Calcitrol (1α,25-dihydroxyvitamin D₃) inhibits androgen glucuronidation in prostate cancer cells

Prostate cancer is the second most common cause of cancer death in Americans (1). Over the last several years, the active metabolite of vitamin D, 1α,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃, calcitriol], has been identified as a promising anticancer agent (2). In addition to its classic targets (the regulation of calcium homeostasis and bone metabolism), calcitriol exhibits both antiproliferative and prodifferentiating actions in several malignant cells and tumors, including those from prostate cancer (3–5). Furthermore, in clinical trials, calcitriol shows synergistic and/or additive effects with chemotherapy, radiation, or other anticancer treatments (6, 7). The mechanisms underlying the antiproliferative activity of calcitriol are varied and are cell type specific (8). Nevertheless, calcitriol exerts antiproliferative actions in a variety of prostate cancer cell lines (9–11) as well as in primary cultures of normal and cancer cells (12). Most of its biological effects are mediated by direct transcriptional regulation of specific target genes. The vitamin D derivative binds to and activates the nuclear receptor VDR, which in turn regulates the transcriptional activity of responsive genes (13).

The androgens testosterone and dihydrotestosterone (DHT) are required for prostate growth but also play an important role in prostate cancer development. Both experimental and epidemiologic data indicate that androgens are among the main factors controlling the development, maintenance, and progression of prostate cancer (13, 14). Thus, a first-line treatment strategy for advanced prostate cancer is androgen deprivation. DHT is extensively metabolized in the prostate (reviewed in refs. 15, 16) to ultimately form androstane-3α-diol-17glucuronide (3α-diol-17G) and androsterone-3glucuronide (ADT-3G), which are the major androgen metabolic end products found in circulation (17, 18). Glucuronidation is a conjugation reaction, in which the glucuronosyl group from the cofactor UDP-glucuronic acid is transferred to an acceptor molecule catalyzed by enzymes of the UDP-glucuronosyltransferase (UGT) superfamily (19). Among the 18 functional human UGT isoforms identified thus far, only UGT2B7, UGT2B15, and UGT2B17 are able to conjugate androgens and only UGT2B15 and UGT2B17 are expressed in the human prostate (15). The amino acid sequence identity between these two enzymes is very high (95%), but their specificity towards 5α-reduced androgens differs significantly. The conjugating activity of UGT2B15 is limited to the formation of 3α-diol-17G and DHT-17G, whereas UGT2B17 forms 3α-diol-17G, DHT-17G, and ADT-3G (20, 21).

Although several genes regulated by calcitriol have been identified in prostate cancer cells, the particular effect of this active vitamin D metabolite on androgen metabolism has not been investigated. Considering the fact that a microarray study in LNCaP cells revealed a significant reduction of UGT2B15 expression after treatment with calcitriol (22) and that the major role of glucuronidation is the inactivation of androgens, the present study was undertaken to elucidate the effect of calcitriol on the expression of UGT2B15 and UGT2B17 and on their capacity to inactivate 5α-reduced androgens. The androgen-dependent prostate cancer cell line LNCaP is considered a classic model for the study of androgen metabolism in the prostate (18) and was therefore used in this study. LNCaP cells express a functional androgen receptor (AR; ref. 23) and VDR (24, 25) as well as the two androgen-glucuronidating UGT isoforms characterized in the prostate, UGT2B15 and UGT2B17 (26).

1,25(OH)₂D₃, UDP-glucuronic acid, poly-l-lysine, cycloheximide (CHX), the anti-actin antibody, and the steroid substrates (3α-diol, ADT, and DHT) were obtained from Sigma. 3α-Diol-17G, ADT-3G, and DHT-17G were purchased from Steraloids and 3α-diol-3G was purchased from Research Plus. Casodex was provided by Endorecherche, and the synthetic AR agonist, R1881, was from Perkin-Elmer. Protein assay reagents were obtained from Bio-Rad Laboratories. Cell culture materials and transfection reagents were obtained from Invitrogen. SYBR Green PCR mix was purchased from Applied Biosystems. The anti-AR antibody was purchased from Santa Cruz Biotechnology, the anti-UGT2B15 and anti-UGT2B17 antibodies were described previously (26, 27), and the secondary antibody against rabbit IgG was purchased from Amersham.

Human prostate cancer LNCaP, 22Rv1, PC-3, and DU145 cells were obtained from the American Type Culture Collection. LNCaP and 22Rv1 cells were grown in RPMI 1640 supplemented with 10% FBS; PC-3 cells in Ham’s F-12K medium supplemented with 10% FBS and DU145 cells were cultured in MEM supplemented with 10% FBS. For RNA and protein analyses, LNCaP cells were seeded in six-well plates and cultured in RPMI 1640 supplemented with 0%, 0.2%, or 10% FBS, as indicated, for 48 h to allow cell attachment. 22Rv1, PC-3, and DU145 cells were seeded in 12-well plates in their respective medium supplemented with 0.2% FBS and cultured overnight. All cell lines were then treated for the indicated duration with vehicle (DMSO or ethanol, 0.1%, v/v), calcitriol at indicated concentrations, R1881 (1 nmol/L), casodex (10 μmol/L), and/or CHX (20 μg/mL) in their respective medium supplemented with 0%, 0.2%, or 10% FBS as indicated. For glucuronidation assays, LNCaP cells were seeded in 10-cm culture dishes and cultured until 70% confluency. After a FBS deprivation period of 24 h in RPMI 1640 supplemented with 0.2% FBS, cells were treated with vehicle (ethanol) or calcitriol (10 nmol/L) for 48 or 96 h in the same medium (medium was changed after 48 h for the 72- and 96-h treatment periods). Finally, cells were harvested and resuspended in PBS (pH 7.4, 50 mmol/L) containing DTT (0.5 mmol/L) and homogenized by pipetting.

Glucuronidation assays were conducted using 20 μg total protein from homogenized LNCaP cells as described previously (28). The formation of glucuronide conjugates was analyzed by high-performance liquid chromatography using identification by mass spectrometry as already reported (28, 29).

Total RNA isolation and reverse transcription (RT) were done as described previously (28, 30). Real-time PCRs were done using an ABI Prism 7500 instrument from Applied Biosystems. Each 20 μL reaction was composed of 10 μL SYBR Green PCR mix, 2 μL of each primer, and 6 μL RT product (1:500 dilution). The cycling variables for real-time PCR were as follows: initial heating 95°C for 10 min followed by 40 cycles of 95°C for 15 s and the indicated temperature for 60 s. The following primers were used for amplification of the indicated cDNA: UGT2B15 forward 5-GTGTTGGGAATATTATGACTACAGTAAC-3 and reverse 5-GGGTATGTTAAATAGTTCAGCCAGT-3 (125 nmol/L each, 56°C), UGT2B17 forward 5-TGACTTTTGGTTTCAAGC-3 and reverse 5-TTCCATTTCCTTAGGCAA-3 (300 nmol/L each, 56°C), TMPRSS2 forward 5-GTGATTTCTCATCCAAATTA-3 and reverse 5-TCCAGCAGAGCTGTTCTGGC-3 (300 nmol/L each, 60°C), and 28S forward 5-AAACTCTGGTGGAGGTCCGT-3 and reverse 5-CTTACCAAAGTGGCCCACTA-3 (200 nmol/L each, 56°C or 60°C). The specific amplification of each UGT cDNA was ensured by testing PCR strategies with all human UGT2B cDNAs (1.0 ng) and by direct sequencing of PCR products. C_T values were analyzed using the comparative C_T (ΔΔC_T) method as described by the manufacturer (Applied Biosystems). The amount of target (2^-ΔΔCT) was obtained by normalizing to the endogenous reference 28S and was relativized to the vehicle-treated baseline control. For each gene, the amplification efficiency and the accuracies of the ΔΔC_T for UGTs versus 28S rRNA were tested using a 2 to 5 log of concentrations of cDNA produced from LNCaP cell purified mRNA.

Total protein from LNCaP cells was purified according to the Tri-Reagent acid/phenol protocol. For Western blot experiments, 1 μg human liver microsomes (BD Biosciences Discovery Labware) and 10 μg total protein were size separated by 10% SDS-PAGE and immunoblotted with anti-UGT2B15 or anti-UGT2B17 antibody as reported (26, 27), an anti-AR antibody (1:4,000 dilution), or an anti-actin antibody (1:1,000 dilution). Relative expression levels were quantified using BioImage Visage 110s (Genomic Solution).

The luciferase constructs containing 2,401 bp of the UGT2B15 and UGT2B17 promoters and their deletions constructs (1,236, 876, 436, 326, 235, and 109 bp) were described previously (31, 32). To remove the putative 59-bp VDRE (-113 to -171 bp) insert (2B15-i59) from the UGT2B15 promoter and to introduce it into the UGT2B17 promoter, the pGL3-based 2.4-kb promoter constructs of UGT2B15 and UGT2B17 were digested by NdeI and NcoI. Fragments were separated by gel electrophoresis, and the digestion product from UGT2B15 was cloned into the NdeI/NcoI-digested pGL3-based UGT2B17 promoter construct and vice versa. In addition to the 59-bp insert, three nonidentical bases (at positions -382/-322, -314/-254, and -214/-154 in the promoters of UGT2B15/UGT2B17) were introduced in reporter constructs during the subcloning processes. Three changes, numbered mut1-mut3, were removed afterwards by site-directed mutagenesis using the QuickChange site-directed mutagenesis kit (Stratagene) and the following PCR primers (mutated nucleotides are italicized): UGT2B15 mut1 (-382 bp), 5-GATATTAAAAAATGCCGTTTGAGTTGTAT-3′; UGT2B15 mut2 (-314 bp), 5-AGCCTCTCACTTGCCACTGTTCTTA-3′; UGT2B15 mut3 (-214 bp), 5-GAGTAATTGTAACATAAAAGAACACC-3′; UGT2B17 mut1 (-322 bp), 5-GATATTAAAAAATGGCGTTTGAGTGTAT-3′; UGT2B17 mut2 (-254 bp), 5-AGCCTCTCACCTGCCACTGTTCTTA-3′; and UGT2B17 mut3 (-154 bp), 5-GAGTAATTGTAAATATAAAAGAACACC-3′. Mutations of the putative VDRE in the UGT2B15 -2,401-bp construct were also introduced using the same kit. The forward primers used for mutagenesis were as follows: UGT2B15-VDRE mutation, 5′-CACACTAAAATAAATATGATTTTATCAATCTTTTGTTGGTCTCC-3′. DNA sequencing was done for all constructs to verify the correct mutations and to ensure that no spurious mutations occurred elsewhere.

Lipofectin or LipofectAMINE 2000 (Invitrogen) was used as transfection reagent for LNCaP and 22Rv1 cells or PC-3 and DU145 cells, respectively. The indicated pGL3 reporter (100 ng) was cotransfected with 30 ng pCMV-β-GAL as an internal control. All samples were complemented with pBS-SK+ (Stratagene) to an identical amount of 500 ng/well. At 6 h post-transfection, the transfection medium was removed and cells were incubated in the presence of calcitriol at the indicated concentrations or vehicle (0.1%, v/v) in serum-free medium for 24 h. Subsequently, cells were harvested and cell lysates (20 μL) were assayed for luciferase and β-galactosidase activities.

Nontargeting or AR-specific small interfering RNAs (siRNA) were ON-TARGET plus SMART pool reagents from Dharmacon. LNCaP cells were transfected with siRNA using Lipofectin transfection reagent according to the manufacturer’s instructions. At 24 h post-transfection, the medium was changed to RPMI 1640 supplemented with 0.2% FBS for 24 h. Cells were then treated with calcitriol (10 nmol/L) or R1881 (1 nmol/L) in RPMI 1640 supplemented with 0.2% FBS for another 24 h. The AR protein levels and TMPRSS2 and UGT mRNA levels were analyzed as described above.

All data are presented as mean ± SD. Statistical analysis was done using a two-tailored Student’s t test JMP V4.0.2 software (SAS Institute). In all cases, P ≤ 0.05 was considered statistically significant.

To determine whether calcitriol affects androgen glucuronidation in prostate cancer cells, LNCaP cells were treated with vehicle or calcitriol (10 nmol/L) for 48 or 96 h, and glucuronidation assays were then done with cell homogenates in the presence of 3α-diol, ADT, or DHT (Fig. 1). The formation of all glucuronide conjugates was significantly reduced after 48 h of treatment with calcitriol, and the effect was even more pronounced after 96 h (Fig. 1). The strongest effect was observed for ADT-3G (56% reduction after 96 h) followed by DHT-17G (50% reduction after 96 h). These results show that calcitriol inhibits the glucuronide conjugation of DHT and its 5α-reduced metabolites 3α-diol and ADT in LNCaP cells.

UGT2B15 and UGT2B17 are the two major androgen-glucuronidating enzymes in LNCaP cells (15). To investigate whether calcitriol modulates the expression of these two genes, LNCaP cells were treated with 10 nmol/L calcitriol for 24 or 96 h and mRNA levels of UGT2B15 and UGT2B17 were determined by real-time RT-PCR. Both UGT2B15 and UGT2B17 mRNA levels were significantly reduced after 24 and 96 h of calcitriol treatment (Fig. 2A and B); however, the reduction of UGT2B15 was more pronounced (65% reduction after 96 h) than that of UGT2B17 (50% reduction after 96 h). Subsequent analysis of UGT2B15 and UGT2B17 protein levels in calcitriol-treated LNCaP cells confirmed the real-time RT-PCR results (Fig. 2C and D). Calcitriol treatment resulted in strong inhibition of UGT2B15 mRNA levels and protein expression (88% reduction after 96 h; Fig. 2D) and somewhat weaker inhibition of UGT2B17 mRNA levels and protein expression (78% reduction after 96 h; Fig. 2D).

Additional time-course experiments (6-96 h) revealed that UGT2B15 mRNA levels decreased more rapidly than UGT2B17 (Supplementary Data 1).⁴

The VDR agonist calcitriol has both genomic and nongenomic effects on cancer cells (4). To determine the manner in which VDR affects UGT2B15 and UGT2B17 expression, LNCaP cells were treated with calcitriol (10 nmol/L) in the presence or absence of CHX, an inhibitor of protein synthesis (Fig. 3A and B). Cotreatment with CHX only slightly interfered with the calcitriol-dependent repression of UGT2B15 expression (Fig. 3A) but completely abolished the effect on UGT2B17 (Fig. 3B). This observation indicates that UGT2B17 suppression by calcitriol requires the synthesis of at least one mediator protein, whereas the inhibition of UGT2B15 is independent of de novo protein synthesis.

To confirm these observations, transient transfections of LNCaP cells were carried out using plasmid constructs in which luciferase transcription was driven by the 2.4-kb UGT2B15 or UGT2B17 promoters (Fig. 3C and D). As expected (33), transfection of the UGT2B15 and UGT2B17 promoter plasmids resulted in strong basal luciferase activity. Treatment with increasing concentrations of calcitriol led to a significant dose-dependent decrease of the UGT2B15 promoter activity (Fig. 3C). However, no reduction of the promoter activity was observed for the UGT2B17 construct in the presence of calcitriol (Fig. 3D). Taken together, these observations suggest that the suppression of the two UGT genes by calcitriol involves different mechanisms.

Previous studies reported that a series of VDR target genes are affected by calcitriol only in LNCaP cells cultured in the presence of serum (13). Most of the experiments described above (except transient transfections) were done with cells cultured in the presence of 0.2% FBS. To determine the importance of FBS for UGT regulation in prostate cancer cells, LNCaP cells were treated for 24 h with calcitriol (10 nmol/L) in the absence or presence of 0.2% or 10% FBS. UGT2B15 mRNA levels were significantly decreased by calcitriol in all conditions (Supplementary Data 2).⁴ However, the more serum the medium contained, the more the expression of UGT2B15 was inhibited. Likewise, UGT2B17 expression was more strongly inhibited with 10% FBS than with 0.2% FBS. More interestingly, calcitriol failed to inhibit UGT2B17 expression in the absence of serum (Supplementary Data 2).⁴ Although the presence of serum has a synergistic effect for both UGT2B15 and UGT2B17 calcitriol-dependent inhibition, these results indicate that serum is absolutely required for the calcitriol-mediated repression of UGT2B17 expression.

AR plays an important role in the serum-dependent antiproliferative effects of calcitriol (10), and treatment with AR agonists also significantly reduces UGT mRNA levels in LNCaP cells (29, 34). These observations suggest a role for AR in the regulation of UGT2B17 expression by calcitriol. To evaluate this hypothesis, LNCaP cells grown in 0.2% FBS were treated with calcitriol in the absence or presence of the AR antagonist casodex. As expected (34), casodex abolished the reduction in the levels of UGT2B15 (Fig. 4A) and UGT2B17 (Fig. 4B) mRNA caused by the synthetic AR agonist, R1881. However, the reduction of UGT2B15 mRNA levels by calcitriol was not significantly affected by casodex (Fig. 4A), whereas it abolished the reduction of UGT2B17 mRNA levels by calcitriol (Fig. 4B).

We confirmed that AR is required for the suppressive effect of calcitriol on UGT2B17 expression by specific blockage of endogenous AR signaling using an AR-specific siRNA. The AR siRNA was effective in strongly reducing AR protein levels in LNCaP cells in comparison with the nontargeted control siRNA (Fig. 4C). Furthermore, the stimulation of TMPRSS2 expression, a positively regulated AR target gene (35), by R1881 was reduced by 73% by the AR siRNA (Fig. 4D). These data confirm the ability of AR siRNA to interfere with the androgen signaling pathway. Calcitriol and R1881 were fully capable of reducing UGT2B15 (Fig. 4E) and UGT2B17 (Fig. 4F) mRNA levels in LNCaP cells transfected with the nontargeted control siRNA (29, 34). However, the transfection of AR siRNA abolished the inhibitory effect of R1881 on both UGT isoforms (Fig. 4E and F), whereas the effect of calcitriol was only blocked for UGT2B17 (Fig. 4F). In contrast, the calcitriol-induced inhibition of UGT2B15 expression was maintained even in the presence of AR siRNA (Fig. 4E).

Taken together, these different experimental approaches clearly show that calcitriol essentially requires a functional AR to inhibit the expression of the UGT2B17 gene, whereas it exerts its suppressive effects on the UGT2B15 gene in an AR-independent manner.

Our results using CHX treatment and transient transfections (Fig. 3) indicate that calcitriol directly controls UGT2B15 expression. To localize the potential VDR-responsive region within the UGT2B15 promoter, LNCaP cells were transiently transfected with plasmid constructs in which successive 5′ deletions of the UGT2B15 promoter drive the expression of the luciferase gene. Cells were then treated with calcitriol. Deletion of the promoter to -235 bp sustained promoter activity and the inhibitory response to calcitriol treatment (Fig. 5A; data not shown). However, the deletion of the promoter to -109 bp eliminated basal activation and the subsequent response to calcitriol (Fig. 5A). Thus, a putative VDR-responsive region was identified in the UGT2B15 promoter region between -235 and -109 bp.

Considering the high identity of the nucleotide composition between the proximal UGT2B15 and UGT2B17 gene promoters (∼91% identity for the region between -1,662 and the transcriptional start site; ref. 31), we expected that minor nucleotide changes between the two genes gave rise to a VDRE-like sequence within the UGT2B15 promoter. Alignment of the sequences revealed the presence of a 59-bp insert (-113 to -171) within the UGT2B15 promoter (B15-i59) that was absent in the UGT2B17 promoter (Fig. 5B). Additionally, the distal boundary of this insert (-157 to -177) also contained a highly conserved inverted palindrome-9 motif (AGGTCAtcaatctttTGTTGG), a sequence that has been shown previously to function as a VDRE (36). To test whether the 59-bp insert was responsive to calcitriol, the B15-i59 fragment was cut out of the UGT2B15 -2,401-bp construct and cloned into the UGT2B17 -2,401-bp promoter at the equivalent position. Transfection of the UGT2B15 construct lacking B15-i59 revealed significant reduction of the basal promoter activity (Fig. 5C) to a similar level as the UGT2B15 -235-bp construct (Fig. 5A). However, in contrast to this wild-type promoter, activity of the UGT2B15 construct lacking the 59-bp insert was not affected by the presence of calcitriol (Fig. 5C). Furthermore, insertion of the B15-i59 fragment into the UGT2B17 promoter yielded a stronger basal activation compared with the wild-type construct and a significant response to calcitriol treatment (Fig. 5D), indicating that the B15-i59 insert may be important for the direct negative response of the UGT2B15 promoter to calcitriol.

To further confirm the importance of the putative VDRE distal to B15-i59 in the UGT2B15 promoter, half of the inverted palindrome-9 sequence (-172 to -177) was modified by site-directed mutagenesis to generate the UGT2B15 -2,401 mutant VDRE construct. As expected, the mutation abolished the response of the UGT2B15 promoter to calcitriol treatment and also reduced the basal promoter activity (Fig. 5C). Taken together, these results indicate that the UGT2B15 promoter region between -113 and -171 contains a VDR-responsive element that mediates the effects of calcitriol on the expression of UGT2B15.

The androgen-dependent LNCaP cell line is a widely used classic model for prostate cancer (18), but one knows that this type of cancer is a heterogeneous and multistage disease. Thus, to determine whether calcitriol affects androgen-conjugating UGT enzymes in other cell models, we have compared its effects on UGT2B15 and UGT2B17 gene expression in (a) 22Rv1 cells: an AR-positive, androgen-independent cell line that responds to androgen treatment (14, 37), and (b) PC-3 and DU145 cells: two AR-negative, androgen-independent cell lines (37). All these cell lines have been reported to express VDR (10, 24, 38).

First of all, these experiments have revealed that UGT2B15 and UGT2B17 transcripts are absent from PC-3 or DU145 cells (data not shown), indicating that these enzymes are not expressed in these AR-negative prostate cancer cells. By contrast, 22Rv1 cells exhibited elevated levels of UGT2B15 and UGT2B17 mRNA, and treatment with calcitriol resulted in a significant reduction of these levels (Fig. 6). Interestingly, and similarly to what we observed with LNCaP cells (Supplementary Data 1),⁴ the reduction of UGT2B17 expression was delayed when compared with the response of the UGT2B15 gene. Indeed, the reduction of UGT2B15 mRNA concentration was significant after 24 h of treatment (Fig. 6A), whereas UGT2B17 mRNA levels were decreased only when cells were incubated with calcitriol for 72 h (Fig. 6B). In addition, the effect of calcitriol on the UGT promoters was also investigated in 22Rv1 through transient transfection experiments. The two 2.4-kb UGT promoter constructs displayed strong basal activity (Fig. 6C and D), and as observed in LNCaP cells (Fig. 3C), calcitriol treatment strongly reduced UGT2B15 promoter activity (by >50%; Fig. 6C), whereas UGT2B17 was affected only at lower levels (reduction of 20% and 28% in the presence of 10 and 100 nmol/L calcitriol, respectively; Fig. 6D). In the PC-3 and DU145 cells, no basal activity was observed with these constructs, which is consistent with the absence of UGT2B15 and UGT2B17 gene expression (data not shown).

Overall these observations indicate that calcitriol is able to reduce androgen glucuronidation in prostate cancer cells where UGT2B15 and UGT2B17 enzymes are expressed.

In this study, human UGT2B15 and UGT2B17 were identified as negatively regulated target genes of the endogenous vitamin D metabolite, calcitriol, in prostate cancer LNCaP and 22Rv1 cells. We have also established that this negative regulation causes a drastic reduction of androgen glucuronidation. These observations are of clinical importance because vitamin D derivatives are considered to be promising therapeutic compounds for prostate cancer treatment (6), whereas androgens play a major role in the pathogenesis of this disease (16).

The observation that calcitriol drastically reduces androgen glucuronidation in UGT-expressing prostate cancer cells suggests that treatment with this molecule may have profound consequences for androgen homeostasis and activity in androgen-sensitive prostate cancer cells. Indeed, reduction of glucuronidation in LNCaP cells causes an accumulation of free androgens and higher cell proliferation rates (28, 39). Thus it is reasonable to speculate that, by reducing UGT expression, calcitriol will also result in DHT accumulation. In addition, DHT is a potent inhibitor of UGT2B15 and UGT2B17 expression and of its own glucuronidation (26, 28, 29). Therefore, the reduction of DHT glucuronidation caused by calcitriol may be further amplified through this negative autoregulatory loop. However, calcitriol also exerts numerous androgen-independent antiproliferative effects in prostate cancer cells and has been shown to be efficient for dampening prostate cancer cell proliferation (4, 7, 40, 41). Nevertheless, the reduction of UGT gene expression and activity may be interpreted as a plausible limitation of the antineoplastic properties of calcitriol in androgen-sensitive prostate cancer cells. These observations also suggest that other vitamin D derivatives having androgen-independent antiproliferative effects that do not modulate UGT enzymes may display stronger antiproliferative effects in LNCaP cells than calcitriol. Interestingly, a recent study revealed that vitamin D_5, a 1α-hydroxylated metabolite of vitamin D, showed significantly different molecular effects in bone and prostate cells (40, 42, 43). In particular, this compound is a stronger reducer of LNCaP cell proliferation than calcitriol, and it fails to activate AR in this cell line (43). These data suggest that vitamin D₅ may not affect UGT2B17 expression, and it will therefore be of importance to determine the effect of such novel vitamin D derivatives on androgen glucuronidation in prostate cancer cells.

When the effects of calcitriol treatment on UGT expression was investigated in other prostate cancer cell lines, our primary observation was that UGT2B15 and UGT2B17 enzymes are not expressed in the AR-negative PC-3 and DU145 cells. While this observation impairs to show the AR requirement for the calcitriol-induced UGT2B17 gene repression in recurrent prostate cancers, it is consistent with the previously reported absence of androgen glucuronidation in these cells (44). Interestingly, a similar pattern of response to calcitriol treatment was observed in LNCaP and 22Rv1 cells, thus indicating that the inhibitory effect of calcitriol on androgen glucuronidation is not limited to LNCaP cells but may correspond to a more general phenomenon in UGT-positive prostate cancer cells.

In addition to androgen inactivation, UGT2B15 and/or UGT2B17 also exerts important roles for drug metabolism (45). The nonsteroidal anti-inflammatory drug ibuprofen is metabolized by both UGT isoforms (46), and the antiestrogen tamoxifen is potently glucuronidated by UGT2B15 (47). Interestingly, ibuprofen cooperates with calcitriol in an additive manner to inhibit LNCaP cell growth (6). Likewise, tamoxifen cooperates to inhibit mammary carcinoma MCF7 cell growth (6). It is therefore likely that calcitriol, by reducing UGT expression, also affects the metabolism of these drugs and thereby reinforces their beneficial properties. Taken together, these observations suggest that the calcitriol-dependent inhibition of UGT2B15 and UGT2B17 gene expression could be a therapeutic advantage when calcitriol is used concomitantly with other drugs conjugated by one or both enzymes.

Using several experimental approaches, we have shown that the calcitriol-dependent reduction of UGT2B15 and UGT2B17 gene expression occurs at the transcriptional level. However, the two genes are differentially affected in the presence of the VDR activator: calcitriol down-regulates UGT2B17 expression in an AR-dependent manner, whereas it interferes more directly with the UGT2B15 gene through a VDR-responsive promoter region. The androgen-dependent manner by which UGT2B17 is regulated is consistent with the previous observations that VDR up-regulates AR (48) and that AR reduces UGT2B17 expression (29, 32). In addition to VDR and AR, this regulatory cascade may also involve other unidentified factors because both the regulation of AR by VDR (48) and UGT2B17 by AR (34) have been shown to be indirect. The implication of such a multifactor regulatory process is further supported by the delay observed for UGT2B17 down-regulation when compared with UGT2B15. Furthermore, the AR-dependent manner in which calcitriol regulates UGT2B17 expression is also consistent with previous observations that many regulatory and antiproliferative effects of vitamin D compounds in LNCaP cells are mediated through AR (13).

Contrariwise, the modulation of UGT2B15 appears to occur in a simple manner through a VDR-responsive region in the UGT2B15 promoter that is directly modulated in the presence of calcitriol. This region contains a highly conserved inverted palindrome-9 sequence, reported previously to function as a negative element for regulation of transcription in the presence of the ligand-activated VDR (36). Site-directed mutagenesis of this sequence abolishes the calcitriol-dependent reduction of the UGT2B15 promoter activity. However, electrophoretic mobility shift assays and chromatin immunoprecipitation experiments failed to show the formation of such a complex (data not shown). Together, these data suggest that activation of VDR may initiate the formation of a protein-protein complex comprising VDR and an unidentified positive regulator of UGT2B15. In such case, VDR sequesters the positive factor(s), which is no more available to bind to and activate the VDR-responsive region within the UGT2B15 promoter. Similar direct protein-protein interactions between VDR and other transcription factors have been reported in case of VDR negatively regulated genes, such as human parathyroid hormone or the osteoblast endopeptidase PHEX (49, 50).

In conclusion, the present study shows that, through differential regulation of UGT2B15 and UGT2B17 expression, calcitriol treatment results in reduced inactivation of androgens in LNCaP cells, an effect that potentially limits its antineoplastic properties in androgen-sensitive prostate cancer cells, such as LNCaP and 22Rv1 cells.

We thank Jean-Sebastien Carrier for providing the previously reported UGT2B15 and UGT2B17 promoter constructs, Jocelyn Trottier for helpful discussion, and Dr. Virginie Bocher for critical reading of this work.