Direct Progesterone Receptor and Indirect Androgen Receptor Interactions with the Kallikrein-Related Peptidase 4 Gene Promoter in Breast and Prostate Cancer

Breast cancer and prostate cancer are significant diseases in Western countries, accounting for ∼30% of all newly diagnosed cancers for 2006 (1). Although the etiology of both diseases is not fully understood, there are some similarities in the epidemiology and pathogenesis between breast and prostate cancers. For example, the incidence of breast and prostate cancers are reported to be higher in Caucasians and African Americans when compared with women and men with Asian ancestry (2, 3). However, the most significant association is that in a majority of cases, both breast and prostate cancers are regulated by hormones in the initial stages of disease and current therapeutic options involve targeting the estrogen receptor (ER) and androgen receptor (AR) signaling axes, respectively (4, 5). Consequently, much of the research into breast and prostate cancers has focused on genes that are involved in the estrogen and androgen signaling pathways. However, despite extensive research efforts, the precise molecular events leading to the initiation and progression of these diseases are still largely unknown.

Steroid hormones, such as estrogen, progesterone, and androgens, mediate their cellular effects through their cognate receptors. Upon ligand binding, steroid receptors participate in a sequence of events that ultimately result in the translocation of the receptor into the cell nucleus whereupon it binds to hormone response elements (HRE) in the promoters of target genes to initiate transcription. One such family of hormone target genes that are regulated by steroid hormones in prostatic, breast, and endometrial cells, and which are thought to play a role in the progression to neoplasia, are the tissue kallikrein-related (KLK) peptidases (6, 7). The KLKs are a multigene family of serine proteases that are involved in the posttranslational processing of polypeptide precursors to their biologically active forms (6, 7), a function that is central to a number of biological events. Additionally, KLKs can activate proteases such as urokinase-type plasminogen activator and cleave structural proteins of the extracellular matrix (6, 7), which are critical events in tumorigenesis, invasion, and metastasis.

KLK4 is a relatively recently described KLK family member (8-11). Although the porcine and mouse orthologue of KLK4, enamel matrix serine protease-1 (EMSP-1), plays a role in tooth development by degrading amelogenin, which is the predominant protein in the enamel matrix of developing teeth (12), the physiologic role of KLK4 remains to be established. By analogy, KLK4 may have similar extracellular matrix remodeling properties and potentially facilitate in local invasion and/or metastatic progression of carcinomas. In support of this notion, KLK4 can activate pro-PSA and single-chain urokinase-type plasminogen activator (13) and recombinant KLK4 also degrades prostatic acid phosphatase (13) and members of the insulin-like growth factor binding protein family (14). Collectively, this suggests that KLK4 may have an important role in prostate tumorigenesis.

At the mRNA level, KLK4 has been shown to be up-regulated by androgens, estradiol, progesterone, and glucocorticoids in prostate cancer cells (10, 15) and by the synthetic progestin (norgestrel) in BT-474 breast cancer cells (16). Further, KLK4 protein levels are increased by as much as 4-fold in response to estradiol in ovarian cancer cells (17) and by 40-fold in response to estradiol and progesterone in endometrial cancer cells (18). However, functional HREs have yet to be characterized for the KLK4 promoter, although putative AREs have been predicted using in silico modeling (11). In fact, functional HREs mediating hormone-induced gene expression of the KLKs have only been identified in the promoter and enhancers for the KLK2 and KLK3 (prostate-specific antigen, PSA) genes (19-27).

To date, the promoter of the KLK4 gene has not been well defined with two groups mapping alternative transcription initiation sites (TIS) using 5′-random amplification of complementary ends (5′-RACE; refs. 9, 15). Initially, the full-length KLK4 transcript was reported to be derived from the classic five exons (10). However, subsequent studies have identified a variant transcript that begins in exon 2 (15).

The studies reported here were designed to further characterize the hormonal gene regulation of KLK4 in breast cancer (T47D) and prostate (LNCaP and 22RV1) cancer cells. We confirm that the predominantly used TIS for KLK4 gene expression in breast and prostate cancer cells is similar to that previously reported in exon 2 (15). We have shown that the KLK4 promoter region encompassing 2.8 kb from this TIS is not responsible for the observed androgen regulation of the KLK4 gene expression in LNCaP and 22Rv1 prostate cancer cells. However, our analysis suggests that a progesterone-responsive region in this 2.8 kb of the KLK4 promoter regulates progesterone-induced KLK4 gene expression in the T47D breast cancer cells.

The association of KLK4 gene expression with progesterone receptor (PR) and AR status was assessed for four breast and seven prostate cell lines using reverse transcription-PCR (RT-PCR). Figure 1A shows that KLK4 is expressed in one breast (T47D) and three prostate (LNCaP, 22Rv1, and MDA-PCa-2b) cell lines. Importantly, KLK4 was found to be only expressed in PR-positive breast and AR-positive prostate cells (Fig. 1A). Consequently, the progesterone and androgen regulation of KLK4 was further investigated using quantitative RT-PCR in T47D and LNCaP cells, respectively.

Treatment of T47D cells with 10 nmol/L of progesterone over 24 hours resulted in a 1.38-fold (P < 0.01) increase in KLK4 mRNA and a 1.65-fold (P < 0.01) response for HSD11B2, a prototypical progesterone-regulated gene (Fig. 1B). Further, the attenuation of KLK4 and HSD11B2 response by the PR antagonist RU486 shows that regulation of KLK4 and HSD11B2 by progesterone was mediated through the PR (Fig. 1B).

In LNCaP cells, KLK4 expression was up-regulated by 3.2-fold (P = 0.03) after the cells were treated for 24 hours with 1 nmol/L of the synthetic androgen, R1881, when compared with the vehicle control (Fig. 1C). Androgen-induced up-regulation of KLK4 expression was attenuated when cells were treated with 1 μmol/L bicalutamide, an AR antagonist, which suggests that the observed regulation is mediated through the AR (Fig. 1B). The KLK4 response to androgen ± antagonist was similar to the PSA/KLK3 gene (positive control for androgen regulation), although PSA had a much higher response to androgen treatment (∼8-fold, P < 0.05; Fig. 1C).

To further validate the hormonal regulation of KLK4 gene expression, quantitative RT-PCR (RT-qPCR) was also carried out on T47D and LNCaP cells that were cultured in hormone-depleted medium. After maintaining both T47D and LNCaP cells in medium containing 10% charcoal-stripped FCS for 4 days, KLK4 expression was found to be down-regulated by 4.2-fold (P = 0.01) and 2.4-fold (P < 0.01), respectively, compared with cells grown in regular FCS (Fig. 1D and E).

To identify potential HREs in the KLK4 promoter, it was essential to first map the location of the KLK4 TIS for breast cancer (T47D) and prostate cancer (LNCaP) cells. Using 5′-RLM-RACE, a KLK4 TIS in LNCaP cells (TIS2b) was found downstream of the originally reported TIS1 (9) and 43 bp upstream of the KLK4 exon 2 ATG. In T47D cells, the TIS (TIS2a) was 35 bp longer and situated 78 bp upstream of the exon 2 ATG (Fig. 2A-C). Transcripts generated from these TISs would be derived from just four exons and thus are shorter than the published full-length transcript (Fig. 2D) but similar to that reported in LNCaP cells (15). The relative expression of transcripts arising from each TIS, as well as the hormonal regulation of these variants in T47D and LNCaP cells, was assessed using RT-qPCR. Figure 2E shows that transcripts using TIS1 have markedly lower expression when compared with transcripts that use TIS2a or TIS2b for T47D or LNCaP cells. Yet, despite their differing abundance, similar fold increases of both variants were observed with androgen and progesterone treatment.

Given our observations that KLK4 gene expression is regulated by progesterone and androgens in breast and prostate cancer cells, we carried out in silico analyses to identify HREs within a region encompassing ∼3 kb of the KLK4 promoter from TIS2b that may mediate these responses. Using four different gene analysis programs, we identified two putative HRES (K4.pPRE and K4.pARE) that have consensus-like half sites, which could interact directly with class I steroid hormone receptors (androgen, progesterone, mineralocorticoid, and glucocorticoid receptors; Fig. 3A). The K4.pPRE sequence (5′-AGAACAtgagagAGAACA-3′) consists of a direct repeat motif that is separated by six nucleotides and is located 2,419 bp upstream of the TIS2b identified here in LNCaP cells and 397 bp upstream from the previously reported TIS1. The K4.pARE sequence (5′-GGTGCAggaGATTGT-3′), located at −1005 bp from TIS2b (1017 bp downstream from TIS1), more closely resembles the classic HRE motif that is traditionally characterized by a palindromic hexameric repeat sequence that is separated by three nucleotides.

Electrophoretic mobility shift assays (EMSA) were carried out with the K4.pARE to determine if the element bound to endogenous AR or recombinant AR-DBD. Incubation of K4.pARE with LNCaP nuclear extracts from cells that were treated with R1881 for 12 hours resulted in the formation of two specific shifts (Fig. 3B). These shifts were also observed in experiments where cells were incubated with PSA AREI. A decrease in shift intensity, particularly for the lower shift, in experiments where cells were incubated with 200- and 400-fold excess unlabeled K4.pARE suggests that these complexes are specific (Fig. 3B). To determine whether these shifts may represent AR/ARE complexes, blocked shift analysis using 1 and 2 μg of AR antibody was carried out. Incubation of both the PSA ARE I and the K4.pARE, with increasing amounts of AR antibody, decreased the shift intensity for both shifts (Fig. 3B). The blocking of these shifts by the AR antibody suggests these complexes may be formed by interaction of the AR with K4.pARE. The specificity of K4.pARE was further shown by abrogating the binding of the lower receptor complex with the AR antibody but not by the nonspecific rabbit IgG antibody in experiments in which LNCaP cells were treated for 24 hours with 1 nmol/L R1881 (Fig. 3C). Both shifts for K4.pARE were identical to the PSA ARE I control and further supports the hypothesis that K4.pARE is forming similar DNA/receptor complexes to those observed for PSA ARE I. The differences in shift intensity and blocking patterns between the two blocking shift experiments likely represent the dynamic receptor/DNA complexes formed from the different treatment times and the consequent accessibility of the AR antibody to the AR. These blocked shifts were also observed using the N-20 AR antibody (Supplementary Fig. S1). However, EMSA analysis of K4.pARE with purified AR-DBD did not result in the formation of any high molecular complexes that are indicative of receptor/DNA complexes, although there was a clear PSA ARE I/AR-DBD complex formed (Fig. 3D). Further, competition of the PSA ARE I reactions with increasing amounts (5- to 40-fold molar excess) of unlabeled K4.pARE did not result in any observed abrogation of PSA ARE I binding with the AR-DBD.

Reporter assays for androgen responsiveness of K4.pARE were carried out in both LNCaP and 22Rv1 cells as the blocked shift assays suggest that K4.pARE may be interacting with the AR in some way, although not directly with the AR-DBD. Three tandem repeats of the putative K4.pARE and its flanking sequences were cloned into the pGL3-Promoter vector (K4.pAREX3-Luc) to enhance the hormonal responsiveness for this element (Fig. 4A). In this assay, the pGL3-Promoter was not significantly regulated by R1881 in either LNCaP (P = 0.074) or 22Rv1 (P = 0.12) cells. For PSA AREIX3-A-Luc, treatment with R1881 resulted in a 3.9- and 4.1-fold up-regulation of promoter activity over the vehicle control in 22Rv1 (P = 0.011) and LNCaP cells (P = 0.037), respectively, confirming that the cells were appropriately sensitive to androgens (Fig. 4B). For K4.pAREX3-Luc, there was no difference in promoter activity in LNCaP or 22Rv1 cells (Fig. 4B).

Seven deletion luciferase promoter constructs that encompass up to 2.8 kb of the KLK4 promoter from TIS2b were generated (Fig. 4C) to identify if other hormone-responsive regions could be identified and to determine whether K4.pARE needs to act synergistically with other cis-elements to mediate the KLK4 gene expression seen at the mRNA level. Androgen regulation of the KLK4 promoter was assessed in both LNCaP and 22Rv1 cells. R1881 (1 nmol/L) stimulated promoter activity for pGL3-Basic was 0.85- and 0.75-fold increased over vehicle control (0.1% ethanol) in 22Rv1 and LNCaP cells, respectively (Fig. 4D). Promoter activity for PSA 5.8-A-Luc was 9.7-fold increased in 22Rv1 cells and 12.2-fold increased in LNCaP cells after treatment with R1881. However, none of the seven KLK4 promoter constructs transfected in either LNCaP or 22Rv1 cells (transfected for 6 hours and treated for 24 hours) were significantly regulated (P < 0.05) by R1881 when compared with cells that were treated with vehicle control.

As there was no androgen regulation of the KLK4 promoter in prostate cancer cells, the focus of the hormonal studies was shifted to the progesterone responsiveness observed in T47D cells. Initial analysis was carried out using the K4.898, K4.2000, and K4.2875 promoter constructs (Fig. 5A). Significant basal promoter activity was observed for all three constructs (K4.2875, ∼4-fold, P < 0.05; K4.2000, ∼8-fold, P < 0.05; K4.898, ∼3-fold, P < 0.05), when compared with the insert-less pGL3-Basic control (Fig. 5B). There was no significant increase in promoter activity for the K4.2000 and K4.898 constructs after cells were treated with 10 nmol/L progesterone for 24 hours. However, a modest but significant (∼1.2-fold, P < 0.05) increase in K4.2875 promoter activity, which was comparable with that seen at the mRNA level, was observed after cells were treated with 10 nmol/L progesterone (Fig. 5B). Further, a trend of increased transcriptional activity of K4.2875 was observed in T47D cells that were treated with increasing amounts of progesterone (Fig. 5C).

As the progesterone-induced luciferase data support the in silico analysis, suggesting that the K4.pPRE may be functional, EMSA analysis was then carried out to assess the direct interaction of K4.pPRE with the PR. Two shifts were observed with a time-specific shift of the upper band seen when the K4.pPRE was incubated with nuclear extracts from T47D cells that were treated with 10 nmol/L progesterone over 24 hours (Fig. 6A). This shift is likely to represent progesterone response element (PRE)/PR complexes as no shifts were observed for experiments that used nuclear proteins from the PR-negative breast cancer cell line MDA-MB-231 (Fig. 6B). Further, addition of PR but not AR antibodies in blocked shift experiments were able to abrogate binding of the higher molecular DNA/protein complexes, although no supershift was observed (Fig. 6C).

Chromatin immunoprecipitation (ChIP) assays were carried out to assess the active recruitment of the PR to the KLK4 PRE promoter region in a more cellular context. Thus, T47D cell lines were grown in the absence of progesterone for 72 hours followed by treatment with or without saturating levels (100 nmol/L) of progesterone for 0 to 24 hours. The PR antibody effectively immunoprecipitated the progesterone-responsive region (−2496 to −2283) of the KLK4 promoter in a manner that was dependent on progesterone treatment and time (Fig. 7, I). In contrast, the PR antibody failed to immunoprecipitate a region of the KLK4 promoter that did not contain K4.pPRE (−2012 to −1839) or the chromatin DNA from a non–progesterone-responsive gene, β2-microglobulin (Fig. 7, II and V, respectively). The respective input DNA controls (Fig. 7, II, IV, and VI) amplified the appropriate product, indicating the specificity of the PCR. As expected, the genomic DNA PCR controls were positive for all panels, indicating the specificity of the primer pairs and validating the PCR.

It has long been established that hormones play an important role in the pathogenesis of breast and prostate cancer. Consequently, the identification of downstream target genes with functions that may contribute to neoplasia is an important area of research to better understand the etiology of these two cancers. One such family of hormone-target genes are the tissue kallikrein-related proteases (KLK), which are a multigene family of serine proteases that are highly expressed in many hormone-dependent cancers (6, 7, 28). The kallikrein 4 gene (KLK4), a more recently cloned member of the KLK gene family, has been reported to be regulated by progestins and androgens in breast and prostate cancer cells, respectively (10, 15, 16). Although the (patho)physiologic role of KLK4 is yet to be established, it has been proposed that this gene may be important in the pathogenesis and progression of hormone-dependent cancers as it is aberrantly expressed in cancer cells when compared with benign cells (17, 29-31), and has known or putative functions that are important in cancer biology (12-14, 32-34). In this study, we have further characterized the hormonal regulation of the KLK4 gene in breast and prostate cancer by interrogating the KLK4 promoter to identify putative HREs that may mediate this response.

Initial profiling on a panel of breast and prostate cells showed that KLK4 is only present in breast cells expressing PR and prostate cells expressing AR, suggesting that progestins and androgens play an active role in mediating KLK4 gene expression. Consequently, we further assessed the hormonal regulation of KLK4 gene expression using quantitative RT-PCR in breast and prostate cancer cells, particularly as previous studies have only taken semiquantitative or qualitative approaches (10, 15, 16). The results from these experiments show that KLK4 is moderately up-regulated in T47D cells that were treated with progesterone for 24 hours (∼1.38-fold). In LNCaP cells, KLK4 was also moderately up-regulated (3.2-fold) in response to androgens. The observed up-regulation of KLK4 by progesterone and androgens are likely to be mediated through their cognate receptors as these responses were attenuated in cells that were treated with hormones and their associated receptor antagonists. Further, these findings are consistent with previous studies that report similar up-regulation of KLK4 mRNA in response to progesterone in breast cancer cells (8) and androgens in prostate cancer cells (10, 15). Conversely, the down-regulation of KLK4 gene expression (4.2-fold for T47D cells and 2.4-fold for LNCaP cells) after cells were maintained in hormone-depleted medium further highlights the importance of hormones in mediating KLK4 expression, despite the apparent modest level of these responses.

To date, the KLK4 TIS, and consequently the promoter region of KLK4, have been ill defined. Early gene characterization studies predicted the KLK4 gene to comprise of up to five exons based on sequencing of EST libraries and/or comparative exon prediction analyses (8, 10, 11). More recently, two groups identified two alternative KLK4 TISs that would result in the expression of one transcript (TIS1 variant) that would be derived from six exons (9) and an alternative truncated transcript (TIS2a/b variant) that would be derived from four exons (15). The biological importance of each variant has been a source of recent controversy (35, 36). Nevertheless, our combined 5′-RACE and RT-qPCR results indicate that whereas the transcript arising from TIS2a/b is more abundant, both transcripts are expressed in T47D and LNCaP cells. Further, both TIS1 and TIS2a/b variants are hormonally regulated in LNCaP and T47D cells to a similar degree. Although there are no consensus TATA box within −25 to −30 bp upstream of TIS2a/b, the promoter region for TIS2a/b does have a cluster of GC box sites (Sp1 sites, CCCgCCC), which may play a role in mediating basal KLK4 gene expression (data not shown). Indeed, a large-scale study of 1,031 genes found that only 32% of promoters harbored a consensus TATA box at −25 to −40 bp from the TIS (37). Further, 97% of these 1,031 genes contained a GC box (−74 to −45 bp from the TIS) and suggests that GC boxes play an important role in mediating transcription of TATA-less genes. Indeed, this is supported in the T47D luciferase reporter assays that showed higher basal activity of the minimal KLK4 promoter construct (K4.898) when compared with the promoter-less pGL3-Basic vector.

Classically, the PR, AR, and glucocorticoid receptor recognizes HREs that comprise of a palindromic hexameric repeat sequence that is separated by three nucleotides to mediate target-gene transactivation (38). However, more recent studies suggest that the second zinc finger of the AR-DBD plays an important role in the ability of the AR to interact specifically with HREs that comprise of a direct hexameric repeat motif as opposed to the classic palindromic sequence (39-43). Significantly, reduced fertility and development of male reproductive organs are some of the outcomes for mice knock-in models that have a mutated AR-DBD, which results in loss of AR transactivation of AR-selective HREs (44). This highlights the importance of both the primary structure of the PR, AR, and glucocorticoid receptor, and the nature of the HREs within target genes in not only mediating gene transcription but also in disease pathology such as cancer. It was thus important to identify HREs within the KLK4 promoter, which may mediate progesterone and androgen up-regulation of KLK4, using either receptor-selective or nonselective motifs. Using in silico modeling, we have identified two consensus-like motifs, one (K4.pARE) located in the promoter region of TIS2a/b (∼1 kb) and another (K4.pPRE) located a further 1.5 kb upstream. EMSA analyses of K4.pARE resulted in the formation of two specific shifts that are likely to represent the formation of K4.pARE/AR complexes (as well as complexes incorporating other AR coregulators, which may account for the multiple shifts) as supported by the shifts being of the same molecular weight as shifts formed in reactions that were probed with PSA ARE I, and that abrogation of receptor/DNA complexes were observed in experiments that were incubated with AR antibodies. The specificity of K4.pARE and PSA ARE I shifts were further confirmed by the inability of the nonspecific IgG antibody to block these DNA receptor complexes. The differences in AR blocking patterns between experiments using LNCaP nuclear proteins from different R1881 treatment times also supports the hypothesis (45) that higher shifts represent higher-order AR/DNA complexes. However, binding assays of the putative K4.pARE with purified AR-DBD did not result in the formation of any shifts. Further, molar excess of K4.pARE was not able to outcompete binding of PSA AREI/AR-DBD complexes. It is possible that binding of the AR to K4.pARE may be ligand dependent. Indeed, the AR is ligand bound and dimerized when bound to the promoters of target genes (46). Moreover, formation of K4.pARE/AR complexes may require the association of other AR coregulators to permit the formation of energetically stable complexes. Finally, it is possible that K4.pARE is interacting indirectly with the AR through other AR coregulators.

To address these issues, luciferase reporter assays were carried out to assess if this element is able to mediate androgen-induced transactivation in prostate cancer cells. However, luciferase reporter assays using three tandem copies of K4.pARE (to enhance any potential signal) showed that this element was not responsive to androgens in either LNCaP or 22Rv1 prostate cancer cells. Luciferase reporter assays were then carried out on seven deletion promoter constructs spanning up to 2.8 kb of the KLK4 promoter from TIS2b (853 bp from TIS1) to assess if androgen regulation of K4.pARE requires interaction with other cis-elements and to identify other potential HREs using a functional approach. However, androgen regulation experiments carried out on all KLK4 promoter constructs in both LNCaP and 22Rv1 cells showed that the 2.8 kb of the KLK4 promoter region is not responsive to androgens. Given our observations in the blocked shift experiments using AR antibodies for K4.pARE, these promoter assays support our hypothesis that this element is interacting indirectly with the AR. According to the androgen receptor mutation database,¹

Given the lack of androgen responsiveness of the KLK4 promoter, further in vitro analyses were focused on the progesterone regulation of K4.pPRE and the KLK4 promoter in breast cancer cells. The EMSA analysis suggests that K4.pPRE is interacting with the PR as shown by the formation of a higher molecular weight shift in experiments that used nuclear proteins from T47D cells treated with progesterone. Further, binding of K4.pPRE/receptor complex was abrogated upon addition of PR antibodies in blocked shift experiments, supporting our hypothesis that these shifts represent PRE/PR complexes. To further characterize the K4.pPRE element and flanking sequences, progesterone-induced luciferase reporter assays were then performed in T47D cells using three promoter constructs, two (K4.2000 and K4.898) that did not harbor K4.pPRE and one (K4.2875) that did. Consistent with the location of the K4.pPRE, only the K4.2875 luciferase construct was significantly up-regulated (∼1.4-fold, P = 0.002) by progesterone. Although the K4.2875 construct showed moderate progesterone response, this was consistent with the marginal response observed at the mRNA level using quantitative RT-qPCR, and similar to other progesterone-responsive promoters in T47D cells, such as c-myc (∼3-fold), an important regulator of breast cell proliferation, differentiation, transformation, and apoptosis (49). Other HREs, such as those found in the PSA promoter (ARE I), have also been shown to have minimal sensitivity to androgens on its own (24, 25). Like the cooperative regulation of PSA AREs in androgen-mediated gene expression (19, 20), the KLK4 PRE may act synergistically with other enhancer or intronic PREs. The ChIP data support our hypothesis that this element is a bona fide HRE as the PR was actively recruited to the KLK4 promoter region, which harbors the PRE element. Further, recruitment of the PR to the KLK4 PRE region seems to occur in a time-dependent manner that is consistent with other nuclear hormone receptor recruitment studies. For example, studies on the estrogen receptor α (ERα) and the cathepsin D promoter have suggested that the cyclic recruitment of ER to target promoters may represent a mechanism that facilitates continuous monitoring of the external environment (50).

In conclusion, we have quantified and confirmed previous reports that KLK4 is exquisitely regulated by progesterone and androgens in breast and prostate cancer cells, respectively. We show that both KLK4 variants are similarly hormone regulated despite differences in their abundance. We have also shown that the androgen-responsive region of KLK4 is unlikely to lie within the ∼2.8 kb promoter region from TIS2a/b investigated in this study, although the progesterone regulation of KLK4 gene expression in breast cancer cells is likely mediated, at least in part, by the PRE identified at −2419 bp in the KLK4 promoter. These studies lay the foundation for future analysis of the hormonal regulation of the KLK4 promoter.

Progesterone was obtained from Sigma Chemical Co., and the synthetic androgen R1881 was from Perkin-Elmer. The PR antagonist (RU486) was obtained from Sigma, and the AR antagonist (bicalutamide) was a gift from Prof. Wayne Tilley (Hanson Institute, Adelaide, Australia). The PR (PR-C19) and AR (C-19) antibodies used in EMSA and ChIP experiments were obtained from Santa Cruz Biotechnology. The rabbit IgG antibody was obtained from Zymed Laboratories.

All cell lines used were obtained from American Type Culture Collection. Cells were maintained in either RPMI 1640 (LNCaP, 22Rv1, DU145, and PC-3) or phenol-red free DMEM/F12 (MCF7, MDA-MB-231, and T47D) medium (Invitrogen), supplemented with 10% FCS (Invitrogen) and 50 units/mL penicillin G and 50 μg/mL streptomycin (CSL Biosciences). RWPE-1 and RWPE-2 cells were grown in keratinocyte serum-free medium with 50 μg/mL bovine pituitary extract and 5 ng/mL recombinant human epidermal growth factor (Invitrogen). MDA-PCa-2B cells were maintained in BRFF-HPC1 medium from AthenaES (Sapphire Biosciences). MCF-10A cells were maintained in DMEM/F12 medium (Invitrogen) supplemented with 5% horse serum (Invitrogen), 10 μg/mL insulin (Sigma), 20 ng/mL epidermal growth factor (Invitrogen), 100 ng/mL cholera enterotoxin (Sigma), 0.5 μg/mL hydrocortisone (Sigma), and 100 units/mL penicillin/streptomycin. For steroid treatment experiments, cells were cultured until ∼70% confluent and then maintained in 2% charcoal-stripped serum (HyClone) for 48 h followed by addition of steroid hormones (1 nmol/L R1881, 10 nmol/L progesterone) for an additional 24 h. In antagonist experiments, cells were pretreated with either 1 μmol/L bicalutamide or 1 μmol/L RU486 for 2 h before the addition of hormones and antagonists. For hormone starvation experiments, cells were maintained in phenol red–free medium containing 10% charcoal-stripped serum for 4 d before harvesting and RNA extraction.

Total RNA was extracted using the TRI Reagent (Sigma) according to the manufacturer's protocol. RNA was then treated with DNaseI (Roche Diagnostics), purified through an RNeasy column (Qiagen) and electrophoresed to determine the integrity of the RNA before use in 5′-RACE experiments. Complementary DNA (cDNA) was synthesized from 2 μg of total RNA using random hexamers (Proligo) and SuperScript III Reverse Transcriptase (Invitrogen). RT-PCR was carried out on a panel of four breast cell lines (T47D, MCF7, MCF-10A, MDA-MB-231) and seven prostate cell lines (LNCaP, 22Rv1, MDA-PCa-2b, RWPE-1, RWPE-2, DU145, PC-3) using primers targeting the KLK4 (K4Ex2QS: 5′-ggcactggtcatggaaaacga-3′ and K4Ex3QAS: 5′-tcaagactgtgcaggcccagcc-3′), PR (F: 5′-gattcagaagccagccagag-3′, R: 5′-tgcctctcgcctagttgatt-3′), AR (F: 5′-atcaggggcgaagtagagcatc-3′, R: 5′-agccccactgaggggacaacc-3′; ref. 51), and β2-microglobulin (β2M-F: 5′-tgaattgctatgtgtctgggt-3′, β2M-R: 5′-cctccatgatgctgcttacat-3′) genes. RT-PCR was performed using PCR annealing temperatures of 60°C for KLK4, AR and PR, and 55°C for β2M.

KLK4, 18S rRNA, prostate-specific antigen (PSA), and 11β-hydroxysteroid dehydrogenase type 2 enzyme (HSD11B2) transcripts were analyzed by RT-qPCR in 20 μL reactions with 1× Sybr Green (ABI, Applied Biosystems), 150 nmol/L of forward and reverse primers (see below), and the relevant cDNA (1 μL of a 1:20 dilution for KLK4, PSA, and HSD11B2 and 1:100 dilution for 18S). PCR was done on an ABI Prism 7000 sequence detection system (Applied Biosystems) and cDNA was quantitated using known DNA concentration standards. Gene expression was calculated as the amount of KLK4, PSA, and HSD11B2 cDNA relative to 18S cDNA copy number, and data are represented as the fold change in hormone-induced gene expression over vehicle control (0.1% ethanol). The primer sequences used in the RT-qPCR analyses are as follows: KLK4 (K4Ex2QS and K4Ex3QAS from above), 18S (18S-F: 5′-ttcggaactgaggccatgat-3′; 18S-R: 5′-cgaacctccgactttcg-3′), PSA (PSA-F: 5′-agtgcgagaagcattcccaac-3′; PSA-R: 5′-ccagcaagatcacgcttttgtt-3′), and HSD11B2 (HSD11B2-F: 5′-tggcgctactcatggacacat-3′; HSD11B2-R: 5′-tttcccactgacccacgtttc-3′).

Exon-specific RT-qPCR was carried out using primers K4Ex1QS and K4Ex2QAS and K4Ex2QS and K4Ex3QAS to detect KLK4 variants that use either TIS1 or TIS2a/b (29). Standard curves of known copy number using amplicons cloned into pGEMT were used to quantify gene expression. Exon 1-2 values represent the TIS1 variant, whereas the expression of TIS2a/b is defined as (exon 2-3)-(exon1-2) because exons 2 and 3 are present in all KLK4 variants. A two-tailed Student's t test was used to calculate statistical significance from three independent experiments that were each carried out in duplicate or triplicate.

To identify the TIS, RNA ligase–mediated RACE (FirstChoice RLM-RACE kit, Ambion) was carried out as per the manufacturer's instructions, except that Superscript II (instead of AMV) reverse transcriptase was used for cDNA synthesis. The first- and second-round KLK4-specific PCR primers used were 5′-agcccgatggtgtaggagtt-3′ and 5′-cgatggtgtaggagttctggaaacagtg-3′, respectively. PCR cycling variables were 94°C for 5 min followed by 40 cycles of 94°C for 30 s, 55°C or 62°C for 30 s (first- and second-round PCR, respectively), 72°C for 30 s, and a final 72°C extension for 10 min. PCR amplicons were cloned into pGEMT easy vector (Promega) and clones were sequenced using the ABI PRISM Dye Terminator 2 protocol (Applied Biosystems) at the Australia Genome Research Facility, University of Queensland, Brisbane, Australia.

Four independent gene analysis programs, Cister, SigScan, MatInspector (52-54), and ConSite,²

Soluble nuclear protein fractions were isolated from progesterone-treated T47D cells and R1881-treated LNCaP and 22Rv1 cells using the NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce Biotechnology) according to the manufacturer's instructions. Protein concentrations were quantified using the bicinchoninic acid protein assay (Pierce).

Nucleotide sequences encompassing a putative PRE (K4.pPRE) and a putative ARE (K4.pARE) in the KLK4 promoter were synthesized (Proligo) with their complements. The forward sequences for K4.pPRE (5′-aaaaagagagaaAGAACATgagagAGAACAggagagaatgag-3′) and K4.pARE (5′-tcgaaagccgagGGTGCAggaGATTGTGcttcc-3′) were end labeled with biotin (Pierce), then annealed to their labeled complementary oligomers. The putative HREs are underlined and the half sites are capitalized. PSA ARE I-A DNA probes (5′-tcgacttgcAGAACAgcaAGTACTagctg-3) were also used in EMSA to serve as a positive control for interaction with the AR. Binding reactions were carried out using 5 μg of nuclear proteins or 0.2 μg of purified androgen receptor DNA binding domain (AR-DBD; ref. 45) with the LightShift Chemiluminescent EMSA kit (Pierce) as instructed. Blocked shift experiments were done using 1 or 2 μg of either PR antibody (PR-C19) or AR antibody (C-19). PR and AR antibodies were incubated with nuclear proteins for 24 h at 4°C before addition to the labeled K4.pPRE or K4.pARE oligomer. Nuclear extracts from the PR-negative breast cancer cell line, MDA-MB-231, was used as a negative control to assess binding specificity in PR experiments. Controls using molar excess of unlabeled DNA probes were also included in EMSA analyses to assess for specificity of AR shifts.

A bacterial artificial chromosome clone (BC85745) and Cosmid clone (R28781; Lawrence Livermore Laboratory) were both used as template to generate seven deletion constructs encompassing 2,875 bp of the KLK4 promoter. XhoI and HindIII restriction sites were added to the primers to facilitate orientation-specific cloning into the pGL3-Basic vector (Promega). Forward primer sequences were K4.170, 5′-tgtgctcgagctgctcctgaacctctgacc-3′; K4.445, 5′-cacctcgagctaccctgaatccctgacca-3′; K4.688, 5′-cacctcgagaaaacggtgttttggtgtgc-3′; K4.1298, 5′-cacctcgaggtgtgtgtgtctgaccgtgct-3′; K4.898, 5′-ctcgagcaaacggtgttttggtgtg-3′; K4.2000, 5′-ctcgagagcagtggaatccaggagc-3′ and K4.2875, 5′-cacaagcttcagtcctcgccgtttatgat-3′, and PCR was carried out using a common reverse primer (K4.Anchor, 5′-aagcttcagtcctcgccgtttatgat-3′). The KLK4 promoter inserts were amplified using the “proofreading” enzyme Pfx or high-fidelity platinum Taq (Invitrogen) DNA polymerase and subcloned into the pGEM-T Easy (Promega) vector before cloning into pGL3-Basic (Promega). All constructs were sequenced to verify their sequence fidelity as described above. KLK4 luciferase promoter constructs are designated in Results according to their corresponding forward primer name.

Oligonucleotides encoding three tandem copies of a putative ARE in the KLK4 promoter (KLK4.pARE) and its native flanking sequences were synthesized with BglII and KpnI restriction site overhangs to facilitate orientation-specific ligation into the multiple cloning site of the luciferase pGL3-Promoter vector (Promega) and designated as KLK4.pAREX3-sense (5′-ccgagGGTGCAggaGATTGTgcttcccgagGGTGCAggaGATTGTgcttcccgagGGTGCAggaGATTGTgcttcca-3′) and KLK4.pAREX3-antisense (5′-gatctggaagcACAATCtccTGCACCctcgggaagcACAATCtccTGCACCctcgggaagcACAATCtccTGCACCctcgggtac-3′). The putative AREs are underlined and the hexameric half sites are capitalized. The oligonucleotides were annealed together by denaturing at 95°C for 2 min and left at room temperature for 1 h before ligation with the pGL3-Promoter vector. The KLK4.pARE luciferase reporter construct is designated as K4.pAREX3-Luc. The PSA 5.8-A-Luc and PSA AREIX3_A-Luc promoter constructs (45) were also used in luciferase reporter assays to serve as a positive control for androgen response.

Cells were seeded in 24-well plates at a density of ∼1 × 10⁵ per well and the culture medium was changed to phenol red–free DMEM (T47D) or RPMI 1640 (LNCaP and 22Rv1), 2% charcoal-stripped FCS for 48 hours before transfection. Transient transfection was carried out with 0.5 μg of either pGL3-Basic, pGL3-Promoter, PSA promoter construct, or KLK4 promoter construct using Opti-MEM I reduced Serum Medium (Invitrogen) and 3 μL of Lipofectamine 2000 per well (Invitrogen). Renilla (0.3 μg) was used as an internal control to monitor for transfection efficiency. After 6 h of transfection, cells were treated with either vehicle control (0.1% ethanol), 10 nmol/L progesterone (T47D), or 1 nmol/L R1881 (LNCaP and 22Rv1) for a further 24 h. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) on a PolarStar plate reader (BMG, Labtech). Data are expressed as Luciferase activity normalized to Renilla activity and represented as the SE from three independent experiments that were each carried out in triplicate, unless otherwise stated.

The ChIP assay was carried out on progesterone-treated T47D cells as previously described (19). Immunoprecipitation of PR/DNA complexes was done using 2 μg of PR-19 antibody. All DNA samples were first purified (High Pure, Roche), and PCR was done using primers that are (a) located proximal to the K4.pPRE (5′-ggaaatttgctggagaagca-3′ and 5′-tgcctgtatctctcattttctc-3′); (b) within the KLK4 promoter but not proximal to the K4.pPRE (5′-gcctgagagagttgagctgg-3 and 5′-agaaggcacagaggctgagaa-3′), and (c) in a non–progesterone-regulated promoter (β2-microglobulin; 5′-gccgatgtacagacagcaaa-3′ and 5′-tgctgtcagcttcagga atg-3′). PCR was performed in the linear range of amplification. The variables were as follows: 94°C for 5 min, then 30 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 30 s followed by a final 10 min, 72°C extension.

No potential conflicts of interest were disclosed.

We thank Dr. Linda Ashworth at Lawrence Livermore Laboratories, CA, for the gift of the chromosome 19 BACS and COSMID clones and Dr. A. Whitbread for the critical review of the manuscript.