Androgens stimulate bone formation and play an important role in the maintenance of bone mass. Clinical observations suggest that both gonadal and adrenal androgens contribute to the positive impact of androgenic steroids on bone metabolism. We investigated the mechanism of action of the adrenal androgen dehydroepiandrosterone (DHEA) and its sulfated compound dehydroepiandrosterone sulfate (DHEAS) on human osteoblastic cells (HOCs) in vitro. The DHEA- and DHEAS-induced effects were analyzed in parallel with the actions elicited by the gonadal androgen dihydrotestosterone (DHT). There was no qualitative difference between the effects of gonadal and adrenal androgens on HOC metabolism in vitro. Both were stimulatory as regards cell proliferation and differentiated functions, but the gonadal androgen DHT was significantly more potent than DHEA. The actions of DHT and DHEA on HOC proliferation and alkaline phosphatase (ALP) production could be prevented by the androgen receptor antagonist hydroxyflutamide and inhibitory transforming growth factor β antibodies (TGF-βab), respectively, but were not affected by the presence of the 3β-hydroxysteroid dehydrogenase (3βHSD) and 5-α-reductase (5-AR) inhibitor 17β-N,N-diethylcarbamoyl-4-methyl-4aza-5α-androstan-3-one (4-MA). This indicates that DHT and DHEA (1) exert their mitogenic effects by androgen receptor–mediated mechanisms, (2) stimulate ALP production by increased TGF-β expression, (3) that the action of DHT is not affected by the presence of 4-MA, and that (4) DHEA does not need to be metabolized by 3βHSD or 5-AR first to exert its effects on HOCs in vitro.
The gonadal androgens testosterone and dihydrotestosterone (DHT) are important regulators of bone cell activity and bone mass in animals1,2 and humans.3,4 Additionally, declining serum levels of gonadal androgens are associated with osteopenia in vivo.5 Gonadal androgens are known to act directly on osteoblasts, stimulating growth and differentiation of osteoblastic cells in vitro by binding to an androgen receptor.6,7 Adrenal androgens also play a role in maintaining bone mass. Significant positive correlations have been observed between serum levels and bone mass in both animal and clinical studies.8,9 Dehydroepiandrosterone (DHEA), the major adrenal androgen, reduces the ovariectomy-induced osteopenia in rats10 and DHEA serum levels are positively correlated with bone mass in aging women and patients with deficient adrenal androgen production.11–13 Although a recent study has demonstrated that gonadal and adrenal androgens stimulate alkaline phosphatase production by human bone cells by the same mechanism,14 the effects of adrenal androgens on other aspects of bone cell metabolism, and the mechanism by which they exert their effects on bone cells, remain unknown.
In view of the growing interest in DHEA and its sulphated metabolite (DHEAS) as anti-aging drugs,15–20 the present studies were carried out to determine the effects of the adrenal androgen DHEA on the growth and differentiated functions of human osteoblastic cells (HOCs) in both early and late stages of culture to identify a possible mechanism of action of adrenal androgens on HOC metabolism.
MATERIALS AND METHODS
Human bone cell cultures
HOCs were obtained from bone biopsies of healthy, male patients (18, 19, 27, 43, 54, and 68 years) undergoing elective orthopedic surgery. The experimental protocol was approved by the local ethics committee of the University of Heidelberg. HBCs were harvested and characterized as HOCs as described previously.21 Briefly, cortical bone chips from the femoral shaft were thoroughly cleaned from periosteal tissue and bone marrow. After 10–20 days of continuous culture in Dulbeccos's modified Eagle's medium (DMEM) (GIBCO, Grand Island, NY, U.S.A.) with 10% bovine calf serum (BCS) (Hyclone, Logan, Utah, U.S.A.) and 1% penicillin (100 U/ml)/streptomycin (100 μg/ml) solution (PS) (Irvine Scientific, Santa Ana, CA, U.S.A.), cells were digested from the bone chips by trypsinization and identified as bone cells on the basis of osteocalcin secretion, formation of mineral in vitro, type I collagen mRNA expression, and 1,25-dihydroxyvitamin D3–inducible alkaline phosphatase production. HOCs were grown at 37°C in humidified air (5% CO2/95% air) and only HOCs of the first and second passage were used in experiments. Twenty-four hours and 1 h prior to the addition of the steroids, the culture medium was changed to phenol red–free DMEM containing 1% charcoal-treated BCS and 1% PS. HOCs were enumerated with a hemocytometer after 48 and 72 h. Alkaline phosphatase (ALP) activity, cellular protein, the number of ALP positive staining (ALP+) HOCs, and osteocalcin secreted into the culture medium were determined after 72 h of continuous steroid treatment. For mineralization assays, HOCs were cultured in phenol red–free DMEM containing 10% charcoal-treated BCS and 1% PS. After reaching confluency, the culture medium was changed to fresh medium containing 10 nM β-glycerophosphate, 50 μg/ml ascorbic acid, and steroids. For all experiments, control cultures received ethanol vehicle (0.01% v/v) but did not receive the steroid. The adrenal androgens (DHEA, DHEAS) were generous gifts of Dr. Jan Stepan, Praque, Czech Republic; DHT was purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.) 17β-N,N-diethylcarbamoyl-4-methyl-4aza-5α-androstan-3-one (4-MA) was obtained from Dr. Rasmusson (Merck, Sharp and Dohme Research Laboratories, Rahway, NJ, U.S.A.).
Statistical significance between groups was analyzed by the two-tailed Student's t-test or two-way analysis of variance (ANOVA) as indicated. The results were expressed as percent of mean (n = 6) of controls (mean ± SD). All experiments were repeated at least twice.
DNA synthesis assay
[3H]thymidine incorporation was measured to detect the early effects of androgens on DNA synthesis. HOCs were cultured for 24 h in experimental medium in the presence of androgens, and [3H]thymidine was added at 1 μCi/well for the last 6 h of the incubation period. After incubation with [3H]thymidine, the medium was removed, the cell layer rinsed with phosphate buffered saline (PBS) twice, and precipitated with 0.5 ml of 5% trichloroacetic acid (TCA). After rinsing with 5% TCA, the cells were solubilized using 0.5 ml 0.25 M NaOH. One-half milliliter NaOH cell lysates were transferred into a 3.5 ml scintillation cocktail and measured in a β-scintillation counter.
Specific ALP activity and ALP staining
After 72 h of continuous steroid treatment, cultures were rinsed with PBS, extracted with 0.01% Triton X-100 (Sigma Chemical Co.) containing 0.01% azide, 12.5 mM Tris buffer and 12.5 mM sodium bicarbonate at pH 10.3. The extracts were then frozen at −200°C overnight. The ALP activity was determined in the Triton X-100 cell extracts by spectrophotometrically measuring the breakdown of paranitrophenol-phosphate by ALP. The protein content in the Triton X-100 cell extracts was determined using the bicionic acid (BCA) protein assay (Pierce, Rockford, IL, U.S.A.) and specific ALP activity was expressed as U/g of protein.
For ALP+ staining, HOCs were rinsed with PBS and stained using an azo-dye capture technique employing Naphthol AS-TR phosphate as substrate, fast blue violet LB salt, and 100 μl of 10 mM magnesium chloride in 0.05 M tris buffer, pH 8.6 (Sigma Chemical Co.). After stain removal and a second rinse with PBS, the stained cells were counted in rasters across the culture wells under an inverted Leitz microscope (Leitz, Wetzlar, Germany) equipped with an eyepiece reticule.
The concentration of osteocalcin in the culture media was measured by radioimmunoassay (OSCAtest assay, Henning Berlin GmbH, Berlin, Germany). The amount of osteocalcin measured in the culture media was corrected for cell protein in Triton extracts of the culture wells.
HOC populations were cultured in mineralization medium for 7 days before adding 1 μCi/ml Ca45 (ICN, Irvine, CA, U.S.A.). After 24 h, the culture wells were thoroughly rinsed twice with PBS and the mineral dissolved in 1 N HCl. The hydrolized mineral was centrifuged at 14,000 rpm for 5 minutes, and the amount of Ca45 incorporated into mineral was measured by liquid scintillography in an aliquot of the supernatant.
To examine DHEA binding in confluent HOC populations, we performed displacement binding experiments in a whole-cell binding assay using the androgenic steroids DHT, methyltrienolone (= R1881), and DHEA. Briefly, 24 h prior to binding analysis of confluent HOC populations, the culture medium was changed to serum-phenol red–free DMEM. For the binding experiments, cultures were rinsed twice with cold PBS, and the culture medium was changed to serum-phenol red–free DMEM. Cultures were incubated with 1 nM of the radiolabeled androgenic compound3H-methyltrienolone (3H-R1881 obtained from New England Nuclear, Danvers, MA, U.S.A., specific activity 86 Ci/mmol) alone (= control) or together with various doses of unlabeled DHEA, DHT, or R1881 for 1 h at 4°C. The androgenic steroid methyltrienolone (R1881) was utilized for the whole-cell binding experiment because R1881 has a higher affinity to the androgen receptor (AR) than physiological compounds (e.g., testosterone) and cannot be metabolized by osteoblasts. After 1 h of incubation at 4°C, the cultures were rinsed twice with PBS, and 3 ml of 0.25 N NaOH were added. After solubilization of the cell layers, the cell lysates were added to 15 ml of scintillation fluid and evaluated in a β-scintillation counter.
Mitogenic action of gonadal and adrenal androgens
The observation of positive correlations between serum levels of adrenal androgens and bone mass13 and the finding that the number of differentiated osteoblasts determines the amount of newly formed bone22 suggest that adrenal androgens increase the number of bone cells. So far, gonadal androgens are the only androgenic steroids having been shown to exert direct proliferative action on HBCs via androgen receptor–mediated mechanisms. To examine the possible mitogenic effects of the adrenal steroids DHEA and DHEAS and to compare the effects with the action of gonadal androgens, HBC populations were treated in parallel with adrenal androgens and DHT. DHEA and DHT stimulated HOC proliferation in a dose-dependent manner after treatment for 48 and 72 h (Table 1a). We did not find a reproducible stimulatory effect of DHEAS on HOC proliferation (data not shown). There was no qualitative difference between the effects of DHEA and DHT on HOC growth, but DHT, at doses at or greater than 0.1 nM, elicited a significantly more potent mitogenic response than DHEA after 48 h of treatment (Tables 1a and 2). The finding of a greater mitogenic effect of DHT compared with DHEA may be attributable to a greater affinity of DHT to the androgen receptor. To test this possibility, displacement binding experiments were performed as shown in Figure 1. DHT was more than 20-fold as potent in competing with the synthetic androgen R1881 for binding to androgen binding sites when compared with DHEA. There was no major difference between DHT and R1881 with regards to competing with labeled R1881 for binding to androgen binding sites of HOC. Therefore, the observation of a greater mitogenic effect of DHT (when compared with the mitogenic effect of DHEA on HOCs) may be related to a greater affinity of DHT to the androgen receptor.
Table Table 1a. Effects of DHT and DHEA on Human Osteoblastic Cell Proliferation
Previously, we showed that the positive effect of DHT on HBC proliferation can be inhibited by the androgen receptor antagonist hydroxyflutamide (OHFlu).6 To test whether the mitogenic effect of DHEA can also be inhibited by OHFlu, the dose response experiment was repeated in the presence of the androgen receptor antagonist OHFlu. OHFlu alone had no significant effect on HOC proliferation, but the addition of OHFlu to DHEA abolished the proliferative effect of DHEA on HOCs (Table 1a). Therefore, the mitogenic actions of both DHT and DHEA on HBCs are mediated by the androgen receptor.
If DHT and DHEA exert their positive effects on HBC proliferation by an androgen receptor–mediated mechanism, it should be possible to demonstrate an additive effect of DHEA and DHT on HOC growth in the lower concentration range. At higher androgen doses, there should be no additive interaction because all available androgen receptors are occupied by either compound. The observed pattern of responses was consistent with the expected results. After incubation for 24 h, either with 1 nM DHT and cotreatment with increasing concentrations of DHEA or with 1 nM DHEA and cotreatment with increasing concentrations of DHEA, it was found that DHEA and DHT were additive on HOC growth in the lower dose range but were not additive at the higher doses tested (Table 1b).
Table Table 1b. Effects of DHEA and DHT on DNA Synthesis
An alternative explanation for the direct mitogenic effect of DHEA is a possible metabolization of DHEA by HOCs into androstenedione, testosterone, and DHT, which then binds to the androgen receptor and elicits the observed proliferative effects. To investigate this possibility, we repeated the experiments testing the mitogenic effects of DHEA and DHT, but this time in the absence and presence of 4-MA, a 5-α-reductase inhibitor, which also has a potent inhibitory effect on 3β-hydroxysteroid dehydrogenase (3β-HSD). Inhibition of 3β-HSD activity prevents the conversion of DHEA to androstenedione and testosterone, which could be further metabolized into aromatized compounds (e.g., 17β-estradiol).23 The presence of 4-MA did not affect the mitogenic actions of a wide dose range of either DHEA or DHT on HOC proliferation after 48 h of exposure (Table 2). This observation indicates that DHEA does not need to be metabolized by the actions of 3β-HSD or 5-α-reductase to exert a direct proliferative effect on HOC proliferation.
Table Table 2. Effects of 4-MA on the Mitogenic Action of DHT and DHEA
Effects of adrenal and gonadal androgens on differentiated bone cell functions
A committed osteoblastic stem cell developing into a mature osteoblast expresses the ALP gene before the osteocalcin gene is turned on.24 ALP stimulates mineral formation,25 whereas the exact role of osteocalcin in the mineralization process is not known.24 Because gonadal androgens (e.g., DHT) have been shown to stimulate ALP production, we examined the possibility that DHT and DHEA have different effects on ALP and osteocalcin expression and on the formation of mineral in vitro, respectively.
Previously, we reported that the positive effect of the gonadal androgen DHT on ALP appears to be mediated by transforming growth factor beta 2 (TGF-β2)26 which was also observed by Bodine et al.14 Therefore, we sought to confirm that (1) DHEA does also increase ALP activity in HOCs and (2) that the positive effects of both DHT and of the adrenal androgen DHEA on ALP are mediated by TGF-β2. To this end, HOCs were treated with various doses of DHT and DHEA in the presence or absence of inhibitory TGFβ-antibodies (TGF-βab's). DHT and DHEA were found to have biphasic stimulatory effects on the ALP content of cells with maximal effective doses ranging between 1–10 nM (Table 3). There was no significant quantitative difference between the effects of DHT and DHEA on ALP activity. In the presence of inhibitory TGF-βab, the stimulatory effects of DHT and DHEA on ALP were abolished. This observation is consistent with the hypothesis that the positive effects of DHT and DHEA on ALP are both mediated by increased expression of TGF-β.
Table Table 3. Effect of Inhibitory TGF-βAntibodies on the Stimulatory Action of DHT and DHEA on Specific ALP Activity
The effect of DHEA on ALP activity was somewhat (although not significantly) smaller compared with the action of DHT on ALP. This could be due to an indirect effect of DHEA, which needs to be metabolized into another androgenic compound first. This androgenic compound (e.g., testosterone) could then elicit a positive effect on ALP activity. To investigate this possibility, HOCs were treated with DHT and DHEA with and without 4-MA being present, and this time the number of ALP+ staining cells was determined. DHT and DHEA significantly increased the number of ALP+ HOCs, whereby we did not find a reproducible quantitative difference between the positive effects of DHT and DHEA on the number of ALP+ HOCs. The positive effects of DHT and DHEA on ALP+ cells were not changed by the presence of 4-MA (Table 4). Thus, similar to the mitogenic action, the stimulatory effect of DHT and DHEA on ALP is also direct and does not require 5-α-reductase or 3β-HSD activity to convert the androgen into another active compound.
Table Table 4. Effect of 4-MA on the Action of DHT and DHEA on the Number of ALP Positive Staining HOC
Table Table 5. Effects of DHT and DHEA on Osteocalcin Secretion and Mineralization in Culture
Increased osteocalcin secretion by bone cells occurs late in the osteoblastic differentiation process.24 In vitro, increased osteocalcin secretion can be observed concomitant to the initiation of mineral formation. Osteocalcin secretion and mineral formation were measured after 3 and 7 days, respectively, to determine the effects of DHT and DHEA on the late osteoblastic differentiation processes. DHT and DHEA dose-dependently stimulated osteocalcin secretion with a maximum effect at 100 nM (Table 5). Both DHT and DHEA also significantly increased the formation of mineral in vitro. DHT was more potent than DHEA in this respect. Therefore, DHT and DHEA exhibit stimulatory effects on metabolic functions of HOCs (i.e., ALP production, osteocalcin secretion, and mineral formation) occurring at different stages of the osteoblastic differentiation process.
The role of adrenal androgens in regulating bone metabolism and their possible mechanism of action is not clear. Only recently have some of the effects of adrenal androgens on human bone cells been investigated in greater detail.14 We sought to extend our knowledge about the mechanism of action of adrenal androgens on human bone cells by comparing the effects of gonadal and adrenal androgens on human bone cell growth and osteoblastic functions. The results presented in this study demonstrate that DHT and the adrenal androgen DHEA have similar stimulatory actions on human bone cell proliferation and differentiated cell functions in vitro. Both DHT and DHEA increase human bone cell growth and stimulate osteoblastic functions occurring early (i.e., alkaline phosphatase expression) and late (i.e., osteocalcin secretion and mineralization) during osteoblastic differentiation.24 Therefore, gonadal and adrenal androgens have similar effects on the regulation of bone cell metabolism. Quantitative differences between DHT and DHEA with regard to their mitogenic and differentiation stimulating effects may be related to differences in their binding affinities to the androgen receptor, which has previously been reported for the effects of DHT and testosterone.32
Similar to the findings of Bodine et al.,14 we also observed similar effects of DHEA and DHT on ALP production and osteocalcin secretion. The finding of similar dose- and time-dependent effects of gonadal and adrenal androgens on HOC metabolism suggests a similar mechanism of action for both androgens. An additive effect of DHEA and DHT was observed on HOC proliferation in the lower concentration range. At higher androgen doses, there was no additive interaction, presumably because all available androgen receptors were occuppied by either compound. If DHT and DHEA act through the same androgen receptor–mediated mechanism, the effects of DHEA should also be blocked by the androgen receptor antagonist hydroxyflutamide, which has previously been demonstrated for DHT-induced effects on HOCs.6 The nonsteroidal androgen receptor antagonist hydroxyflutamide abolished the positive effect of DHEA on HOC proliferation, indicating that DHEA binds directly to the androgen receptor of HOC. Contrary to the present finding, Grover and Odell27 did not observe DHEA binding to the cytosolic androgen receptor in rat prostate extracts. There are three possible explanations for this discrepency. First, there may be a species-related difference in the binding characteristics of androgen receptors. The exchange of a single amino acid residue can have a major impact on the binding pattern of steroid receptors, as recently demonstrated for the murine glucocorticoid receptor.28 Second, the protein extraction procedure of Grover and Odell may have altered the androgen receptor structure, for example by separating components of the androgen receptor heterotetrameric complex,29 which may affect the binding characteristics of the androgen receptor protein.30,31 Third, DHEA may be converted into androstenedione, testosterone, and DHT, which could then bind to the androgen receptor and elicit the observed effects. In addition, these compounds may in turn be further metabolized by 3β-hydroxysteroid-dehydrogenase into aromatized bone-active compounds, such as 17β-estradiol.23 To examine the third possibility, HOCs were treated with DHEA in the presence of 4-MA, which blocks 5-α-reductase and 3β-hydroxysteroid dehydrogenase activity23 and had no effect either on the stimulatory effects of DHEA on HOC proliferation or on the induced increase in the number of ALP positive staining bone cells. Consequently, the combined treatment with DHEA and 4-MA prevents a possible conversion of DHEA into testosterone, DHT, and estrogens. The presence of 4-MA did not affect the stimulatory effects of DHEA on HOC proliferation and on the increase in the number of ALP+ staining bone cells. This demonstrates that DHEA acts directly on HOCs via binding to the androgen receptor and does not need to be metabolized first to testosterone and DHT.
The conclusion that both DHT and DHEA regulate HOC metabolism by the same mechanism is supported by the observation that both androgens induce TGF-β2 secretion in HOC populations that stimulate ALP production in a parakrine fashion,14 and by our finding that inhibitory TGF-β antibodies abolished stimulatory effects of both DHT and DHEA on ALP production in HOC populations. This observation is consistent with the hypothesis that the positive effects of DHT and DHEA on ALP are mediated by TGF-β2. In contrast to the findings of Bodine et al.,14 we observed equipotent effects for DHEA and DHT on HOC metabolism. Possibly this difference is due to differences in the homogeneities of the HOC populations studies. We used only male HOCs for our experiments, whereas Bodine et al. used pooled HOC populations of male and female origin. Whether male HOCs are more responsive than female HOCs to gonadal and adrenal androgens is currently under investigation. Certainly gonadal and adrenal androgens are potent regulators of growth and osteoblastic functions in males and females, and androgens are responsible for the sexual dimorphism observed in the skeleton.
Considering the relatively high serum concentrations of adrenal androgens compared with gonadal androgens and estrogens in both sexes, it is conceivable that adrenal androgens could play a more important role in the maintenance of bone mass than previously realized. Adrenal androgens may be particularly important in aging males and postmenopausal females when serum levels of testosterone and estrogens decline.
We are grateful for expert technical assistance by Mrs. Uli Sommer. This work was supported by grants from the Deutsche Forschungsgemeinschaft, Germany (Ka 682/2–2 and Ka 682/2–3).