Global Methylation Pattern of Genes in Androgen-Sensitive and Androgen-Independent Prostate Cancer Cells Free

Epigenetic mechanism is a vital event in the transcriptional regulation of various genes in eukaryotes. It includes aberrant DNA methylation and histone modification at multiple levels and plays important role in normal developmental processes, gene imprinting, and human carcinogenesis (1). Aberrant DNA methylation of CpG (cytosine preceding guanosine) sites is among the earliest and most frequent alterations in cancer (2, 3). This modification has important regulatory effects on gene expression, especially when involving CpG-rich areas known as CpG islands, located in the promoter regions of many genes. In many cases, aberrant methylation of the CpG island in genes has been correlated with a loss of gene expression and function (4, 5). Markers for aberrant methylation may represent a promising avenue for monitoring the onset and progression of cancer. Identification of promoter methylation of several genes in small biopsies and bodily fluids of cancer patients has proven to be useful as a molecular tool for cancer detection and progression (6).

Prostate cancer is the second leading cause of cancer death among men in the United States and Europe (7). This disease is associated with considerable morbidity and mortality, but curative treatment (radical prostatectomy or radiotherapy) is feasible for patients with the early-stage disease (8). However, our understanding of the epigenetic changes that underlie the progression of this disease remains at an early stage. Association studies predicted several genetic factors associated with risk of prostate cancer in different populations, but there is still scarcity of data on the epigenetic events. A better understanding of the molecular and epigenetic changes in prostate cancer is likely to contribute to improved diagnosis, clinical management, and better treatment outcomes. A recent study has shown that smoking influences aberrant CpG hypermethylation of several genes in prostate cancer (9).

There are several molecular methods such as methylation-specific PCR (MSP), methylation-specific digestion, and bisulfite sequencing for methylation analysis. However, they can screen only a few sites of methylation at one time. In our current study, we screened the promoter methylation pattern of 82 genes (at one time) using an array method in three prostate cancer cell lines (LNCaP, DU145, and PC3) and in one normal prostate cell line (RWPE1). We investigated promoter methylation patterns in various categories of genes, such as cell cycle regulators, transcription factors, tumor suppressor, and genes involved in tumor growth and progression. Moreover, an analysis was done to explore the difference in methylation pattern between the androgen-sensitive and androgen-independent prostate cancer cells. A distinct methylation pattern was observed between normal and prostate cancer cell lines as well as androgen-sensitive and androgen-independent cell lines. This study will provide the opportunity to investigate the potential role and mechanism of novel genes in prostate cancer development and progression.

Human prostate normal and cancer cell lines (RWPE1, RWPE2, LNCaP, DU145, and PC3) were obtained from the American Type Culture Collection (ATCC). RWPE1 and RWPE2 are normal prostatic epithelial cell lines, which were isolated from a histologically normal adult human prostate. These cells are androgen receptor (AR), p53, and pRb positive. Monolayer cultures were maintained in epithelial cell enrichment medium (Epi-media: DMEM/F12 with 5% horse serum, 20 ng/mL epidermal growth factor, 10 μg/mL insulin, 100 ng/mL cholera toxin, 500 ng/mL hydrocortisone, and 15 mmol/L HEPES). Prostate cancer cells, LNCaP, are androgen-sensitive (AR-positive) cells derived from a patient with metastatic site to the left supraclavicular lymph node. DU145 and PC3 are androgen-independent (AR-negative) cells. They secrete prostate-specific antigen and kallikrein-2. PC3 and DU145 cells are derived from a patient with grade 4 prostatic adenocarcinoma that had metastasized to the brain and bone, respectively. These DU145 and PC3 cells were maintained in DMEM/F12 containing 10% fetal bovine serum and 4 mmol/L l-glutamine. Cells were harvested when they reached 80% confluence.

Genomic DNA from cell lines was isolated by QIAamp DNA Mini kit (Qiagen, Inc.). Methylation profile study was carried out using Promoter Methylation Array kit (Panomics, Inc.). Briefly, 2 μg of genomic DNA were digested with 10 units MseI (New England Biolabs) to produce small fragments of DNA, which retain the CpG islands and were purified. These DNA fragments were further ligated with linker DNA for PCR analysis. In the next step, DNA was incubated with methylation binding protein (MBP) in the presence of binding buffer at 15°C for 30 min, which forms a protein/DNA complex. The methylated DNA fragments were isolated by centrifugation using a separation column. Next, the purified methylated fragments were converted into the biotinylated probe by PCR amplification in the presence of biotin-dCTP for 30 cycles at 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min. During PCR, hybridization membranes were pretreated with pretreatment buffer (provided with the kit) at room temperature and then biotinylated probes were hybridized to the methylation array membrane in the presence of hybridization buffer (provided with kit) at 50°C for overnight in a rotating hybridization oven. The next day, the membrane was washed and incubated in blocking buffer for 15 min and then streptavidin–horseradish peroxidase conjugate was added to the membrane. After washing, the membrane was incubated with 1× detection buffer for 5 min at room temperature and then the substrate was added as provided with the kit. The images were developed with chemifluorescence reagent. Spot intensities on the membrane were determined by using Quantity One Fluor-S Imaging device (Bio-Rad). The promoter microarray contains duplicate spots of 82 gene promoter sequences, 28 positive controls, and 12 nonpromoter controls (Table 1). Microarray data were normalized within array normalization using positive controls (mean volume absorbance of gene of interest/mean volume absorbance of positive controls). A relative methylation fold was obtained in LNCaP, PC3, and DU145 compared with the corresponding gene methylation absorbance in RWPE1 cell line.

Promoter methylation PCR was done to validate the quantitative results obtained from the Panomics array. We randomly selected up to seven genes and did the promoter methylation PCR. Genomic DNA (2 μg) was digested with 10 units MseI and incubated with MBP to form a protein/DNA complex (as above). Methylated DNA was further separated through column and amplified with promoter-specific primers (Table 2) for CASP8, CD14, HOXA2, MASPIN, MDR1, TMS1, and STAT1 at the following program: 94°C for 5 min, 94°C for 1 min, 56°C for 1 min, and 72°C for 2 min for 35 cycles.

Furthermore, to confirm and validate the methylation results from Panomics array, we did the MSP of CASP8, MASPIN, MDR1, and TMS1 genes followed by bisulfite sequencing of TMS1 and MDR1 genes. Bisulfite modification of 2 μg of DNA was done using EZ DNA Methylation kit (Zymo Research). Briefly, DNA was mixed with 130 μL of CT conversion reagent and kept in PCR machine at 98°C for 10 min, 64°C for 2.5 h, and finally at 4°C for 1 h. Next, these samples were added to Zymo-Spin column containing 600 μL of M-binding buffer. Samples were centrifuged at 13,000 × g for 30 s and washed with M-wash buffer. For desulfonation, 200 μL of M-Desulfonation buffer was added to column, incubated for 20 min, and again centrifuged at 13,000 × g for 30 s. The column was washed twice with washing buffer and then modified DNA was eluted in 10 μL of M-elution buffer. The modified DNA was used for MSP for CASP8, TMS1, MASPIN, and MDR1 and for bisulfite sequencing for TMS1.

RNA was extracted by Trizol (Invitrogen) method. cDNA was prepared using Thermoscript RT kit (Invitrogen) and 2 μL of each cDNA were used for real-time PCR assay using SYBR Green. Gene expression of few genes, such as CASP8, TMS1, MDR1, and MASPIN, was analyzed before and after the treatment with 5 μmol/L of demethylating agent 5-aza-2′-deoxycytidine (AZC) for 72 h.

We have calculated the mean ± SD (absorbance) of methylation fold of each gene in normal prostate and prostate cancer cell lines.

The Panomics promoter methylation array technique enabled us to screen the methylation status of 82 genes in prostate cancer cell lines at one time. This technique was quite simple and sensitive. In this genome-wide global methylation assay, >50% of the 82 genes were methylated in prostate cancer cell lines compared with 13% methylated genes in normal prostate cell line (i.e., RWPE1). Androgen-independent cell lines DU145 and PC3 expressed 64.6% and 61% methylated genes, respectively, compared with 53.7% methylated genes in androgen-sensitive cell line (i.e., LNCaP cells; Fig. 1). Furthermore, when the methylated genes were subgrouped according to their function, we found an increased methylation in number of genes belonging to tumor suppressor category followed by transcription factors, cell cycle, angiogenesis, immune-related, apoptosis, and transporter genes.

Normal prostate cells (RWPE1 and RWPE2) also expressed methylated genes albeit in smaller numbers: 2.4% tumor suppressor and oncogenes, 1.2% immune-related and apoptotic genes, and ∼4.9% belonging to other categories. No hypermethylation was observed in cell cycle, angiogenic, and transporter genes in normal prostate cells. The androgen-sensitive prostate cancer cell LNCaP showed a significant increase in hypermethylation of genes that belong to tumor suppressor, transcription factors (9.8%), immune-related (6.1%), oncogenes (3.6%), cell cycle (3.6%), angiogenesis (4.9%), and transporter genes (3.7%). AR-negative prostate cancer cells DU145 and PC3 showed similar pattern as the AR-sensitive prostate cancer cells, with the exception of a greater number of hypermethylated genes belonging to the transcription factor (11–12%) and cell cycle (6–7%) categories. These observations suggest that there are selective differences in patterns between normal and cancer cells and androgen-sensitive and androgen-independent prostate cancer cells.

Furthermore, we compared the mean hypermethylation intensity in androgen-sensitive and androgen-independent prostate cancer cells in reference to noncancer prostate cell RWPE1 (cutoff value, 5-fold). The data from RWPE2 normal prostate cells were similar to RWPE1; hence, these data are not shown. Of 82 genes screened, 58 were hypermethylated to different levels in androgen-sensitive and androgen-independent prostate cancer cells. Some of the candidate genes belonging to different functional categories according to their function are listed in Table 3.

We observed a considerable difference in the promoter methylation array pattern between normal prostate cells (RWPE1) and prostate cancer cells (LNCaP, DU145, and PC3). Figure 2A represents a heat map of all the 82 genes showing extent of methylation in all cancer cell lines compared with normal RWPE1. The genes are grouped according to their relative methylation fold: ≥10-fold, 5- to 10-fold, 2- to 5-fold, and <2-fold. Although their relative intensity of methylation varied from one cell type to other, they exhibited similar pattern of methylation except for few genes. Tumor suppressor genes such as TMS1, RB, RBL1, DAPK, SFN, 14-3-3σ, PAX6, FHIT, and SRBC were highly methylated (>10-fold), whereas MASPIN and DBCCR1 were moderately methylated (5- to 6-fold). Transcription factors such as STAT5, MYOD, NPAT, ATF2, WT1, HOXA2, SYBL1, and STAT1 and cell cycle and DNA repair genes such as p27KIP1, TP73, CDKN2A, p21, and K-RAS were also highly to moderately methylated in prostate cancer cell lines. Genes belonging to transporters (MDR1, MLC1, CFTR, and MTX-1), immune system (IFN, KIR2DL4, GAGE1, BAGE, LAGE-1, and IL4), and angiogenesis (TIMP3, TSP1, and PAI-1) were moderately to highly methylated in prostate cancer cells. Methylation in some of these genes, such as PAX6, SRBC, ATF2, STAT5, LAGE-1, NPAT, CFTR, MTX-1, IL4, and RIOK3, in prostate cancer was not reported thus far.

Increased DNA methylation was observed in AR-negative cells (DU145 and PC3) compared with the AR-positive cell (LNCaP). CASP8, GPC3, CD14, MGMT, IFN, GATA3, H-RAS, SIM2, MDR1, IGRP, WT1, CDKN2A, and PDGF-B genes showed >2-fold methylation in AR-independent cell lines compared with androgen-sensitive cell lines. In contrast, decreased methylation was observed in some tumor suppressor genes (RBL1, RB, TMS1, and DAPK), transcription factors (HOXA2, STAT5, and STAT1), cell cycle genes (CyclinD2 and K-RAS), and angiogenic genes (PAI-1 and TIMP3; Fig. 2B).

Promoter methylation PCR by Panomics kit confirms that genes present on the array are methylated (Fig. 3). CASP8, HOXA2, and MDR1 shows the presence of some methylated copies in RWPE1 cells. Our bisulfite MSP results showed that MASPIN and MDR1 are completely unmethylated, whereas CASP8 and TMS1 are expressed in both methylated and unmethylated forms in normal cell line (RWPE1; Fig. 4A). These sets of genes are strongly methylated in cancer cell lines compared with normal cell line (RWPE1). Bisulfite sequence analysis of TMS1 also suggested that the promoter is unmethylated in RWPE1 cell lines, completely methylated in LNCaP and DU145, and partially methylated in PC3 (Fig. 4C and D). Bisulfite sequence analysis of MDR1 gene also showed one methylated site in RWPE1, two methylated sites in DU145, and four methylated sites in PC3 cells. LNCaP cells showed only partially methylated sites. After 48 hours of treatment with AZC, several genes became demethylated; however, the reversal to demethylation was not complete (Fig. 4B). The presence of methylated bands suggests that genes were not completely demethylated after 48 hours of treatment.

One of the most obvious outcome from DNA methylation is reduced mRNA expression of that gene. Our RNA expression data support that these genes are methylated in cancer cell lines with a corresponding decrease in their gene expression compared with normal RWPE1 cell line (Fig. 5A). Gene expression was also restored for MDR1, TMS1, and MASPIN genes after AZC treatment (Fig. 5B). Only CASP8 gene showed nonsignificant restoration in LNCaP and DU145 cell lines (Fig. 5B), whereas in PC3 its expression was reduced.

We have shown the effect of DNA methylation on target gene MDR1 in AR-independent (PC3 and DU145) and AR-positive sensitive cell (LNCaP) lines, as well as in a normal prostate cell line RWPE1 (Fig. 6). Our data clearly show functional differences in reexpression of MDR1 on inhibition of DNA methylation by AZC and subsequent treatment with docetaxel. RWPE1 nontumorigenic prostate cells express good levels of constitutive MDR1 protein, as assessed by immunofluorescent analysis. Pretreatment with AZC followed by docetaxel treatment had no significant effect on MDR1 expression. In contrast, AR-negative prostate cancer cells PC3 and DU145, which had significant loss of MDR1 due to its promoter methylation, showed a significant restoration of MDR1 on AZC treatment. These cells showed further increase in MDR1 expression in response to AZC plus docetaxel treatment. Interestingly, LNCaP (AR-positive or AR-responsive cells) did show some response in restoring MDR1 expression. However, the restoration was not as marked as for PC3 and DU145 (AR-negative cells). These observations suggest that there are functional differences between AR-positive and AR-negative prostate cancer cells in their response to inhibitors of DNA methylation and chemotherapeutic agent docetaxel.

Because several target genes, including MDR1, were silenced due to DNA methylation, we examined the effect of docetaxel on normal and prostate cancer cells, where MDR1 was restored by inhibiting DNA methylation. Supplementary Fig. S4 shows that docetaxel alone or after treatment with AZC had no effect on growth inhibition of normal prostate cells RWPE1. In addition, docetaxel alone had very little effect on growth inhibition of prostate cancer cells PC3 and DU145. However, when prostate cancer cells were pretreated with AZC for 24 hours, there was a significant increase in docetaxel-induced growth inhibition of both prostate cancer cells.

In this study, we investigated the differential DNA methylation pattern between cancer versus noncancer and androgen-sensitive versus androgen-independent prostate cancer cells by a TranSignal promoter methylation array technology (Panomics). This technique provides a high-throughput analysis of promoter methylation of 82 genes simultaneously. The principle behind this technique is the isolation of methylated DNA from whole genomic DNA using MBP and hybridization with a DNA array containing complementary oligos with regions of the corresponding promoter gene. The membrane consist of sets of genes, representing specific cellular functions such as tumor suppressor, transcription factor, cell cycle, and angiogenesis. This is a rapid method designed to detect methylation status of 82 of genes at one time using less amount of DNA compared with other methods of methylation analysis such as MSP and bisulfite sequencing (10, 11). Thus, this high-throughput screening method has distinct advantage in analyzing abnormal gene expression in various human cancers and diseases. It will help in identifying epigenetic markers involved in cancer development and progression and also provide targets for epigenetic therapy.

The molecular events involved in neoplastic initiation and progression of prostate cancer are poorly understood despite the recognition of various events during prostate cancer tumorigenesis. Each cancer cell type has several methylated genes but the methylation pattern of individual type of tumor is different (12). In prostate cancer, most of the studies reported methylation in genes such as GSTP1, APC, RASSF1A, RARb2, CRBP1, TIMP3, MGMT, and PTGS2 and their association with progression of prostate cancer (13–17). These genes can be seen as only a partial picture of the methylation changes; there may be many more genes that need to be deciphered.

Our data showed methylation of various tumor suppressor genes, such as TMS1, RB, RBL1, PAX6, FHIT, DAPK, SRBC, and SFN. These genes were highly methylated in both androgen-sensitive as well as in androgen-independent cells. TMS1 (target of methylation-induced silencing), also known as ASC [apoptosis speck-like protein containing a caspase recruitment domain (CARD)], encodes for a CARD-containing regulatory protein and has been shown to promote apoptosis directly and by activation of downstream caspases (18). Das et al. (19) showed that methylation-mediated silencing of TMS1/ASC is a frequent event in prostate cancer. RB, RBL1, DAPK, and FHIT are tumor suppressor genes that showed mild to high methylation in cancer cell lines and in prostate cancer tissues. SRBC is a newly identified tumor suppressor gene that is also hypermethylated in prostate cancer cell lines. Various studies reported the hypermethylation of p27/KIP1, p21, and CDKN2A in prostate cancer (20, 21). Our study compliments these observations.

Angiogenesis is an important step in tumor progression. We observed a high to moderate methylation in angiogenic genes PAI-1, TIMP3, THBS2, and TSP1. Yegnasubramanian et al. (22) also reported hypermethylation in TIMP3 in PC3 and DU145 cell lines. Kang et al. (15) observed a moderate methylation in THBS1, thus supporting our results.

Our finding suggests a greater hypermethylation in transcription factor genes STAT1, ATF2, HOXA2, MYOD, STAT5, and SYBL1 in prostate cancer cells. Although there are no reports on STAT1 methylation in prostate cancer, a promoter hypermethylation was observed in squamous cell carcinogenesis (23). POU3F1, NPAT, and CIITA were also moderately methylated in prostate cancer cell lines. Epigenetic regulation of MYOD in colorectal cancer (24) and CIITA (25) has been reported but no study reported the methylation in POU3F1, NPAT, STAT5, and HOXA2 thus far in prostate cancer.

Membrane transporters and metabolizing genes also play a major role in transport of drugs across the plasma membrane and their subsequent metabolism. We observed >10-fold of methylation in MDR1 gene in AR-negative cell lines (DU145 and PC3). MDR1 is a well-known multidrug resistant gene. Our study confirms other studies that showed hypermethylation of MDR1 in prostate cancer (22, 26). MDR1 regulates the trafficking of drugs, peptides, xenobiotics, and ions across cell membranes. Its expression correlates with resistance to hormone therapy and is thought to be important in the progression of primarily hormone-sensitive malignancies such as prostate cancer. CFTR, MTX, GPC3, and MLC1 were also found to be methylated in our study. As far as metabolizing genes are concerned, most of the studies found hypermethylation in GSTP1 gene in 90% of prostate tumors (15–17, 22). Our MSP and bisulfite sequencing data also confirm methylation of GSTP1 in all three cancer cells but not in normal prostate cells RWPE1. In addition, we observed the methylation in G6PD gene, which is associated with glucose metabolism.

Our data show methylation of immune-related genes and hormone receptors such as IFN-γ, IL4, BAGE, GAGE1, KIR2DL4, LAGE-1, and PR in prostate cancer cells. Other interesting genes, such as RIOK3, NES-1, and POMC, were hypermethylated in prostate cancer cells. However, their exact function needs further elucidation. The biochemical and physiologic significance of methylation of these genes in prostate cancer needs to be determined.

We also observed differences in methylation pattern between two AR-independent cell lines (DU145 and PC3). These differences may be associated with their tumor signatures based on their origin, individual phenotype, and genotype. Differential methylation pattern between androgen-sensitive and androgen-independent cancer cell suggests an important avenue for targeting receptor-associated genes in cancer. Wang et al. (27) also observed a difference in methylation pattern of TGFBR2 between LNCaP and PC3 cell lines. Yamada et al. (28) reported a difference in methylation pattern in MFPC7 gene between LNCaP and DU145. A study by Yu et al. (29) in prostate cancer samples and LNCaP, DU145, and PC3 cell lines also observed a dramatic difference in methylation pattern. Thus, our study compliments these reports showing a difference in methylation pattern between androgen-sensitive versus androgen-independent cell lines. In contrast, Yegnasubramanian et al. (22) analyzed methylation pattern by real-time MSP in prostate tissues and cell lines. These authors observed no significant difference in methylation pattern between various genes in different prostate cancer cell lines. However, we observed some differences between two androgen-independent prostate cells PC3 and DU145. Yu et al. (29) also observed a difference in methylation pattern between PC3 and DU145 cell lines in few genes such as CSPG4, CTDP1, and DSIN; however, they did not analyze their data quantitatively.

In conclusion, our study identified various candidate genes belonging to tumor suppressor, cell cycle, transcription factor, and angiogenesis by a promoter methylation array that allows the rapid detection of methylation status of 82 genes at the same time.

Identification of differences in methylation pattern between androgen-sensitive and androgen-independent prostate cells may provide opportunities to investigate mechanisms of target genes in the diagnosis, prognosis, and treatment of prostate cancer patients. However, additional investigations on the role of DNA methylation and its effect on transcriptional and translational level of regulatory genes should be done on prostate cancer and normal tissues from patients. An important limitation of our study is analysis of methylation profile in in vitro ATCC cells. Methylation pattern may vary in patient's samples that are influenced by a variety of additional factors, such as age, ethnicity, tumor stage, disease progression, and treatment. Hence, there is a need for confirming methylation status of these genes in prostate cancer specimens. Further study on tissue specimens and correlation of the pattern of methylation of these genes with pattern of gene expression silencing may help to define the underlying mechanism of tumor development and switching of prostate cells from AR sensitive to AR independent. Thus, the identification of methylation pattern in different genes may open opportunities for epigenetic therapy. Regulation studies confirmed that inhibition of DNA methylation by AZC can restore target gene and its protein expression. As an example, we investigated the expression of MDR1, an important multidrug transporter. MDR1, which was lost in prostate cancer cells, was restored on AZC treatment, and these cells became more sensitive to growth inhibition by docetaxel.

No potential conflicts of interest were disclosed.

Grant Support: NIH/National Cancer Institute grants U56 CA101599-01 (J.V. Vadgama) and CA15083-25S3 (J.V. Vadgama); NIH/National Institutes of Diabetes, Digestive and Kidney Diseases grant R25DK067015-01 (J.V. Vadgama); and Department of Defense Breast Cancer Research Program grant BC043180 (J.V. Vadgama).

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.