Identification of gene pathways altered by deletion of the androgen receptor specifically in mineralizing osteoblasts and osteocytes in mice

in Journal of Molecular Endocrinology

Androgens play a key role in skeletal growth and maintenance in males and can mediate their actions, at least in part, via the androgen receptor (AR) in osteoblasts. To investigate the mechanisms by which androgens exert their effects via the AR in mineralizing osteoblasts and osteocytes, we identified gene targets/pathways regulated by the AR using targeted gene expression and microarray approaches on bone isolated from mice in which the AR is specifically deleted in mineralizing osteoblasts and osteocytes (mOBL-ARKOs). Gene ontology mining indicated a number of biological processes to be affected in the bones of mOBL-ARKOs including skeletal and muscular system development and carbohydrate metabolism. All genes identified to have altered expression in the bones of mOBL-ARKOs were confirmed by Q-PCR for their androgen responsiveness in an androgen deprivation and replacement mouse model. The osteoblast genes Col1a1 and Bglap and the osteoclast genes Ctsk and RANKL (Tnfs11) were upregulated in the bones of mOBL-ARKOs, consistent with the increased matrix synthesis, mineralization, and bone resorption observed previously in these mice. Of significant interest, we identified genes involved in carbohydrate metabolism (adiponectin and Dpp4) and in growth and development (GH, Tgfb (Tgfb2), Wnt4) as potential targets of androgen action via the AR in mineralizing osteoblasts.

Abstract

Androgens play a key role in skeletal growth and maintenance in males and can mediate their actions, at least in part, via the androgen receptor (AR) in osteoblasts. To investigate the mechanisms by which androgens exert their effects via the AR in mineralizing osteoblasts and osteocytes, we identified gene targets/pathways regulated by the AR using targeted gene expression and microarray approaches on bone isolated from mice in which the AR is specifically deleted in mineralizing osteoblasts and osteocytes (mOBL-ARKOs). Gene ontology mining indicated a number of biological processes to be affected in the bones of mOBL-ARKOs including skeletal and muscular system development and carbohydrate metabolism. All genes identified to have altered expression in the bones of mOBL-ARKOs were confirmed by Q-PCR for their androgen responsiveness in an androgen deprivation and replacement mouse model. The osteoblast genes Col1a1 and Bglap and the osteoclast genes Ctsk and RANKL (Tnfs11) were upregulated in the bones of mOBL-ARKOs, consistent with the increased matrix synthesis, mineralization, and bone resorption observed previously in these mice. Of significant interest, we identified genes involved in carbohydrate metabolism (adiponectin and Dpp4) and in growth and development (GH, Tgfb (Tgfb2), Wnt4) as potential targets of androgen action via the AR in mineralizing osteoblasts.

Keywords:

Introduction

Androgens are essential for skeletal growth and bone accrual during puberty (Vanderschueren et al. 2004) and for bone maintenance post-puberty in males (Basaria & Dobs 2001), determining both the strength and size of adult bone (Wiren 2005). We and others have shown that global deletion of the androgen receptor (AR) in mice results in bones of reduced size and cortical thickness, with decreased trabecular bone volume compared with control males (Yeh et al. 2002, Kawano et al. 2003, Venken et al. 2006, MacLean et al. 2010), indicating that these important actions of androgens are via the AR in bone and not through aromatization to estrogen and action via the estrogen receptor. However, the exact mechanism by which androgens exert these actions via the AR on bone remains unclear.

There is convincing evidence in the literature indicating that the effects of androgens on bone are mediated, at least in part, by the AR expressed on osteoblasts (Colvard et al. 1989, Kasperk et al. 1989). There are three distinct stages of osteoblast development: proliferation, matrix development and maturation, and mineralization of the bone matrix (Owen et al. 1990). During the bone formation process, a proportion of osteoblasts are embedded within the mineralized bone matrix, becoming osteocytes (Franz-Odendaal et al. 2006), which also express the AR (Abu et al. 1997). It is therefore conceivable that androgens may have different actions on osteoblasts at different stages of their development.

Insight into the role of androgens via the AR in osteoblasts at specific stages of their development has been gained by the use of genetically modified mouse models. Transgenic mice over-expressing the AR from the proliferating stage of osteoblast development, including those located at the periosteum, using the 3.6 kb Col1a1 promoter, have increased periosteal apposition (Wiren et al. 2004). In contrast, over-expression of the AR in osteoblasts from late in the matrix development and maturation stage using the 2.3 kb Col1a1 promoter has no effect on bone size as measured by periosteal circumference (Wiren et al. 2008). Both AR transgenic models do, however, have a common phenotype of reduced formation at endocortical surfaces and increased trabecular bone volume as a result of reduced bone turnover (Wiren et al. 2004, 2008). Together, these data suggest that activation of the AR in immature osteoblasts mediates its anabolic effects on the periosteum, while AR action in both immature and mature osteoblasts inhibits bone formation at the endocortical surfaces while downregulating bone turnover within trabecular bone. Consistent with these findings, we have previously shown that deletion of the AR specifically in mineralizing osteoblasts and osteocytes in mOBL-ARKOs using the Cre/loxP system leads to trabecular and cortical bone loss due to increased bone resorption and an uncoupling of the bone matrix synthesis and mineralization processes, while bone size was not affected (Chiang et al. 2009).

We now extend our studies on the mOBL-ARKO mice in order to identify target genes/pathways in bone that are regulated either directly or indirectly by androgens acting via the AR in mineralizing osteoblasts and osteocytes. To achieve this, we used two approaches. First, we used a targeted approach in which we determined the level of expression of genes expressed by osteoblasts, at defined stages of their development, and by osteoclasts, in the bones of mOBL-ARKOs and wild-type (WT) littermate controls. The second approach was to use microarray analysis to identify novel gene targets/pathways in bone regulated by the AR in mineralizing osteoblasts.

Materials and methods

Mice

The generation of the mOBL-ARKO mice used in this study has been described previously (Chiang et al. 2009). Briefly, mOBL-ARKOs were generated by breeding floxed AR mice (Notini et al. 2005) with osteocalcin-Cre mice (Chiang et al. 2009). mOBL-ARKOs were on a congenic C57BL/6 background. WT littermates were used as controls in all experiments. Genomic DNA isolated from tail biopsy was used as a template for PCR genotyping. The primer sequences used to detect the Cre transgene and the floxed AR allele have been described previously (Notini et al. 2007). Mice were supplied with water and standard chow containing 1.2% calcium and 0.96% phosphorous (Ridley Agriproducts, Corowa, NSW, Australia) ad libitum and were housed at 22 °C in a 12 h light:12 h darkness cycle in standard cages. All procedures involving animals were approved by the Austin Health Animal Ethics Committee.

Serum and bone collection

WT and mOBL-ARKO mice were killed at 6 and 12 weeks of age by isoflurane anesthesia and cervical dislocation following a 4 or 12 h fast for 6- and 12-week-old mice respectively. Both tibia and femora were dissected and growth plates and all surrounding soft and connective tissue was removed. Cardiac blood was collected and serum was stored at −20 °C. In order to test the androgen responsiveness of those genes identified to be regulated in mOBL-ARKOs by microarray and gene expression analyses, 7-week-old WT male mice were randomly allocated into sham, orchidectomy (ORX) plus vehicle, or ORX plus dihydrotestosterone (DHT) treatment experimental groups. Empty control implants or DHT implants (3×1 cm; Axell et al. 2006) were placed in the intra-abdominal fat layer at the time of surgery. At 10 weeks posttreatment, mice were killed by i.p. injection of 100 mg/kg sodium pentobarbitone and both tibiae were dissected and all surrounding soft and connective tissue was removed.

RNA isolation and cDNA synthesis

Total RNA was isolated from whole tibia and femur, including bone marrow but excluding growth plates, of WT and mOBL-ARKO male mice at 6 and 12 weeks of age and both tibiae of sham, ORX, and ORX plus DHT groups as described previously (Davey et al. 2000). Total RNA (2 μg) was treated with two units of DNase I according to the manufacturer's instructions (DNA-Free Kit; Ambion, Austin, TX, USA). cDNA was synthesized from 1 μg DNase-treated RNA using random hexamers (Promega) and M-MLV reverse transcriptase, according to the manufacturer's instructions (Promega).

Microarray analysis

Microarray analysis was performed by the Australian Genome Research Facility using the Affymetrix mouse genome 430 v2.0 array. This array uses more than 45 000 probe sets to measure gene expression of more than 39 000 transcripts and variants, including 34 000 well-characterized genes. Five micrograms of total RNA from WT and mOBL-ARKO males at 6 and 12 weeks of age (n=3 per group) were analyzed in duplicate. Normalization and differential gene expression analysis were carried out by the Australian Genome Research Facility using Partek Genomic Suite (Copyright Partek Inc., St Louis, MO, USA) and R/Bioconductor software (http://www.bioconductor.org). Genes were considered to be androgen responsive if expression was increased or decreased compared with controls by ≥1.5-fold (P<0.05). Pathways, processes, or functions that were over-represented in the androgen-responsive gene set were identified using the online Ingenuity pathway analysis tool (http://www.ingenuity.com/).

Quantitative real-time PCR

The mRNA levels of a number of genes known to be important in osteoblast and osteoclast development were determined in bone RNA samples of WT and mOBL-ARKOs (n=15–20 per group) using quantitative real-time PCR (Q-PCR). The expression of genes of interest, identified to be androgen responsive by microarray analysis, was also confirmed in these bone RNA samples. Q-PCR was performed in duplicate using 10–75 ng cDNA per 20 μl reaction on an Applied Biosystems 7500 Real-Time PCR System, using Applied Biosystems TaqMan gene expression assays (Supplementary Table 1, see section on supplementary data given at the end of this article). Absolute expression was calculated using the CT method with values for the gene of interest normalized to eukaryotic 18S rRNA and expressed relative to a reference sample.

Serum biochemistry

Serum DHT was analyzed by organic solvent extraction and RIA (Beckman Coulter, Fullerton, CA, USA) as described previously (MacLean et al. 2008). The sensitivity of the assay is 0.01 nM, the intra-assay coefficient of variation range is ≤6.2%, the inter-assay coefficient of variation is ≤8.5%, and the assay is highly specific for DHT. Serum adiponectin was measured by ELISA according to the manufacturer's instructions (Quantikine; R&D Systems, Inc., Minneapolis, MN, USA). Fasting serum glucose levels were measured using a GM7 Analox glucose analyser (Helena Laboratories, Beaumont, TX, USA).

Statistical analysis

To compare the means of two groups, the unpaired Student's t-test was used, with Levene's test for homogeneity of variance. To compare more than two groups, data were analyzed by one-way ANOVA, with Tukey's or Tamhane's post hoc tests for equal and unequal variance respectively. All data were analyzed with SPSS for MacOSX except the GO χ2 analysis, which was calculated using the Ingenuity online tool. A P value of <0.05 was considered significant.

Results

Targeted approach

To identify genes regulated directly or indirectly via the AR, we first used a targeted approach to determine the mRNA levels of genes expressed by osteoblasts at defined stages of development and by osteoclasts in mOBL-ARKOs and WT controls. The mRNA levels of the osteoblast-specific genes, Col1a1 (P=0.05) and Bglap (P<0.05), were increased twofold in the bones of mOBL-ARKOs compared with WT mice at 6 and 12 weeks of age respectively (Figs 1 and 2). mRNA expression of the osteoclast genes, Ctsk and RANKL, was increased in the bones of mOBL-ARKO by 3.4-fold at 6 weeks of age and by twofold at 12 weeks of age, respectively (P<0.05), while the ratio of RANKL:OPG did not differ from WT (Fig. 2).

Figure 1
Figure 1

mRNA levels of osteoblast genes in the bones of WT and mOBL-ARKO mice at 6 and 12 weeks of age by Q-PCR. Runx2, runt-related transcription factor 2 (cbfa1); Col1a1, type 1 α1 collagen; Alp1, alkaline phosphatase 1; Spp1, osteopontin; Bglap, osteocalcin. Mean+s.e.m., n=16–20 per group, *P<0.05 vs WT within age group, unpaired Student's t-test.

Citation: Journal of Molecular Endocrinology 49, 1; 10.1530/JME-12-0014

Figure 2
Figure 2

mRNA levels of osteoclast genes in the bones of WT and mOBL-ARKO mice at 6 and 12 weeks of age by Q-PCR. Ctsk, cathepsin K; Acp5, tartrate-resistant acid phosphatase; RANK, receptor activator of nuclear factor κB; RANKL, receptor activator of nuclear factor κB ligand; OPG, osteoprotegerin. Mean+s.e.m., n=16–20 per group; *P<0.05 vs WT within age group, unpaired Student's t-test.

Citation: Journal of Molecular Endocrinology 49, 1; 10.1530/JME-12-0014

Microarray analysis

To identify novel gene targets/pathways activated by the AR in osteoblasts, we used microarray analysis. In the bones of mOBL-ARKOs, microarray analysis identified 24 genes that were regulated by ≥1.5-fold at 6 weeks of age, 374 genes regulated at 12 weeks of age, with 13 genes regulated at both 6 and 12 weeks of age compared with WT (data available at http://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-1123). Gene ontology mining performed on the androgen-responsive genes identified a number of significantly over-represented functional processes/disorders including cellular growth and proliferation, cellular development, skeletal and muscular system development and function, and carbohydrate metabolism (Tables 1 and 2). Candidate genes representative of these biological processes were chosen for further investigation based on the evidence provided from PubMed searches to support their potential role in the regulation of bone. The regulation of key candidate genes in the bones of mOBL-ARKOs compared with WT mice was confirmed by Q-PCR (n=16–20 per group). A number of genes of interest, whose expression was identified to be altered in mOBL-ARKOs by microarray, could not be confirmed by Q-PCR analysis in a larger cohort of bone cDNA samples (data not shown). In the bones of mOBL-ARKOs, the mRNA expression of genes associated with bone growth and development was upregulated; Gh mRNA expression was increased 3.6-fold (P<0.005), Wnt4 increased by 2.3-fold (P<0.05), Irf4 increased by 2.6-fold (P<0.001), and Tgfb2 increased by twofold (P<0.05) compared with WT at 12 weeks of age (Fig. 3). ltbp2 gene expression was upregulated 3.4- and 2.4-fold at 6 (P<0.005) and 12 weeks (P<0.05) of age, respectively, in mOBL-ARKOs compared with controls (Fig. 3). Expression of genes involved in carbohydrate metabolism, Dpp4 and Adipoq, were upregulated in mOBL-ARKOs by 2.6-fold at 6 weeks of age (P<0.05) and by 2.7-fold at 12 weeks of age (P<0.005) (Fig. 4).

Table 1

Summary of Top Bio Functions identified by GO analysis from microarray analysis of 24 genes to be over-represented in bones of mOBL-ARKOs at 6 weeks of age (≥1.5-fold)

NameP value
Molecular and cellular functions
 Cellular growth and proliferation4.7×10−5 to 4.0×10−2
 Cellular development1.2×10−4 to 3.7×10−2
 Gene expression6.2×10−4 to 2.5×10−2
 Cell death1.9×10−3 to 4.6×10−2
 Antigen presentation6.8×10−3 to 8.7×10−3
Physiological system development and function
 Hematological system development and function4.7×10−5 to 3.7×10−2
 Humoral immune response4.7×10−5 to 2.8×10−2
 Tissue development4.7×10−5 to 9.9×10−3
 Hematopoiesis1.2×10−4 to 3.7×10−2
 Connective tissue development and function1.9×10−3 to 1.8×10−2
Table 2

Summary of Top Bio Functions identified by GO analysis from microarray analysis of 374 genes to be over-represented in bones of mOBL-ARKOs at 12 weeks of age (≥1.5-fold)

NameP value
Molecular and cellular functions
 Carbohydrate metabolism1.3×10−7 to 1.3×10−2
 Cellular development3.3×10−6 to 1.3×10−2
 Cellular growth and proliferation5.3×10−6 to 1.3×10−2
 Cell death2.0×10−5 to 1.3×10−2
 Small molecule biochemistry2.9×10−5 to 1.3×10−2
Physiological system development and function
 Tissue morphology8.5×10−16 to 1.3×10−2
 Skeletal and muscular system development and function2.0×10−15 to 1.3×10−2
 Hematological system development and function4.2×10−6 to 1.3×10−2
 Humoral immune response4.2×10−6 to 1.3×10−2
 Hematopoiesis5.9×10−6 to 1.3×10−2
Figure 3
Figure 3

mRNA levels of genes involved in growth and development in the bones of WT and mOBL-ARKO mice at 6 and 12 weeks of age by Q-PCR. Tgfb2, TGFB2; Wnt4, wingless-related MMTV integration site 4; ltbp2, latent transforming growth factor binding protein 2; Irf4, interferon regulatory factor 4. Mean+s.e.m., n=16–20 per group; *P<0.05, **P<0.005, ***P<0.001 vs WT within age group, unpaired Student's t-test.

Citation: Journal of Molecular Endocrinology 49, 1; 10.1530/JME-12-0014

Figure 4
Figure 4

mRNA levels of genes involved in carbohydrate metabolism in the bones of WT and mOBL-ARKOs at 6 and 12 weeks of age by Q-PCR. Adipoq, adiponectin; Dpp4, dipeptidyl peptidase 4. Mean+s.e.m., n=16–20 per group; *P<0.05, **P<0.005 vs WT within age group, unpaired Student's t-test.

Citation: Journal of Molecular Endocrinology 49, 1; 10.1530/JME-12-0014

Androgen deprivation and replacement experiment

Genes confirmed to be regulated in mOBL-ARKOs were further investigated for their response to androgen deprivation and replacement using an ORX and DHT treatment mouse model. Serum DHT levels were decreased in ORX WT males compared with sham controls (mean±s.e.m.; sham: 1.55±0.49 nM, n=11, vs ORX+vehicle: 0.45±0.07 nM, n=11, P<0.05). Treatment of ORX males with DHT implants for 10 weeks increased serum DHT levels by 40-fold (mean±s.e.m.; ORX+vehicle: 0.45±0.07 nM, n=11, vs ORX+DHT: 17.79±2.37 nM, n=13, P<0.001). As serum estradiol is below the detection limit of 5 pg/ml in male mice, such measurements were not performed as they would not be informative in ORX mice with or without DHT treatment. A number of genes confirmed to be upregulated in mOBL-ARKOs by Q-PCR were also shown to be increased in WT males following ORX (Fig. 5a, b, c and d). This comprised osteoblast genes, Col1a1 and Bglap (Fig. 5a), and those identified by microarray analysis to be involved in growth and development, Gh and Irf4 (Fig. 5c), and carbohydrate metabolism, Adipoq (Fig. 5d). Treatment of ORX males with DHT restored the mRNA levels of these genes to those observed in WT sham-operated controls (Fig. 5a, b, c and d).

Figure 5
Figure 5

mRNA levels of genes associated with (a) osteoblast development, (b) osteoclasts, (c) growth and development, and (d) carbohydrate metabolism that were altered in mOBL-ARKOs in response to androgen deprivation (ORX WT males+vehicle) and androgen replacement (ORX+DHT). Mean+s.e.m., n=9–11 per group; *P<0.05 vs WT sham, **P<0.005 vs ORX+vehicle, ANOVA and post hoc test.

Citation: Journal of Molecular Endocrinology 49, 1; 10.1530/JME-12-0014

Serum adiponectin and glucose

Serum adiponectin was decreased in mOBL-ARKOs at 6 weeks of age compared with WT (P<0.01) but was unchanged at 12 weeks of age (Fig. 6a). Serum glucose was unaffected in mOBL-ARKOs at 6 or 12 weeks of age (Fig. 6b).

Figure 6
Figure 6

Serum (a) adiponectin (μg/ml) and (b) glucose (mmol/l) in wild-type (WT) and mOBL-ARKOs at 6 and 12 weeks of age. Mean+s.e.m., n=9–14 per group, **P<0.005 vs WT within age group, unpaired Student's t-test.

Citation: Journal of Molecular Endocrinology 49, 1; 10.1530/JME-12-0014

Discussion

It is well established that androgens play an important role in skeletal growth and bone maintenance in males. We have previously shown that the actions of androgens on bone, at least in part, are mediated via the AR located in osteoblasts (Notini et al. 2007, Chiang et al. 2009, MacLean et al. 2010). We have shown using our mOBL-ARKO mouse model that androgens act through the AR in mineralizing osteoblasts and osteocytes to maintain cortical and trabecular bone by controlling the coordination of bone matrix synthesis and mineralization and by indirectly regulating bone resorption (Chiang et al. 2009). We now extend our studies on male mOBL-ARKO mice in order to identify target genes/pathways within bone that are regulated either indirectly or directly by AR action in mineralizing osteoblasts.

To achieve this aim, we used two approaches. The first was a targeted approach whereby we determined the levels of genes expressed specifically by osteoblasts at different stages of their development and also genes known to be highly expressed by osteoclasts. The second approach was microarray analysis on RNA extracted from whole bone of mOBL-ARKOs and controls to identify possible novel gene targets and pathways responsible for the bone phenotype observed in these mice (Chiang et al. 2009). Our rationale for determining the gene expression profiles in both immature (6 weeks of age) and adult (12 weeks of age) mOBL-ARKOs is that the bone loss in mOBL-ARKOs was most marked at 6 weeks of age, with an unexpected compensatory increase in osteoblast activity observed in adult mice at 12 weeks of age (Chiang et al. 2009). The rationale for performing microarray analysis on RNA collected from whole bone was to ensure that we had the ability to identify genes that are regulated either indirectly or directly by AR action in mineralizing osteoblasts in both cortical and trabecular bone and in all bone cell types. This was particularly important given that we have previously shown that deletion of the AR specifically in mineralizing osteoblasts in mOBL-ARKOs has not only direct effects on bone formation but also indirect effects on osteoclasts to increase bone resorption, with these effects being observed in both trabecular and cortical bone (Chiang et al. 2009).

Gene ontology mining performed on the androgen-responsive genes identified a number of significantly over-represented functional processes/disorders including those involved in growth and development, in particular those involved in skeletal and muscular system development and function and carbohydrate metabolism. A number of genes identified to have altered expression in the bone of mOBL-ARKOs by microarray analysis could not be confirmed by Q-PCR, highlighting the importance of confirming all microarray analysis gene regulation data in a larger cohort of RNA samples before drawing conclusions. All the genes discussed below were confirmed to have altered expression in RNA isolated from the whole bones of mOBL-ARKOs compared with controls by Q-PCR with 16–20 samples per group. Given the heterogeneous nature of bone together with the intricate communication network that exists between bone cells, this approach has advantages over using in vitro cultures of isolated osteoblasts from mOBL-ARKOs, as it is a better representation of the bone microenvironment in vivo and also allows the identification of genes indirectly altered in other bone cell types as a result of loss of AR action in mineralizing osteoblasts, such as osteoclasts. As a proof of principal, all genes confirmed to be regulated following their identification by either the targeted or microarray approach were investigated for their androgen responsiveness in an androgen deprivation and replacement male mouse model. ORX led to negligible levels of serum DHT. As mice do not produce adrenal androgens (Payne & Hales 2004), the ORX mice are completely androgen null and the low levels of DHT detected in these mice must be due to the low level of cross-reactivity of the anti-DHT antibody with other steroid hormones (Yarrow et al. 2010). Treatment of the ORX mice with the nonaromatizable androgen, DHT, resulted in supraphysiological levels of DHT, 40-fold higher than those observed in sham controls. As expected, the expression of a number of genes identified to be upregulated in mOBL-ARKOs (Col1a1, Bglap, Gh, Irf4, and Adipoq) was increased following ORX, and treatment of these mice with DHT restored the expression levels to those observed in sham-operated controls. The upregulation of Col1a1 and Bglap expression in the bones of ORX males is consistent with the well-established increase in bone turnover and bone loss following ORX in male rodents, which is prevented with DHT treatment (Vanderschueren et al. 1992). As DHT cannot be aromatized to estradiol, these data provide further evidence to support the regulation of the expression of these genes by androgens acting via the AR in mineralizing osteoblasts and osteocytes.

Abnormalities in bone mineralization previously observed in the mOBL-ARKOs were characterized by increased unmineralized bone matrix (osteoidosis) and a reduction in the number of mineralizing surfaces in skeletally immature mOBL-ARKOs. The osteoidosis persisted throughout adulthood in the mOBL-ARKOs despite a compensatory increase in both the extent and the rate of mineralization (Chiang et al. 2009). Consistent with these findings, we identified increases in the expression of genes encoding type 1 α1 collagen (Col1a1) and osteocalcin (Bglap), which are expressed during the matrix development and maturation and mineralization stages of osteoblast development, respectively, in addition to upregulation of the gene encoding latent transforming growth factor binding protein (ltbp2), a structural component of bone matrix. Consistent with the increased bone resorption and subsequent loss of trabeculae and thinning of cortical bone observed in the mOBL-ARKOs (Chiang et al. 2009), the expression of the osteoclast genes, Ctsk and RANKL, were increased in the bones of mOBL-ARKOs compared with controls at 6 weeks of age.

We previously hypothesized that the compensatory increase in mineralization observed in the absence of androgen action in the mineralizing osteoblasts of adult mOBL-ARKOs may be mediated by increased mechanical forces exerted on the fewer remaining trabeculae and/or action via the GH/IGF1 axis (Chiang et al. 2009). In support of this hypothesis, we identified a number of genes known to be associated with increased bone formation that were upregulated in the adult mOBL-ARKOs, including the growth factors, Gh and TGFB (Tgfb2), and the osteoblastogenesis stimulatory factor, WNT4 (Wnt4).

GH in the circulation is derived primarily from the anterior pituitary (Harvey 2010) but is also made locally within bone and acts in an autocrine or paracrine manner to stimulate bone growth (Baker et al. 1992, Harvey 2010). The upregulation of Gh expression in the bones of mOBL-ARKOs is consistent with the increase in GH secretion that occurs following AR blockade in males during late puberty (Metzger & Kerrigan 1993).

Osteoblasts secrete TGFB and deposit it into the bone matrix, which they can then respond to in an autocrine manner (Robey et al. 1987). The major action of TGFB in bone is to increase the proliferation and early differentiation of osteoblasts, while inhibiting terminal differentiation into osteocytes and apoptosis (Bonewald & Dallas 1994, Karsdal et al. 2002). The increase in Tgfb2 expression in the bones of mOBL-ARKOs is consistent with the increased bone mineral matrix and mineralization in these mice. A possible role of androgens via the AR in the regulation of Tgfb2 expression is supported by the observation that androgen withdrawal increases Tgfb2 mRNA in LNCap cells (Knabbe et al. 1993).

Wnt4, which was also upregulated in adult mOBL-ARKOs, is expressed by bone lining cells and osteocytes and stimulates osteogenic differentiation of mesenchymal stem cells via the noncanonical signaling pathway in vitro and bone formation in vivo (Chang et al. 2007). While it has been shown that WNT4 is the first signaling molecule to influence the sex determination cascade to regulate the development of the female reproductive tract by antagonizing the production of testosterone (Biason-Lauber & Konrad 2008), to our knowledge, this is the first evidence for a role of androgens acting via the AR in osteoblasts to regulate Wnt4 gene expression within bone.

Of significant interest, we identified two genes involved in carbohydrate metabolism to be upregulated by deletion of the AR in mineralizing osteoblasts and osteocytes. The first was Adipoq, which encodes adiponectin. The CNS contributes directly to the regulation of bone homeostasis via efferent neural pathways (Wong et al. 2008). It has recently been proposed that adiponectin, predominantly a lipid-derived adipokine, is one of the key candidates responsible for the cross talk between bone and fat. Supportive evidence for this theory includes the expression of adiponectin receptors by osteoblasts (Berner et al. 2004) and osteoclasts (Oshima et al. 2005), and over-expression of adiponectin in mice leads to increased trabecular bone volume due to decreased bone resorption (Oshima et al. 2005). Interestingly, serum adiponectin levels are inversely related to visceral fat mass (Arita et al. 1999). It has been proposed that this paradoxical relationship is maintained by the inhibition of adiponectin secretion by cytokines and hormones including testosterone (Swarbrick & Havel 2008) evidenced by the elevated levels of circulating adiponectin in hypogonadal men and ORX mice, which decrease following treatment with testosterone (Lanfranco et al. 2004, Page et al. 2005, Xu et al. 2005). The upregulation of Adipoq expression in bone of the mOBL-ARKOs is consistent with an action of androgens, acting via the AR in mineralizing osteoblasts to negatively regulate the expression of adiponectin. The increased Adipoq expression within bone of the mOBL-ARKOs together with the increased mineralization observed in these mice is also consistent with the positive actions through the autocrine/paracrine pathway of locally produced adiponectin within bone to stimulate bone formation (Berner et al. 2004, Shinoda et al. 2006).

Based on our previous observation of increased adiposity and serum adiponectin in global AR knockout mice (Rana et al. 2011), we determined the serum levels of adiponectin and fasting glucose in mOBL-ARKOs. In mOBL-ARKOs, serum glucose levels were unaffected; however, adiponectin was decreased at 6 weeks of age, which returned to normal by adulthood. This is perhaps not surprising given that the levels of adiponectin found in the serum do not necessarily reflect the local concentrations found within the bone microenvironment (Berner et al. 2004). While mOBL-ARKOs have normal body weight (Chiang et al. 2009), the mass of different fat pads and the local concentrations of adiponectin within the bone have not been determined. Nonetheless, these findings of the regulation of both circulating adiponectin and Adipoq expression within bone in the absence of AR action in mineralizing osteoblasts are intriguing and warrant further investigation.

The second gene involved in carbohydrate metabolism identified to be upregulated in mOBL-ARKOs was Dpp4, which encodes dipeptidyl peptidase 4 (DPP4). DPP4 plays a major role in glucose metabolism by degrading incretins, the gastrointestinal hormones that increase insulin secretion from β-cells after eating, before the rise in blood glucose (Drucker & Nauck 2006). Evidence for a role of DPP4 in the regulation of bone metabolism, to date, is limited. DPP4 has been shown to be expressed by cells of the monocyte–macrophage lineage, osteosarcomas, and osteoclasts (Dohi et al. 2009), while different DPP4 inhibitors have been shown to have opposite effects on bone mineral density in mice (Kyle et al. 2011). While low circulating levels of testosterone have been implicated to play a role in the pathogenesis of diabetes in humans (Grossmann et al. 2010) and DPP4 has been identified to have actions on bone, the potential role of androgens in the regulation of DPP4 within bone is unknown.

This study is the first step in identifying genes within bone that have altered expression as a result of loss of androgen action via the AR in mineralizing osteoblasts and osteocytes. The targeted and microarray approaches used in this study identified a number of genes of significant interest. Many of the genes shown to be upregulated in mOBL-ARKOs were consistent with their bone phenotype and the action of androgens via the AR in mineralizing osteoblasts to regulate matrix synthesis and mineralization (Col1a1, Bglap, Gh, and Tgfb2) and to indirectly inhibit bone resorption (Cstk and RANKL; Chiang et al. 2009), thereby validating our experimental approach. Importantly, this approach also enabled the identification of novel AR gene targets within bone, such as Wnt4, Adipoq, and Dpp4.

In conclusion, androgens act directly via the AR in mineralizing osteoblasts to regulate the expression of genes involved in the development and regulation of osteoblasts, osteoclasts, and skeletal growth, in addition to those involved in energy metabolism. Future studies targeting these genes and pathways within specific bone cell types using knockdown and over-expression approaches will allow further investigation into the downstream targets of the AR in osteoblasts.

Supplementary data

This is linked to the online version of the paper at http://dx.doi.org/10.1530/JME-12-0014.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This study was funded by the National Health and Medical Research Council Australia (Project Grant 566503).

Acknowledgements

The authors would like to thank Dr Helen MacLean, Dr Cherie Chiang, and Maria Chiu for their assistance with this study.

References

  • AbuEOHornerAKusecVTriffittJTCompstonJE1997The localization of androgen receptors in human bone. Journal of Clinical Endocrinology and Metabolism8234933497. doi:10.1210/jc.82.10.3493.

    • Search Google Scholar
    • Export Citation
  • AritaYKiharaSOuchiNTakahashiMMaedaKMiyagawaJHottaKShimomuraINakamuraTMiyaokaK1999Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochemical and Biophysical Research Communications2577983. doi:10.1006/bbrc.1999.0255.

    • Search Google Scholar
    • Export Citation
  • AxellAMMacLeanHEPlantDRHarcourtLJDavisJAJimenezMHandelsmanDJLynchGSZajacJD2006Continuous testosterone administration prevents skeletal muscle atrophy and enhances resistance to fatigue in orchidectomized male mice. American Journal of Physiology. Endocrinology and Metabolism291E506E516. doi:10.1152/ajpendo.00058.2006.

    • Search Google Scholar
    • Export Citation
  • BakerARHollingsheadPGPitts-MeekSHansenSTaylorRStewartTA1992Osteoblast-specific expression of growth hormone stimulates bone growth in transgenic mice. Molecular and Cellular Biology1255415547. doi:10.1128/MCB.12.12.5541.

    • Search Google Scholar
    • Export Citation
  • BasariaSDobsAS2001Hypogonadism and androgen replacement therapy in elderly men. American Journal of Medicine110563572. doi:10.1016/S0002-9343(01)00663-5.

    • Search Google Scholar
    • Export Citation
  • BernerHSLyngstadaasSPSpahrAMonjoMThommesenLDrevonCASyversenUReselandJE2004Adiponectin and its receptors are expressed in bone-forming cells. Bone35842849. doi:10.1016/j.bone.2004.06.008.

    • Search Google Scholar
    • Export Citation
  • Biason-LauberAKonradD2008WNT4 and sex development. Sexual Development2210218. doi:10.1159/000152037.

  • BonewaldLFDallasSL1994Role of active and latent transforming growth factor beta in bone formation. Journal of Cellular Biochemistry55350357. doi:10.1002/jcb.240550312.

    • Search Google Scholar
    • Export Citation
  • ChangJSonoyamaWWangZJinQZhangCKrebsbachPHGiannobileWShiSWangCY2007Noncanonical Wnt-4 signaling enhances bone regeneration of mesenchymal stem cells in craniofacial defects through activation of p38 MAPK. Journal of Biological Chemistry2823093830948. doi:10.1074/jbc.M702391200.

    • Search Google Scholar
    • Export Citation
  • ChiangCChiuMMooreAJAndersonPHGhasem-ZadehAMcManusJFMaCSeemanEClemensTLMorrisHA2009Mineralization and bone resorption are regulated by the androgen receptor in male mice. Journal of Bone and Mineral Research24621631. doi:10.1359/jbmr.081217.

    • Search Google Scholar
    • Export Citation
  • ColvardDSEriksenEFKeetingPEWilsonEMLubahnDBFrenchFSRiggsBLSpelsbergTC1989Identification of androgen receptors in normal human osteoblast-like cells. PNAS86854857. doi:10.1073/pnas.86.3.854.

    • Search Google Scholar
    • Export Citation
  • DaveyRAHahnCNMayBKMorrisHA2000Osteoblast gene expression in rat long bones: effects of ovariectomy and dihydrotestosterone on mRNA levels. Calcified Tissue International677579. doi:10.1007/s00223001100.

    • Search Google Scholar
    • Export Citation
  • DohiOOhtaniHHatoriMSatoEHosakaMNaguraHItoiEKokubunS2009Histogenesis-specific expression of fibroblast activation protein and dipeptidylpeptidase-IV in human bone and soft tissue tumours. Histopathology55432440. doi:10.1111/j.1365-2559.2009.03399.x.

    • Search Google Scholar
    • Export Citation
  • DruckerDJNauckMA2006The incretin system: glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. Lancet36816961705. doi:10.1016/S0140-6736(06)69705-5.

    • Search Google Scholar
    • Export Citation
  • Franz-OdendaalTAHallBKWittenPE2006Buried alive: how osteoblasts become osteocytes. Developmental Dynamics235176190. doi:10.1002/dvdy.20603.

    • Search Google Scholar
    • Export Citation
  • GrossmannMGianattiEJZajacJD2010Testosterone and type 2 diabetes. Current Opinion in Endocrinology Diabetes & Obesity17247256. doi:10.1097/MED.0b013e32833919cf.

    • Search Google Scholar
    • Export Citation
  • HarveyS2010Extrapituitary growth hormone. Endocrine38335359. doi:10.1007/s12020-010-9403-8.

  • KarsdalMALarsenLEngsigMTLouHFerrerasMLochterADelaisseJMFogedNT2002Matrix metalloproteinase-dependent activation of latent transforming growth factor-beta controls the conversion of osteoblasts into osteocytes by blocking osteoblast apoptosis. Journal of Biological Chemistry2774406144067. doi:10.1074/jbc.M207205200.

    • Search Google Scholar
    • Export Citation
  • KasperkCHWergedalJEFarleyJRLinkhartTATurnerRTBaylinkDJ1989Androgens directly stimulate proliferation of bone cells in vitro. Endocrinology12415761578. doi:10.1210/endo-124-3-1576.

    • Search Google Scholar
    • Export Citation
  • KawanoHSatoTYamadaTMatsumotoTSekineKWatanabeTNakamuraTFukudaTYoshimuraKYoshizawaT2003Suppressive function of androgen receptor in bone resorption. PNAS10094169421. doi:10.1073/pnas.1533500100.

    • Search Google Scholar
    • Export Citation
  • KnabbeCKleinHZugmaierGVoigtKD1993Hormonal regulation of transforming growth factor beta-2 expression in human prostate cancer. Journal of Steroid Biochemistry and Molecular Biology47137142. doi:10.1016/0960-0760(93)90067-7.

    • Search Google Scholar
    • Export Citation
  • KyleKAWillettTLBaggioLLDruckerDJGrynpasMD2011Differential effects of PPAR-{gamma} activation versus chemical or genetic reduction of DPP-4 activity on bone quality in mice. Endocrinology152457467. doi:10.1210/en.2010-1098.

    • Search Google Scholar
    • Export Citation
  • LanfrancoFZitzmannMSimoniMNieschlagE2004Serum adiponectin levels in hypogonadal males: influence of testosterone replacement therapy. Clinical Endocrinology60500507. doi:10.1111/j.1365-2265.2004.02007.x.

    • Search Google Scholar
    • Export Citation
  • MacLeanHEChiuWSMaCMcManusJFDaveyRACameronRNotiniAJZajacJD2008A floxed allele of the androgen receptor gene causes hyperandrogenization in male mice. Physiological Genomics33133137. doi:10.1152/physiolgenomics.00260.2007.

    • Search Google Scholar
    • Export Citation
  • MacLeanHEMooreAJSastraSAMorrisHAGhasem-ZadehARanaKAxellAMNotiniAJHandelsmanDJSeemanE2010DNA-binding-dependent androgen receptor signaling contributes to gender differences and has physiological actions in males and females. Journal of Endocrinology20693103. doi:10.1677/JOE-10-0026.

    • Search Google Scholar
    • Export Citation
  • MetzgerDLKerriganJR1993Androgen receptor blockade with flutamide enhances growth hormone secretion in late pubertal males: evidence for independent actions of estrogen and androgen. Journal of Clinical Endocrinology and Metabolism7611471152. doi:10.1210/jc.76.5.1147.

    • Search Google Scholar
    • Export Citation
  • NotiniAJDaveyRAMcManusJFBateKLZajacJD2005Genomic actions of the androgen receptor are required for normal male sexual differentiation in a mouse model. Journal of Molecular Endocrinology35547555. doi:10.1677/jme.1.01884.

    • Search Google Scholar
    • Export Citation
  • NotiniAJMcManusJFMooreABouxseinMJimenezMChiuWSGlattVKreamBEHandelsmanDJMorrisHA2007Osteoblast deletion of exon 3 of the androgen receptor gene results in trabecular bone loss in adult male mice. Journal of Bone and Mineral Research22347356. doi:10.1359/jbmr.061117.

    • Search Google Scholar
    • Export Citation
  • OshimaKNampeiAMatsudaMIwakiMFukuharaAHashimotoJYoshikawaHShimomuraI2005Adiponectin increases bone mass by suppressing osteoclast and activating osteoblast. Biochemical and Biophysical Research Communications331520526. doi:10.1016/j.bbrc.2005.03.210.

    • Search Google Scholar
    • Export Citation
  • OwenTAAronowMShalhoubVBaroneLMWilmingLTassinariMSKennedyMBPockwinseSLianJBSteinGS1990Progressive development of the rat osteoblast phenotype in vitro: reciprocal relationships in expression of genes associated with osteoblast proliferation and differentiation during formation of the bone extracellular matrix. Journal of Cellular Physiology143420430. doi:10.1002/jcp.1041430304.

    • Search Google Scholar
    • Export Citation
  • PageSTHerbstKLAmoryJKCovielloADAnawaltBDMatsumotoAMBremnerWJ2005Testosterone administration suppresses adiponectin levels in men. Journal of Andrology268592.

    • Search Google Scholar
    • Export Citation
  • PayneAHHalesDB2004Overview of steroidogenic enzymes in the pathway from cholesterol to active steroid hormones. Endocrine Reviews25947970. doi:10.1210/er.2003-0030.

    • Search Google Scholar
    • Export Citation
  • RanaKFamBCClarkeMVPangTPZajacJDMacLeanHE2011Increased adiposity in DNA binding-dependent androgen receptor knockout male mice associated with decreased voluntary activity and not insulin resistance. American Journal of Physiology. Endocrinology and Metabolism301E767E778. doi:10.1152/ajpendo.00584.2010.

    • Search Google Scholar
    • Export Citation
  • RobeyPGYoungMFFlandersKCRocheNSKondaiahPReddiAHTermineJDSpornMBRobertsAB1987Osteoblasts synthesize and respond to transforming growth factor-type beta (TGF-beta) in vitro. Journal of Cell Biology105457463. doi:10.1083/jcb.105.1.457.

    • Search Google Scholar
    • Export Citation
  • ShinodaYYamaguchiMOgataNAkuneTKubotaNYamauchiTTerauchiYKadowakiTTakeuchiYFukumotoS2006Regulation of bone formation by adiponectin through autocrine/paracrine and endocrine pathways. Journal of Cellular Biochemistry99196208. doi:10.1002/jcb.20890.

    • Search Google Scholar
    • Export Citation
  • SwarbrickMMHavelPJ2008Physiological, pharmacological, and nutritional regulation of circulating adiponectin concentrations in humans. Metabolic Syndrome and Related Disorders687102. doi:10.1089/met.2007.0029.

    • Search Google Scholar
    • Export Citation
  • VanderschuerenDVan HerckESuikerAMVisserWJSchotLPBouillonR1992Bone and mineral metabolism in aged male rats: short and long term effects of androgen deficiency. Endocrinology13029062916. doi:10.1210/en.130.5.2906.

    • Search Google Scholar
    • Export Citation
  • VanderschuerenDVandenputLBoonenSLindbergMKBouillonROhlssonC2004Androgens and bone. Endocrine Reviews25389425. doi:10.1210/er.2003-0003.

    • Search Google Scholar
    • Export Citation
  • VenkenKDe GendtKBoonenSOphoffJBouillonRSwinnenJVVerhoevenGVanderschuerenD2006Relative impact of androgen and estrogen receptor activation in the effects of androgens on trabecular and cortical bone in growing male mice: a study in the androgen receptor knockout mouse model. Journal of Bone and Mineral Research21576585. doi:10.1359/jbmr.060103.

    • Search Google Scholar
    • Export Citation
  • WirenKM2005Androgens and bone growth: it's location, location, location. Current Opinion in Pharmacology5626632. doi:10.1016/j.coph.2005.06.003.

    • Search Google Scholar
    • Export Citation
  • WirenKMZhangXWToombsARKasparcovaVGentileMAHaradaSJepsenKJ2004Targeted overexpression of androgen receptor in osteoblasts: unexpected complex bone phenotype in growing animals. Endocrinology14535073522. doi:10.1210/en.2003-1016.

    • Search Google Scholar
    • Export Citation
  • WirenKMSemiraleAAZhangXWWooATommasiniSMPriceCSchafflerMBJepsenKJ2008Targeting of androgen receptor in bone reveals a lack of androgen anabolic action and inhibition of osteogenesis: a model for compartment-specific androgen action in the skeleton. Bone43440451. doi:10.1016/j.bone.2008.04.026.

    • Search Google Scholar
    • Export Citation
  • WongIPZenginAHerzogHBaldockPA2008Central regulation of bone mass. Seminars in Cell and Developmental Biology19452458. doi:10.1016/j.semcdb.2008.08.001.

    • Search Google Scholar
    • Export Citation
  • XuAChanKWHooRLWangYTanKCZhangJChenBLamMCTseCCooperGJ2005Testosterone selectively reduces the high molecular weight form of adiponectin by inhibiting its secretion from adipocytes. Journal of Biological Chemistry2801807318080. doi:10.1074/jbc.M414231200.

    • Search Google Scholar
    • Export Citation
  • YarrowJFConoverCFLipinskaJASantillanaCAWronskiTJBorstSE2010Methods to quantify sex steroid hormones in bone: applications to the study of androgen ablation and administration. American Journal of Physiology. Endocrinology and Metabolism299E841E847. doi:10.1152/ajpendo.00384.2010.

    • Search Google Scholar
    • Export Citation
  • YehSTsaiMYXuQMuXMLardyHHuangKELinHYehSDAltuwaijriSZhouX2002Generation and characterization of androgen receptor knockout (ARKO) mice: an in vivo model for the study of androgen functions in selective tissues. PNAS991349813503. doi:10.1073/pnas.212474399.

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

 

      Society for Endocrinology

Sept 2018 onwards Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 823 224 29
PDF Downloads 157 72 9
  • View in gallery

    mRNA levels of osteoblast genes in the bones of WT and mOBL-ARKO mice at 6 and 12 weeks of age by Q-PCR. Runx2, runt-related transcription factor 2 (cbfa1); Col1a1, type 1 α1 collagen; Alp1, alkaline phosphatase 1; Spp1, osteopontin; Bglap, osteocalcin. Mean+s.e.m., n=16–20 per group, *P<0.05 vs WT within age group, unpaired Student's t-test.

  • View in gallery

    mRNA levels of osteoclast genes in the bones of WT and mOBL-ARKO mice at 6 and 12 weeks of age by Q-PCR. Ctsk, cathepsin K; Acp5, tartrate-resistant acid phosphatase; RANK, receptor activator of nuclear factor κB; RANKL, receptor activator of nuclear factor κB ligand; OPG, osteoprotegerin. Mean+s.e.m., n=16–20 per group; *P<0.05 vs WT within age group, unpaired Student's t-test.

  • View in gallery

    mRNA levels of genes involved in growth and development in the bones of WT and mOBL-ARKO mice at 6 and 12 weeks of age by Q-PCR. Tgfb2, TGFB2; Wnt4, wingless-related MMTV integration site 4; ltbp2, latent transforming growth factor binding protein 2; Irf4, interferon regulatory factor 4. Mean+s.e.m., n=16–20 per group; *P<0.05, **P<0.005, ***P<0.001 vs WT within age group, unpaired Student's t-test.

  • View in gallery

    mRNA levels of genes involved in carbohydrate metabolism in the bones of WT and mOBL-ARKOs at 6 and 12 weeks of age by Q-PCR. Adipoq, adiponectin; Dpp4, dipeptidyl peptidase 4. Mean+s.e.m., n=16–20 per group; *P<0.05, **P<0.005 vs WT within age group, unpaired Student's t-test.

  • View in gallery

    mRNA levels of genes associated with (a) osteoblast development, (b) osteoclasts, (c) growth and development, and (d) carbohydrate metabolism that were altered in mOBL-ARKOs in response to androgen deprivation (ORX WT males+vehicle) and androgen replacement (ORX+DHT). Mean+s.e.m., n=9–11 per group; *P<0.05 vs WT sham, **P<0.005 vs ORX+vehicle, ANOVA and post hoc test.

  • View in gallery

    Serum (a) adiponectin (μg/ml) and (b) glucose (mmol/l) in wild-type (WT) and mOBL-ARKOs at 6 and 12 weeks of age. Mean+s.e.m., n=9–14 per group, **P<0.005 vs WT within age group, unpaired Student's t-test.