Direct Agonist/Antagonist Functions of Dehydroepiandrosterone
http://www.100md.com
《内分泌学杂志》
Department of Molecular Endocrinology (F.C., K.K., E.B., J.F., H.W., A.S., S.H., L.P.F., A.A.R.), Merck Research Laboratories, West Point, Pennsylvania 19486
Centro de Investigación Básica (M.M., C.T.M.), Merck, Sharp & Dohme de Espaa, Madrid 28027, Spain
Abstract
Dehydroepiandrosterone (DHEA) exhibits peak adrenal secretion in the fetus at term and around age 30 yr in the adult. Levels then progressively decline, which is associated with decreased levels of testosterone, dihydrotestosterone, and estrogen in peripheral tissues. DHEA supplementation in postmenopausal women increases bone formation and density, an effect mainly attributed to peripheral conversion to sex hormones. In this study, we tested DHEA for direct effects on the androgen (AR) and estrogen (ER) receptors. DHEA bound to AR with a Ki of 1 μM, which was associated with AR transcriptional antagonism on both the mouse mammary tumor virus and prostate-specific antigen promoters, much like the effects of bicalutamide. Unlike bicalutamide, DHEA stimulated, rather than inhibited, LNCaP cell growth, suggesting possible interaction with other hormone receptors. Indeed DHEA bound to ER and ER, with Ki values of 1.1 and 0.5 μM, respectively. Despite the similar binding affinities, DHEA showed preferential agonism of ER with an EC50 of approximately 200 nM and maximal activation at 1 μM. With ER we found 30–70% agonism at 5 μM, depending on the assay. Physiological levels of DHEA are approximately 30 nM and up to 90 nM in the prostate. DHEA at 30 nM is actually sufficient to activate ER transcription to the same degree as estrogen at its circulating concentration, and additive effects are seen when the two were combined. Taken together, DHEA has the potential for physiologically relevant direct activation of ER. With peak levels at term and age 30 yr, there is also a potential for antagonist effects on AR and partial agonism of ER.
Introduction
DEHYDROEPIANDROSTERONE (DHEA) and DHEA sulfate (DHEA-S) are natural steroids secreted by the adrenal zona reticularis. Their circulating levels change during fetal development, normal human growth, and aging. Two major peaks occur at birth and around age 30 yr for both sexes, which results in the highest circulating concentration for DHEA-S at approximately 8 μM. The beginning of a second peak at age 6–8 yr serves as an indicator of adrenarche (1, 2). Throughout human life, the normal range of DHEA is 7–31 nM, whereas DHEA-S ranges from 1.3 to 6.8 μM (3). Local concentrations of DHEA can reach as high as 90 nM in prostate tissue (4, 5). By age 50 yr, only 50% of the peak concentrations can be detected in serum. By age 70 yr, serum levels fall to about 20% of the peak (1, 2, 6).
DHEA-S serves as reservoir for DHEA and is converted via a cellular sulfatase (3, 5). DHEA can then serve as a precursor for androstenediol, androstenedione, testosterone, estrone, testosterone (T), dihydrotestosterone (DHT), and 17-estradiol (E2) (2, 7). Physiological roles for DHEA and DHEA-S have been assigned whereby these serve as precursors for peripheral synthesis of DHT, T, and E2. Levels of these precursors determine the DHT/T and E2 concentrations in peripheral tissues, although they have little effect on circulating levels of the sex hormones (1, 5). Results from animal studies have suggested that DHEA may be beneficial in preventing age-related obesity, diabetes, cancer, heart disease and immune dysfunction (8, 9, 10, 11). Dietary supplementation of DHEA has also been used to counter age-related declines in sexual function and bone mass, among others (12).
The discovery of peripheral synthesis of androgens/estrogens from DHEA and DHEA-S (2, 5) led to the discovery of intracrinology, which holds that androgens and estrogens can, within the same cell, be synthesized from DHEA and/or androstenedione and exert their biological function (2). Biological function is mediated through the action of these sex hormones on androgen (AR) or estrogen receptor (ER), which belong to the nuclear receptor superfamily/steroid receptor subfamily (13). To date, only one AR gene has been identified with affinity to DHT and T in the subnanomolar range (14). Meanwhile, E2 binds with subnanomolar affinity to either of two ER isoforms, ER and ER (15). All three receptors function as hormone-regulated transcription factors. Recently interest has arisen in identifying possible direct effects of hormone metabolites or precursors that may also act via these receptors (16, 17). For instance, androstenediol binds to ER and ER with equilibrium binding constants (Kis) of 3.6 and 0.9 nM, respectively (18). Meanwhile, androstenedione showed no apparent affinity to either receptor. Furthermore, we found that both androstenedione and androstenediol bind and activate the endogenous AR with either a transfected reporter or an endogenous gene readout [prostate-specific antigen (PSA)] (Ref.17 and our unpublished data). The available information indicates a clear role for certain intermediate metabolites or precursors of DHEA to exhibit direct biological functions in addition to serving as precursors of sex hormones.
In the present study, we investigated whether DHEA might possess direct biological functions through interaction with AR and ER using various in vitro and cell-based assays. Our results clearly indicated that DHEA could function directly via ER, especially ER, as an agonist and via AR as an antagonist.
Materials and Methods
Reagents
Human mammary gland tumor cell lines, MDA-MB453 and T47D, monkey kidney cell line COS-1, and human prostate cancer cell line LNCaP were purchased from American Type Culture Collection (Manassas, VA). Stable human kidney cell lines of human embryonic kidney (HEK)293/human (h)ER and HEK293/hER containing estrogen response element (ERE)2-alkaline phosphatase (ALP) were from CIBE (19). LipofectAmine 2000, phenol red-free DMEM, MEM, RPMI 1640, Gentamicin, L-glutamine, HEPES, trypsin, regular and charcoal-stripped fetal bovine serum (FBS) were all from Gibco BRL/Invitrogen (Carlsbad, CA). Hydroxyapatite was from Calbiochem (La Jolla, CA). Human insulin, polyethylimine, sodium molybdate, triamcinolone acetonide, DHEA, DHEA-S, T, DHT, and E2 were all from Sigma (St. Louis, MO). R1881, [3H]R1881, and [3H]E2 were from NEN Life Science Products (Boston, MA). Proteinase inhibitor and FuGENE6 transfection reagent were from Roche Molecular Biochemicals (Indianapolis, IN). The Dual-Glo luciferase assay system and phRL-TK vector were from Promega (Madison, WI). Unifilter-96 GF/B and MICROSCINT were from Packard (Boston, MA). The PSA ELISA and DHEA enzyme immunoassay kits were from Diagnostic Systems Laboratories (Webster, TX). Cytostar 96-well scintillating microplates and [methyl-14C] thymidine CFA532 were from Amersham (Piscataway, NJ).
Preparation of nuclear receptors
Endogenous receptors for AR and glucocorticoid receptor (GR) were from the MDA-MB453 human breast cell line and were prepared as previously described (17). Endogenous AR from LNCaP cells and progesterone receptor (PR) from T47D cells were also prepared using the same method, although LNCaP and T47D cells were cultured in RPMI 1640 medium without insulin.
Both hER (accession no. X03635) and ER (accession no. af051427) were expressed in COS-1 cells. COS-1 cells were cultured in DMEM containing 2 mM L-glutamine, 10% FBS, and 20 μg/ml gentamicin. Cells were transfected with 15 μg DNA per 5 x 106 cells in 150-mm plates using LipofectAmine 2000 (Gibco BRL/Invitrogen) as per the manufacturer’s protocol. Two to three days after transfection, receptors were harvested as described above.
Cells were suspended at approximately 107 cells/ml in TEGM/PI [10 mM Tris-HCl, 1 mM EDTA, 10% glycerol, 1 mM -mercaptoethanol, 10 mM sodium molybdate (pH 7.2), and one pellet of protease inhibitors in 50 ml of buffer] and then snap frozen in ethanol/dry ice and stored at –80 C. Immediately before performing binding assays, frozen cells were thawed on ice. Cell lysates were collected by centrifugation at 10,000 x g at 4 C for 20 min after thawing the cell suspension. The lysate containing receptor was titrated before use for radioligand displacement assays.
Hydroxyapatite-based nuclear receptor binding assays
The nuclear receptor binding assays were performed as previously described (17). IC50 values for each compound were generated using MRLCalc (Merck & Co., West Point, PA). IC50s were converted to equilibrium binding constants (Kis) using the equation: Ki = Kd x IC50/(Kd + L), where L is the concentration of radioligand. For AR, Kd was calculated as the equilibrium dissociation constant for [3H]R1881 binding to either MDA-MB453 (MDA) hAR (0.45 nM) or LNCaP hAR (0.56 nM). For ER, the Kd for binding of [3H]estradiol to hER was 0.12 nM, whereas for ER, the Kd was 0.35 nM.
MDA-MB453 endogenous hAR and transfected mouse mammary tumor virus (MMTV) luciferase (luc) reporter assays
For each 96-well assay plate, 2 x 106 MDA cells were transfected with 6 μg MMTVluc (17), which was divided equally into each of the wells. The phRL-TK Renilla luciferase reporter, included in each transfection for signal normalization, was used at 60 ng per 96-well plate. On the day after transfection, cell media were replaced by fresh OptiMEM medium (Invitrogen) containing the test ligands. Cells were treated with compounds either at a single dose or across a concentration range, as described in Results. Luciferase and Renilla luciferase activities were measured 24 h after initiation of treatment as per manufacturer’s protocols and using a 1450 MicrobetaJet (PerkinElmer, Shelton, CT).
PSA induction and cell proliferation of LNCaP cells
Freshly prepared LNCaP cells were suspended in RPMI 1640 medium containing 10% charcoal-stripped FBS at 2 x 105 cells/ml for PSA assays and 1 x 105 cells/ml for proliferation assays. Cells in 100 μl of medium were seeded into each well of 96-well plates. For plates used in the proliferation study, 0.5 μCi/ml of [methyl-14C]thymidine were added to the medium. Two to four hours after seeding, cells were treated with compounds, which were added in a 25 μl volume of medium in either the presence (antagonist mode) or absence (agonist mode) of 0.5 nM DHT or R1881, as indicated. PSA levels in the medium were measured 24 h after initiation of compound treatment using a PSA ELISA kit. Cell proliferation was measured using a 1450 MicrobetaJet (Wallac) at 24, 48, 72, or 120 h.
Transcription assays using the Gal4 chimeric system
Gal4 DNA binding domain (DBD) derived from pM vector of CLONTECH was inserted downstream of the cytomegalovirus promoter of pcDNA3.1(+) (Invitrogen). Individual ligand binding domains (LBDs) of recombinant hAR, hER, or hER were fused in frame with the Gal4 DBD. The Gal4 luciferase reporter was constructed by inserting five copies of consensus Gal4 binding sites from pG5CAT (Clontech) upstream of the luciferase reporter gene in pGL3 basic (Promega). COS-1 cells at 1 x 106 per plate were transfected with 2.0 μg receptor, 6.0 μg reporter, and 50 ng phRL/96-well plate using FuGene 6. For the Gal4AR system, 1.5 μg of the glucocorticoid interacting protein 1 coactivator (20) were included.
Measurement of DHEA concentrations
For transcription assays, serially diluted DHEA samples were collected after 24 h treatment of assay cells. The same set of DHEA samples was also collected from parallel plates that were treated in a similar fashion, except the wells lacked cells. DHEA concentrations were measured using a DHEA enzyme immunoassay kit as per the manufacturer’s protocol.
Statistic analysis
All data are presented as mean ± SD. Statistical significance was calculated using PRISM4 (GraphPad Inc., San Diego, CA) by one-way ANOVA. Significance is indicated (*, P < 0.05; **, P < 0.01; or ***, P < 0.001).
Results
Binding of DHEA to AR
Because DHEA-S is converted to DHEA, which can be further metabolized to form T, E2, and DHT in peripheral tissues, such as prostate and mammary gland (2), preliminary studies examined possible direct effects of both DHEA-S and DHEA on transcription. However, we observed no detectable binding of DHEA-S to any of the tested hormone receptors, nor was there detectable transcriptional induction (see Fig. 6A and data not shown). We thus focused on a possible role for DHEA as a ligand for the sex hormone receptors, beginning with AR. Binding affinity of DHEA for AR was performed using endogenous receptor from either MDA or LNCaP cells (Table 1). As a counterscreen, we also examined endogenous GR from MDA and endogenous PR from T47D cells (data not shown). With tested concentrations as high as 5 μM, there was not even minimal detectable DHEA binding to either GR or PR. However, DHEA did competitively inhibit R1881 binding to AR from both cell lines with an average Ki of 1.2 μM. For comparison, binding constants for T and DHT displacement of R1881 binding to AR averaged 0.5 and approximately 0.2 nM, respectively (Table 1).
Transcriptional activity of DHEA on AR targets
T and DHT are downstream metabolites of DHEA, which act as pure agonists for the AR. Because DHEA did bind to AR, albeit with micromolar affinity, we hypothesized that this ligand might also exhibit transcriptional activity, possibly serving as an agonist. To test this, we used a semiendogenous transcription system using endogenous hAR and a transfected MMTVluc reporter in MDA cells. We previously validated this as an effective readout for AR transcriptional activity for ligands that do not cross-react with GR (17). Initial experiments were run in agonist-mode, whereby DHT and T elicited a full transcriptional activation with EC50s in the 0.5–1 nM range (Fig. 1). Unlike T and DHT, DHEA showed little or no transcriptional activation of the AR, even when dosed into the mircomolar range. Like DHEA, the AR antagonist, bicalutamide, also showed a complete lack of AR agonism in this setting. Owing to the relatively high affinity of E2 for AR (Table 1), we also tested this ligand for putative agonist behavior. Interestingly, despite its low nanomolar affinity, E2 displayed no agonist behavior until dosing was increased to near 1 μM. Within this high concentration range, E2 displayed a full agonist profile with transcriptional activation 100% that of T or DHT. Therefore, there was a separation between the low nanomolar binding affinity and the transcriptional activation by E2.
Within the tested dose range, DHEA demonstrated no agonist behavior, as noted above. Rather than further increase the DHEA dose (as was necessary for E2-mediated agonist effects), we tested the possibility that this ligand might instead function as a transcriptional antagonist (Fig. 2). This was achieved by coincubating the reporter-transfected MDA cells with 0.5 nM DHT along with either 2 or 5 μM DHEA. In addition, bicalutamide was used as a positive control for AR antagonism at its most efficacious dose in these cells, 1 μM (17). All data were normalized to the transcriptional activity of DHT at 10 nM, which was included as an independent, maximally efficacious control (100% activity). At 0.5 nM, DHT elicited a approximately 80% maximal activation. As expected, bicalutamide reduced transcriptional activity to approximately 20% that of maximum, which represented a 75% inhibitory effect (in comparison with the original 80% activation). Whereas there was slight but significant inhibitory effect of DHEA at 2 μM, there was a more profound 75% inhibition at 5 μM. Therefore, DHEA was as efficacious as bicalutamide in inhibiting the transcriptional activity induced by 0.5 nM DHT, although its potency was around 5-fold lower.
We next investigated DHEA agonist and antagonist activity in a fully endogenous transcription system (Figs. 3 and 4). We first measured PSA production in LNCaP cells (Fig. 3), whose AR encodes a T877A mutation known to broaden ligand specificity (21). Overexpression of this mutant in CV1 cells was reported to allow a 15-fold agonist response to DHEA (vs. 60- to 100-fold for T and DHT, respectively), whereas there was no agonist response to DHEA when wild-type AR was overexpressed (22). The possibility that DHEA acts as an antagonist on LNCaP or any other form of AR has not yet been established. We began by treating LNCaP cells with a single maximally efficacious dose of each ligand for 24 h, and the PSA levels in the medium were then measured by ELISA. Consistent with the semiendogenous transcriptional responses seen in MDA cells, DHT and T were fully efficacious in LNCaP cells at 10 and 100 nM, respectively. E2 also elicited a strong agonist response at 2 μM, although the maximal effect was only 70% that of the androgens. This E2 response is consistent with that described in a previous report (21). Consistent with the mild agonist response of the T877A mutant AR in CV1 cells, DHEA at 5 μM elicited an agonist response at about 12% that of 10 nM DHT. Meanwhile, 1 μM bicalutamide showed absolutely no agonism, and it even slightly repressed the basal activity (Fig. 3). The modest agonist effect of DHEA, in comparison with the pure antagonist response of bicalutamide, is reminiscent of the modest agonism elicited by the otherwise pure antagonist, cyproterone acetate, in these cells (21). Hydroxyflutamide has also been shown to exhibit a mild agonist response on the T877A mutant, but not wild-type AR, when overexpressed in CV1 cells (22).
The higher level of DHEA-induced agonism in LNCaP cells (vs. that seen in MDA cells) could reflect an altered response due to the AR mutation (as noted above) or a possible low level of metabolism of DHEA to form DHT (23). In the former case, it has not yet been established whether the partial agonist response of the T877A mutant to DHEA is also accompanied by an antagonist response. In the latter case, the DHEA might act to antagonize the effects of the downstream metabolite. To test the hypothesis that DHEA acted primarily as an antagonist in the LNCaP cells, we examined its ability to inhibit the transcriptional activation of AR by either the natural ligand, DHT, or the synthetic pure agonist, R1881 (Fig. 4). LNCaP cells were treated with either DHT or R1881 (both 0.5 nM) in combination with DHEA (5 μM) or other reference compounds, as described above. Again, because the initial concentrations of DHT or R1881 were set at submaximum doses, we observed an additive effect when 10 nM DHT (10.5 nM final concentration) or 100 nM T were added in combination with the initial 0.5 nM dose. Consistent with its 80% maximal activation at 2 μM, E2 actually showed a small, 20–25% antagonism in these experiments. This suggested that E2 was actually a mixed agonist/antagonist in these cells. As expected, bicalutamide showed substantial 30 and 60% inhibition of DHT- and R1881-induced transcription, respectively. In further support of its potential for AR antagonism, DHEA elicited a 60–70% inhibition of either agonist on AR-mediated transcription in the LNCaP cells. This suggests that, although the T877A mutant allows some modest agonist responses to DHEA, this ligand acts primarily as an antagonist in the LNCaP background.
Collectively the data suggest that DHEA may be a naturally occurring antagonist of AR, albeit when used in the low micromolar range. However, because the reported circulation level for DHEA is at 7–31 nM and higher in specific tissues, it may be difficult to attribute this effect any physiologically meaningful effect under most conditions. Under these circumstances, peripheral conversion of DHEA to T or DHT could result in the generation of sufficient concentrations to elicit an agonist response that could not be overcome by the relatively low concentration of DHEA. Meanwhile, although the DHEA reservoir molecule, DHEA-S, does circulate at 1.3–6.8 μM, it remained transcriptionally inactive in all assays tested (see Fig. 6A and data not shown).
Proliferative effects of DHEA on LNCaP cells
In consideration of the antagonistic behavior of DHEA on AR in both MDA and LNCaP cells alone, one might hypothesize that it may also exhibit an antimitogenic effect on cell growth. Consistent with our earlier transcription approaches, we tested this by measuring DHEA effects on LNCaP cell proliferation in agonist mode in comparison with DHT, T, E2, and bicalutamide, each dosed at their transcriptionally active concentrations (Fig. 5A). As expected, both DHT (10 nM) and T (100 nM) enhanced the growth rate vs. the vehicle-treated control, with approximately 5 and 25% more cells seen at 48 and 72 h, respectively. It was previously reported that E2 stimulates LNCaP cell proliferation (24). Indeed, E2 was even more effective in stimulating cell proliferation than the androgens, with cell number reaching 45% above vehicle effects by 72 h. As expected, bicalutamide had no effect on cell growth, and cell number remained identical with that seen with the vehicle control. Unexpectedly, and contrary to its antagonist effects on PSA production in these cells, DHEA was a full agonist for cell growth, and stimulation was identical with that of DHT or T by 72 h. This response distinguished DHEA from bicalutamide, which otherwise shared a common ability to antagonize AR-mediated transcription, as discussed above.
To investigate a possible connection to ER, we expanded the analyses by directly comparing DHEA with E2 across a concentration curve over a 5-d treatment period (Fig. 5B). Over this more extended time, E2 elicited a 100% greater number of cells than was seen in the control group with an EC50 of approximately 50 nM and a maximal induction at 100 nM or more. Thus, activation of proliferation by E2 is seen at concentrations well below those required for activation of AR-mediated transcription, which is consistent with an ER-mediated effect. The magnitude of the response to DHEA over the same 5-d period was comparable with that of E2, with a 100% increase in cell growth at the top-tested dose of 5 μM. However, the EC50 was approximately 100-fold higher than that for E2. In consideration of the transcriptional antagonism of the AR in LNCaP cells by DHEA in contrast to its mitogenic effects, we therefore hypothesized that the cell proliferation was mediated via a non-AR mechanism, possibly ER based.
Binding of DHEA to ER and ER
As noted above, E2 treatment was previously shown to elicit a mitogenic affect on LNCaP cell growth. Interestingly, this effect was associated with its action on ER (24) and not ER, which is the key driver for mitogenic effects in breast cancer cells. We therefore extended our initial receptor binding assays to include ER and ER (Table 1). In these analyses, we examined the ability of DHEA to competitively inhibit E2 binding to receptor that was transiently expressed in COS-1 cells. This method was chosen over the use of endogenously expressed receptor, owing to the lack of readily available cells with selective and substantial expression of either ER or ER alone. DHEA exhibited a similar binding affinity for ER to that for AR (1.1 μM), whereas it bound to ER with a slightly lower Ki of 0.5 ± 0.3 μM (Table 1). Similar results were observed with stable transfectants (data not shown).
Transcriptional activity of DHEA on ER targets
Binding analyses indicated that DHEA can interact with both ER isoforms with affinities that extend into the midnanomolar range. We therefore investigated potential transcriptional activation, based on the proliferative effects on LNCaP cells, which were comparable with those of E2. For these analyses, we examined activation function 2-dependent transcription using Gal4-DBD/ER-LBD chimeras (Gal4ER) (Fig. 6A), which we transiently expressed in COS-1 cells. A Gal4-DBD/AR-LBD chimera (Gal4AR) was also tested for comparison. For these analyses, percent activation was compared with an independent control using 200 nM E2 (data not shown). A sample of E2 at the same concentration as the control was also assessed as an independent test sample, which yielded a value of approximately 115% vs. the reference control, which is nonstatistically different (Fig. 6A). Consistent with its mitogenic potential in the LNCaP system, DHEA, at both 2 and 5 μM, also exhibited a strong agonist activation of Gal4ER, and activity was similar to or somewhat higher (150%) than that elicited by E2 in the control group. Although E2 exhibited an essentially equal stimulation of both ER and ER in these analyses, DHEA exhibited a clear preference for ER, with no activation on ER at the lower concentration of 2 μM. At 5 μM, the 55% stimulation of ER was roughly one third the level of activity seen with ER at the same concentration. Using the activation function 2-dependent system, we observed no transcriptional activation of AR by either DHEA or E2 in COS-1 cells. DHEA-S was inactive against all receptors, whereas DHT and T activated only Gal4AR (Fig. 6A).
In separate analyses, the effects of DHEA on Gal4ER-mediated transcription were tested in the absence or presence of the pure estrogen antagonist, ICI-182780 (Fig. 6B). Two concentrations of DHEA (1 and 5 μM) and E2 (200 nM and 1 μM) were tested. All test samples were compared with an independent control stimulated with E2 at 200 nM. In antagonist control experiments, ICI-182,780 (100 nM) was fully efficacious in blocking E2-mediated transcription at the lower tested dose. However, when the concentration of E2 was increased 5-fold, ICI-182,780 lost its efficacy. Meanwhile ICI-182780 fully blocked DHEA-mediated transcription at either dose. Together, these data suggest that DHEA-induced, ER-mediated transcription is achieved through direct binding of DHEA to this receptor’s LBD.
In initial binding analyses of ER and ER (Table 1), DHEA appeared to exhibit a slight preference for ER. To explore this further, dose-response effect on transcription was examined. For these analyses, COS-1 cells were transfected with GAL4ERs and GAL4-luc reporter (Fig. 7A) or HEK293 cells containing stably expressed intact receptors and ERE2-ALP reporter were used (Fig. 7B). Dose response curves compared DHEA with the native ligand E2. In the GAL4 chimeric system, DHEA maintained ER specificity, with an EC50 of around 200 nM. With ER, approximately 40% activation was seen at the substantially higher concentration of 5 μM. Essentially full activation of the full-length ER receptor was observed in the HEK293 background with 90% activation seen at 5 μM, although there was little activation at 1 μM. The preference for ER was also maintained in this background, with an EC50 of about 200 nM, as seen in the Gal4 chimeric system in COS-1 cells. These data support a model for DHEA activation of ER with an approximate 5- to 10-fold higher transcriptional specificity for the ER receptor vs. ER.
Unlike the transcriptional antagonism of DHEA on the AR, with activity seen only above 2 μM, this ligand exhibited transcriptional activation of ER in the midnanomolar range. This suggests a greater likelihood that DHEA could activate ER at physiologically relevant concentrations. The circulating levels of DHEA are between 7 and 31 nM, with even higher levels found in peripheral tissues, such as the prostate. For E2, circulating levels range between 70 and 220 pM in the basal state and can reach as high as 740 pM during ovulation (3). To test for activation of ER at more physiological levels, DHEA (30 nM) was compared with, or combined with, E2 (320 pM). This E2 level is roughly 50% above basal levels outside ovulation. Neither ligand alone elicited very much transcriptional activity, with levels reaching around 2–4% of the maximum E2 effect seen at higher doses (data not shown). This suggests that even E2 at its physiological levels has only modest effects on ER. However, when E2 was combined with DHEA, the total activity doubled (P < 0.01 vs. E2 or DHEA alone), suggesting the real possibility that a physiological function of ER may come as a result of the combined activities of E2 and DHEA.
Discussion
Evidence from DNA coding for the various steroid receptors has suggested that steroidogenic and steroid-inactivating enzymes and steroid receptors have coevolved (7). The steroid metabolism enzymes have played key roles in achieving ligand specificity that has accompanied steroid receptor evolution, which likely began with one receptor that might bind to various steroid ligands. Sequence analysis suggests that the original ER was the first evolutionary steroid receptor. The modern evolved ER isoforms and other steroid receptors can thus be traced back to this ancestral ER and likely arose via gene duplication and mutational selection. Evolutionary analysis also suggests the possibility that 5 steroids, such as DHEA and its derivatives, could have served as the original ligands for the ancestral ER (7). Our data on DHEA binding to and activation of ER and ER lend credence to this model.
A biological role for DHEA as a precursor for sex hormones generated in peripheral tissues is already clearly established, and it is the sex hormones themselves that are believed to carry out the biological functions of this precursor (2). In this study, we found evidence to suggest that DHEA itself may also act as a sex hormone with specificity for the ER receptor. In competition binding analyses, the data indicated that DHEA exhibited affinity to AR and ERs with some preference for ER. Affinities for expressed ERs fell into the 0.2–1.5 μM range. There was no measurable affinity for endogenously expressed GR or PR. The affinity of DHEA to the ERs is generally consistent with a previous report (18), although transcriptional activities were not previously described. Our transcription analyses demonstrated that DHEA activates ER with an maximum response similar to or slightly greater than that seen with E2.
Our initial analyses focused on DHEA as a potential androgen because it has long been considered an androgenic steroid precursor, exerting its action via peripheral conversion to more potent androgens, such as T and DHT (2, 25). DHEA did exhibit binding affinity to AR at micromolar concentration, but its affinity was 2- to 3-fold lower than that for ER. DHEA transactivation function on AR was tested using a semiendogenous reporting system with endogenously expressed hAR in MDA-MB453 cells and minimal amounts of transfected MMTVluc promoter (17). Whereas this reporting system accurately measured DHT, T, and E2 transcriptional activation, it showed that DHEA, like bicalutamide, had undetectable agonist activity. Instead, DHEA, like bicalutamide, behaved more like an AR antagonist. This was seen in antagonist mode transcriptional profiling, which used DHT at levels yielding submaximal activity, whereby DHEA at 5 μM repressed DHT activity to the same extent as that seen with bicalutamide. That DHEA behaved more like an antagonist on endogenous AR was further established in its effects on LNCaP PSA production, in which this ligand exhibited minimal agonist activity and essentially maximal inhibitory activity in suppressing agonist-induced PSA expression. Indeed the antagonist effect was comparable with bicalutamide in both the MDA and LNCaP systems.
Our findings are in contrast to previous reports of DHEA androgenic activity on AR using transient cotransfection of AR along with an androgen responsive element-luciferase reporter (16, 21). The discrepancy between our current data and those from previous publications may stem from the difference between transiently expressed vs. endogenous AR, the latter being used here. Under normal levels of expression, AR resides in an unliganded state within the cytoplasm, in which it forms complexes with numerous chaperones, including heat shock proteins. When expressed in a transient transfection system, AR levels tend to be overexpressed, which can cause an imbalance between receptor and chaperone (26). Chaperones keep the receptor in a transcriptionally inactive state, and agonist binding induces the release of the chaperones in the process of inducing an active receptor conformation (27). Because the receptor lacking the accompanying chaperones may be more readily activated, substantial overexpression might better detect very weak agonist activity, which would not otherwise exist with normal receptor levels. In the present study, we used endogenously expressed AR in each of two different cell lines, and readouts included either a minimally expressed reporter (MMTV promoter) or an endogenous marker (PSA). The consistency in the results from the different reporter systems suggested predominantly antagonist effects of DHEA on AR-mediated transcription. It is notable, however, that there was a mild agonist response seen in the production of PSA in LNCaP cells in response to DHEA, possibly due to the mutant form of AR or, less likely, to a low level of metabolism of the ligand. There was a more profound DHEA-stimulated antagonism observed in both the PSA and MMTV readouts, suggesting that the primary role for the unmetabolized ligand is to serve as an AR antagonist. One should note, however, that because these effects required quite high concentrations, any physiological function(s) of DHEA through AR in vivo may be limited to certain stages of development.
A probable role for DHEA as a direct steroid receptor ligand at physiological concentrations would more likely be mediated through ER. With an EC50 of approximately 200 nM, there is a strong suggestion that ER could be the most selective target. Based on the binding affinity of DHEA for this receptor and the ability of ICI-182780 to antagonize DHEA-induced transcription, this activity is most probably mediated through direct binding to ER. ER could also be involved when concentrations rise high enough. Weak DHEA activity toward ER has also been reported previously in a study examining transcriptional effects but not receptor binding (28). Based on the physiological levels of DHEA, which reach to 30 nM or higher at certain developmental stages, it is unlikely that this ligand could exert a significant biological function via ER (EC50 > 1 μM) under most conditions. In contrast, because DHEA is a more potent transcriptional activator of ER, there exists a real possibility for a physiologically meaningful interaction. Indeed, the concentrations of DHEA in circulation and local tissues are high enough to stimulate ER to the same extent as that seen with circulating levels of E2 when tested at somewhat above its top normal level (outside ovulation). The ER activity resulting from DHEA regulation could even rise higher during the fetal development, at adrenarche, and at around age 30 yr, because concentrations at these life stages are higher. The amount of estrogenic activity from DHEA could therefore contribute significantly to the total estrogenic milieu in the body.
Because of the perceived beneficial effects of DHEA, as outlined in the introductory text, it has been increasingly used as a dietary supplement, in an effort to retard the effects of aging (4, 12, 29). However, concerns over its safety, especially in breast or prostate cancer patients, suggests the need for caution (4, 5, 30, 31). In addition to being a direct regulator of ER, DHEA can be metabolized into more potent estrogen and androgens in peripheral tissues, such as mammary gland and prostate, as has already been clearly established (2). Many research groups have reported stimulatory effects of androgens and estrogens on cancer cells (21, 31). We showed here that DHEA exhibited minimal agonist activity on the endogenous PSA promoter, and it actually antagonized subnanomolar DHT and R1881 activity. Nonetheless, our analyses of LNCaP cell proliferation demonstrated that at the indicated dose, DHEA was actually as effective as DHT and T in increasing LNCaP cell proliferation. Meanwhile, E2 appeared to be even more efficacious after three days of treatment, although by 5 d, the maximum effect of DHEA and E2 were comparable.
These observations are in general agreement with a previous report (4), although there were differences in potency that may be traced to methodology. In any event, our and the previous studies both indicate that DHEA, like androgens and estrogens, increase LNCaP cell proliferation, which is also the case for endothelial cell proliferation, as reported in an independent study (32). All three studies thus agree on the proliferative potential of DHEA. Whereas we suggest that the effect might be elicited via ER, others have drawn no conclusion or suggested that neither ER nor AR is the responsible target (4, 32). As to whether the DHEA-ER interaction is stimulatory for LNCaP growth should be the subject of further study. Such a study should include not only an analysis of how ER and AR antagonists might affect DHEA-mediated mitogenic stimulation but also potential metabolism of DHEA to sex steroids in these cells (23) as well as potential nongenomic effects, as have been suggested in other studies of endothelial cells (32, 33). Regardless of the mechanism of DHEA-mediated stimulation of LNCaP growth, our observation led us to test for possible effects of DHEA on ER, and our ultimate conclusions led us to ER as a physiologically relevant target of this ligand.
Identifying physiological roles for DHEA as a direct ligand for ER is also complicated by the fact that DHEA can be converted to estrogen, among other metabolites, in cells. However, certain parallels in the actions of DHEA and ER-selective ligands in in vivo models of inflammation suggest that direct action on this receptor is possible. ER-selective ligands have been demonstrated to show efficacy in inflammatory bowel disease, adjuvant-induced arthritis, and the treatment of experimental endometriosis (34, 35). Thus, a role for ER in suppressing inflammation is likely. Such a role for DHEA has also been suggested. Clinical studies have shown that DHEA (and DHEA-S) levels are reduced in subjects with rheumatoid arthritis and systemic lupus erythematosus, among other diseases (reviewed in Ref.11). Moreover, DHEA exhibits antiinflammatory effects in an animal model of antigen-induced arthritis, including an inhibition of the formation of autoantibodies (36). One possible reason for the beneficial effects of DHEA in arthritis could include its conversion to downstream metabolite(s). One can likely rule out 7-hydroxy-DHEA as a probable candidate because this metabolite has been correlated with increased severity of arthritis progression (37). This is important to note because others have reported that 7-hydroxy-DHEA is a partial agonist for ER, albeit with greater than 10-fold lower affinity vs. that for DHEA (as reported here) (38). As to whether 7-hydroxy-DHEA can serve as an antagonist that could help to drive the inflammatory response remains to be established.
Based on the complexity of DHEA metabolism to various potential ligands for AR and the ERs vs. its direct effects on these receptors, it may be difficult to specifically attribute DHEA’s direct transcriptional actions to any specific beneficial effect. Nonetheless, there are several reasons to suggest DHEA as a direct ligand for ER and to a lesser extent ER and AR. First, our receptor binding assays were performed using an in vitro cell-free system, in which metabolism of DHEA to other more potent derivatives is negligible. The affinity for AR and ERs therefore likely reflects an accurate measure of binding to these receptors. Second, the transcriptional effects of DHEA were observed in a concentration range consistent with affinity for each respective receptor. Third, we measured possible DHEA metabolism in the HEK293 system, whereby ER activation was clearly observed, by quantifying the concentration in cells 24 h post treatment against parallel DHEA samples incubated in the absence of any cells. There was no observable changes in DHEA concentrations in either sets of samples (data not shown), which suggests that DHEA was not significantly metabolized under our experimental conditions. Fourth, DHEA exhibited both agonist (ER and ER) and antagonist (AR) behavior in the various transcription assays, whereas its major metabolites (T and E2) serve only as agonists. Fifth, in the case of the DHEA metabolite, 7-hydroxy-DHEA, which is ER-selective in COS-1 cells, this metabolite displays only partial (50%) agonism of ER and only at a substantially higher concentration of 20 μM (38). Sixth, with regard to ER vs. ER, E2 showed equipotent activation of these receptors, whereas DHEA strictly favored ER. Overall, the unique characteristics of DHEA-mediated transcriptional effects on ER, ER, and AR cannot be replicated by any known downstream metabolite. Thus, the data suggest direct action of DHEA on these receptors with a preference for ER vs. ER or AR.
In summary, the composite data suggest a role for DHEA as a direct ligand for ER in addition to its role as a precursor for sex hormone production. Although the observed activity at physiologically relevant concentrations of DHEA is low and so is the activity of E2. Whereas DHEA has the potential to act as an agonist of ER or an antagonist of AR, this requires substantially higher concentrations than needed to activate ER. Theoretically, suitable concentrations might be achieved for in vivo effects on ER and AR, although this would likely be limited to the fetus at term and perhaps one or two developmental stages after birth. The increased concentrations of DHEA at these stages would also make it a more efficacious agonist on ER. Although metabolism of DHEA makes difficult the process of providing absolute proof of its direct action on ER in vivo, the preponderance of the evidence leaves little reason to conclude otherwise.
Acknowledgments
The authors gratefully appreciate the technical support from Ms. Qin Su and Mr. Carlo Gambone. Mr. Robert Kearney (Basic Research System) provided assistance on the software for data analysis.
Footnotes
Abbreviations: ALP, Alkaline phosphatase; AR, androgen receptor; DBD, DNA binding domain; DHEA, dehydroepiandrosterone; DHEA-S, DHEA sulfate; DHT, dihydrotestosterone; E2, 17-estradiol; ER, estrogen receptor; ERE, estrogen response element; FBS, fetal bovine serum; GR, glucocorticoid receptor; h, human; HEK, human embryonic kidney; Kd, equilibrium dissociation constant; Ki, equilibrium binding constant; LBD, ligand binding domain; luc, luciferase; MDA, MDA-MB453; MMTV, mouse mammary tumor virus; PR, progesterone receptor; PSA, prostate-specific antigen; T, testosterone.
References
Auchus RJ, Rainey WE 2004 Adrenarche—physiology, biochemistry and human disease. Clin Endocrinol (Oxf.) 60:288–296
Labrie F, Luu-The V, Labrie C, Simard J 2001 DHEA and its transformation into androgens and estrogens in peripheral target tissues: intracrinology. Front Neuroendocrinol 22:185–212
Stewart P 2002 The adrenal cortex. Philadelphia: Saunders
Arnold JT, Le H, McFann KK, Blackman MR 2005 Comparative effects of DHEA vs. testosterone, dihydrotestosterone, and estradiol on proliferation and gene expression in human LNCaP prostate cancer cells. Am J Physiol Endocrinol Metab 288:E573–E584
Mohler JL, Gregory CW, Ford 3rd OH, Kim D, Weaver CM, Petrusz P, Wilson EM, French FS 2004 The androgen axis in recurrent prostate cancer. Clin Cancer Res 10:440–448
Dharia S, Parker Jr CR 2004 Adrenal androgens and aging. Semin Reprod Med 22:361–368
Baker ME 2004 Co-evolution of steroidogenic and steroid-inactivating enzymes and adrenal and sex steroid receptors. Mol Cell Endocrinol 215:55–62
Valenti G 2004 Neuroendocrine hypothesis of aging: the role of corticoadrenal steroids. J Endocrinol Invest 27:62–63
Libe R, Barbetta L, Dall’Asta C, Salvaggio F, Gala C, Beck-Peccoz P, Ambrosi B 2004 Effects of dehydroepiandrosterone (DHEA) supplementation on hormonal, metabolic and behavioral status in patients with hypoadrenalism. J Endocrinol Invest 27:736–741
Arlt W 2004 Dehydroepiandrosterone replacement therapy. Semin Reprod Med 22:379–388
Chen CC, Parker Jr CR2004 Adrenal androgens and the immune system. Semin Reprod Med 22:369–377
Legrain S, Girard L 2003 Pharmacology and therapeutic effects of dehydroepiandrosterone in older subjects. Drugs Aging 20:949–967
Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P1995 The nuclear receptor superfamily: the second decade. Cell 83:835–839
McPhaul MJ, Young M 2001 Complexities of androgen action. J Am Acad Dermatol 45:S87–S94
Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson JA 1996 Cloning of a novel receptor expressed in rat prostate and ovary. Proc Natl Acad Sci USA 93:5925–5930
Holterhus PM, Piefke S, Hiort O 2002 Anabolic steroids, testosterone-precursors and virilizing androgens induce distinct activation profiles of androgen responsive promoter constructs. J Steroid Biochem Mol Biol 82:269–275
Chen F, Knecht K, Leu C, Rutledge SJ, Scafonas A, Gambone C, Vogel R, Zhang H, Kasparcova V, Bai C, Harada S, Schmidt A, Reszka A, Freedman L 2004 Partial agonist/antagonist properties of androstenedione and 4-androsten-3,17-diol. J Steroid Biochem Mol Biol 91:247–257
Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S, Gustafsson JA 1997 Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors and . Endocrinology 138:863–870
Barkhem T, Carlsson B, Nilsson Y, Enmark E, Gustafsson J, Nilsson S 1998 Differential response of estrogen receptor and estrogen receptor to partial estrogen agonists/antagonists. Mol Pharmacol 54:105–112
Hong H, Kohli K, Trivedi A, Johnson DL, Stallcup MR 1996 GRIP1, a novel mouse protein that serves as a transcriptional coactivator in yeast for the hormone binding domains of steroid receptors. Proc Natl Acad Sci USA 93:4948–4952
Steketee K, Timmerman L, Ziel-van der Made AC, Doesburg P, Brinkmann AO, Trapman J 2002 Broadened ligand responsiveness of androgen receptor mutants obtained by random amino acid substitution of H874 and mutation hot spot T877 in prostate cancer. Int J Cancer 100:309–317
Tan J, Sharief Y, Hamil KG, Gregory CW, Zang DY, Sar M, Gumerlock PH, DeVere White RW, Pretlow TG, Harris SE, Wilson EM, Mohler JL, French FS 1997 Dehydroepiandrosterone activates mutant androgen receptors expressed in the androgen-dependent human prostate cancer xenograft CWR22 and LNCaP cells. Mol Endocrinol 11:450–459
Koh E, Kanaya J, Namiki M 2001 Adrenal steroids in human prostatic cancer cell lines. Arch Androl 46:117–125
Maggiolini M, Recchia AG, Carpino A, Vivacqua A, Fasanella G, Rago V, Pezzi V, Briand PA, Picard D, Ando S 2004 Oestrogen receptor is required for androgen-stimulated proliferation of LNCaP prostate cancer cells. J Mol Endocrinol 32:777–791
Rehman KS, Carr BR 2004 Sex differences in adrenal androgens. Semin Reprod Med 22:349–360
Caplan AJ, Jackson S, Smith D 2003 Hsp90 reaches new heights. Conference on the Hsp90 chaperone machine. EMBO Rep 4:126–130
Georget V, Terouanne B, Nicolas JC, Sultan C 2002 Mechanism of antiandrogen action: key role of hsp90 in conformational change and transcriptional activity of the androgen receptor. Biochemistry 41:11824–11831
Bruder JM, Sobek L, Oettel M 1997 Dehydroepiandrosterone stimulates the estrogen response element. J Steroid Biochem Mol Biol 62:461–466
Lamberts SW, van den Beld AW, van der Lely AJ 1997 The endocrinology of aging. Science 278:419–424
Algarte-Genin M, Cussenot O, Costa P 2004 Prevention of prostate cancer by androgens: experimental paradox or clinical reality. Eur Urol. 46:285–294; discussion 294–295
Missmer SA, Eliassen AH, Barbieri RL, Hankinson SE 2004 Endogenous estrogen, androgen, and progesterone concentrations and breast cancer risk among postmenopausal women. J Natl Cancer Inst 96:1856–1865
Williams MR, Dawood T, Ling S, Dai A, Lew R, Myles K, Funder JW, Sudhir K, Komesaroff PA 2004 Dehydroepiandrosterone increases endothelial cell proliferation in vitro and improves endothelial function in vivo by mechanisms independent of androgen and estrogen receptors. J Clin Endocrinol Metab 89:4708–4715
Simoncini T, Mannella P, Fornari L, Varone G, Caruso A, Genazzani AR 2003 Dehydroepiandrosterone modulates endothelial nitric oxide synthesis via direct genomic and nongenomic mechanisms. Endocrinology 144:3449–3455
Harris HA, Bruner-Tran KL, Zhang X, Osteen KG, Lyttle CR 2005 A selective estrogen receptor- agonist causes lesion regression in an experimentally induced model of endometriosis. Hum Reprod 20:936–941
Harris HA, Albert LM, Leathurby Y, Malamas MS, Mewshaw RE, Miller CP, Kharode YP, Marzolf J, Komm BS, Winneker RC, Frail DE, Henderson RA, Zhu Y, Keith Jr JC 2003 Evaluation of an estrogen receptor- agonist in animal models of human disease. Endocrinology 144:4241–4249
Rontzsch A, Thoss K, Petrow PK, Henzgen S, Brauer R 2004 Amelioration of murine antigen-induced arthritis by dehydroepiandrosterone (DHEA). Inflamm Res 53:189–198
Dulos J, Verbraak E, Bagchus WM, Boots AM, Kaptein A 2004 Severity of murine collagen-induced arthritis correlates with increased CYP7B activity: enhancement of dehydroepiandrosterone metabolism by interleukin-1. Arthritis Rheum 50:3346–3353
Martin C, Ross M, Chapman KE, Andrew R, Bollina P, Seckl JR, Habib FK 2004 CYP7B generates a selective estrogen receptor agonist in human prostate. J Clin Endocrinol Metab 89:2928–2935(Fang Chen, Kristin Knecht)
Centro de Investigación Básica (M.M., C.T.M.), Merck, Sharp & Dohme de Espaa, Madrid 28027, Spain
Abstract
Dehydroepiandrosterone (DHEA) exhibits peak adrenal secretion in the fetus at term and around age 30 yr in the adult. Levels then progressively decline, which is associated with decreased levels of testosterone, dihydrotestosterone, and estrogen in peripheral tissues. DHEA supplementation in postmenopausal women increases bone formation and density, an effect mainly attributed to peripheral conversion to sex hormones. In this study, we tested DHEA for direct effects on the androgen (AR) and estrogen (ER) receptors. DHEA bound to AR with a Ki of 1 μM, which was associated with AR transcriptional antagonism on both the mouse mammary tumor virus and prostate-specific antigen promoters, much like the effects of bicalutamide. Unlike bicalutamide, DHEA stimulated, rather than inhibited, LNCaP cell growth, suggesting possible interaction with other hormone receptors. Indeed DHEA bound to ER and ER, with Ki values of 1.1 and 0.5 μM, respectively. Despite the similar binding affinities, DHEA showed preferential agonism of ER with an EC50 of approximately 200 nM and maximal activation at 1 μM. With ER we found 30–70% agonism at 5 μM, depending on the assay. Physiological levels of DHEA are approximately 30 nM and up to 90 nM in the prostate. DHEA at 30 nM is actually sufficient to activate ER transcription to the same degree as estrogen at its circulating concentration, and additive effects are seen when the two were combined. Taken together, DHEA has the potential for physiologically relevant direct activation of ER. With peak levels at term and age 30 yr, there is also a potential for antagonist effects on AR and partial agonism of ER.
Introduction
DEHYDROEPIANDROSTERONE (DHEA) and DHEA sulfate (DHEA-S) are natural steroids secreted by the adrenal zona reticularis. Their circulating levels change during fetal development, normal human growth, and aging. Two major peaks occur at birth and around age 30 yr for both sexes, which results in the highest circulating concentration for DHEA-S at approximately 8 μM. The beginning of a second peak at age 6–8 yr serves as an indicator of adrenarche (1, 2). Throughout human life, the normal range of DHEA is 7–31 nM, whereas DHEA-S ranges from 1.3 to 6.8 μM (3). Local concentrations of DHEA can reach as high as 90 nM in prostate tissue (4, 5). By age 50 yr, only 50% of the peak concentrations can be detected in serum. By age 70 yr, serum levels fall to about 20% of the peak (1, 2, 6).
DHEA-S serves as reservoir for DHEA and is converted via a cellular sulfatase (3, 5). DHEA can then serve as a precursor for androstenediol, androstenedione, testosterone, estrone, testosterone (T), dihydrotestosterone (DHT), and 17-estradiol (E2) (2, 7). Physiological roles for DHEA and DHEA-S have been assigned whereby these serve as precursors for peripheral synthesis of DHT, T, and E2. Levels of these precursors determine the DHT/T and E2 concentrations in peripheral tissues, although they have little effect on circulating levels of the sex hormones (1, 5). Results from animal studies have suggested that DHEA may be beneficial in preventing age-related obesity, diabetes, cancer, heart disease and immune dysfunction (8, 9, 10, 11). Dietary supplementation of DHEA has also been used to counter age-related declines in sexual function and bone mass, among others (12).
The discovery of peripheral synthesis of androgens/estrogens from DHEA and DHEA-S (2, 5) led to the discovery of intracrinology, which holds that androgens and estrogens can, within the same cell, be synthesized from DHEA and/or androstenedione and exert their biological function (2). Biological function is mediated through the action of these sex hormones on androgen (AR) or estrogen receptor (ER), which belong to the nuclear receptor superfamily/steroid receptor subfamily (13). To date, only one AR gene has been identified with affinity to DHT and T in the subnanomolar range (14). Meanwhile, E2 binds with subnanomolar affinity to either of two ER isoforms, ER and ER (15). All three receptors function as hormone-regulated transcription factors. Recently interest has arisen in identifying possible direct effects of hormone metabolites or precursors that may also act via these receptors (16, 17). For instance, androstenediol binds to ER and ER with equilibrium binding constants (Kis) of 3.6 and 0.9 nM, respectively (18). Meanwhile, androstenedione showed no apparent affinity to either receptor. Furthermore, we found that both androstenedione and androstenediol bind and activate the endogenous AR with either a transfected reporter or an endogenous gene readout [prostate-specific antigen (PSA)] (Ref.17 and our unpublished data). The available information indicates a clear role for certain intermediate metabolites or precursors of DHEA to exhibit direct biological functions in addition to serving as precursors of sex hormones.
In the present study, we investigated whether DHEA might possess direct biological functions through interaction with AR and ER using various in vitro and cell-based assays. Our results clearly indicated that DHEA could function directly via ER, especially ER, as an agonist and via AR as an antagonist.
Materials and Methods
Reagents
Human mammary gland tumor cell lines, MDA-MB453 and T47D, monkey kidney cell line COS-1, and human prostate cancer cell line LNCaP were purchased from American Type Culture Collection (Manassas, VA). Stable human kidney cell lines of human embryonic kidney (HEK)293/human (h)ER and HEK293/hER containing estrogen response element (ERE)2-alkaline phosphatase (ALP) were from CIBE (19). LipofectAmine 2000, phenol red-free DMEM, MEM, RPMI 1640, Gentamicin, L-glutamine, HEPES, trypsin, regular and charcoal-stripped fetal bovine serum (FBS) were all from Gibco BRL/Invitrogen (Carlsbad, CA). Hydroxyapatite was from Calbiochem (La Jolla, CA). Human insulin, polyethylimine, sodium molybdate, triamcinolone acetonide, DHEA, DHEA-S, T, DHT, and E2 were all from Sigma (St. Louis, MO). R1881, [3H]R1881, and [3H]E2 were from NEN Life Science Products (Boston, MA). Proteinase inhibitor and FuGENE6 transfection reagent were from Roche Molecular Biochemicals (Indianapolis, IN). The Dual-Glo luciferase assay system and phRL-TK vector were from Promega (Madison, WI). Unifilter-96 GF/B and MICROSCINT were from Packard (Boston, MA). The PSA ELISA and DHEA enzyme immunoassay kits were from Diagnostic Systems Laboratories (Webster, TX). Cytostar 96-well scintillating microplates and [methyl-14C] thymidine CFA532 were from Amersham (Piscataway, NJ).
Preparation of nuclear receptors
Endogenous receptors for AR and glucocorticoid receptor (GR) were from the MDA-MB453 human breast cell line and were prepared as previously described (17). Endogenous AR from LNCaP cells and progesterone receptor (PR) from T47D cells were also prepared using the same method, although LNCaP and T47D cells were cultured in RPMI 1640 medium without insulin.
Both hER (accession no. X03635) and ER (accession no. af051427) were expressed in COS-1 cells. COS-1 cells were cultured in DMEM containing 2 mM L-glutamine, 10% FBS, and 20 μg/ml gentamicin. Cells were transfected with 15 μg DNA per 5 x 106 cells in 150-mm plates using LipofectAmine 2000 (Gibco BRL/Invitrogen) as per the manufacturer’s protocol. Two to three days after transfection, receptors were harvested as described above.
Cells were suspended at approximately 107 cells/ml in TEGM/PI [10 mM Tris-HCl, 1 mM EDTA, 10% glycerol, 1 mM -mercaptoethanol, 10 mM sodium molybdate (pH 7.2), and one pellet of protease inhibitors in 50 ml of buffer] and then snap frozen in ethanol/dry ice and stored at –80 C. Immediately before performing binding assays, frozen cells were thawed on ice. Cell lysates were collected by centrifugation at 10,000 x g at 4 C for 20 min after thawing the cell suspension. The lysate containing receptor was titrated before use for radioligand displacement assays.
Hydroxyapatite-based nuclear receptor binding assays
The nuclear receptor binding assays were performed as previously described (17). IC50 values for each compound were generated using MRLCalc (Merck & Co., West Point, PA). IC50s were converted to equilibrium binding constants (Kis) using the equation: Ki = Kd x IC50/(Kd + L), where L is the concentration of radioligand. For AR, Kd was calculated as the equilibrium dissociation constant for [3H]R1881 binding to either MDA-MB453 (MDA) hAR (0.45 nM) or LNCaP hAR (0.56 nM). For ER, the Kd for binding of [3H]estradiol to hER was 0.12 nM, whereas for ER, the Kd was 0.35 nM.
MDA-MB453 endogenous hAR and transfected mouse mammary tumor virus (MMTV) luciferase (luc) reporter assays
For each 96-well assay plate, 2 x 106 MDA cells were transfected with 6 μg MMTVluc (17), which was divided equally into each of the wells. The phRL-TK Renilla luciferase reporter, included in each transfection for signal normalization, was used at 60 ng per 96-well plate. On the day after transfection, cell media were replaced by fresh OptiMEM medium (Invitrogen) containing the test ligands. Cells were treated with compounds either at a single dose or across a concentration range, as described in Results. Luciferase and Renilla luciferase activities were measured 24 h after initiation of treatment as per manufacturer’s protocols and using a 1450 MicrobetaJet (PerkinElmer, Shelton, CT).
PSA induction and cell proliferation of LNCaP cells
Freshly prepared LNCaP cells were suspended in RPMI 1640 medium containing 10% charcoal-stripped FBS at 2 x 105 cells/ml for PSA assays and 1 x 105 cells/ml for proliferation assays. Cells in 100 μl of medium were seeded into each well of 96-well plates. For plates used in the proliferation study, 0.5 μCi/ml of [methyl-14C]thymidine were added to the medium. Two to four hours after seeding, cells were treated with compounds, which were added in a 25 μl volume of medium in either the presence (antagonist mode) or absence (agonist mode) of 0.5 nM DHT or R1881, as indicated. PSA levels in the medium were measured 24 h after initiation of compound treatment using a PSA ELISA kit. Cell proliferation was measured using a 1450 MicrobetaJet (Wallac) at 24, 48, 72, or 120 h.
Transcription assays using the Gal4 chimeric system
Gal4 DNA binding domain (DBD) derived from pM vector of CLONTECH was inserted downstream of the cytomegalovirus promoter of pcDNA3.1(+) (Invitrogen). Individual ligand binding domains (LBDs) of recombinant hAR, hER, or hER were fused in frame with the Gal4 DBD. The Gal4 luciferase reporter was constructed by inserting five copies of consensus Gal4 binding sites from pG5CAT (Clontech) upstream of the luciferase reporter gene in pGL3 basic (Promega). COS-1 cells at 1 x 106 per plate were transfected with 2.0 μg receptor, 6.0 μg reporter, and 50 ng phRL/96-well plate using FuGene 6. For the Gal4AR system, 1.5 μg of the glucocorticoid interacting protein 1 coactivator (20) were included.
Measurement of DHEA concentrations
For transcription assays, serially diluted DHEA samples were collected after 24 h treatment of assay cells. The same set of DHEA samples was also collected from parallel plates that were treated in a similar fashion, except the wells lacked cells. DHEA concentrations were measured using a DHEA enzyme immunoassay kit as per the manufacturer’s protocol.
Statistic analysis
All data are presented as mean ± SD. Statistical significance was calculated using PRISM4 (GraphPad Inc., San Diego, CA) by one-way ANOVA. Significance is indicated (*, P < 0.05; **, P < 0.01; or ***, P < 0.001).
Results
Binding of DHEA to AR
Because DHEA-S is converted to DHEA, which can be further metabolized to form T, E2, and DHT in peripheral tissues, such as prostate and mammary gland (2), preliminary studies examined possible direct effects of both DHEA-S and DHEA on transcription. However, we observed no detectable binding of DHEA-S to any of the tested hormone receptors, nor was there detectable transcriptional induction (see Fig. 6A and data not shown). We thus focused on a possible role for DHEA as a ligand for the sex hormone receptors, beginning with AR. Binding affinity of DHEA for AR was performed using endogenous receptor from either MDA or LNCaP cells (Table 1). As a counterscreen, we also examined endogenous GR from MDA and endogenous PR from T47D cells (data not shown). With tested concentrations as high as 5 μM, there was not even minimal detectable DHEA binding to either GR or PR. However, DHEA did competitively inhibit R1881 binding to AR from both cell lines with an average Ki of 1.2 μM. For comparison, binding constants for T and DHT displacement of R1881 binding to AR averaged 0.5 and approximately 0.2 nM, respectively (Table 1).
Transcriptional activity of DHEA on AR targets
T and DHT are downstream metabolites of DHEA, which act as pure agonists for the AR. Because DHEA did bind to AR, albeit with micromolar affinity, we hypothesized that this ligand might also exhibit transcriptional activity, possibly serving as an agonist. To test this, we used a semiendogenous transcription system using endogenous hAR and a transfected MMTVluc reporter in MDA cells. We previously validated this as an effective readout for AR transcriptional activity for ligands that do not cross-react with GR (17). Initial experiments were run in agonist-mode, whereby DHT and T elicited a full transcriptional activation with EC50s in the 0.5–1 nM range (Fig. 1). Unlike T and DHT, DHEA showed little or no transcriptional activation of the AR, even when dosed into the mircomolar range. Like DHEA, the AR antagonist, bicalutamide, also showed a complete lack of AR agonism in this setting. Owing to the relatively high affinity of E2 for AR (Table 1), we also tested this ligand for putative agonist behavior. Interestingly, despite its low nanomolar affinity, E2 displayed no agonist behavior until dosing was increased to near 1 μM. Within this high concentration range, E2 displayed a full agonist profile with transcriptional activation 100% that of T or DHT. Therefore, there was a separation between the low nanomolar binding affinity and the transcriptional activation by E2.
Within the tested dose range, DHEA demonstrated no agonist behavior, as noted above. Rather than further increase the DHEA dose (as was necessary for E2-mediated agonist effects), we tested the possibility that this ligand might instead function as a transcriptional antagonist (Fig. 2). This was achieved by coincubating the reporter-transfected MDA cells with 0.5 nM DHT along with either 2 or 5 μM DHEA. In addition, bicalutamide was used as a positive control for AR antagonism at its most efficacious dose in these cells, 1 μM (17). All data were normalized to the transcriptional activity of DHT at 10 nM, which was included as an independent, maximally efficacious control (100% activity). At 0.5 nM, DHT elicited a approximately 80% maximal activation. As expected, bicalutamide reduced transcriptional activity to approximately 20% that of maximum, which represented a 75% inhibitory effect (in comparison with the original 80% activation). Whereas there was slight but significant inhibitory effect of DHEA at 2 μM, there was a more profound 75% inhibition at 5 μM. Therefore, DHEA was as efficacious as bicalutamide in inhibiting the transcriptional activity induced by 0.5 nM DHT, although its potency was around 5-fold lower.
We next investigated DHEA agonist and antagonist activity in a fully endogenous transcription system (Figs. 3 and 4). We first measured PSA production in LNCaP cells (Fig. 3), whose AR encodes a T877A mutation known to broaden ligand specificity (21). Overexpression of this mutant in CV1 cells was reported to allow a 15-fold agonist response to DHEA (vs. 60- to 100-fold for T and DHT, respectively), whereas there was no agonist response to DHEA when wild-type AR was overexpressed (22). The possibility that DHEA acts as an antagonist on LNCaP or any other form of AR has not yet been established. We began by treating LNCaP cells with a single maximally efficacious dose of each ligand for 24 h, and the PSA levels in the medium were then measured by ELISA. Consistent with the semiendogenous transcriptional responses seen in MDA cells, DHT and T were fully efficacious in LNCaP cells at 10 and 100 nM, respectively. E2 also elicited a strong agonist response at 2 μM, although the maximal effect was only 70% that of the androgens. This E2 response is consistent with that described in a previous report (21). Consistent with the mild agonist response of the T877A mutant AR in CV1 cells, DHEA at 5 μM elicited an agonist response at about 12% that of 10 nM DHT. Meanwhile, 1 μM bicalutamide showed absolutely no agonism, and it even slightly repressed the basal activity (Fig. 3). The modest agonist effect of DHEA, in comparison with the pure antagonist response of bicalutamide, is reminiscent of the modest agonism elicited by the otherwise pure antagonist, cyproterone acetate, in these cells (21). Hydroxyflutamide has also been shown to exhibit a mild agonist response on the T877A mutant, but not wild-type AR, when overexpressed in CV1 cells (22).
The higher level of DHEA-induced agonism in LNCaP cells (vs. that seen in MDA cells) could reflect an altered response due to the AR mutation (as noted above) or a possible low level of metabolism of DHEA to form DHT (23). In the former case, it has not yet been established whether the partial agonist response of the T877A mutant to DHEA is also accompanied by an antagonist response. In the latter case, the DHEA might act to antagonize the effects of the downstream metabolite. To test the hypothesis that DHEA acted primarily as an antagonist in the LNCaP cells, we examined its ability to inhibit the transcriptional activation of AR by either the natural ligand, DHT, or the synthetic pure agonist, R1881 (Fig. 4). LNCaP cells were treated with either DHT or R1881 (both 0.5 nM) in combination with DHEA (5 μM) or other reference compounds, as described above. Again, because the initial concentrations of DHT or R1881 were set at submaximum doses, we observed an additive effect when 10 nM DHT (10.5 nM final concentration) or 100 nM T were added in combination with the initial 0.5 nM dose. Consistent with its 80% maximal activation at 2 μM, E2 actually showed a small, 20–25% antagonism in these experiments. This suggested that E2 was actually a mixed agonist/antagonist in these cells. As expected, bicalutamide showed substantial 30 and 60% inhibition of DHT- and R1881-induced transcription, respectively. In further support of its potential for AR antagonism, DHEA elicited a 60–70% inhibition of either agonist on AR-mediated transcription in the LNCaP cells. This suggests that, although the T877A mutant allows some modest agonist responses to DHEA, this ligand acts primarily as an antagonist in the LNCaP background.
Collectively the data suggest that DHEA may be a naturally occurring antagonist of AR, albeit when used in the low micromolar range. However, because the reported circulation level for DHEA is at 7–31 nM and higher in specific tissues, it may be difficult to attribute this effect any physiologically meaningful effect under most conditions. Under these circumstances, peripheral conversion of DHEA to T or DHT could result in the generation of sufficient concentrations to elicit an agonist response that could not be overcome by the relatively low concentration of DHEA. Meanwhile, although the DHEA reservoir molecule, DHEA-S, does circulate at 1.3–6.8 μM, it remained transcriptionally inactive in all assays tested (see Fig. 6A and data not shown).
Proliferative effects of DHEA on LNCaP cells
In consideration of the antagonistic behavior of DHEA on AR in both MDA and LNCaP cells alone, one might hypothesize that it may also exhibit an antimitogenic effect on cell growth. Consistent with our earlier transcription approaches, we tested this by measuring DHEA effects on LNCaP cell proliferation in agonist mode in comparison with DHT, T, E2, and bicalutamide, each dosed at their transcriptionally active concentrations (Fig. 5A). As expected, both DHT (10 nM) and T (100 nM) enhanced the growth rate vs. the vehicle-treated control, with approximately 5 and 25% more cells seen at 48 and 72 h, respectively. It was previously reported that E2 stimulates LNCaP cell proliferation (24). Indeed, E2 was even more effective in stimulating cell proliferation than the androgens, with cell number reaching 45% above vehicle effects by 72 h. As expected, bicalutamide had no effect on cell growth, and cell number remained identical with that seen with the vehicle control. Unexpectedly, and contrary to its antagonist effects on PSA production in these cells, DHEA was a full agonist for cell growth, and stimulation was identical with that of DHT or T by 72 h. This response distinguished DHEA from bicalutamide, which otherwise shared a common ability to antagonize AR-mediated transcription, as discussed above.
To investigate a possible connection to ER, we expanded the analyses by directly comparing DHEA with E2 across a concentration curve over a 5-d treatment period (Fig. 5B). Over this more extended time, E2 elicited a 100% greater number of cells than was seen in the control group with an EC50 of approximately 50 nM and a maximal induction at 100 nM or more. Thus, activation of proliferation by E2 is seen at concentrations well below those required for activation of AR-mediated transcription, which is consistent with an ER-mediated effect. The magnitude of the response to DHEA over the same 5-d period was comparable with that of E2, with a 100% increase in cell growth at the top-tested dose of 5 μM. However, the EC50 was approximately 100-fold higher than that for E2. In consideration of the transcriptional antagonism of the AR in LNCaP cells by DHEA in contrast to its mitogenic effects, we therefore hypothesized that the cell proliferation was mediated via a non-AR mechanism, possibly ER based.
Binding of DHEA to ER and ER
As noted above, E2 treatment was previously shown to elicit a mitogenic affect on LNCaP cell growth. Interestingly, this effect was associated with its action on ER (24) and not ER, which is the key driver for mitogenic effects in breast cancer cells. We therefore extended our initial receptor binding assays to include ER and ER (Table 1). In these analyses, we examined the ability of DHEA to competitively inhibit E2 binding to receptor that was transiently expressed in COS-1 cells. This method was chosen over the use of endogenously expressed receptor, owing to the lack of readily available cells with selective and substantial expression of either ER or ER alone. DHEA exhibited a similar binding affinity for ER to that for AR (1.1 μM), whereas it bound to ER with a slightly lower Ki of 0.5 ± 0.3 μM (Table 1). Similar results were observed with stable transfectants (data not shown).
Transcriptional activity of DHEA on ER targets
Binding analyses indicated that DHEA can interact with both ER isoforms with affinities that extend into the midnanomolar range. We therefore investigated potential transcriptional activation, based on the proliferative effects on LNCaP cells, which were comparable with those of E2. For these analyses, we examined activation function 2-dependent transcription using Gal4-DBD/ER-LBD chimeras (Gal4ER) (Fig. 6A), which we transiently expressed in COS-1 cells. A Gal4-DBD/AR-LBD chimera (Gal4AR) was also tested for comparison. For these analyses, percent activation was compared with an independent control using 200 nM E2 (data not shown). A sample of E2 at the same concentration as the control was also assessed as an independent test sample, which yielded a value of approximately 115% vs. the reference control, which is nonstatistically different (Fig. 6A). Consistent with its mitogenic potential in the LNCaP system, DHEA, at both 2 and 5 μM, also exhibited a strong agonist activation of Gal4ER, and activity was similar to or somewhat higher (150%) than that elicited by E2 in the control group. Although E2 exhibited an essentially equal stimulation of both ER and ER in these analyses, DHEA exhibited a clear preference for ER, with no activation on ER at the lower concentration of 2 μM. At 5 μM, the 55% stimulation of ER was roughly one third the level of activity seen with ER at the same concentration. Using the activation function 2-dependent system, we observed no transcriptional activation of AR by either DHEA or E2 in COS-1 cells. DHEA-S was inactive against all receptors, whereas DHT and T activated only Gal4AR (Fig. 6A).
In separate analyses, the effects of DHEA on Gal4ER-mediated transcription were tested in the absence or presence of the pure estrogen antagonist, ICI-182780 (Fig. 6B). Two concentrations of DHEA (1 and 5 μM) and E2 (200 nM and 1 μM) were tested. All test samples were compared with an independent control stimulated with E2 at 200 nM. In antagonist control experiments, ICI-182,780 (100 nM) was fully efficacious in blocking E2-mediated transcription at the lower tested dose. However, when the concentration of E2 was increased 5-fold, ICI-182,780 lost its efficacy. Meanwhile ICI-182780 fully blocked DHEA-mediated transcription at either dose. Together, these data suggest that DHEA-induced, ER-mediated transcription is achieved through direct binding of DHEA to this receptor’s LBD.
In initial binding analyses of ER and ER (Table 1), DHEA appeared to exhibit a slight preference for ER. To explore this further, dose-response effect on transcription was examined. For these analyses, COS-1 cells were transfected with GAL4ERs and GAL4-luc reporter (Fig. 7A) or HEK293 cells containing stably expressed intact receptors and ERE2-ALP reporter were used (Fig. 7B). Dose response curves compared DHEA with the native ligand E2. In the GAL4 chimeric system, DHEA maintained ER specificity, with an EC50 of around 200 nM. With ER, approximately 40% activation was seen at the substantially higher concentration of 5 μM. Essentially full activation of the full-length ER receptor was observed in the HEK293 background with 90% activation seen at 5 μM, although there was little activation at 1 μM. The preference for ER was also maintained in this background, with an EC50 of about 200 nM, as seen in the Gal4 chimeric system in COS-1 cells. These data support a model for DHEA activation of ER with an approximate 5- to 10-fold higher transcriptional specificity for the ER receptor vs. ER.
Unlike the transcriptional antagonism of DHEA on the AR, with activity seen only above 2 μM, this ligand exhibited transcriptional activation of ER in the midnanomolar range. This suggests a greater likelihood that DHEA could activate ER at physiologically relevant concentrations. The circulating levels of DHEA are between 7 and 31 nM, with even higher levels found in peripheral tissues, such as the prostate. For E2, circulating levels range between 70 and 220 pM in the basal state and can reach as high as 740 pM during ovulation (3). To test for activation of ER at more physiological levels, DHEA (30 nM) was compared with, or combined with, E2 (320 pM). This E2 level is roughly 50% above basal levels outside ovulation. Neither ligand alone elicited very much transcriptional activity, with levels reaching around 2–4% of the maximum E2 effect seen at higher doses (data not shown). This suggests that even E2 at its physiological levels has only modest effects on ER. However, when E2 was combined with DHEA, the total activity doubled (P < 0.01 vs. E2 or DHEA alone), suggesting the real possibility that a physiological function of ER may come as a result of the combined activities of E2 and DHEA.
Discussion
Evidence from DNA coding for the various steroid receptors has suggested that steroidogenic and steroid-inactivating enzymes and steroid receptors have coevolved (7). The steroid metabolism enzymes have played key roles in achieving ligand specificity that has accompanied steroid receptor evolution, which likely began with one receptor that might bind to various steroid ligands. Sequence analysis suggests that the original ER was the first evolutionary steroid receptor. The modern evolved ER isoforms and other steroid receptors can thus be traced back to this ancestral ER and likely arose via gene duplication and mutational selection. Evolutionary analysis also suggests the possibility that 5 steroids, such as DHEA and its derivatives, could have served as the original ligands for the ancestral ER (7). Our data on DHEA binding to and activation of ER and ER lend credence to this model.
A biological role for DHEA as a precursor for sex hormones generated in peripheral tissues is already clearly established, and it is the sex hormones themselves that are believed to carry out the biological functions of this precursor (2). In this study, we found evidence to suggest that DHEA itself may also act as a sex hormone with specificity for the ER receptor. In competition binding analyses, the data indicated that DHEA exhibited affinity to AR and ERs with some preference for ER. Affinities for expressed ERs fell into the 0.2–1.5 μM range. There was no measurable affinity for endogenously expressed GR or PR. The affinity of DHEA to the ERs is generally consistent with a previous report (18), although transcriptional activities were not previously described. Our transcription analyses demonstrated that DHEA activates ER with an maximum response similar to or slightly greater than that seen with E2.
Our initial analyses focused on DHEA as a potential androgen because it has long been considered an androgenic steroid precursor, exerting its action via peripheral conversion to more potent androgens, such as T and DHT (2, 25). DHEA did exhibit binding affinity to AR at micromolar concentration, but its affinity was 2- to 3-fold lower than that for ER. DHEA transactivation function on AR was tested using a semiendogenous reporting system with endogenously expressed hAR in MDA-MB453 cells and minimal amounts of transfected MMTVluc promoter (17). Whereas this reporting system accurately measured DHT, T, and E2 transcriptional activation, it showed that DHEA, like bicalutamide, had undetectable agonist activity. Instead, DHEA, like bicalutamide, behaved more like an AR antagonist. This was seen in antagonist mode transcriptional profiling, which used DHT at levels yielding submaximal activity, whereby DHEA at 5 μM repressed DHT activity to the same extent as that seen with bicalutamide. That DHEA behaved more like an antagonist on endogenous AR was further established in its effects on LNCaP PSA production, in which this ligand exhibited minimal agonist activity and essentially maximal inhibitory activity in suppressing agonist-induced PSA expression. Indeed the antagonist effect was comparable with bicalutamide in both the MDA and LNCaP systems.
Our findings are in contrast to previous reports of DHEA androgenic activity on AR using transient cotransfection of AR along with an androgen responsive element-luciferase reporter (16, 21). The discrepancy between our current data and those from previous publications may stem from the difference between transiently expressed vs. endogenous AR, the latter being used here. Under normal levels of expression, AR resides in an unliganded state within the cytoplasm, in which it forms complexes with numerous chaperones, including heat shock proteins. When expressed in a transient transfection system, AR levels tend to be overexpressed, which can cause an imbalance between receptor and chaperone (26). Chaperones keep the receptor in a transcriptionally inactive state, and agonist binding induces the release of the chaperones in the process of inducing an active receptor conformation (27). Because the receptor lacking the accompanying chaperones may be more readily activated, substantial overexpression might better detect very weak agonist activity, which would not otherwise exist with normal receptor levels. In the present study, we used endogenously expressed AR in each of two different cell lines, and readouts included either a minimally expressed reporter (MMTV promoter) or an endogenous marker (PSA). The consistency in the results from the different reporter systems suggested predominantly antagonist effects of DHEA on AR-mediated transcription. It is notable, however, that there was a mild agonist response seen in the production of PSA in LNCaP cells in response to DHEA, possibly due to the mutant form of AR or, less likely, to a low level of metabolism of the ligand. There was a more profound DHEA-stimulated antagonism observed in both the PSA and MMTV readouts, suggesting that the primary role for the unmetabolized ligand is to serve as an AR antagonist. One should note, however, that because these effects required quite high concentrations, any physiological function(s) of DHEA through AR in vivo may be limited to certain stages of development.
A probable role for DHEA as a direct steroid receptor ligand at physiological concentrations would more likely be mediated through ER. With an EC50 of approximately 200 nM, there is a strong suggestion that ER could be the most selective target. Based on the binding affinity of DHEA for this receptor and the ability of ICI-182780 to antagonize DHEA-induced transcription, this activity is most probably mediated through direct binding to ER. ER could also be involved when concentrations rise high enough. Weak DHEA activity toward ER has also been reported previously in a study examining transcriptional effects but not receptor binding (28). Based on the physiological levels of DHEA, which reach to 30 nM or higher at certain developmental stages, it is unlikely that this ligand could exert a significant biological function via ER (EC50 > 1 μM) under most conditions. In contrast, because DHEA is a more potent transcriptional activator of ER, there exists a real possibility for a physiologically meaningful interaction. Indeed, the concentrations of DHEA in circulation and local tissues are high enough to stimulate ER to the same extent as that seen with circulating levels of E2 when tested at somewhat above its top normal level (outside ovulation). The ER activity resulting from DHEA regulation could even rise higher during the fetal development, at adrenarche, and at around age 30 yr, because concentrations at these life stages are higher. The amount of estrogenic activity from DHEA could therefore contribute significantly to the total estrogenic milieu in the body.
Because of the perceived beneficial effects of DHEA, as outlined in the introductory text, it has been increasingly used as a dietary supplement, in an effort to retard the effects of aging (4, 12, 29). However, concerns over its safety, especially in breast or prostate cancer patients, suggests the need for caution (4, 5, 30, 31). In addition to being a direct regulator of ER, DHEA can be metabolized into more potent estrogen and androgens in peripheral tissues, such as mammary gland and prostate, as has already been clearly established (2). Many research groups have reported stimulatory effects of androgens and estrogens on cancer cells (21, 31). We showed here that DHEA exhibited minimal agonist activity on the endogenous PSA promoter, and it actually antagonized subnanomolar DHT and R1881 activity. Nonetheless, our analyses of LNCaP cell proliferation demonstrated that at the indicated dose, DHEA was actually as effective as DHT and T in increasing LNCaP cell proliferation. Meanwhile, E2 appeared to be even more efficacious after three days of treatment, although by 5 d, the maximum effect of DHEA and E2 were comparable.
These observations are in general agreement with a previous report (4), although there were differences in potency that may be traced to methodology. In any event, our and the previous studies both indicate that DHEA, like androgens and estrogens, increase LNCaP cell proliferation, which is also the case for endothelial cell proliferation, as reported in an independent study (32). All three studies thus agree on the proliferative potential of DHEA. Whereas we suggest that the effect might be elicited via ER, others have drawn no conclusion or suggested that neither ER nor AR is the responsible target (4, 32). As to whether the DHEA-ER interaction is stimulatory for LNCaP growth should be the subject of further study. Such a study should include not only an analysis of how ER and AR antagonists might affect DHEA-mediated mitogenic stimulation but also potential metabolism of DHEA to sex steroids in these cells (23) as well as potential nongenomic effects, as have been suggested in other studies of endothelial cells (32, 33). Regardless of the mechanism of DHEA-mediated stimulation of LNCaP growth, our observation led us to test for possible effects of DHEA on ER, and our ultimate conclusions led us to ER as a physiologically relevant target of this ligand.
Identifying physiological roles for DHEA as a direct ligand for ER is also complicated by the fact that DHEA can be converted to estrogen, among other metabolites, in cells. However, certain parallels in the actions of DHEA and ER-selective ligands in in vivo models of inflammation suggest that direct action on this receptor is possible. ER-selective ligands have been demonstrated to show efficacy in inflammatory bowel disease, adjuvant-induced arthritis, and the treatment of experimental endometriosis (34, 35). Thus, a role for ER in suppressing inflammation is likely. Such a role for DHEA has also been suggested. Clinical studies have shown that DHEA (and DHEA-S) levels are reduced in subjects with rheumatoid arthritis and systemic lupus erythematosus, among other diseases (reviewed in Ref.11). Moreover, DHEA exhibits antiinflammatory effects in an animal model of antigen-induced arthritis, including an inhibition of the formation of autoantibodies (36). One possible reason for the beneficial effects of DHEA in arthritis could include its conversion to downstream metabolite(s). One can likely rule out 7-hydroxy-DHEA as a probable candidate because this metabolite has been correlated with increased severity of arthritis progression (37). This is important to note because others have reported that 7-hydroxy-DHEA is a partial agonist for ER, albeit with greater than 10-fold lower affinity vs. that for DHEA (as reported here) (38). As to whether 7-hydroxy-DHEA can serve as an antagonist that could help to drive the inflammatory response remains to be established.
Based on the complexity of DHEA metabolism to various potential ligands for AR and the ERs vs. its direct effects on these receptors, it may be difficult to specifically attribute DHEA’s direct transcriptional actions to any specific beneficial effect. Nonetheless, there are several reasons to suggest DHEA as a direct ligand for ER and to a lesser extent ER and AR. First, our receptor binding assays were performed using an in vitro cell-free system, in which metabolism of DHEA to other more potent derivatives is negligible. The affinity for AR and ERs therefore likely reflects an accurate measure of binding to these receptors. Second, the transcriptional effects of DHEA were observed in a concentration range consistent with affinity for each respective receptor. Third, we measured possible DHEA metabolism in the HEK293 system, whereby ER activation was clearly observed, by quantifying the concentration in cells 24 h post treatment against parallel DHEA samples incubated in the absence of any cells. There was no observable changes in DHEA concentrations in either sets of samples (data not shown), which suggests that DHEA was not significantly metabolized under our experimental conditions. Fourth, DHEA exhibited both agonist (ER and ER) and antagonist (AR) behavior in the various transcription assays, whereas its major metabolites (T and E2) serve only as agonists. Fifth, in the case of the DHEA metabolite, 7-hydroxy-DHEA, which is ER-selective in COS-1 cells, this metabolite displays only partial (50%) agonism of ER and only at a substantially higher concentration of 20 μM (38). Sixth, with regard to ER vs. ER, E2 showed equipotent activation of these receptors, whereas DHEA strictly favored ER. Overall, the unique characteristics of DHEA-mediated transcriptional effects on ER, ER, and AR cannot be replicated by any known downstream metabolite. Thus, the data suggest direct action of DHEA on these receptors with a preference for ER vs. ER or AR.
In summary, the composite data suggest a role for DHEA as a direct ligand for ER in addition to its role as a precursor for sex hormone production. Although the observed activity at physiologically relevant concentrations of DHEA is low and so is the activity of E2. Whereas DHEA has the potential to act as an agonist of ER or an antagonist of AR, this requires substantially higher concentrations than needed to activate ER. Theoretically, suitable concentrations might be achieved for in vivo effects on ER and AR, although this would likely be limited to the fetus at term and perhaps one or two developmental stages after birth. The increased concentrations of DHEA at these stages would also make it a more efficacious agonist on ER. Although metabolism of DHEA makes difficult the process of providing absolute proof of its direct action on ER in vivo, the preponderance of the evidence leaves little reason to conclude otherwise.
Acknowledgments
The authors gratefully appreciate the technical support from Ms. Qin Su and Mr. Carlo Gambone. Mr. Robert Kearney (Basic Research System) provided assistance on the software for data analysis.
Footnotes
Abbreviations: ALP, Alkaline phosphatase; AR, androgen receptor; DBD, DNA binding domain; DHEA, dehydroepiandrosterone; DHEA-S, DHEA sulfate; DHT, dihydrotestosterone; E2, 17-estradiol; ER, estrogen receptor; ERE, estrogen response element; FBS, fetal bovine serum; GR, glucocorticoid receptor; h, human; HEK, human embryonic kidney; Kd, equilibrium dissociation constant; Ki, equilibrium binding constant; LBD, ligand binding domain; luc, luciferase; MDA, MDA-MB453; MMTV, mouse mammary tumor virus; PR, progesterone receptor; PSA, prostate-specific antigen; T, testosterone.
References
Auchus RJ, Rainey WE 2004 Adrenarche—physiology, biochemistry and human disease. Clin Endocrinol (Oxf.) 60:288–296
Labrie F, Luu-The V, Labrie C, Simard J 2001 DHEA and its transformation into androgens and estrogens in peripheral target tissues: intracrinology. Front Neuroendocrinol 22:185–212
Stewart P 2002 The adrenal cortex. Philadelphia: Saunders
Arnold JT, Le H, McFann KK, Blackman MR 2005 Comparative effects of DHEA vs. testosterone, dihydrotestosterone, and estradiol on proliferation and gene expression in human LNCaP prostate cancer cells. Am J Physiol Endocrinol Metab 288:E573–E584
Mohler JL, Gregory CW, Ford 3rd OH, Kim D, Weaver CM, Petrusz P, Wilson EM, French FS 2004 The androgen axis in recurrent prostate cancer. Clin Cancer Res 10:440–448
Dharia S, Parker Jr CR 2004 Adrenal androgens and aging. Semin Reprod Med 22:361–368
Baker ME 2004 Co-evolution of steroidogenic and steroid-inactivating enzymes and adrenal and sex steroid receptors. Mol Cell Endocrinol 215:55–62
Valenti G 2004 Neuroendocrine hypothesis of aging: the role of corticoadrenal steroids. J Endocrinol Invest 27:62–63
Libe R, Barbetta L, Dall’Asta C, Salvaggio F, Gala C, Beck-Peccoz P, Ambrosi B 2004 Effects of dehydroepiandrosterone (DHEA) supplementation on hormonal, metabolic and behavioral status in patients with hypoadrenalism. J Endocrinol Invest 27:736–741
Arlt W 2004 Dehydroepiandrosterone replacement therapy. Semin Reprod Med 22:379–388
Chen CC, Parker Jr CR2004 Adrenal androgens and the immune system. Semin Reprod Med 22:369–377
Legrain S, Girard L 2003 Pharmacology and therapeutic effects of dehydroepiandrosterone in older subjects. Drugs Aging 20:949–967
Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P1995 The nuclear receptor superfamily: the second decade. Cell 83:835–839
McPhaul MJ, Young M 2001 Complexities of androgen action. J Am Acad Dermatol 45:S87–S94
Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson JA 1996 Cloning of a novel receptor expressed in rat prostate and ovary. Proc Natl Acad Sci USA 93:5925–5930
Holterhus PM, Piefke S, Hiort O 2002 Anabolic steroids, testosterone-precursors and virilizing androgens induce distinct activation profiles of androgen responsive promoter constructs. J Steroid Biochem Mol Biol 82:269–275
Chen F, Knecht K, Leu C, Rutledge SJ, Scafonas A, Gambone C, Vogel R, Zhang H, Kasparcova V, Bai C, Harada S, Schmidt A, Reszka A, Freedman L 2004 Partial agonist/antagonist properties of androstenedione and 4-androsten-3,17-diol. J Steroid Biochem Mol Biol 91:247–257
Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S, Gustafsson JA 1997 Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors and . Endocrinology 138:863–870
Barkhem T, Carlsson B, Nilsson Y, Enmark E, Gustafsson J, Nilsson S 1998 Differential response of estrogen receptor and estrogen receptor to partial estrogen agonists/antagonists. Mol Pharmacol 54:105–112
Hong H, Kohli K, Trivedi A, Johnson DL, Stallcup MR 1996 GRIP1, a novel mouse protein that serves as a transcriptional coactivator in yeast for the hormone binding domains of steroid receptors. Proc Natl Acad Sci USA 93:4948–4952
Steketee K, Timmerman L, Ziel-van der Made AC, Doesburg P, Brinkmann AO, Trapman J 2002 Broadened ligand responsiveness of androgen receptor mutants obtained by random amino acid substitution of H874 and mutation hot spot T877 in prostate cancer. Int J Cancer 100:309–317
Tan J, Sharief Y, Hamil KG, Gregory CW, Zang DY, Sar M, Gumerlock PH, DeVere White RW, Pretlow TG, Harris SE, Wilson EM, Mohler JL, French FS 1997 Dehydroepiandrosterone activates mutant androgen receptors expressed in the androgen-dependent human prostate cancer xenograft CWR22 and LNCaP cells. Mol Endocrinol 11:450–459
Koh E, Kanaya J, Namiki M 2001 Adrenal steroids in human prostatic cancer cell lines. Arch Androl 46:117–125
Maggiolini M, Recchia AG, Carpino A, Vivacqua A, Fasanella G, Rago V, Pezzi V, Briand PA, Picard D, Ando S 2004 Oestrogen receptor is required for androgen-stimulated proliferation of LNCaP prostate cancer cells. J Mol Endocrinol 32:777–791
Rehman KS, Carr BR 2004 Sex differences in adrenal androgens. Semin Reprod Med 22:349–360
Caplan AJ, Jackson S, Smith D 2003 Hsp90 reaches new heights. Conference on the Hsp90 chaperone machine. EMBO Rep 4:126–130
Georget V, Terouanne B, Nicolas JC, Sultan C 2002 Mechanism of antiandrogen action: key role of hsp90 in conformational change and transcriptional activity of the androgen receptor. Biochemistry 41:11824–11831
Bruder JM, Sobek L, Oettel M 1997 Dehydroepiandrosterone stimulates the estrogen response element. J Steroid Biochem Mol Biol 62:461–466
Lamberts SW, van den Beld AW, van der Lely AJ 1997 The endocrinology of aging. Science 278:419–424
Algarte-Genin M, Cussenot O, Costa P 2004 Prevention of prostate cancer by androgens: experimental paradox or clinical reality. Eur Urol. 46:285–294; discussion 294–295
Missmer SA, Eliassen AH, Barbieri RL, Hankinson SE 2004 Endogenous estrogen, androgen, and progesterone concentrations and breast cancer risk among postmenopausal women. J Natl Cancer Inst 96:1856–1865
Williams MR, Dawood T, Ling S, Dai A, Lew R, Myles K, Funder JW, Sudhir K, Komesaroff PA 2004 Dehydroepiandrosterone increases endothelial cell proliferation in vitro and improves endothelial function in vivo by mechanisms independent of androgen and estrogen receptors. J Clin Endocrinol Metab 89:4708–4715
Simoncini T, Mannella P, Fornari L, Varone G, Caruso A, Genazzani AR 2003 Dehydroepiandrosterone modulates endothelial nitric oxide synthesis via direct genomic and nongenomic mechanisms. Endocrinology 144:3449–3455
Harris HA, Bruner-Tran KL, Zhang X, Osteen KG, Lyttle CR 2005 A selective estrogen receptor- agonist causes lesion regression in an experimentally induced model of endometriosis. Hum Reprod 20:936–941
Harris HA, Albert LM, Leathurby Y, Malamas MS, Mewshaw RE, Miller CP, Kharode YP, Marzolf J, Komm BS, Winneker RC, Frail DE, Henderson RA, Zhu Y, Keith Jr JC 2003 Evaluation of an estrogen receptor- agonist in animal models of human disease. Endocrinology 144:4241–4249
Rontzsch A, Thoss K, Petrow PK, Henzgen S, Brauer R 2004 Amelioration of murine antigen-induced arthritis by dehydroepiandrosterone (DHEA). Inflamm Res 53:189–198
Dulos J, Verbraak E, Bagchus WM, Boots AM, Kaptein A 2004 Severity of murine collagen-induced arthritis correlates with increased CYP7B activity: enhancement of dehydroepiandrosterone metabolism by interleukin-1. Arthritis Rheum 50:3346–3353
Martin C, Ross M, Chapman KE, Andrew R, Bollina P, Seckl JR, Habib FK 2004 CYP7B generates a selective estrogen receptor agonist in human prostate. J Clin Endocrinol Metab 89:2928–2935(Fang Chen, Kristin Knecht)