CYP7B1-mediated metabolism of dehydroepiandrosterone and 5α-androstane-3β,17β-diol – potential role(s) for estrogen signaling

Authors


M. Norlin, Department of Pharmaceutical Biosciences, Division of Biochemistry, University of Uppsala, Box 578, S-751 23 Uppsala, Sweden
Fax: +46 18 558 778
Tel: +46 18 471 4331
E-mail: maria.norlin@farmbio.uu.se

Abstract

CYP7B1, a cytochrome P450 enzyme, metabolizes several steroids involved in hormonal signaling including 5α-androstane-3β,17β-diol (3β-Adiol), an estrogen receptor agonist, and dehydroepiandrosterone, a precursor for sex hormones. Previous studies have suggested that CYP7B1-dependent metabolism involving dehydroepiandrosterone or 3β-Adiol may play an important role for estrogen receptor β-mediated signaling. However, conflicting data are reported regarding the influence of different CYP7B1-related steroids on estrogen receptor β activation. In the present study, we investigated CYP7B1-mediated conversions of dehydroepiandrosterone and 3β-Adiol in porcine microsomes and human kidney cells. As part of these studies, we compared the effects of 3β-Adiol (a CYP7B1 substrate) and 7α-hydroxy-dehydroepiandrosterone (a CYP7B1 product) on estrogen receptor β activation. The data obtained indicated that 3β-Adiol is a more efficient activator, thus lending support to the notion that CYP7B1 catalysis may decrease estrogen receptor β activation. Our data on metabolism indicate that the efficiencies of CYP7B1-mediated hydroxylations of dehydroepiandrosterone and 3β-Adiol are very similar. The enzyme catalyzed both reactions at a similar rate and the Kcat/Km values were in the same order of magnitude. A high dehydroepiandrosterone/3β-Adiol ratio in the incubation mixtures, similar to the ratio of these steroids in many human tissues, strongly suppressed CYP7B1-mediated 3β-Adiol metabolism. As the efficiencies of CYP7B1-mediated hydroxylation of dehydroepiandrosterone and 3β-Adiol are similar, we propose that varying steroid concentrations may be the most important factor determining the rate of CYP7B1-mediated metabolism of dehydroepiandrosterone or 3β-Adiol. Consequently, tissue-specific steroid concentrations may have a strong impact on CYP7B1-dependent catalysis and thus on the levels of different CYP7B1-related steroids that can influence estrogen receptor β signaling.

Abbreviations
3β-Adiol

5α-androstane-3β,17β-diol

DHEA

dehydroepiandrosterone

DHT

dihydrotestosterone

ER

estrogen receptor

ERE

estrogen response element

HEK

human embryonic kidney

The steroid hydroxylase CYP7B1, a member of the cytochrome P450 enzyme family, has attracted increasing interest in recent years due to its multiple reported roles for key events in cellular physiology [1–9]. CYP7B1 is widely expressed in tissues of human and other species and metabolizes several steroids involved in hormonal signaling and other processes. Substrates for CYP7B1 include 5α-androstane-3β,17β-diol (3β-Adiol), an estrogen receptor (ER) agonist, and dehydroepiandrosterone (DHEA), an essential precursor for androgens and estrogens. In addition to its role as sex hormone precursor, DHEA is reported to affect a number of processes in various tissues, including central nervous system function, immune system, lipid profiles and cellular growth [8–13].

The action of CYP7B1 in various tissues results in the formation of 7- and/or 6-hydroxymetabolites, which can be eliminated from the cell, thereby decreasing intracellular levels of CYP7B1 substrates. Some reports, however, suggest that CYP7B1-mediated catalysis might lead to formation of active hormones with impact on several processes, including cellular growth, immune system and brain function [7–9]. In view of its high catalytic activity towards sex hormone precursors, as well as towards certain estrogens, the action of CYP7B1 may affect hormonal signaling in several ways [5,7,14].

Recent studies have indicated that CYP7B1-dependent metabolism may play an important role for ERβ-mediated signaling. The manner in which CYP7B1 affects this, however, remains unclear. The results of some studies indicate that CYP7B1-mediated catalysis leads to formation of an ERβ ligand, whereas other studies have proposed that CYP7B1 catalysis instead counteracts ERβ ligand activation [5,7]. As ERβ is considered to affect a wide range of biological systems throughout the body, events regulating its function are of considerable interest [5,15].

In the present study, we used porcine tissues and human kidney cells to investigate and compare CYP7B1-mediated conversions of DHEA and 3β-Adiol, both of which are reported to affect ERβ activation. The pig is a useful animal model for studies of CYP7B1-mediated catalytic reactions due to the high CYP7B1 content in pig tissues and the closer similarity of porcine and human cytochrome P450 enzymes compared with rodent isoforms [16]. Our findings indicate that CYP7B1 action is subject to age- and tissue-specific differences. Furthermore, the data indicate that tissue-specific steroid concentrations may have a large impact on CYP7B1-dependent catalysis and thus on the levels of different CYP7B1-related steroids that can influence ERβ signaling.

Results

Tissue- and age-specific differences in porcine CYP7B1-mediated hydroxylase activities towards DHEA and 3β-Adiol

In previous studies, we described CYP7B1-mediated 7α-hydroxylation of DHEA in microsomes from pig liver and kidney [3,6]. In the current study, CYP7B1-mediated formation of hydroxymetabolites from DHEA and 3β-Adiol was analyzed in microsomes prepared from various tissues obtained from pigs of different ages. Because this is the first study on CYP7B1-mediated 3β-Adiol metabolism in the pig, GC/MS analysis was carried out to determine the structure of the 3β-Adiol hydroxymetabolite formed. Previous studies report the formation of both 6- and 7-hydroxymetabolites from 3β-Adiol [17–19]. The GC/MS analysis carried out in the present study showed that the main product formed from 3β-Adiol in pig liver is 5α-androstane-3β,7α,17β-triol (for GC/MS chromatogram, see Supplementary material). Only minor amounts of a 6-hydroxy derivative were observed. Also, trace amounts of 5α-androstane-3β,7β,17β-diol were detected by GC/MS. From present and previous findings, it appears that CYP7B1 is capable of carrying out both 6- and 7-hydroxylation of 3β-Adiol. It is possible that the main product formed from 3β-Adiol may vary in different species and different cellular environments.

The results of the analyses of porcine DHEA and 3β-Adiol hydroxylation in liver, kidney and lung are shown in Table 1. The data indicate marked tissue-specific differences between younger and older animals. In liver, 7α-hydroxylation of both substrates increased with age, whereas the rate of hydroxylation decreased with age in the kidney. These findings of age-dependent differences are in agreement with previous data on DHEA metabolism [6] and indicate a similar pattern for 7α-hydroxylation of 3β-Adiol in pig tissues.

Table 1.   Difference in CYP7B1-mediated hydroxylase activity between adult male pig and male piglet. The animals were approximately 10 months (adult) and 5 days (piglet) of age. Microsome fractions were prepared from tissues and the catalytic activity was measured by incubation with radiolabeled substrates and analysis by RP-HPLC, as described in the Experimental procedures. Hydroxylase activity is displayed as pmol·mg−1 of microsomal protein × min ± SD (= 7 samples). Incubations without NADPH were used as negative controls. Product formation in negative controls corresponded to an activity of ≤ 10 pmol·mg−1 protein × min. A comparison of hydroxyproduct formation in tissues obtained from piglets and sexually mature animals indicates marked tissue-specific differences between younger and older animals. There is a significant difference in hydroxyproduct formation between older and younger animals in the kidney and liver for both substrates. In the liver, the activity is higher in older animals than in piglets whereas, in the kidney, the activity decreased with age (< 0.05, one-way ANOVA). BLD, below limit of detection.
 Substratepmol·mg−1 protein × min
LiverKidneyLung
Adult maleDHEA441 ± 127BLD82 ± 10
Male pigletDHEA244 ± 158130 ± 64104 ± 46
Adult male3β-Adiol443 ± 6615 ± 1033 ± 19
Male piglet3β-Adiol76 ± 4062 ± 3143 ± 22

In a separate set of experiments, microsomes were prepared from tissues of a 2.5-year-old boar to examine male reproductive tissues in an older individual. In this animal, hepatic hydroxylase activities towards DHEA and 3β-Adiol were 591 and 659 pmol·mg−1 microsomal protein × min, respectively. Hydroxylase activities in testicle and prostate were approximately 5% of that in liver. In liver, kidney, testicle and prostate, the catalytic activities towards DHEA and 3β-Adiol were of the same order of magnitude (data not shown).

CYP7B1-mediated activities towards DHEA and 3β-Adiol in different sexes

We also compared CYP7B1-mediated hydroxylation of DHEA and 3β-Adiol in tissues from male and female pigs. As the analysis of metabolism in pig tissues is often carried out with organs from castrated pigs, available from slaughterhouses, we also included castrated male pigs in these studies. The results from incubations with microsomes from kidneys, livers and lungs obtained from normal males, females (gilts) and castrated males are shown in Table 2. No significant differences between sexes were observed for 7α-hydroxylation of DHEA or 3β-Adiol in any of the tissues analyzed. Also, formation of 7α-hydroxyproducts from DHEA and 3β-Adiol was found to be of a similar magnitude in normal and castrated males.

Table 2.   CYP7B1-mediated hydroxylase activity towards DHEA and 3β-Adiol in different sexes. The animals were approximately 10 months of age. Microsome fractions were prepared from tissues and the catalytic activity was measured by incubation with radiolabeled substrates and analysis by RP-HPLC, as described in the Experimental procedures. Experiments were carried out in three sets of triplicate incubations based on material from three different individuals per group. Hydroxylase activity is displayed as pmol·mg−1 of microsomal protein × min ± SD (= 9 samples). Incubations without NADPH were used as negative controls. Product formation in negative controls corresponded to an activity of ≤ 10 pmol·mg−1 protein × min. There is no significant difference between the different groups for either of the two substrates. In lung tissue, there is a significant difference between the hydroxylase activity towards DHEA and 3β-Adiol. There is no significant different between the hydroxylase activity towards the two substrates in liver tissue (< 0.05, two-way ANOVA). BLD, below limit of detection.
 Substratepmol·mg−1 protein × min
LiverKidneyLung
Male castratedDHEA290 ± 73BLD80 ± 15
FemaleDHEA254 ± 24BLD78 ± 37
MaleDHEA441 ± 127BLD82 ± 10
Male castrated3β-Adiol448 ± 11112 ± 1039 ± 8
Female3β-Adiol309 ± 3016 ± 531 ± 11
Male3β-Adiol443 ± 6615 ± 1033 ± 19

Kinetic analysis of CYP7B1-mediated metabolism of DHEA and 3β-Adiol

To study which of these two CYP7B1 substrates is most efficiently metabolized and to obtain more information regarding which reaction might be of the most physiological importance, we determined kinetic parameters for 7α-hydroxylation of DHEA and 3β-Adiol using pig liver microsomes. Kinetic parameters, including Kcat and apparent Km values for these reactions, are summarized in Table 3. Data were obtained by incubation with various amounts of DHEA or 3β-Adiol as described in the Experimental procedures and the parameters were determined by nonlinear regression fitting of the data to the Michaelis–Menten equation or calculated from double reciprocal plots. The catalytic efficiencies for hydroxylation of DHEA and 3β-Adiol as described by the Kcat/Km values were in same order of magnitude (Table 3). The Km value for porcine DHEA 7α-hydroxylation in the present study is in agreement with our previously reported data on this reaction in purified protein fractions [3]. From the present study, it may be concluded that the efficiencies of CYP7B1-mediated 7α-hydroxylation of DHEA and 3β-Adiol appear to be very similar.

Table 3.   Kinetic parameters for CYP7B1-mediated 7α-hydroxylation of DHEA and 3β-Adiol in pig liver. Experiments to determine kinetic parameters were carried out by assay of catalytic activities in adult pig liver microsomes with 1, 2, 3, 4 or 5 μm of substrate (DHEA or 3β-Adiol). Parameters were determined by nonlinear regression fitting of the data to the Michaelis–Menten equation or calculated from double reciprocal plots. Experiments were performed in triplicates or quadruplicates. Catalytic activity was measured by RP-HPLC as described in the Experimental procedures.
 Kmm)Vmax (pmol·mg−1 protein × min)Kcat (nmol·nmol−1 P450 × min)Kcat/Km
7α-hydroxylation of DHEA5100020.4
7α-hydroxylation of 3β-Adiol350010.33

CYP7B1-mediated activities towards DHEA and 3β-Adiol in human kidney (HEK293) cells

HEK293 cells are known to have a high endogenous CYP7B1 expression, although, to our knowledge, no data have been reported on 3β-Adiol metabolism in these cells. In the present study, we examined the endogenous hydroxylase activities towards 3β-Adiol and DHEA in HEK293 cells. The endogenous 7α-hydroxylase activity towards DHEA in HEK293 cells in these experiments was approximately 300 pmol·mg−1 cell protein × 24 h. Analysis of 3β-Adiol metabolism indicated the formation of more than one hydroxymetabolite in HEK293 cells. GC/MS analysis of metabolites demonstrated the formation of both 6- and 7-hydroxy derivatives under these conditions. The mass spectrum of the main metabolite formed from 3β-Adiol in HEK293 cell cultures was consistent with a 6-hydroxy derivative, i.e. 5α-androstane-3β,6ξ,17β-triol (for GC/MS chromatogram, see Supplementary material). We were unable to distinguish between α or β orientation of the 6-hydroxygroup, as indicated by the letter ξ (Greek letter ‘xi’, which corresponds to ‘x’ in our alphabet, representing an unknown configuration). Small amounts of 5α-androstane-3β,7α,17β-triol were also formed. Endogenous hydroxylase activity towards 3β-Adiol in HEK293 cells was approximately 450 pmol·mg−1 cell protein × 24 h.

We also examined the activities towards 3β-Adiol and DHEA in HEK293 cells transfected with an expression vector containing human CYP7B1 cDNA. As expected, overexpression of human CYP7B1 significantly increased the hydroxylation of both 3β-Adiol and DHEA in HEK293 cultures. HEK293 cells transfected with the CYP7B1 expression vector displayed three- to six-fold higher hydroxylase activity than cells transfected with empty vector (data not shown). GC/MS analysis of 3β-Adiol metabolites formed in CYP7B1-transfected cell cultures also showed formation of both 6- and 7-hydroxy derivatives with 5α-androstane-3β,6ξ,17β-triol as the main metabolite.

Effects of 3β-Adiol on CYP7B1-mediated hydroxylation of DHEA

As the studies indicated that the rate of hydroxylation and affinity of CYP7B1 for DHEA and 3β-Adiol are similar, we conducted experiments to study how the concentration of 3β-Adiol would affect the CYP7B1-mediated metabolism of DHEA and vice versa. These experiments were carried out with both porcine microsomes and human HEK293 cells. The main reason for our interest in the effects of steroid concentrations on the rate of metabolic reactions is that the concentrations of these steroids vary considerably in different tissues and species. For example, the concentration of DHEA in human prostate is reported to be approximately 10-fold higher than that of 3β-Adiol [20].

In the first set of experiments, we studied the effect of 3β-Adiol on DHEA hydroxylation. Hydroxylation of DHEA was measured with radiolabeled DHEA in the presence of increasing amounts of unlabeled 3β-Adiol. The results of these experiments are shown in Fig. 1. 3β-Adiol inhibited DHEA hydroxylation by approximately 60–70% when both steroids were present at equimolar concentrations. Further, a 10-fold higher concentration of 3β-Adiol than of DHEA in the incubation mixture resulted in the suppression of DHEA hydroxylation by 80%. In HEK293 cells, the suppressive effect of 3β-Adiol of DHEA metabolism was statistically significant also at a 10-fold lower concentration of 3β-Adiol than of DHEA in the incubation mixture (Fig. 1).

Figure 1.

 Effects of 3β-Adiol on DHEA hydroxylation in pig liver microsomes (black bars) and HEK293 cells (grey bars). Experiments were carried out with a constant concentration (52 μm) of radiolabeled DHEA to study the inhibitory effect by varying levels of 3β-Adiol on DHEA hydroxylation. Concentrations are shown as ratio between added inhibitor (3β-Adiol) and substrate (DHEA). Analysis of the catalytic activity was carried out as described in the Experimental procedures. Catalytic activity is shown as percent of the rate of DHEA hydroxylation in controls (± SD) (= 4–9). Controls consisted of incubations without added inhibitor [0/1]. 3β-Adiol inhibited DHEA hydroxylation by approximately 60–75% when both steroids were present at equimolar concentrations, and when the concentration of 3β-Adiol was increased to 10-fold, compared with the concentration of DHEA, the hydroxylation of DHEA decreased by approximately 80%. *Statistically significant suppression compared with control (< 0.05, one-way ANOVA).

Analysis of the type of inhibition by 3β-Adiol on DHEA hydroxylation was carried out by a series of incubations with pig liver microsomes containing 1–5 μm of DHEA in the presence of 0, 5, 10, or 20 μm unlabeled 3β-Adiol (added as inhibitor). Lineweaver–Burk and Dixon plots were constructed using linear regression fitting of the data. The data obtained indicated that 3β-Adiol is a mixed inhibitor of DHEA hydroxylation (for a Lineweaver–Burk plot showing inhibition by 3β-Adiol on DHEA hydroxylation, see Supplementary material). Thus, the inhibition includes competition at the active site but involves both competitive and uncompetitive components. The patterns of Lineweaver–Burk and Dixon plots were not consistent with a pure noncompetitive or uncompetitive type of inhibition (data not shown). The Ki value for the inhibition of 3β-Adiol on DHEA hydroxylation, calculated from the Dixon plot, was 6 μm.

Effects of DHEA on CYP7B1-mediated 3β-Adiol hydroxylation

We also carried out corresponding studies on the effect of different concentrations of DHEA on the 6- and 7-hydroxylation of 3β-Adiol. In these experiments, we measured 3β-Adiol hydroxylation in pig liver microsomes and human HEK293 cells with radiolabeled 3β-adiol in the presence of increasing amounts of unlabeled DHEA (Fig. 2). The most efficient inhibition of 3β-Adiol metabolism was obtained at high amounts of DHEA, although, in experiments with HEK293 cells, statistically significant effects were observed also at lower concentrations. A 10-fold higher concentration of DHEA than of 3β-Adiol in the incubation mixture decreased the rate of 3β-Adiol hydroxylation by 70–90% (Fig. 2).

Figure 2.

 Effects of DHEA on 3β-Adiol hydroxylation in pig liver microsomes (black bars) and HEK293 cells (grey bars). Experiments were carried out with a constant concentration (52 μm) of radiolabeled 3β-Adiol to study the inhibitory effect by varying levels of DHEA on 3β-Adiol hydroxylation. Concentrations are shown as ratio between added inhibitor (DHEA) and substrate (3β-Adiol). Analysis of the catalytic activity was carried out as described in the Experimental procedures. Catalytic activity is shown as percent of the rate of 3β-Adiol hydroxylation in controls (± SD) (= 3–9). Controls consisted of incubations without added inhibitor [0/1]. DHEA inhibited 3β-Adiol hydroxylation by approximately 70–90% when the concentration of DHEA was increased to 10-fold compared with the concentration of 3β-Adiol. *Statistically significant suppression compared with control (< 0.05, one-way ANOVA).

Similarly, as for the inhibition of DHEA hydroxylation, we carried out experiments to analyze the type of inhibition of 3β-Adiol hydroxylation by DHEA. Incubations were performed with pig liver microsomes and radiolabeled 3β-Adiol (1–5 μm) in the presence of unlabeled DHEA (0, 5, 10, or 20 μm, added as inhibitor). Construction of Lineweaver–Burk and Dixon plots from the data indicated that DHEA is a mixed inhibitor of 3β-Adiol hydroxylation, similar to that found for the inhibition of DHEA hydroxylation by 3β-Adiol (for a Lineweaver–Burk plot showing inhibition by DHEA on 3β-Adiol hydroxylation, see Supplementary material). The patterns of Lineweaver–Burk and Dixon plots were not consistent with a pure noncompetitive or uncompetitive type of inhibition. The Ki for inhibition of 3β-Adiol hydroxylation by DHEA was 24 μm, a higher value than the Ki for inhibition of DHEA hydroxylation by 3β-Adiol (6 μm; see above). Thus, it appears that 3β-Adiol may be a more efficient inhibitor of the hydroxylation of DHEA than vice versa.

Some inhibition experiments were also carried out with DHEA-sulfate, the sulfated ester of DHEA, which is present in even higher amounts than DHEA in human plasma. However, DHEA-sulfate had no significant effect on the rate of CYP7B1-mediated hydroxylation of either DHEA or 3β-Adiol, even at 10-fold higher concentrations of DHEA-sulfate in the incubation mixture (data not shown). Testosterone, which is formed from DHEA and is present in high levels in the same cell types, also did not affect CYP7B1-mediated hydroxylase activity. It may therefore be concluded that the steroid inhibition described in the present study is specific and not a generalized effect by related steroids of a similar structure.

Analysis of ERβ activation by 3β-Adiol and 7α-hydroxy-DHEA

Previous studies indicate that ERβ is activated by 3β-Adiol, whereas others report that this receptor is activated by 7α-hydroxy-DHEA [5,7]. In the present study, we compared the effects of these two steroids on ERβ activation in the same experiment, using the same methodology. Receptor activation was studied by reporter assay with an ER-responsive luciferase reporter vector transfected in HEK293 cells overexpressed with ERβ, in a similar fashion to that previously described [14,21]. This ER-responsive vector contains a strong estrogen response element (ERE) coupled to luciferase. Thus, luciferase expression levels are determined by activation of ER and its subsequent binding to the ERE. ERβ-transfected cells were treated with 3β-Adiol or 7α-hydroxy-DHEA in different concentrations and the levels of luciferase in steroid-treated cells were compared with the luciferase levels in cells treated with the same volume of vehicle (ethanol). Treatment with 17β-estradiol, a known ER agonist, was used as positive control. The results of these experiments (Fig. 3) indicate that 3β-Adiol is a more efficient ERβ activator than 7α-hydroxy-DHEA.

Figure 3.

 Effects of estradiol (E2), 7α-OH-DHEA (7HD) or 3β-Adiol on an ER-responsive ERE luciferase reporter vector. HEK293 cells were transiently transfected with ERβ (1 μg·well−1) and treated with steroids as described in the Experimental procedures. Data are presented as the fold stimulation compared with controls (treated with vehicle). *Statistically significant stimulation compared with control (< 0.05, one-way ANOVA).

Discussion

In the present study, we examined CYP7B1-mediated metabolism of 3β-Adiol and DHEA in porcine microsomal fractions and human kidney cells. The current data indicate that the efficiencies of these two reactions are very similar. As the kcat/Km values are within the same order of magnitude for both substrates, we propose that relative formation of different CYP7B1-related steroids may be dependent on the concentration of substrate(s) present in each tissue. Consequently, cellular steroid levels may have a strong impact on the resulting physiological levels of CYP7B1 substrates as well as of CYP7B1-formed hydroxysteroids.

The concentrations of DHEA and 3β-Adiol and related steroids vary considerably between different tissues [10,20,22]. For example, concentrations of DHEA in tissue samples are reported to be 20-fold higher in human prostate (100 pmol·g−1) than in muscle (5 pmol·g−1). In general, physiological levels of 3β-Adiol are lower than those of DHEA in both humans and pigs, particularly in human adults where plasma concentrations are reported to be approximately 1.5 nm for 3β-Adiol and approximately 10- to 30-fold higher for DHEA [20,22,23]. In human plasma, DHEA-sulfate is present in higher concentrations than DHEA [23,24]. The plasma levels of these steroids are subject to developmental variation, depending on the stage of sexual maturation [10,24]. From the results of the present study, it is clear that high concentrations of DHEA, but not of DHEA-sulfate, strongly suppress CYP7B1-mediated metabolism of 3β-Adiol. The current data imply that human CYP7B1-mediated metabolism of 3β-Adiol should be comparatively lower than that of DHEA in many human tissues due to the generally higher tissue concentrations of DHEA, which most likely would suppress 3β-Adiol hydroxylation during physiological conditions.

CYP7B1-mediated metabolism of DHEA and 3β-Adiol are of interest in connection with several physiological processes. One of these concerns the proposed role of this enzyme for ER-mediated signaling and the potential ER-mediated control of cellular growth, particularly of the prostate [5,25]. However, the available data regarding the effect of CYP7B1 on ERβ activation are conflicting. Data obtained by Weihua et al. [5,26] in rodents indicate that 3β-Adiol binds strongly to ERβ and that CYP7B1-mediated metabolism of 3β-Adiol therefore should lead to induced growth by abolishing ERβ action. By contrast, Martin et al. [7], who studied DHEA 7α-hydroxylation in human prostate cells, report that the CYP7B1-formed product 7α-hydroxy-DHEA is a ligand of ERβ and conclude that CYP7B1 metabolism therefore should activate this receptor. The contrasting findings on the role of CYP7B1 in connection with ERβ action were obtained in different species using different CYP7B1 substrates and methodologies.

In the present study, we carried out experiments to compare effects of 3β-Adiol and 7α-hydroxy-DHEA on activation of ERβ in the same assay and cell type. Our experiments indicate that 3β-Adiol is a more efficient activator than 7α-hydroxy-DHEA. These data are in agreement with the findings obtained by Martin et al. [7] who observed activation of ERβ by 7α-hydroxy-DHEA only at 5 μm or higher concentrations of this metabolite.

DHEA concentrations are generally higher in humans than in rodents and exceed, by at least 10-fold, the concentration of 3β-Adiol in human prostate [22]. Even though DHEA circulates in micromolar concentrations in plasma, DHEA levels measured in human prostate tissue are reported to be approximately 100 pmol·g−1 [20,22]. Thus, the amount of 7α-hydroxy-DHEA needed for activation of ERβ appears to be approximately 50- to 100-fold higher than the reported content of DHEA in human prostate tissue. The concentration of DHEA-sulfate, which can be converted to DHEA by a sulfatase, is in the same order of magnitude in prostate tissue as that of DHEA despite the abundance of DHEA-sulfate in blood [10,22]. It is therefore unlikely that the concentration of a DHEA-metabolite such as 7α-hydroxy-DHEA should reach a high enough concentration to be of physiological relevance for ERβ activation, at least in this tissue. However, Weihua et al. [26] reported that very low concentrations of 3β-Adiol are able to bind ERβ. Ligand-competition experiments with [125I]estradiol indicated a Ki of 2 nm for 3β-Adiol, a value closer to the reported physiological levels of this steroid, which are in the nanomolar range [22,26,27].

From the data on the catalytic properties of CYP7B1 obtained in the present study, it is clear that high concentrations of DHEA strongly suppress CYP7B1-mediated 3β-Adiol metabolism. Although 3β-Adiol is able to compete for the active site of CYP7B1 to some degree even at a 10-fold higher DHEA concentrations, as the present data indicate, the rate of its hydroxylation was decreased by 70–90% under these conditions. It is therefore likely that the comparatively higher DHEA content in prostate and many other tissues may increase the level of 3β-Adiol, and thus of 3β-Adiol-mediated ERβ activation, by suppression of 3β-Adiol metabolism. A proposed effect of high cellular DHEA/3β-Adiol ratios, such as in human prostate, on ERβ activation is outlined in Fig. 4. At high DHEA levels, CYP7B1-mediated 3β-Adiol hydroxylation is strongly suppressed, resulting in higher 3β-Adiol levels and increased ERβ activation (Fig. 4). The steroid suppression of CYP7B1-mediated metabolism observed in the present study is apparently not an unspecific steroid effect because our present and previous data have shown that neither DHEA-sulfate, nor testosterone are able to inhibit CYP7B1-mediated catalysis [3].

Figure 4.

 Suggested effects of a high DHEA/3β-Adiol ratio (such as in prostate tissue) on 3β-Adiol-mediated ERβ activation. High DHEA levels strongly suppresses CYP7B1-mediated 3β-Adiol metabolism, resulting in higher 3β-Adiol levels and increased ERβ activation. The CYP7B1-mediated 7α-OH-DHEA metabolite (not shown) is most likely not formed in sufficient amounts to compete with 3β-Adiol for ERβ binding.

In conclusion, the results obtained in the present study indicate that the efficiencies of DHEA and 3β-Adiol hydroxylation by CYP7B1 are similar, but that a high DHEA/3β-Adiol ratio (similar to the ratio of these steroids in many human tissues) strongly suppresses CYP7B1-mediated 3β-Adiol metabolism. Our data indicate that tissue-specific steroid concentrations may have a large impact on CYP7B1-dependent catalysis and thus on the levels of different CYP7B1-related steroids that can influence ERβ-signaling.

Experimental procedures

Materials

Human embryonic kidney cells (HEK293) (ATCC CRL 1573) were purchased from ATCC (Rockville, MD, USA). DMEM (containing 1000 mg·L−1 glucose), fetal bovine serum, antibiotics/antimycotics, non-essential amino acids and trypsin were obtained from Invitrogen (Carlsbad, CA, USA). The pCMV6 vector containing cDNA encoding for human CYP7B1 was kindly provided by D. W. Russell (University of Texas, Dallas, TX, USA). The ERβ expression vector and the ERE luciferase reporter vector were generous gifts from P. Chambon (Institut de génétique et de biologie moléculaire et cellulaire, Strasbourg, France) and K. Arcaro (University of Massachusetts, MA, USA), respectively. 3β-Adiol, 5α-androstane-3α,17β-diol, DHEA, DHEA-sulfate, dihydrotestosterone (DHT) and hydroxysteroid dehydrogenase (from Pseudomonas testosteroni) were from Sigma Chemical Co. (St Louis, MO, USA). 3H-Labeled DHEA and DHT were from Perkin Elmer Life Sciences (Waltham, MA, USA). All remaining chemicals were of analytical grade and purchased from commercial sources.

Animals and tissue sample collection

Liver, kidney and lung tissues from adult female, male and castrated male pigs (aged 10 months) and liver, kidney, testicle and prostate tissues from an adult domestic boar (aged 2.5 years) were obtained from the Funbo-Lövsta Research Center [Department of Animal Breeding and Genetics, Swedish University of Agricultural Sciences (SLU), Ultuna, Sweden]. Liver, kidney and lung tissues from male uncastrated piglets (aged 5 days) were a generous gift from P. Wallgren [Department of Ruminant and Porcine Diseases, National Veterinary Institute (SVA), Uppsala, Sweden]. All the animals were healthy and untreated at the time of euthanasia. The domestic pig is considered to reach sexual maturity at approximately 6 months of age. All organ tissue samples were stored at −80 °C until microsomal preparation was performed.

Microsomal preparation

The tissue samples were weighed and minced, respectively, in sucrose buffer containing 0.25 m sucrose, 10 mm Tris–Cl (pH 7.4) and 1 mm EDTA to a 20% suspension. Microsomes were prepared from the tissues according to standard methods [28], and were suspended in 50 mm Tris-acetate buffer (pH 7.4) containing 20% glycerol and 0.1 mm EDTA and stored at −80 °C until incubation. Protein contents of the microsomes were assayed by the method of Lowry et al. [29].

Preparation of 3β-Adiol

3H-Labeled 3β-Adiol was prepared from 3H-labeled DHT (Perkin Elmer) by bioconversion using a commercially available hydroxysteroid dehydrogenase from P. testosteroni (Sigma H8879) [30]. A mixture of labeled (100 μCi) and unlabeled (10 μg) DHT was incubated with 0.32 U of hydroxysteroid dehydrogenase and 1 μmol of NADH in a total volume of 1 mL of 0.1 m potassium phosphate buffer containing 0.1 mm EDTA (pH 7.4) for 10 min at 37 °C in a water bath. The reaction was quenched and extracted with 5 mL of ethyl acetate. The organic phase was collected, evaporated under nitrogen gas, dissolved in a small amount of acetone and applied on a silica gel TLC plate. The TLC plate was developed three times in toluene/methanol (90 : 10, v/v). TLC plates with unlabeled DHT, 3β-Adiol and 5α-androstane-3α,17β-diol were used as references and developed together with the sample plate. The sample TLC plate was scanned for localization of the radioactive products, using a Berthold Tracemaster 20 TLC scanner (Berthold/Frieske GmbH, Karlsruhe-Durlach, Germany). The reference TLC plates were exposed to iodine vapours (o/n) to visualize the steroids and the retention times of reference compounds were compared with those of the sample plate. Under these conditions, the main product formed was 3β-Adiol. The yield of 3β-Adiol under the conditions used was approximately 70%. Very small amounts of the 3α-hydroxyderivative, 5α-androstane-3α,17β-diol, were also formed in the reaction. The formed radioactive 3β-Adiol was extracted from the silica gel with ethyl acetate. The extraction procedure was repeated twice. The obtained solution of 3H-labeled 3β-Adiol was evaporated under nitrogen gas and dissolved in 100 μL of ethyl acetate. Radioactivity (c.p.m.·μL−1) was determined by injection of an aliquot on a RP-HPLC (125 × 4 mm LiChrosphere RP 18 column, 5 μm; Merck, Darmstadt, Germany). Elution was monitored by a Radiomatic 150TR Flow Scintillation Analyzer (Hewlett-Packard, Palo Alto, CA, USA). The solution of 3H-labeled 3β-Adiol was diluted to a working concentration of 50 000 c.p.m.·μL−1.

Incubations with microsomes

Incubations with microsomes (0.5–1.0 mg of microsomal protein) were carried out at 37 °C for 20 or 30 min. The substrates DHEA (52 μm, 1 μCi) or 3β-Adiol (52 μm, 200 000 c.p.m.) dissolved in 25 μL of acetone, were incubated with 1 μmol of NADPH in a total volume of 1 mL of 50 mm Tris-acetate buffer (pH 7.4) containing 20% glycerol and 0.1 mm EDTA. Incubations were performed under conditions where the enzyme was saturated with substrate. The incubations were quenched and extracted with 5 mL of ethyl acetate. The organic phase was collected and stored at −20 °C until analysis. Incubations without NADPH were performed at the same time and used as negative controls.

For incubations where steroids were added as potential inhibitors, incubations with labeled substrates were carried out as described above, except that various amounts of unlabeled 3β-Adiol, DHEA, DHEA-sulfate or testosterone were added to the incubation mixtures. For kinetic analysis, pig liver microsomes were incubated with 1, 2, 3, 4 or 5 μm of radiolabeled substrate (DHEA or 3β-Adiol) in the presence of 0, 5, 10 or 20 μm of unlabeled DHEA or 3β-Adiol added as inhibitors. Experiments were carried out in triplicates or quadruplicates and kinetic parameters were determined by nonlinear regression fitting of the data to the Michaelis–Menten equation or calculated from Lineweaver–Burk and Dixon plots.

Cultures of HEK293 cells

HEK293 cells were seeded at approximately 7.5 × 105 cells per 60 mm tissue culture dish in DMEM supplemented with 10% fetal bovine serum and antibiotics/antimycotics. Endogenous enzymatic activity towards DHEA or 3β-Adiol in these cells was examined by addition of 15 μg of substrate dissolved in dimethyl sulfoxide to the medium and incubation for 3, 6, 12 or 24 h at 37 °C with 5% CO2. Most experiments were carried out with an incubation time of 24 h. Following incubations with substrate, the medium was collected and extracted and the organic phase was analyzed for hydroxylated metabolites, as described below. Incubations terminated immediately after addition of substrate (corresponding to an incubation time of 0 h) were used as negative controls. Protein contents of the HEK293 cells were assayed by the method of Lowry et al. [29].

Incubations where steroids were added as potential inhibitors were carried out as described above, except that 1.5–150 μg of unlabeled 3β-Adiol, DHEA or DHEA-sulfate were added to the cell media together with the labeled substrates. Inhibition experiments in HEK293 cell cultures were generally carried out using incubation times of 24 h. In a separate set of experiments, cells were incubated with inhibitors for 12 h instead of 24 h to examine whether a difference in incubation time might influence the results. Effects of steroid inhibitors, however, were found to be similar with 12 h of incubation as with 24 h of incubation.

Overexpression of recombinant human CYP7B1 in HEK293 cells

HEK293 cells were cultured as described above and transfected with the pCMV6 vector containing cDNA encoding for human CYP7B1 [31]. In control experiments, cells were transfected with the same amount of empty pCMV vector without the CYP7B1 insert. Transfection was carried out by electroporation in 0.4 cm cuvettes (Gene Pulser II; Bio-Rad, Hercules, CA, USA), using a single pulse of 0.4 kV and 100 μF. In each experiment, 20 × 106 cells were transfected with 20 μg of DNA in a volume of 0.8 mL of phosphate-buffered saline containing calcium chloride and magnesium chloride (Dulbecco’s, Life Technologies, Inc., Grand Island, NY, USA). After transfection, the cells were cultured for 24 h on 60 mm plates in medium containing 3β-Adiol or DHEA (15 μg) dissolved in dimethylsulfoxide. Following incubations with substrate, the medium was collected and extracted and the organic phase was analyzed for hydroxylated metabolites, as described below.

Analysis of incubations with DHEA

Incubations with 3H-labeled DHEA were analyzed as previously described [3,32]. The organic phase of the incubations with microsomes or HEK293 cell cultures was evaporated under nitrogen gas, dissolved in 100 μL of mobile phase methanol/water (50 : 50, v/v) and subjected to radio RP-HPLC using a 125 × 4 mm LiChrosphere RP 18 column (5 μm; Merck). Elution of labeled steroids was monitored by a Radiomatic 150TR Flow Scintillation Analyzer (Hewlett-Packard). The RP-HPLC mobile phase system consisted of methanol/water (50 : 50, v/v) for 10 min, a linear gradient from 50–100% methanol for the next 10 min and then 100% methanol for the remaining 10 min [3,32]. The retention times were 8–9 min for 7α-hydroxy-DHEA and 19–20 min for DHEA.

HPLC analysis of incubations with 3β-Adiol

Incubations with 3β-Adiol were analyzed either by radio RP-HPLC using 3H-labeled 3β-Adiol, prepared as described above, or by GC/MS analysis (see below) using unlabeled 3β-Adiol (Sigma).

Incubations with 3H-labeled 3β-Adiol were analyzed using a similar system as for incubations with 3H-labeled DHEA. The organic phase of the incubations was evaporated, dissolved in 50% methanol and subjected to RP-HPLC using a 125 × 4 mm LiChrosphere RP 18 column (5 μm; Merck). The RP-HPLC mobile phase system was the same as for incubations with DHEA. The retention times were 7–8 min for 5α-androstane-3β,7α,17β-triol and 5α-androstane-3β,6ξ,17β-triol and 19–20 min for 3β-Adiol. An additional polar metabolite with a retention time of 4–5 min was also detected in the incubations with porcine liver microsomes, but we were unable to characterize this compound further.

GC/MS: identification of 3β-Adiol metabolites

Metabolites of 3β-Adiol present in incubation mixtures or fractions collected from the (radio) HPLC system were identified by GC/MS. Prior to GC/MS analysis, the incubation mixtures or HPLC fractions were extracted using ethyl acetate. The ethyl acetate phase was collected and dried under a gentle stream of nitrogen. The residue was then dissolved in 3 mL of methanol followed by addition of 2 mL of water. The solution was passed through a Sep-Pak C18 cartridge containing octadecylsilane bonded silica (Waters Associates Inc, Milford, MA, USA) followed by 5 mL of water. The total effluent was collected and the organic solvent was removed in vacuo. The remaining aqueous phase was then passed through the same unwashed Sep-Pak C18 cartridge again, before washing with 5 mL of water. Steroids were then eluted with 10 mL of 75% aqueous methanol. This solution was passed through a column (40 × 4 mm) of the strong lipophilic anion exchanger, TEAP-LH-20, in bicarbonate form [33]. After elution with an additional 5 mL of methanol, the total effluent from the column was taken to dryness in vacuo. The residue was then transferred with methanol to a stoppered tube, dried under nitrogen, and steroids were trimethylsilylated in 0.2 mL of pyridine/hexamethyldisilazane/trimethylchlorosilane (3 : 2 : 1, v/v/v), by heating at 60 °C for 30 min. The reagents were removed under nitrogen, and the derivatives were redissolved in hexane. GC/MS was performed using a Finnigan (San Jose, CA, USA) SSQ 710 instrument housing a fused silica column (25 × 0.32 mm) coated with a 0.17-μm layer of cross-linked methyl silicone (Ultra 1; Hewlett-Packard) ending in the ion source. An on-column injection device was used. The oven temperature was 50 °C during the injection and, after 3 min, it was rapidly increased to 185 °C, and then programmed to 280 °C at a rate of 5 °C min−1. The electron energy was 50 eV, and repetitive scanning (30 scans min−1) over the m/z range 50–800 was started after a suitable delay. The identification of a steroid was based on the retention time and/or complete mass spectrum, which were compared with those of reference steroids. The retention indices (Kovats) for 3β-Adiol (5α-androstane-3β,17β-diol) and its characterized major metabolites 5α-androstane-3β,7α,17β-triol (formed in incubations with pig liver microsomes) and 5α-androstane-3β,6ξ,17β-triol (formed in incubations with HEK293) were 2595, 2640 and 2755, respectively. In addition, trace amounts of 5α-androstane-3β,7β,17β-triol (Kovats 2785) were also detected in some incubations. The mass spectrum of the isolated 5α-androstane-3β,7α,17β-triol showed the following intense/significant ions: m/z 524 (molecular ion), 509, 434, 393 (base peak), 344, 254, 239 and 129. The same significant ions were seen in the spectrum of isolated 5α-androstane-3β,7β,17β-triol, but the base peak was then m/z 434. These mass spectra were essentially identical to those of the corresponding reference compounds. Major significant ions in the mass spectrum of isolated 5α-androstane-3β,6ξ,17β-triol were m/z 524 (molecular ion), 509, 434, 419, 344, 329, 318, 254, 239, 228 and 129 (base peak). This mass spectrum was very similar to that of authentic 5α-androstane-3β,6β,17β-triol and differed from those of 3β-Adiol with an extra hydroxy group in another position (i.e. 1, 2, 7, 11, 15, 16, 18 or 19 position). However, because we did not have a mass spectrum of 5α-androstane-3β,6α,17β-triol for comparison, a 6α-hydroxy group of the isolated metabolite could not be excluded. Thus, the absolute configuration of the 6-hydroxy group of this metabolite was not established, as indicated by the letter ξ.

Analysis of ERβ activation by ERE-reporter luciferase assay

HEK293 cells were transiently transfected with an ER-responsive luciferase reporter vector together with an ERβ expression vector and a pCMV β-galactosidase plasmid (to control for transfection efficiency) using calcium co-precipitation, as previously described [14]. The ER-responsive vector contains a strong ERE coupled to luciferase [21]. Transfected cells were treated with 3β-Adiol, 7α-hydroxy-DHEA or estradiol (0.1–1 μm), dissolved in ethanol, and the levels of luciferase in steroid-treated cells were compared with the luciferase levels in cells treated with the same volume of vehicle. Luciferase and β-galactosidase activities were assayed as previously described [14]. ERE-reporter luciferase activity is expressed as relative light units divided by β-galactosidase activity (A420).

Statistical analysis

Data are expressed as mean ± SD. Statistical analysis was performed using one- and two-way analysis of variance followed by Dunnet’s post-hoc test, comparing hydroxylation of different substrates and/or groups of sexes. P < 0.05 was considered statistically significant. The software used was minitab®, release 14 (Minitab Ltd, Coventry, UK).

Acknowledgements

The present study was supported by grants from the Swedish Research Council Medicine and the Åke Wiberg foundation.

Ancillary