interacts with all three ERRs, repressing their activity27–29; (2) hypoxia-inducible factor-α (HIF1α), which physically interacts with ERRα to stimulate HIF-induced transcription30–33; and (3) vascular endothelial growth factor (VEGF), which is a direct target gene of the ERRα/PGC1α complex and HIF1α in muscle and in breast cancer cells.34, 35 Several factors, including prostaglandin E1, BMP, and insulin-like growth factor 1, known to regulate ERRα or γ expression and/or activity, also modulate VEGF production in bone cells, suggesting direct VEGF regulation by ERRα/γ in OBs and OCs10, 25, 38, 39; (4) aryl hydrocarbon receptor (AhR), which has been reported to interact with ERα and ERRα 40–43; (5) PPARα, which was reported to be directly regulated by ERRα in cardiac myocytes and skeletal muscles 44–46; and (6) Wnt11, which has been shown to be regulated in breast cancer cells by ERRα/β-catenin complex and in C3H10T1/2 cells through its interaction with PGC1α.12, 47–49 Recently, we found osteoprotegerin (OPG) regulated in breast cancer cells impacting on bone metastases formation.50
Table 2. ERRα-Interacting Factors That May Impact Bone Physiology
ERRα interacting factors
ERRα = estrogen receptor related receptor alpha; OB = osteoblast; OC = osteoclast.
Small heterodimer partner (SHP)
Human cervical carcinoma (HeLa)
Corepresssor; mutated associated to moderate obesity; target genes of ERRγ; expressed in OB; role in bone formation in vivo
Is ERRα, a Regulator of Adipogenesis, Involved in Glucose Homeostasis in Bone?
We hypothesize that ERRα regulates insulin production and glucose metabolism via three potential pathways in bone (Fig. 3). First, transcriptome analysis of ERRα knockout mice, which display reduced body weight, decreased fat mass, and are resistant to high-fat diet–induced obesity,36 demonstrated altered expression of several genes involved in mitochondrial FAO, and oxidative phosphorylation (OXPHOS) in oxidative tissues such as skeletal muscles, kidney, heart, and white and brown adipose tissues.8 Among them, the medium-chain acyl-coenzyme A dehydrogenase (MCAD), which mediates the first step in the mitochondrial β-oxidation of fatty acid, is directly regulated by ERRα, confirming the involvement of ERRα in the regulation of the rate of tissue FAO.51, 52 Defects in mitochondrial FAO may affect the normal insulin-signaling pathway, thereby contributing to insulin resistance. For instance, a group of genes involved in OXPHOS and known to be downregulated in human diabetic muscle, are direct targets of PGC1α, an action that is dependent on the presence of ERRα.53 ERRα may also control mitochondrial glucose oxidation to acetyl–coenzyme A (CoA) via the regulation of the pyruvate dehydrogenase kinases 2 and 4 (PDK2-4) in adipocyte and hepatoma cells, respectively, inducing a decrease in the conversion of pyruvate to acetyl-CoA in mitochondria and implicating ERRα as a negative regulator of the oxidation of glucose to acetyl-CoA54–58 (Fig. 3). ERRα also acts as a transcriptional repressor of a key gluconeogenic enzyme gene, the phosphoenolpyruvate carboxykinase (PEPCK) gene, suggesting that enhancing ERRα activity may have beneficial effects on glucose metabolism by two distinct mechanisms: increasing mitochondrial oxidative capacity in peripheral tissues and liver, and suppressing hepatic glucose production suggesting potential implication of ERRα in insulin production and sensitivity59 (Fig. 3).
The second potential pathway is related to the view that bone is an endocrine organ critical for the global control of glucose homeostasis. When undercarboxylated, OCN (ucOCN) enters the systemic circulation, where it is believed to act as a hormone, altering insulin production by pancreatic β-cells and global glucose homeostasis.60–63 Insulin itself regulates bone metabolism, stimulating osteoblastogenesis and OCN expression, which favors insulin secretion and sensitivity.64 In addition, the skeleton is also a direct target of the adipose tissue–derived hormone leptin, which inhibits bone formation through the β2 adrenergic receptor (Adrβ2) and negatively regulates OCN activity (leptin-central sympathetic-bone pathway).65–67 ERRα, via its function in mitochondrial oxidative capacity in peripheral tissues and liver, and suppression of hepatic glucose production, regulates insulin production and therefore may regulate bone formation and resorption and increase ucOCN production (Fig. 3). Insulin also upregulates ERRα expression in preadipocytes, a regulation that may also occur in OBs and OCs.68 OCN, similar to ERRα, regulates MCAD in muscle, suggesting crosstalk between ERRα and OCN signaling and implicating OCN in FAO in muscle.60 OCN/ucOCN levels have not been reported in ERRα knockout mice, an issue that we believe deserves attention.
The third pathway is related to the dramatic decrease in fat mass in ERRα knockout mice.36 ERRα positively regulates adipogenesis and lipogenesis,8, 69 which one predicts should alter leptin production, although the latter has not been reported in ERRα knockout mice, leaving open the question of whether ERRα affects bone via leptin (Fig. 3). Nevertheless, regulation of leptin by ERRα may help explain the discrepancy observed in the in vivo and in vitro effects of ERRα in bone. Interestingly, the leptin receptor gene (Lepr) in kidney and the adrenergic receptor β3 (Adrβ3) in adipose tissue are modulated by ERRα, suggesting that ERRα may interfere with the leptin central-sympathetic output-bone pathway 65–67 (Fig. 3).
Is ERRα Involved in Skeletal Aging?
Deregulation of mitochondrial function is a common feature in multiple aspects of aging and a decline in mitochondrial energy metabolism is linked to the development of metabolic disease.71, 72 PGC1 members have been implicated in aging, supporting the hypothesis that, similarly to its coactivators, ERRα may be involved in the aging process in part through its role in FAO.73 Moreover, the dysfunction in mitochondrial biogenesis seen in PGC1β or ERRα knockout mice has indicated that mitochondrial function is also essential for OCs and suggests that mitochondrial dysregulation may be involved in the aging process in bone- and age-related diseases such as osteoporosis.15, 19
Increased adipogenesis in many tissues including the bone marrow is also associated with age-related bone loss,74 a phenomenon attributed to the diversion of MSCs into adipocytes at the expense of OBs.75 Experiments in human and mouse MSCs and RC cells implicate ERRα in MSC fate selection, but whether ERRα acts as an inhibitor/activator of OBs or as an activator for adipocytes depends on the study.21, 22, 24 Interestingly, some authors suggest that ERRα may play different roles in bone versus adipose tissues under different physiological conditions, depending on which PGC1 family member is regulated.15, 21, 22
ERRα expression is dysregulated in OBs in a mouse rheumatoid arthritis model and in chondrocytes in human OA, both conditions more prevalent with aging.10, 13 This is notable in regard to the bone phenotype in ERRα knockout mice, which was more markedly manifested in older mice (ie, 10- to 12-month-old mice) when osteopetrosis due to a defect in OCs was observed versus younger mice (4- to 5-month-old mice) where only a slightly increased OB perimeter and a nonstatistically significant increase in OC number was noticed, suggesting that ERRα may play a role in bone aging.15, 22, 23 Reinforcing that hypothesis, marrow fat volume was decreased in 4-month-old female ERRα knockout mice.22 Genes implicated in mitochondrial biogenesis were also downregulated during osteoclastogenesis in cultures of bone marrow cells isolated from ERRα KO mice, suggesting roles for ERRα in the declines in FAO and mitochondrial function seen in aging and contributing to the increased bone destruction process observed in osteoporosis.15, 19, 73, 76
In addition to mitochondrial function, an increased level of oxidative stress and formation of reactive oxygen species (ROS) in OBs is associated with osteoporosis.77 ROS greatly influence survival of OCs/OBs and oxidative defense by the FoxO transcription factors is required for bone maintenance.19, 75, 78–81 Considerable data now support the idea that ERRα, combined with PGC1 family members, regulates ROS production. Indeed, dysregulation of ERRα with the inverse agonist XCT790 enhanced ROS production in differentiated adipocytes, lung cancer cells, and muscle.54, 82, 83 In muscle, increased production of ROS induced by XCT790 increased expression of glucose transporters, leading to an increase in glucose uptake, reinforcing the idea that modulating ERRα activity may have an impact on insulin sensitivity.83 Troglitazone, an agonist of PPARγ, also interferes with the interactions between ERRα/γ and their coactivator PGC1α, functioning as an inverse agonist. Consequently, troglitazone suppressed the expressions of both PGC1α and PGC1β—which are key regulators of mitochondrial function—reduced mitochondrial mass, and suppressed the expression of superoxide dismutase (SOD) to elevate ROS production.84 Interestingly, and in contrast to the previous data, ERRα combined with PGC1β is required in macrophages for innate response to bacterial pathogenesis by inducing ROS production in response of interferon gamma (IFNγ), suggesting a cell-specific or stress condition–specific function of ERRα/PGC1 in ROS production.85 Several genes regulated by ERRα/PGC1 are now known to be involved in oxidative stress resistance and ROS production in OBs, such as ROS-detoxifying enzymes, the mitochondrial SOD2, and the NAD-dependent deacetylase Sirt3.7, 86, 87 No similar data are yet available in OBs, but these results suggest similar transcriptional regulation may occur in OBs in aging.
We have reviewed the increasing data supporting a role for ERRα in regulation of chondrocytes, OB and OC differentiation and function, and in regulation of glucose uptake and insulin sensitivity in highly metabolic tissues. Together, the data suggest that ERRα may act as a regulator of bone through both direct effects on bone cells and indirect effects on bone's endocrine function. The fact that the bone phenotype observed when ERRα activity is ablated is more marked in older mice, together with its dysregulation in age-related bone diseases and its function in mitochondria, FAO, and in oxidative stress as a regulator of ROS production, suggest that ERRα may act as a regulator of the aging process in bone.
Both authors state that they have no conflicts of interest.
This work was supported by grants from the Canadian Institutes of Health Research (CIHR FRN 88104), the Arthritis Society of Canada (TAS), and the Canadian Arthritis Network (all to JEA); and by fellowship support from the Association Jacques Cartier, TAS, société française de rhumatologie (SFR), and CNRS (all to EB).
Authors' roles: EB and JEA were involved in the design of this perspective and in drafting the manuscript. Both authors take responsibility for all aspects of the article's content.