Advanced hormone-independent prostate cancer is characterized by a significant loss of androgen receptor (AR) expression in 20–30% of the tumors. The transcriptional block underlying this phenomenon is not known, but we have proposed that methylation of CpG sites in the AR promoter may reversibly inactivate transcription of the AR (D. F. Jarrard et al., Cancer Res., 58:5310–5314, 1998). In this study, detailed methylation analysis using bisulfite sequencing was performed on a series of AR expression-positive and -negative prostate cancer cells. We found that methylation of several consensus sequences in the AR promoter (from−131 to −121 and +44 to +54) are tightly linked to the loss of AR expression in metastatic hormone-independent prostate cancer cell lines. These consensus sites of methylation correlate with the minimal promoter region critical for AR transcription. In human tissues, no methylation was demonstrated in normal or primary prostate cancers that express the AR. Four of 15 tumors obtained from men who had died from hormone-independent prostate cancer demonstrated a significant loss of AR expression immunohistochemically and two (50%) of these AR-negative tumors contained AR methylation. We conclude that the AR promoter contains specific CpG methylation hot spots that are markers for gene silencing. Furthermore, AR methylation may represent a phenotype important in the development of hormone independence in a subset of advanced prostate cancer in which AR expression is lost. The finding of AR methylation also represents the first report of aberrant methylation on an X-linked gene associated with a somatic male cancer.

The terminal stages of metastatic prostate cancer are characterized by androgen independence and represent a significant therapeutic problem. The average survival after failing hormone ablative or blocking therapy is only 12 months (1). The evolution of prostate tumor progression under the selective pressure of androgen ablation therapy is not fully understood. Many current paradigms relate the finding of AR3expression or overexpression to the development of hormone independence in advanced prostate cancers. Clearly, mechanisms such as AR amplification and mutation may play a role by altering sensitivity of the expressed AR to androgens (1). However, 20–30% of hormone-independent cancers are characterized by the extensive loss of AR expression, including a complete loss of expression in some cancers (2, 3, 4). The loss of AR expression in androgen-independent cells appears to occur at the transcriptional level (5, 6) and does not involve either deletional or mutational mechanisms (7, 8).

In an alternate approach to this problem, we have recently demonstrated that AR methylation at several methylation-sensitive enzyme restriction sites correlates with a loss of expression in AR-negative prostate cancer cell lines (9). DNA hypermethylation is a well-described epigenetic mechanism that involves the addition of methyl groups to deoxycytosines in the palindromic dinucleotide CpG. Dense clusters of CpGs, termed CpG islands, are found at the 5′ end of roughly 60% of genes and typically encompass the transcriptional promoter. Methylation of these CpG islands may function to inactivate gene transcription (10). CpG island methylation plays a critical role during development by establishing haploid gene dosage and in the long-term repression of selected genes. Aberrant methylation has been implicated in the silencing of autosomal tumor and growth suppressor genes in cancer in a manner analogous to mutation or deletion (10).

The mechanisms by which DNA methylation exerts its repressive function have not been completely defined. Islands of methylated DNA exist in an inactive or condensed chromatin structure in association with MeCP2, a protein that selectively binds to methylated DNA sequences, and with histone deacetylase (11). Deacetylation appears to lead to the compaction of chromatin by favoring interactions between adjacent nucleosomes. The association between local methylation and higher order structure that results in a closed chromatin structure is complex(12). Not all potentially methylatable CpGs in an island are required for gene silencing, and, furthermore, methylated DNA segments can confer silencing at a distance in cis(13). The majority of studies that associate gene inactivation with methylation rely on an analysis at one or several CpG sites, typically using methylation-sensitive restriction enzymes. This often produces conflicting or contradictory results (14, 15). Recently, new techniques using sodium bisulfite treatment have been described that permit the examination of methylation at each CpG site within a CpG island (16). Fine methylation mapping of the RB, GSTP-1, and MGMT genes has identified regions of increased methylation density that correlate with known transcription binding sites (14, 17, 18). Other studies confirm that methylation of specific sequences may exert a local repressive effect on transcription by the direct inhibition of transcription factor binding (19, 20). Therefore,determining gene promoter sites that are methylated and associated with the loss of gene transcription may identify sequences important or critical to transcription.

Males have one copy of the AR located at Xq11–12. In females, most genes located on the X chromosome contain CpG islands that undergo methylation on one copy to establish normal gene dosage. The 3-kb AR promoter CpG island extends from roughly −500 through exon 1 and fulfills the criteria for a CpG island (21). Methylation of the AR promoter during X inactivation has been used previously to study clonality in female tumors (22, 23). These studies have focused on methylation at one or several CpG sites located in exon 1 of the AR in cells with two X chromosomes.

The inactivation of single copy X-linked genes in males is capable of causing total loss of gene expression. Deletion or mutation may lead to the inactivation of X-linked disease genes in inherited diseases in males. One example is the X-linked lymphoproliferative syndrome that predisposes individuals to lymphomas (24). An alternate mechanism for the loss of X-linked gene transcription is aberrant methylation. FMR-1 gene expression is lost via hypermethylation and results in a syndrome that includes mental retardation in affected males (25). Thus far, methylation of specific X-linked genes has not been demonstrated to be associated with the development of sporadic cancers in males.

To further define the spatial relationships between AR CpG island methylation and gene silencing, we performed a detailed methylation analysis of prostate cancer cell lines and tumors. We found regions of consistent methylation in AR expression-negative prostate tumor cells that delineate critical sites for gene transcription. Furthermore,these hot spots identify markers for AR gene silencing. Aberrant methylation of the AR was not detected in human normal prostate or primary prostate cancer tissues. Methylation was found in hormone-insensitive prostate cancers and was associated with a loss of AR expression. Many different mechanisms contribute to overcoming the androgen dependence of advanced prostate cancer. AR methylation may represent a phenotype associated with the development of hormone independence in a subset of prostate cancer that does not express the AR. This finding is of therapeutic interest because we have demonstrated previously that AR gene methylation is reversible when prostate cancer cells are exposed to the demethylating agents 5-azacytidine and 5′ aza-2-deoxycytidine (9).

Cell Lines and Tissue Samples.

The prostate cancer cell lines LNCaP, Du145, PC3, DuPro,TSU-PR1, and PPC-1 were derived from metastatic prostate cancers and cultured in RPMI 1640 (Cellgro, Herndon, VA) supplemented with 10%fetal bovine serum as described previously (9). Ten paired samples of normal prostate and primary prostate cancer tissue were obtained from radical prostatectomy samples. Metastases were obtained from 12 patients who had died from hormone-refractory prostate cancer. Metastatic sites included eight lymph nodes, four liver lesions, two bony lesions, and one lung metastasis. Normal tissues included renal,muscle, and ovarian tissues. Fresh tissues were snap-frozen in liquid nitrogen and examined histopathologically.

DNA Extraction, Sodium Bisulfite Reaction, and PCR Amplification.

Genomic DNA was extracted using standard methods. One μg of genomic DNA was treated with sodium bisulfite at 50°C for 18 h using a CpGenome amplification kit (Oncor, Gaithersburg, MD). Five separate regions within the AR CpG island were amplified using 100 ng of sodium bisulfite-modified DNA in 20 μl of reaction buffer containing 1× PCR buffer, 1.5 μm MgCl2, 125μ m deoxynucleotide triphosphates, 10 pmol of primers, and 1 unit of AmpliTaq Gold (Perkin-Elmer, Branchberg, NJ). All PCR primers were designed for top strand amplification (Table 1). PCR conditions were as follows: 94°C for 9 min, followed by 35 cycles of 94°C for 50 s, 50°C–57°C for 1 min, and 72°C for 50 s. An additional extension at 72°C for 4–20 min was then performed for all reactions.

Sequencing and Subcloning.

PCR products were run on 1.5% agarose gels and then excised and purified using the QIAquick Gel Extraction kit (QIAGEN, Valencia,CA). DNA was then sequenced or subcloned. Direct sequencing of the PCR products was performed using a Thermal Cycle Sequenase kit (Amersham Pharmacia Biotech, Piscataway, NJ). To compensate for possible regional biases in PCR after bisulfite treatment (26), mixing experiments were performed for each primer set to calculate the percentage of methylation at individual sites. PCR bias after bisulfite treatment is dependent on the primer set chosen and the region being amplified (26). Methylation density using direct sequencing techniques has been found to be quantitative(18). Completely methylated Du145 was diluted sequentially with unmethylated LNCaP (0%, 25%, 50%, 75%, and 100%) to generate a standard curve, and the sequencing reactions were performed. Samples were then compared (Imagequant; Stratagene, La Jolla, CA) with these dilution controls to determine the percentage of methylation for each region evaluated. Film was developed at both 24 and 48 h to avoid error due to signal saturation. The methylation of specific sites on the AR as measured by this technique was identical to Southern blotting results (9).

To assess the methylation status of individual alleles, PCR fragments surrounding the transcription start site (region 2) were subcloned using the TOPO TA vector system (Invitrogen, Carlsbad, CA). After transformation using Escherichia coli, we randomly selected 10–25 individual E. coli colonies for each cell line assessed. Plasmid DNA was isolated (QIA Miniprep kit; QIAGEN). Ten to 20 plasmids were then sequenced using region 2 primers (Table 1).

Ms-SNuPE Assay.

We surveyed methylation status at 14 CpG sites (Table 1) in human prostate samples and metastatic cancer samples using the Ms-SNuPE assay(27). The reaction was performed using approximately 20 ng of PCR products generated for each region as described above. Products were placed in a 20-μl reaction buffer containing 1× PCR buffer, 10 pmol of each of the Ms-SNUPE primers, 1.5 μmMgCl2, 1 μCi of either[32P]dCTP or [32P]TTP,and 1 unit of AmpliTaq Gold. Ms-SNUPE primers were designed for the top strand. Reaction conditions included denaturing at 94°C for 10 min,annealing at 55°C for 1 min, and single base extension at 72°C for 1 min. Denatured samples were placed in loading dye (95% formamide, 5 mm EDTA, 0.2% XC, and 0.2% bromphenol blue)and electrophoresed on a 15% denatured polyacrylamide gel (7 m urea). Radioactive intensity was measured using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Reactions were performed in triplicate. Controls included known proportions of unmethylated and fully methylated plasmids.

Immunohistochemistry.

Ten metastatic tumor samples were available for immunohistochemical analysis. Deparaffinized slides underwent microwave treatment in 1 mm EDTA (pH 8.0) for 20 s and were then placed in 1×xylene cyanol wash solution (Ventana Biotek Systems, Tucson,AZ). An automated immunostainer (Ventana GenII) was used to incubate slides for 30 min at 37°C with a mouse monoclonal anti-AR antibody(Ab-1, clone AR-441; Neomarkers, Fremont, CA) or mouse IgG as a negative control (4 μg/ml). The Ventana biotin-avidin-horseradish peroxidase detection system with diaminobenzidine substrate was used. Slides were counterstained with hematoxylin, dehydrated, and coverslipped. A known positive control of normal prostate tissue was treated identically. Adjacent sections from each metastasis also stained positively with fatty acid synthetase.4

Direct Sequence Mapping of CpG Methylation Encompassing the Transcription Start Site in Prostate Cancer Cell Lines.

The methylation status of individual CpGs was determined in five regions of the AR promoter by PCR amplification and direct sequencing of bisulfite-treated DNA (CpG dinucleotide sites and other features of the 3-kb AR CpG island are shown in Fig. 1). We initially mapped the prostate cancer cell lines in region 2 to determine whether any consensus regions of methylation correlated with loss of AR expression. Region 2 extended from −200 to +115 and contained 21 CpG dinucleotide sites surrounding the consensus transcription start site (Table 2). This core promoter contains several cis-acting elements important to AR transcription (28) and contains the minimal AR promoter (29). LNCaP and PC3, which express the AR (30), were unmethylated within this region as evidenced by a lack of cytosine dinucleotides after sodium bisulfite sequencing. During sodium bisulfite treatment and PCR, unmethylated cytosines are converted via uridine to thymidine. In female DNA, which contains one transcriptionally inactive X chromosome (and one active X chromosome),all CpG dinucleotides were partially methylated as predicted (data not shown). In contrast, cell lines not that do not express the AR, Du145,PPC-1, DuPro, and TSU-PR1, showed marked methylation across CpG island region 2. Cell line Du145 demonstrated complete, dense methylation within all CpGs.

Variations in the methylation of CpG sites across region 2 were found in DuPro, PPC-1, TSU-PR1, and Du145. Several consistently(preferentially) methylated regions, or hot spots, were demonstrated in these cell lines by direct sequencing profiles of region 2. CpG dinucleotides −131, −125, −123, and −121 were methylated in all AR-negative cell lines. These CpG sites encompassed a BssHll restriction site (CCGG) and are adjacent to a purine-rich region critical for AR transcription as determined by AR promoter deletion studies (28). A second region involving +44, +49, and +54 was also consistently methylated in Du145, DuPro, and PPC-1. These sites are within the minimal region of the AR promoter necessary for AR expression (29).

CpG Methylation of Individual Alleles in the Core Promoter Region(Region 2).

The direct sequence data demonstrated that the hypermethylation in region 2 around the transcription start site was inversely correlated with the AR expression. However, incomplete methylation was detected in all cell lines that do not express the AR (Du145, DuPro, PPC-1,and TSU-PR1) and may have been due to either variations in methylation at individual CpG sites or the presence of unmethylated alleles. We assessed the methylation status of the core promoter region (region 2)for AR-negative cell lines in 10–20 individual alleles (Fig. 2). PP analysis of the methylation patterns on individual alleles in normal female tissue demonstrated that 8 of 18 alleles were completely unmethylated, consistent with the predicted pattern. This data further confirmed a minimal PCR bias in this region. Methylation density on individual X strands varied (50% of potential sites were methylated), but CpGs between −131 and −121 and +44 and +54 contained consistent methylation on at least one site in all methylated alleles. CpG site +12 was also consistently methylated. Methylation at all CpG sites was not required for gene silencing of the AR.

In contrast, Du145 demonstrated dense methylation in all AR alleles through region 2. However, in this cell line, CpG sites −51, −42,+80, and +99 contained partial methylation, consistent with direct sequencing results (Table 2). All alleles of PPC-1 also contained methylation. Only several sites were methylated in several alleles, but methylation consistently involved +44 to +54 and −131 to −121. Methylation of these AR-negative cell lines involved all alleles, and several regions adjacent to the transcription start site were consistently involved.

A second group of cell lines, DuPro and TSU-PR1, demonstrated methylation in only 40–70% of the alleles. The remaining alleles in region 2 were without any detectable methylation. All alleles contained lower amounts of methylation when compared with normal female tissues and other AR-negative lines, PPC-1 and Du145. Methylation was found to preferentially encompass the identified hot spots. Therefore, both variations in methylation at specific sites and the presence of unmethylated alleles may contribute to the partial methylation found in these lines.

Mapping of CpG Methylation in Other CpG Island Promoter Regions.

The methylation status of individual CpGs was determined by direct sequencing in four other regions of the CpG island (Table 2). Region 1,extending from −526 to −259 (eight CpG sites), was tested to map methylation at the 5′ edge of the CpG island. We found methylation in all prostate cancer cell lines, in normal prostate, and in normal renal tissues and lymphocytes. Methylation occurred frequently at the island edge and did not correlate with AR expression. In regions 3 (+130 to+446), 4 (+498 to +668; data not shown), and 5 (+920 to +1408), we also found no consistent sites of methylation that differentiated AR-expressing cell lines from non-AR-expressing cell lines. LNCaP was found to be completely unmethylated in these three regions. PC3, a cell line that expresses AR at low levels, contained several scattered and minimally methylated sites (10–40%), including several that encompassed the translation start site. Transcriptionally inactive cell lines, including Du145, contained methylated and unmethylated CpGs in these 3′ regions. Complete methylation was noted in TSU-PR1 at sites+1216 and +1218, which may indicate the presence of another transcriptionally important downstream region.

AR Methylation in Normal Prostate Tissues, Primary Prostate Cancers, and Metastatic Cancers.

We examined methylation status in 10 paired normal prostate and primary(localized) prostate cancers derived from radical prostatectomy specimens. Fifteen metastatic lesions were also harvested from 12 patients who had died from hormone-independent prostate cancer. We initially surveyed 14 CpG sites using a quantitative methylation assay,Ms-SNuPE (27). CpG sites examined by Ms-SNuPE (listed in Table 1) included the preferentially methylated sites identified from fine-mapping of prostate cancer cell lines. All normal prostate and primary prostate cancers did not contain methylation in the AR core promoter region or around the translation start site. All of these tissues expressed the AR (data not shown. The Ms-SNuPE analysis of 15 metastatic samples detected methylation of the CpG promoter in 1 lymph node (Fig. 3, M1) and 1 cranial (Fig. 3, M2) metastasis from different patients. Sample M1 contained 70% methylation at two CpG sites in region 2 (−177 and −131), with lower levels of methylation found at other sites. Sample M2 contained methylation (35%) at CpG sites +1257 and +1408 downstream from the translation start site. No significant methylation was detected in region 2 sites tested in this sample.

Direct sequence methylation analysis was performed to further map the distribution of methylation in these samples. Consistent with the Ms-SNuPE assay, neither six normal prostate nor primary prostate cancer samples showed any methylation within the promoter examined, except for methylation in region 1 at the 5′ edge of the island. Representative samples from the direct sequencing of these prostate tissues are shown in Table 2. Direct sequencing of three metastatic samples confirmed substantial methylation in M1 in AR regions 1 and 2 adjacent to the transcription start site (Table 2). This methylation (from −177 to−121 and +44 to +108) encompassed the two consensus methylation regions and was similar to methylation seen in AR-negative prostate cancer cell lines. Sample M2 contained 30% methylation extending downstream from +1257 through +1408. M3 demonstrated no methylation and expressed the AR (data not shown). Direct sequencing confirmed the Ms-SNuPE results.

To correlate AR expression with methylation, an immunohistochemistry assay for AR expression was performed in metastatic lesions from which tissue sections could be obtained. AR staining was found to be heterogeneous and present in most samples, including sample M3 (data not shown). Four of 10 samples were found to contain no staining in 50% or more of cells, including samples M1 (75–90% negative) and M2(50% negative). In summary, methylation of the AR promoter occurred in two of four samples containing a significant loss of AR expression.

Methylation of Non-CpG Sequences.

There have been several reports documenting methylation occurring on the outer cytosine of CpNpG sequences (17, 18). We surveyed C to T conversion in more than 4000 cytosines not followed by guanine in region 2 (data not shown). We found four methylation variants: (a) two at CTG; (b) one on the outer cytosine of CpNpCpG; and (c) one not associated with guanine. The CpNpG methylation variants were observed only on the methylated alleles and were reproducible. If this is due to incomplete sodium bisulfite treatment, the frequency of error is <0.1% with our assay.

High-resolution mapping of methylation patterns on individual DNA molecules has contributed to a better understanding of the epigenetic control of gene transcription and its contribution to human disease. These studies have demonstrated epigenetic heterogeneity, but they also confirm consistent methylation of sites important for gene transcription (14, 17, 25). In this study, we generated a methylation profile for the AR gene in AR-expressing and non-AR-expressing prostate cancer cells in culture and from primary and metastatic prostate samples. This detailed map included the methylation status of 80 CpG sites from −500 to +1408 (relative to the transcription start site). Our results demonstrated that the region containing the AR transcription start site (−177 to +108) was preferentially and consistently methylated in AR-negative prostate cancer cells. These sites encompassed the AR minimal promoter and have been found to be critical for AR gene transcription (29). Methylation of this region was also detected in advanced hormone-independent metastatic prostate cancer but not in primary,localized prostate cancers or in normal tissues. This hypermethylation in vivo was associated with a loss of AR expression. This is a notable finding because it may explain the loss of AR expression in a subset of these tumors during progression to hormone independence. Furthermore, mutations and deletions on X-linked genes have been found in somatic cancer (31, 32). Our data document a novel paradigm, that of DNA methylation, as a putative mechanism for the inactivation or suppression of the X-linked AR in prostate cancer.

Regional variations in methylation density occur across the CpG island in AR-negative prostate cancer cell lines. However, it is significant that we were able to identify several consensus regions of methylation adjacent to the AR transcription start site where methylation correlated with loss of AR expression. These methylation hot spots putatively identify promoter regions critical to transcription. In vitro methylation of an identified hot spot in a reporter construct containing the p16/cdk2 gene led to a subsequent decrease in reporter expression (33). Consensus methylation sequences may encompass known transcription factor binding sites, such as AP-2, and methylation may directly block transcription factor binding (19, 20). In the AR,consensus regions of methylation occur within the AR minimal promoter as determined by deletion studies (29). Although investigated widely, the specific cis regulatory elements within this region have not been clearly defined (28, 29, 34). We have identified one hot spot in the AR promoter occurring at −132 to −122 adjacent to a purine-rich region found to positively regulate AR promoter activity by the functional analysis of deletion mutants (28). A second consensus region of methylation was found between +44 and +54. This fragment did not correlate with any previously identified cis-acting elements. In several cell lines and in one metastasis (M2) lacking AR expression, dense methylation downstream from the translation start site may indicate another potential consensus region of methylation. This region was not fully characterized in our study. Methylation of these hot spots may additionally reflect a repressive chromatin structure and nucleosome positioning (35). We suggest that methylation of these hot spots may function in a manner similar to an inactivating mutation in the AR gene.

The analysis of methylation by fine-mapping prostate cancer cells in vitro and in vivo demonstrated a region of relative lack of methylation that directly encompassed the transcription start site (−56 to +20). This was confirmed by an analysis of individual alleles. Transcriptional inactivation associated with methylation does not clearly require every CpG to be methylated in this region. Undermethylation may reflect the binding of proteins to this region, thus protecting the DNA from the action of DNA methyltransferase. This region contains a Sp1 site (−42), and Sp1 sites are generally devoid of methylation in otherwise hypermethylated genes such as the X-linked gene HPRT(36). Sp1 transcription factors may also attract factors that induce the removal of methylated cytosines (37, 38). The finding of regional hypomethylation also emphasizes the necessity of examining all or multiple CpG sites when assessing genes for hypermethylation.

To further address the question of methylation heterogeneity, we isolated and sequenced individual copies of AR DNA. In normal female tissues and prostate cancer cells, we confirmed that methylation occurs consistently within the two identified hot spots. Complete methylation of all CpGs in the minimal promoter was not observed, although one cell line (Du145) contained methylation on >90% of the CpGs in this region. Methylation density, as well as location, appears to be important for gene silencing. Methylation on as few as one or two critical sites is sufficient to inactivate transcription of the Epstein-Barr virus (39). With the exception of the identified hot spots, we noted a marked variation in the methylation of individual CpGs between alleles, even in normal female tissues. It was recently reported that T cells from normal individuals demonstrate a significant amount of diversity in the methylation of specific Notl loci, a methylation-sensitive restriction enzyme site typically found within CpG islands (40). Epigenotypes from peripheral blood also contain significant heterogeneity across the FMR1 locus (25). Variability in methylation found at most sites in the AR further confirms a lack of faithful inheritance of specific methylation patterns during DNA replication,even among molecules from clonal cellular populations. This heterogeneity reflects the mechanisms regulating maintenance methylation and demethylation.

The analysis of methylation in specific alleles revealed that two AR-negative prostate cancer cell lines, TSU-PR1 and DuPro, contained unmethylated alleles in the minimal promoter region of the CpG island. Therefore, methylation may not directly regulate gene silencing in one or both of these cell lines. However, sequencing of each allele was not tested in all regions of the CpG island. Complete methylation may occur in other downstream regions containing important elements for transcription. For example, we noted full methylation of sites +1216 and +1218 in TSU-PR1 and dense methylation of other downstream sites in one metastasis (M2). Our previous data demonstrating that the demethylating agent 5′ aza-2-deoxycytidine induces AR reexpression in these two cell lines support the presence of other methylated regions that are of importance to AR transcription (9). Methylated DNA associates with repressive chromatin characterized by the presence of underacetylated histones. We are able to induce expression of the AR in all AR-methylated cell lines by sequentially exposing prostate cancer cells to 5′ aza-2-deoxycytidine and the histone deacetylation inhibitor trichostatin A.5

Regional methylation within the AR CpG island was associated with silencing of the AR gene both in vitro and in vivo. In the prostate cancer samples examined, methylation occurred only in advanced hormone-independent disease from patients who had died of prostate cancer. A significant loss of AR expression(≥50%) was found in 4 of 10 specimens tested. This rate is similar to that seen in previous studies (2, 3, 4). In certain tissues,such as bone metastases, the loss of AR expression may be greater(3). Methylation clearly encompassed the previously identified hot spots in the transcription start site in one metastatic sample (M1). A second metastatic tumor (M2) contained methylation extending downstream from the translation start site. AR loss of expression due to methylation in this region alone did not clearly lead to a loss of expression on the basis of our data, but we cannot rule out the presence of another critical hot spot in the 3′ region of the CpG island. It has been observed, based on methylation patterns in the E-cadherin and the Von Hippel-Lindau (VHL) genes,that during carcinogenesis, methylation propagates from 5′ and 3′“ de novo methylation centers“ at the edges of the CpG island (41). Associated with this progressive methylation was the loss of gene transcription. Therefore, incomplete methylation may serve as a marker for gene inactivation in some cases.

The loss of AR expression may represent an important phenotype in prostate cancer progression. Growth under these AR-independent conditions appears to be dependent on the acquisition of autocrine pathways, such as activation of epidermal growth factor receptor or transforming growth factor β (42, 43), or the activation of alternate pathways including HER-2/neu tyrosine kinase(44). The loss of AR expression associated with methylation of the gene may represent a tumor with a ‘hypermethylator phenotype’ that has been postulated to occur in colon cancer and other cancers (45). Alternatively, inactivation of the AR may be a selective alteration important for hormone-independent growth in some prostate cancers. Expressed AR may function as a growth suppressor in PC3 and LNCaP (containing a mutant AR) cancer cells (46, 47). A similar “squelching” phenomenon has been proposed to explain the observation that estrogen receptor reexpression inhibits estrogen receptor-negative breast cancer cells (48). These observations suggest that the loss of AR expression, including that associated with AR methylation, may be an important step in the progression of a subset of prostate tumors.

Fig. 1.

Schematic representation of the 5′ region of the human AR gene (Xq11–12) demonstrating the AR CpG island. The AR gene contains a∼3-kb CG-rich region (each CpG is indicated by a vertical tic) extending across the transcription start site (right angle arrow at +1) and through exon 1(black bar). Regions examined by mapping methylation are shown (designated as Regions 1–5). Methylation-sensitive restriction enzyme sites are noted as follows: Smal, S; BssHll, B; and Hpall, H. UTR, untranslated region. ATG,translation start site.

Fig. 1.

Schematic representation of the 5′ region of the human AR gene (Xq11–12) demonstrating the AR CpG island. The AR gene contains a∼3-kb CG-rich region (each CpG is indicated by a vertical tic) extending across the transcription start site (right angle arrow at +1) and through exon 1(black bar). Regions examined by mapping methylation are shown (designated as Regions 1–5). Methylation-sensitive restriction enzyme sites are noted as follows: Smal, S; BssHll, B; and Hpall, H. UTR, untranslated region. ATG,translation start site.

Close modal
Fig. 2.

Determination of the methylation status of individual alleles in the AR core promoter region (region 2). The methylation status of each allele in metastatic prostate cancer cell lines and normal female tissues was assessed by direct sequencing of cloned products from bisulfite-treated DNA. The data are summarized graphically with each tic mark (top)representing a CpG site from −200 to +115. Specific CpG sites in region 2 are listed in Table 2. Each individual allele sequenced is represented by a row of circles. Ten to 20 alleles were sequenced for each cell line. Unmethylated CpG dinucleotide, ○; methylated CpG dinucleotide, •.

Fig. 2.

Determination of the methylation status of individual alleles in the AR core promoter region (region 2). The methylation status of each allele in metastatic prostate cancer cell lines and normal female tissues was assessed by direct sequencing of cloned products from bisulfite-treated DNA. The data are summarized graphically with each tic mark (top)representing a CpG site from −200 to +115. Specific CpG sites in region 2 are listed in Table 2. Each individual allele sequenced is represented by a row of circles. Ten to 20 alleles were sequenced for each cell line. Unmethylated CpG dinucleotide, ○; methylated CpG dinucleotide, •.

Close modal
Fig. 3.

Methylation of the AR promoter in prostate tissues using Ms-SnuPE. Ms-SNuPE was carried out as described previously(27) to quantitatively assess the methylation status at individual CpG sites (Table 1). Genomic DNA was first reacted with sodium bisulfite to convert unmethylated cytosine to uracil while leaving 5-methylcytosine unchanged. Amplification of the desired target sequence was then performed using PCR primers specific for bisulfite-converted DNA, and the resulting product was isolated and used as a template for single nucleotide primer extension. Fifteen hormone-independent prostate metastases and 10 paired normal/primary prostate cancers were tested. None of the normal or primary prostate cancers demonstrated methylation. Representative normal prostate(NP1) and primary prostate cancer (PC1)samples are shown with the individual CpG sites tested. Methylation was detected in two hormone-independent metastatic lesions. M1 was ∼ 70% methylated at CpG core promoter sites −177 and −131. M2, another metastatic lesion,demonstrated ∼30% methylation at +1257 and +1408 in exon 1. UM, unmethylated; M, methylated.

Fig. 3.

Methylation of the AR promoter in prostate tissues using Ms-SnuPE. Ms-SNuPE was carried out as described previously(27) to quantitatively assess the methylation status at individual CpG sites (Table 1). Genomic DNA was first reacted with sodium bisulfite to convert unmethylated cytosine to uracil while leaving 5-methylcytosine unchanged. Amplification of the desired target sequence was then performed using PCR primers specific for bisulfite-converted DNA, and the resulting product was isolated and used as a template for single nucleotide primer extension. Fifteen hormone-independent prostate metastases and 10 paired normal/primary prostate cancers were tested. None of the normal or primary prostate cancers demonstrated methylation. Representative normal prostate(NP1) and primary prostate cancer (PC1)samples are shown with the individual CpG sites tested. Methylation was detected in two hormone-independent metastatic lesions. M1 was ∼ 70% methylated at CpG core promoter sites −177 and −131. M2, another metastatic lesion,demonstrated ∼30% methylation at +1257 and +1408 in exon 1. UM, unmethylated; M, methylated.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by NIH CA76184-01 and a University of Wisconsin Howard Hughes Faculty Development Award.

3

The abbreviations used are: AR, androgen receptor; Ms-SNuPE, methylation-sensitive single nucleotide primer extension.

4

G. Steven Bova, personal communication.

5

Unpublished observations.

Table 1

PCR primers used for the amplification of bisulfite-treated DNA and Ms-SNuPE

A. PCR primers used for the amplification of bisulfite-treated DNA
RegionSense oligonucleotideAntisense oligonucleotidePositionSize (bp)
5′-GGGTAGAAATGATTTTGATTTTTGGG-3′ 5′-CTAACTCTAACCCTAAAAAACCCCTA-3′ −595 to−201 394 
5′-TGGTTTAGGAAATTAGGAGTTATTTAGG-3′ 5′-TCCCTTCGACTCCTATACAACACTA-3′ −257 to 154 384 
5′-GAGGAGGCGATAGAGGGAAAAAGG-3′ 5′-TCAATCCTACCAAACACTTTCCTT-3′  −63 to 287 224 
5′-AGTTTGTAGAGAGGTAATTTTTTTTGGT-3′ 5′-CCTAACTACCTTTTCATCTTTTAATCT-3′ 458 to 693 235 
5′-AAGAGATAGATTGTGAGTTTAGTAGGG-3′ 5′-TTCCTCATCCAAAACCAAATAACCTATAA-3′ 869 to 1436 567 
B. PCR primers used for direct quantitative DNA sequencing of PCR products     
Region Sense oligonucleotide Antisense oligonucleotide Position  
5′-TTTTTGGGGTTGAGGGTTTTTAGAGT-3′  −577  
5′-TGGTTTAGGAAATTAGGAGTTATTTAGG-3′  −257  
 5′-AAAGGAGGTGGGAAGGTAAGGAG-3′  −91  
5′-GGAGGCGATAGAGGGAAAAAGGGT-3′  65  
  5′-TCAATCCTACCAAACACTTTCCTT-3′ 287  
5′-AGTTTGTAGAGAGGTAATTTTTTTTGGT-3′  458  
  5′-CCTAACTACCTTTTCATCTTTTAATCT-3′ 693  
5′-AAGAGATAGATTGTGAGTTTAGTAGGG-3′  869  
 5′-TTTTTAGAATTTGTTTTAGAGCGTG-3′  1181  
  5′-TTCCTCATCCAAAACCAAATAACCTATAA-3′ 1436  
C. Ms-SNuPE primers and sites tested     
Region Primer name Sequence Position  
SnupeAR5330 5′-TTGTAGGAGATTTTTTTATTTTTTTATTTG-3′ −417  
SnupeAR5391 5′-TGGTAGTTAGGAGTAGGTATTTTTAT-3′ −361  
SnupeAR5497 5′-TTTAGGAAATTTGGAGTTTTGG-3′ −259  
SnupeAR5574 5′-GGGTTAGAGTTAGTTTTTTTTGTTTT-3′ −177  
SnupeAR5617 5′-GTATTTGTTTTTTTAAAGTTATTAGGTAGG-3′ −131  
SnupeAR5701 5′-AGGTGGGAAGGTAAGGAGGT-3′ −56  
SnupeAR5750 5′-AAATTGTTGTATTTGTTTTTTATTTTTTAG-3′ +3  
SnupeAR5799 5′-GGGAGTTAGTTTGTTGGGAGAG-3′ +44  
SnupeAR6789 5′-GTGGGTAGTTAGTTGTAG-3′ +1030  
SnupeAR6843 5′-GAGTAAGAGAAGGGGAGG-3′ +1040  
SnupeAR6918 5′-GGGTTGGGAAGGGTTTATTTT-3′ +1162  
SnupeAR6964 5′-GAGTTTTTTAGAATTTGTTTTAGAG-3′ +1212  
SnupeAR7013 5′-GGGTTTTAGGTATTTAGAGGT-3′ +1257  
SnupeAR7160 5′-TGAGGATGGTTTTTTTTAAGTTTAT-3′ +1408  
A. PCR primers used for the amplification of bisulfite-treated DNA
RegionSense oligonucleotideAntisense oligonucleotidePositionSize (bp)
5′-GGGTAGAAATGATTTTGATTTTTGGG-3′ 5′-CTAACTCTAACCCTAAAAAACCCCTA-3′ −595 to−201 394 
5′-TGGTTTAGGAAATTAGGAGTTATTTAGG-3′ 5′-TCCCTTCGACTCCTATACAACACTA-3′ −257 to 154 384 
5′-GAGGAGGCGATAGAGGGAAAAAGG-3′ 5′-TCAATCCTACCAAACACTTTCCTT-3′  −63 to 287 224 
5′-AGTTTGTAGAGAGGTAATTTTTTTTGGT-3′ 5′-CCTAACTACCTTTTCATCTTTTAATCT-3′ 458 to 693 235 
5′-AAGAGATAGATTGTGAGTTTAGTAGGG-3′ 5′-TTCCTCATCCAAAACCAAATAACCTATAA-3′ 869 to 1436 567 
B. PCR primers used for direct quantitative DNA sequencing of PCR products     
Region Sense oligonucleotide Antisense oligonucleotide Position  
5′-TTTTTGGGGTTGAGGGTTTTTAGAGT-3′  −577  
5′-TGGTTTAGGAAATTAGGAGTTATTTAGG-3′  −257  
 5′-AAAGGAGGTGGGAAGGTAAGGAG-3′  −91  
5′-GGAGGCGATAGAGGGAAAAAGGGT-3′  65  
  5′-TCAATCCTACCAAACACTTTCCTT-3′ 287  
5′-AGTTTGTAGAGAGGTAATTTTTTTTGGT-3′  458  
  5′-CCTAACTACCTTTTCATCTTTTAATCT-3′ 693  
5′-AAGAGATAGATTGTGAGTTTAGTAGGG-3′  869  
 5′-TTTTTAGAATTTGTTTTAGAGCGTG-3′  1181  
  5′-TTCCTCATCCAAAACCAAATAACCTATAA-3′ 1436  
C. Ms-SNuPE primers and sites tested     
Region Primer name Sequence Position  
SnupeAR5330 5′-TTGTAGGAGATTTTTTTATTTTTTTATTTG-3′ −417  
SnupeAR5391 5′-TGGTAGTTAGGAGTAGGTATTTTTAT-3′ −361  
SnupeAR5497 5′-TTTAGGAAATTTGGAGTTTTGG-3′ −259  
SnupeAR5574 5′-GGGTTAGAGTTAGTTTTTTTTGTTTT-3′ −177  
SnupeAR5617 5′-GTATTTGTTTTTTTAAAGTTATTAGGTAGG-3′ −131  
SnupeAR5701 5′-AGGTGGGAAGGTAAGGAGGT-3′ −56  
SnupeAR5750 5′-AAATTGTTGTATTTGTTTTTTATTTTTTAG-3′ +3  
SnupeAR5799 5′-GGGAGTTAGTTTGTTGGGAGAG-3′ +44  
SnupeAR6789 5′-GTGGGTAGTTAGTTGTAG-3′ +1030  
SnupeAR6843 5′-GAGTAAGAGAAGGGGAGG-3′ +1040  
SnupeAR6918 5′-GGGTTGGGAAGGGTTTATTTT-3′ +1162  
SnupeAR6964 5′-GAGTTTTTTAGAATTTGTTTTAGAG-3′ +1212  
SnupeAR7013 5′-GGGTTTTAGGTATTTAGAGGT-3′ +1257  
SnupeAR7160 5′-TGAGGATGGTTTTTTTTAAGTTTAT-3′ +1408  
Table 2

Methylation status of AR promoter and 5′-untranslated region in prostate cancer cellsa

Methylation status of AR promoter and 5′-untranslated region in prostate cancer cellsa
Methylation status of AR promoter and 5′-untranslated region in prostate cancer cellsa
1
Koivisto P., Kolmer M., Visakorpi T., Kallioniemi O. P. Androgen receptor gene and hormonal therapy failure of prostate cancer.
Am. J. Pathol.
,
152
:
1
-9,  
1998
.
2
Hobisch A., Culig Z., Radmayr C., Bartsch G., Klocker H., Hittmair A. Androgen receptor status of lymph node metastases from prostate cancer.
Prostate
,
28
:
129
-135,  
1996
.
3
Hobisch A., Culig Z., Radmayr C., Bartsch G., Klocker H., Hittmair A. Distant metastases from prostatic carcinoma express androgen receptor protein.
Cancer Res.
,
55
:
3068
-3072,  
1995
.
4
Van der Kwast T. H., Schalken J., Ruizeveld de Winter J. A., van Vroonhoven C. C., Mulder E., Boersma W., Trapman J. Androgen receptors in endocrine-therapy-resistant human prostate cancer.
Int. J. Cancer
,
48
:
189
-193,  
1991
.
5
Tilley W. D., Wilson C. M., Marcelli M., McPhaul M. J. Androgen receptor gene expression in human prostate carcinoma cell lines.
Cancer Res.
,
50
:
5382
-5386,  
1990
.
6
Wolf D. A., Herzinger T., Hermeking H., Blaschke D., Horz W. Transcriptional and posttranscriptional regulation of human androgen receptor expression by androgen.
Mol. Endocrinol.
,
7
:
924
-936,  
1993
.
7
Quigley C. A., De Bellis A., Marschke K. B., el-Awady M. K., Wilson E. M., French F. S. Androgen receptor defects: historical, clinical, and molecular perspectives.
Endocr. Rev.
,
16
:
271
-321,  
1995
.
8
Dai J. L., Maiorino C. A., Gkonos P. J., Burnstein K. L. Androgenic up-regulation of androgen receptor cDNA expression in androgen-independent prostate cancer cells.
Steroids
,
61
:
531
-539,  
1996
.
9
Jarrard D. F., Kinoshita H., Shi Y., Sandefur C., Hoff D., Meisner L. M., Chang C., Herman J. G., Isaacs W. B., Nassif N. Methylation of the androgen receptor CpG Island is associated with loss of androgen receptor expression in prostate cancer cells.
Cancer Res.
,
58
:
5310
-5314,  
1998
.
10
Baylin S. B., Herman J. G., Graff J. R., Vertino P. M., Issa J. P. Alterations in DNA methylation: a fundamental aspect of neoplasia.
Adv. Cancer Res.
,
72
:
141
-196,  
1998
.
11
Nan X., Ng H. H., Johnson C. A., Laherty C. D., Turner B. M., Eisenman R. N., Bird A. Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex.
Nature (Lond.)
,
393
:
386
-389,  
1998
.
12
Jones P. A., Laird P. W. Cancer epigenetics comes of age.
Nat. Genet.
,
21
:
163
-167,  
1999
.
13
Rideout W. M., III, Eversole-Cire P., Spruck C. H., III, Hustad C. M., Coetzee G. A., Gonzales F. A., Jones P. A. Progressive increases in the methylation status and heterochromatinization of the myoD CpG island during oncogenic transformation.
Mol. Cell. Biol.
,
14
:
6143
-6152,  
1994
.
14
Qian X. C., Brent T. P. Methylation hot spots in the 5′ flanking region denote silencing of the O6-methylguanine-DNA methyltransferase gene.
Cancer Res.
,
57
:
3672
-3677,  
1997
.
15
Lapidus R. G., Nass S. J., Butash K. A., Parl F. F., Weitzman S. A., Graff J. G., Herman J. G., Davidson N. E. Mapping of ER gene CpG island methylation by methylation-specific polymerase chain reaction.
Cancer Res.
,
58
:
2515
-2519,  
1998
.
16
Clark S. J., Harrison J., Paul C. L., Frommer M. High sensitivity mapping of methylated cytosines.
Nucleic Acids Res.
,
22
:
2990
-2997,  
1994
.
17
Stirzaker C., Millar D. S., Paul C. L., Warnecke P. M., Harrison J., Vincent P. C., Frommer M., Clark S. J. Extensive DNA methylation spanning the Rb promoter in retinoblastoma tumors.
Cancer Res.
,
57
:
2229
-2237,  
1997
.
18
Millar D. S., Ow K. K., Paul C. L., Russell P. J., Molloy P. L., Clark S. J. Detailed methylation analysis of the glutathione S-transferase π (GSTP1) gene in prostate cancer.
Oncogene
,
18
:
1313
-1324,  
1999
.
19
Comb M., Goodman H. M. CpG methylation inhibits proenkephalin gene expression and binding of the transcription factor AP-2.
Nucleic Acids Res.
,
18
:
3975
-3982,  
1990
.
20
Iguchi-Ariga S. M., Schaffner W. CpG methylation of the cAMP-responsive enhancer/promoter sequence TGACGTCA abolishes specific factor binding as well as transcriptional activation.
Genes Dev.
,
3
:
612
-619,  
1989
.
21
Gardiner-Garden M., Frommer M. CpG islands in vertebrate genomes.
J. Mol. Biol.
,
196
:
261
-282,  
1987
.
22
Mashal R. D., Fejzo M. L., Friedman A. J., Mitchner N., Nowak R. A., Rein M. S., Morton C. C., Sklar J. Analysis of androgen receptor DNA reveals the independent clonal origins of uterine leiomyomata and the secondary nature of cytogenetic aberrations in the development of leiomyomata.
Genes Chromosomes Cancer
,
11
:
1
-6,  
1994
.
23
Green A. J., Sepp T., Yates J. R. Clonality of tuberous sclerosis harmatomas shown by non-random X-chromosome inactivation.
Hum. Genet.
,
97
:
240
-243,  
1996
.
24
Nichols K. E., Harkin D. P., Levitz S., Krainer M., Kolquist K. A., Genovese C., Bernard A., Ferguson M., Zuo L., Snyder E., Buckler A. J., Wise C., Ashley J., Lovett M., Valentine M. B., Look A. T., Gerald W., Housman D. E., Haber D. A. Inactivating mutations in an SH2 domain-encoding gene in X-linked lymphoproliferative syndrome.
Proc. Natl. Acad. Sci. USA
,
95
:
13765
-13770,  
1998
.
25
Stoger R., Kajimura T. M., Brown W. T., Laird C. D. Epigenetic variation illustrated by DNA methylation patterns of the fragile-X gene FMR1.
Hum. Mol. Genet.
,
6
:
1791
-1801,  
1997
.
26
Warnecke P. M., Stirzaker C., Melki J. R., Millar D. S., Paul C. L., Clark S. J. Detection and measurement of PCR bias in quantitative methylation analysis of bisulphite-treated DNA.
Nucleic Acids Res.
,
25
:
4422
-4426,  
1997
.
27
Gonzalgo M. L., Jones P. A. Rapid quantitation of methylation differences at specific sites using methylation-sensitive single nucleotide primer extension (Ms-SNuPE).
Nucleic Acids Res.
,
25
:
2529
-2531,  
1997
.
28
Mizokami A., Yeh S. Y., Chang C. Identification of 3′,5′-cyclic adenosine monophosphate response element and other cis-acting elements in the human androgen receptor gene promoter.
Mol. Endocrinol.
,
8
:
77
-88,  
1994
.
29
Takane K. K., McPhaul M. J. Functional analysis of the human androgen receptor promoter.
Mol. Cell. Endocrinol.
,
119
:
83
-93,  
1996
.
30
Edelstein R. A., Carr M. C., Caesar R., Young M., Atala A., Freeman M. R. Detection of human androgen receptor mRNA expression abnormalities by competitive PCR.
DNA Cell Biol.
,
13
:
265
-273,  
1994
.
31
Choi C., Kim M. H., Juhng S. W. Loss of heterozygosity on chromosome Xp22.2-p22.13 and Xq26.1-q27.1 in human breast carcinomas.
J. Korean Med. Sci.
,
13
:
311
-316,  
1998
.
32
Glaab W. E., Tindall K. R. Mutation rate at the hprt locus in human cancer cell lines with specific mismatch repair-gene defects.
Carcinogenesis (Lond.)
,
18
:
1
-8,  
1997
.
33
Gonzalgo M. L., Hayashida T., Bender C. M., Pao M. M., Tsai Y. C., Gonzales F. A., Nguyen H. D., Nguyen T. T., Jones P. A. The role of DNA methylation in expression of the p19/p16 locus in human bladder cancer cell lines.
Cancer Res.
,
58
:
1245
-1252,  
1998
.
34
Mizokami A., Chang C. Induction of translation by the 5′-untranslated region of human androgen receptor mRNA.
J. Biol. Chem.
,
269
:
25655
-25659,  
1994
.
35
Patel S. A., Graunke D. M., Pieper R. O. Aberrant silencing of the CpG island-containing human O6-methylguanine DNA methyltransferase gene is associated with the loss of nucleosome-like positioning.
Mol. Cell. Biol.
,
17
:
5813
-5822,  
1997
.
36
Hornstra I. K., Yang T. P. High-resolution methylation analysis of the human hypoxanthine phosphoribosyltransferase gene 5′ region on the active and inactive X chromosomes: correlation with binding sites for transcription factors.
Mol. Cell. Biol.
,
14
:
1419
-1430,  
1994
.
37
Brandeis M., Frank D., Keshet I., Siegfried Z., Mendelsohn M., Nemes A., Temper V., Razin A., Cedar H. Sp1 elements protect a CpG island from de novo methylation.
Nature (Lond.)
,
371
:
435
-438,  
1994
.
38
Macleod D., Charlton J., Mullins J., Bird A. P. Sp1 sites in the mouse aprt gene promoter are required to prevent methylation of the CpG island.
Genes Dev.
,
8
:
2282
-2292,  
1994
.
39
Robertson K. D., Hayward S. D., Ling P. D., Samid D., Ambinder R. F. Transcriptional activation of the Epstein-Barr virus latency C promoter after 5-azacytidine treatment: evidence that demethylation at a single CpG site is crucial.
Mol. Cell. Biol.
,
15
:
6150
-6159,  
1995
.
40
Zhu X., Deng C., Kuick R., Yung R., Lamb B., Neel J. V., Richardson B., Hanash S. Analysis of human peripheral blood T cells and single-cell-derived T cell clones uncovers extensive clonal CpG island methylation heterogeneity throughout the genome.
Proc. Natl. Acad. Sci. USA
,
96
:
8058
-8063,  
1999
.
41
Graff J. R., Herman J. G., Myohanen S., Baylin S. B., Vertino P. M. Mapping patterns of CpG island methylation in normal and neoplastic cells implicate both upstream and downstream regions in de novo methylation.
J. Biol. Chem.
,
272
:
22322
-22329,  
1997
.
42
Gil-Diez de Medina S., Salomon L., Colombel M., Abbou C. C., Bellot J., Thiery J. P., Radvanyi F., Van der Kwast T. H., Chopin D. K. Modulation of cytokeratin subtype, EGF receptor, and androgen receptor expression during progression of prostate cancer.
Hum. Pathol.
,
29
:
1005
-1012,  
1998
.
43
Carruba G., Leake R. E., Rinaldi F., Chalmers D., Comito L., Sorci C., Pavone-Macaluso M., Castagnetta L. A. Steroid-growth factor interaction in human prostate cancer. I. Short-term effects of transforming growth factors on growth of human prostate cancer cells.
Steroids
,
59
:
412
-420,  
1994
.
44
Craft N., Shostak Y., Carey M., Sawyers C. L. A mechanism for hormone-independent prostate cancer through modulation of androgen receptor signaling by the HER-2/neu tyrosine kinase.
Nat. Med.
,
5
:
280
-285,  
1999
.
45
Toyota M., Ahuja N., Ohe-Toyota M., Herman J. G., Baylin S. B., Issa J. P. CpG island methylator phenotype in colorectal cancer.
Proc. Natl. Acad. Sci. USA
,
96
:
8681
-8686,  
1999
.
46
Umekita Y., Hiipakka R. A., Kokontis J. M., Liao S. Human prostate tumor growth in athymic mice: inhibition by androgens and stimulation by finasteride.
Proc. Natl. Acad. Sci. USA
,
93
:
11802
-11807,  
1996
.
47
Heisler L. E., Evangelou A., Lew A. M., Trachtenberg J., Elsholtz H. P., Brown T. J. Androgen-dependent cell cycle arrest and apoptotic death in PC-3 prostatic cell cultures expressing a full-length human androgen receptor.
Mol. Cell. Endocrinol.
,
126
:
59
-73,  
1997
.
48
Levenson A. S., Jordan V. C. Transfection of human estrogen receptor (ER) cDNA into ER-negative mammalian cell lines.
J. Steroid Biochem. Mol. Biol.
,
51
:
229
-239,  
1994
.