CpG methylation at promoter site −140 inactivates TGFβ2 receptor gene in prostate cancer
Article first published online: 13 MAY 2005
Copyright © 2005 American Cancer Society
Volume 104, Issue 1, pages 44–52, 1 July 2005
How to Cite
Zhao, H., Shiina, H., Greene, K. L., Li, L.-C., Tanaka, Y., Kishi, H., Igawa, M., Kane, C. J., Carroll, P. and Dahiya, R. (2005), CpG methylation at promoter site −140 inactivates TGFβ2 receptor gene in prostate cancer. Cancer, 104: 44–52. doi: 10.1002/cncr.21135
- Issue published online: 17 JUN 2005
- Article first published online: 13 MAY 2005
- Manuscript Accepted: 11 FEB 2005
- Manuscript Revised: 1 FEB 2005
- Manuscript Received: 9 NOV 2004
- National Institutes of Health (NIH). Grant Numbers: R01AG21418, R01CA1018447, R01CA1018447, T32DK07790
- Veterans Affairs. Grant Numbers: VA Merit Review, Research Enhancement Award Program (REAP)
- CpG methylation;
- prostate cancer
The action of transforming growth factor β (TGF-β) is mediated through type 1 (TβRI) and type 2 (TβRII) receptors. Prostate cancer cells are often resistant to TGF-β signaling due to loss of TβRII expression. The authors of the current study hypothesized that CpG methylation of the TβRII promoter at the Sp1 binding site −140 mediates this loss of TβRII expression in prostate cancer.
Sixty-seven prostate cancer (PC) samples, 8 benign prostatic hyperplasia (BPH) samples, and 4 prostate cancer cell lines (DUPro, LNCaP, ND-1 and PC-3) were analyzed for 1) TβRII mRNA expression by semiquantitative RT-PCR, 2) TβRII protein expression by immunohistochemistry, and 3) TGFβRII promoter methylation at CpG site −140 by methylation specific PCR and bisulfite DNA sequencing. Prostate cancer cell lines were treated with the demethylating agent 5aza2′deoxycytidine to determine if TβRII gene expression could be increased by blocking promoter methylation.
mRNA and protein expression of TβRII was lower in the PC samples than in the BPH samples. CpG methylation at site −140 was higher in PC than in BPH (P < 0.01). Promoter methylation was inversely correlated with TβRII mRNA expression in the PC and BPH samples (P < 0.0001). PC3, ND1, and DUPro TβRII mRNA expression increased following treatment of cells with 5-aza-2′-deoxycytidine.
CpG methylation of the TβRII promoter at CPG site −140 leads to functional loss of the TβRII gene in prostate cancer. Treatment with 5-aza-2′ deoxycytidine can restore gene expression. The current study results report the first association between prostate cancer and loss of the TGF- β signaling pathway by TβRII DNA promoter methylation. Cancer 2005;. © 2005 American Cancer Society.
Transforming growth factor beta (TGF-β) is a pleiotropic growth factor that inhibits epithelial cell proliferation.1 TGF-β initiates intracellular signaling by inducing the assembly of a heterotetrameric complex of two types of transmembrane receptors known as type 1 and type 2 receptors (TβRI and TβRII).2 Both type 1 and type 2 receptors have an intracellular serine/threonine kinase domain. Upon ligand-induced formation of the heteromeric complex, type 1 kinase is activated by type 2 kinase, which is constitutively active,3 The activated type 1 kinase in turn propagates the signal by phosphorylating specific intracellular proteins (members of the SMAD family).4
In prostate cancer, TGF-β synthesis is enhanced compared to prostatic intraepithelial neoplasia (PIN) and benign prostatic hyperplasia (BPH).5 However, prostate cancer cells frequently acquire resistance to TGF-β mediated growth inhibition6 and increasing levels of TGF-β are associated with prostate cancer progression.7 The mechanism underlying the resistance to TGF-β mediated growth inhibition in prostate cancer cells has not been fully investigated but may be caused by functional defects in the TGF-β signaling pathways.8 Both TβRI and TβRII are required to ensure the proper transduction of TGF-β signaling;3 therefore, functional loss of either receptor would cause resistance to TGF-β mediated growth inhibition.
In several malignancies, including prostate cancer, down-regulation of TβRII has been associated with resistance to TGF-β–mediated cell growth inhibition.9 Down-regulation of TGF-β-receptor expression has also been associated with worse clinical outcomes in patients with prostate cancer.10 Epigenetic-based promoter hypermethylation is a common mechanism by which tumor suppressors are repressed in various human cancers.11–13 However, aberrant TGF-βRII promoter hypermethylation has been detected in only 1 sample of primary esophageal cancer14 and in 3 out of 25 lung cancer cell lines.15 Other studies could not detect TGF-βRII promoter hypermethylation in various human cancer cells.16–18 Here we analyze TGF-βRII regulation in human prostate cancer by determining the methylation status of CpG site −140, a known Sp1 binding site, which lies within the first positive regulatory element of the TGF-βRII promoter.
We hypothesize that CpG island methylation of the TβRII promoter may play an important role in regulating TβRII expression in human prostate cancer. To test this hypothesis, 67 prostate cancer samples obtained from radical prostatectomy (RP) specimens, 48 BPH samples collected by transurethral resection of the prostate (TURP), and 4 prostate cancer cell lines were used to determine the relation between TGFβRII promoter methylation status and TGFβRII protein and mRNA expression. Further, we will determine whether treatment of prostate cancer cell lines with a demethylating agent could then restore or increase expression of the TGFβRII gene.
MATERIALS AND METHODS
Human Prostate Tissues
Sixty-seven clinically localized prostate cancer (PC) samples and 8 benign prostatic hyperplasia (BPH) samples were collected by RP and TURP, respectively, between June 1998 and August 2000 at Shimane Medical University. Forty additional BPH samples were collected by TURP at the San Francisco Veterans Affairs Medical Center. No patient received preoperative hormone therapy. Prostate cancer was classified as organ-confined in 21 patients and locally advanced in 46 patients. Gleason tumor grade was determined according to the General Rule for Clinical and Pathological Studies on Prostate Cancer by the Japanese Urological Association and the Japanese Society of Pathology (The Japanese Urological Association and the Japanese Society of Pathology, 1999). Distribution of Gleason scores for RP samples is as follows: Gleason < 4 (n = 20), Gleason 5–7 (n = 42), and Gleason 8–10 (n = 5). Other postoperative pathologic findings including capsular invasion, seminal vesicle invasion, and venous and lymphatic involvement were evaluated.
Half of each tissue sample was fixed in 10% buffered formalin (pH 7.0) and embedded in paraffin wax. Sections (5 μm thick) were used for hematoxylin and eosin (H & E) staining for histologic evaluation and for TGFβRII immunostaining. The remaining half of each sample from Shimane University was immediately frozen and stored at −80 °C until used. Prostate cancer localization was determined by preoperative imaging studies including transrectal ultrasound (TRUS), color Doppler ultrasound (CDUS), and magnetic resonance imaging (MRI) of the prostate. Prostate cancer specimens were excluded from this study if discordance was identified between preoperative imaging studies and postoperative pathology.
Nucleic Acid Extraction
Genomic DNA from frozen PC (67) and BPH (8) samples was extracted by a commercial DNA extraction kit (Qiagen Inc, Valencia, CA)). Genomic DNA of cell culture samples was extracted using DNAzol reagent (Molecular Research Center Inc, Cincinnati, OH). Total RNA was extracted using TRI reagent (Molecular Research Center Inc, Cincinnati, OH). The RNA pellet obtained after isopropanol and ethanol precipitation was dried and resuspended in 25 μl of RNase-free water. DNA and RNA concentrations were determined by spectrophotometer, and their integrity was checked by gel electrophoresis.
Differential Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) for TGFβRII Messenger RNA Expression
RT-PCR was performed as previously described19 with some modification. Briefly, total RNA (1μg) was reverse transcribed using an oligo-dT primer and M-MLV RT (Promega, Madison, WI) in a 20 μL reaction, followed by a 1:5 dilution. The resulting complementary DNA was amplified by PCR using primers specific for the TβRII gene: hTβRIIf2 (forward); 5′-CCA ATA TCC TCG TGA AGA ACG-3′ and hTβRIIr2 (reverse); 5′-CAG CAT TCT CCA AAT TCA TCC-3′ or β-actin: hβ-actinF (forward); 5′-CCA CGA AAC TAC CTT CAA CTC C-3′ and hβ-actinR (reverse); 5′-CAT ACT CCT GCT TGC TGA TCC-3′. The PCR mixture was prepared in a 20 μL volume containing 2 μL PCR buffer (Sigma Chemical Co., St. Louis, MO), 0.2 μL dNTP (10 μM), 0.2 μl TβRII primers (10 μM), 0.1 μL β-actin primers (10 μM) and 1 μL Taq polymerase (Sigma Chemical Co.). PCR reactions were performed in a PTC-200 thermal cycler (MJ Research Inc., Watertown, MA) under conditions as follows: 94 °C for 2 minutes, 25 cycles at 94 °C for 30 seconds, 60 °C for 30 seconds, and 72 °C for 30 seconds; followed by final extension at 72 °C for 5 minutes.
For semiquantitative analysis of amplified products, the authors prepared three additional dilutions of each sample and performed RT-PCR using these dilutions. We adjusted the PCR cycles until these four sample dilutions were within linear range. PCR products were electrophoresed on 2.0% agarose gels, and the expression level of these genes was evaluated by an AlphaImager 2200 (Alpha Innotech Corporation, San Leandro, CA). TGFβRII expression relative to β-actin expression was quantified and expressed as arbitrary units (AUs).
Immunohistochemistry of TβRII
Immunostaining was performed on formalin-fixed, paraffin-embedded sections obtained from 67 PC and 48 BPH samples. Thick tissue sections measuring 5 μm were deparaffinized by xylene and rehydrated in graded alcohols for 3 minutes. Slides were pretreated by microwave in 10 mM sodium citrate buffer for 20 minutes. After overnight incubation at 4 °C with primary TGFβRII antibody (Santa Cruz Biotech, Santa Cruz, CA) at a 1:100 dilution, the staining process was performed using a commercial kit (Santa Cruz Biotech). As a negative control, the corresponding blocking peptide (Santa Cruz Biotech) was used instead of the primary antibody. Sections were developed with diaminobenzidine (DAB) (Sigma, Saint Louis, MO) and counterstained using hematoxylin. In each specimen, the percentage of cytoplasmic staining was calculated within a total of 1000 cells examined per region by two different observers using a 40× objective lens. Samples were classified as negative (< 5% positive cells), weak (5–30% positive cells), or positive (> 30% positive cells) for immunostain.
Methylation Specific PCR (MSP) and Unmethylation Specific PCR (USP) Using Bisulfite-Modified DNA as Templates
Genomic DNA (1 μg) was modified with sodium bisulfite according to a method previously described.20 Modified DNA was amplified by PCR using the primer sets shown in Figure 1. First round PCR was performed using the universal primers TGFβRIIpF2 (forward: 5′-TTA GGA GTA ATT TGA AGA AAG TTG AGG-3′) and TGFβRIIpR2 (reverse: 5′-TAC CAA TCA TAT TTC CTA AAC CAA C-3′) in a volume of 20 μL. PCR reactions were performed at 94 °C for 2 minutes as an initial denaturing step with the next 40 cycles at 94 °C for 30 seconds, 55 °C for 30 seconds, 72 °C for 30 seconds, and a final extension at 72 °C for 5 minutes.
For the nested second PCR used to detect methylation status at the −140 binding site, the methylation specific primer TβRIIpMR2 for MSP (5′-CTT TCA ACT ACC CCT CAC CG-3′) or the unmethylation specific primer TβRIIpUR2 for USP (5′-CTT TCA ACT ACC CCT CAC CA-3′) was used as the reverse primer, and TβRIIpF2 was the common forward primer for MSP and USP. For semi-quantitative analysis, we prepared four dilutions of each sample and used them for MSP and USP. We determined cycle number by adjusted PCR cycles until these four dilutions of each sample were within linear range.
PCR was performed at 94 °C for 2 minutes; 27 cycles at 94 °C for 30 seconds, 65 °C for 30 seconds, and 72 °C for 30 seconds; followed by a final extension at 72 °C for 5 minutes. The PCR products obtained were electrophoresed on 2.0% agarose gel, and densitometric analysis of MSP and USP bands in each sample was performed using an AlphaImager 2200 (Alpha Innotech Corporation, San Leandro, CA).
Methylation level was determined by M/(M + U), where M and U are the density of each USP and MSP band, respectively, expressed as a percentage. The density of the MSP band (M) reflects the amount of methylated TβRII gene at a specific site, whereas the density of the MSP plus USP bands (M + U) reflects the total amount of TGFβRII gene at a specific site.
Bisulfite DNA Sequencing Analysis
To confirm the methylation level at the CpG site −140 in the prostate cancer cell lines, bisulfite-modified DNA was amplified using a pair of universal primers as follows: forward primer TGFβRIIpF1 (5′-TTA TTT TTG GAT TTT AAT TTG TAA T-3′) and reverse primer TGFβRIIpR1 (5′-AAC TAA ACA AAA CCT CTC TCC-3′). Bisulfite-modified DNA (1 μL) was used as a template in a total volume of 20 μL PCR mixture. The PCR reaction was carried out at 94 °C for 2 minutes, 30 cycles at 94 °C for 30 seconds, annealing at 55 °C for 30 seconds, and extension at 72 °C for 30 seconds, and final extension at 72 °C for 5 minutes. Purified PCR products were used as templates for direct DNA sequencing using a reverse primer (TβRIIpR1 or TβRIIpR2) and the dye terminator method according to the manufacturer's instructions (Applied BioSystems, Foster City, CA). The methylation level at each CpG site was determined by the ratio between peak values of G and A (G/[G + A]). The percentage of methylation was calculated by G/[G + A], where G and A were peak levels on the electropherogram of bisulfite-DNA sequencing.18 Methylation levels were defined as complete (> 90%), partial (10–90%), or none (< 10%).
Cell Culture and 5-aza-2′-deoxycytidine (5-aza-dC) Treatment
Four human prostate cancer cell lines (DuPro, LNCaP, PC3, and ND1) were obtained from the University of California San Francisco Cell Culture Facility (San Francisco, CA). For 5-aza-dC treatment, exponentially growing cells were subcultured on Day 0 in RPMI 1640 medium supplemented with 10% fetal calf serum. Cells were treated with a final concentration of 10 μM 5-aza-dC for 3 nonconsecutive days (i.e., Days 1, 3, and 5) and washed and grown in regular medium on Days 2 and 4. On Day 6, the cells were harvested and nucleic acids were extracted.
The relation between pathologic variables with parametric values was analyzed using the ANOVA test followed by Scheffe multiple comparison test. Correlation between two parametric variables was determined using Pearson coefficient correlation. A P < 0.05 was considered statistically significant.
Immunostaining of TGFβRII in Human Prostate
Typical immunostaining of TGFβRII in BPH and prostate cancer tissue is shown in Figure 1. All 8 BPH tissues showed positive TβRII immunoreactivity, which was predominantly localized to the cytoplasm of the epithelial cells (Fig. 1A). There was very little immunoreactivity of TGFβRII in stromal cells. In PC tissues, 4% (3 out of 67) showed positive staining, 9% (6 out of 67) showed weak staining, and 87% (58 out of 67) showed negative staining.
Effects of 5-aza-dC Treatment on the Expression of TβRII and Its Relation to Density of MSP and USP Bands in Prostate Cancer Cell Lines
Figure 2 shows that after 5-aza-dC treatment, expression of TβRII mRNA was increased in DUPro, ND-1, and PC-3 cell lines. In DuPro, ND-1, and PC-3 cell lines, methylation analysis at the −140 site, a known Sp1 binding site in the first positive regulatory element of the TGFBRII gene, showed increased TβRII mRNA expression after 5-aza-dC treatment. Increased mRNA expression was accompanied by a simultaneous increase in USP density and a decrease in MSP density. Conversely in LNCaP cell lines, there was no expression of TβRII mRNA before or after 5-aza-dC treatment, although a weak MSP band before treatment was lost after 5-aza-dC treatment (Figure 2).
Bisulfite Modified DNA Sequencing in Prostate Cancer Cell Lines and Prostate Cancer Tissues
Bisulfite-DNA sequencing in prostate cancer cell lines before 5-aza-dC treatment is summarized in Figure 3B. In primary prostate cancer tissues, the most frequently affected site of CpG methylation within the promoter region is the −140 nucleotide site (Fig. 3C).
Expression of mRNA Transcript of TβRII Gene in Prostate
The mean expression level of TβRII mRNA was significantly lower in prostate cancer samples (0.08 ± 0.07 AU) than in BPH samples (0.50 ± 0.25 AU)(P < 0.01). There was no significant difference in expression level of TβRII mRNA between organ-confined and locally advanced prostate cancers. Similarly TβRII mRNA expression did not correlate with Gleason score, capsular invasion, seminal vesical invasion, or lymphovascular invasion.
Methylation Analysis at the −140 CpG Site by MSP and USP in Prostate Tissues
All 8 BPH samples showed faint or absent MSP bands and strong USP bands (Fig. 4), whereas in PC tissues, all 67 samples showed both MSP and USP bands (Fig. 5B). Bisulfite-modified DNA sequencing using flanking primers, TβRIIpF2 or TβRIIpR2, confirmed the validity of the MSP and USP results (Fig. 5B). Methylation levels at the −140 site as determined by densitometric analysis (M/[M + U]) was inversely correlated with TβRII mRNA expression (P < 0.0001) (Fig. 6). Significantly higher levels of methylation at this site were found in PC samples when compared to BPH samples (P < 0.01). However, there was no significant difference in methylation status between organ confined and locally advanced prostate cancers (Fig. 5C) or among Gleason scores (Fig. 5D).
Quantification of Methylation Status Determined by Bisulfite DNA Sequencing and by Density of MSP and USP Bands
Methylation levels at the −140 CpG site were quantified by measuring peak levels of G and A on bisulfite DNA sequencing (G/[G + A]) and by density of MSP and USP bands (M/[M + U]) (Fig. 6). These two methods showed positive correlation, although sequencing provided a better quantitative assessment of methylation. Mean methylation levels in BPH and PC samples were 0.2 ± 0.6% and 38.8 ± 20.9%, respectively, at the −140 CpG site.
TGF-β is a potent tumor suppressor that represses cancer proliferation and tumor growth. Many cancers become resistant to TGF-β growth inhibition due to suppression of TβRII, an essential component of the TGF-β signaling pathway.1, 21–23 Several studies have identified somatic mutations in the TβRII gene that repress TβRII expression in human cancers and lead to TGF-β resistance.24–27 Here, we report an additional mechanism by which TβRII is suppressed. We find that in the human prostate gland, the TβRII promoter is aberrantly methylated in a cancer-dependent fashion. We also find that promoter methylation suppresses TβRII expression, demonstrated by the lack of protein expression seen on immunohistochemical studies in prostate cancer samples. Together, these findings imply that promoter methylation mediates cancer-dependent suppression of TβRII in prostate tumors. Such suppression can lead to TGF-β resistance, a common phenomenon in prostate cancer that is associated with advanced cancer stage and poor patient outcome.28–31
Although several other studies have identified a correlation between TβRII levels, Gleason sum, and pathologic stage,9, 32 our results failed to achieve statistical significance. Unlike other studies, however, we correlated mRNA expression rather than immunohistochemical data with clinical parameters. Other explanations for our findings may be the Gleason grade distribution of our cohort, with few patients having high grade tumor (> Gleason 7) or seminal vesical invasion, and differences in grouping Gleason grade for statistical analysis.
We believe that this is the first report to identify TβRII promoter methylation as a common occurrence in human prostate cancer. This finding is, perhaps, surprising because previous reports analyzing cancers at other sites have implied that epigenetic regulation of TβRII is not common.14–18 We, however, observe TβRII promoter methylation in almost all (98%) of prostate cancer specimens analyzed. In addition, TβRII promoter methylation exhibits cancer specificity because it is observed in only 13% of BPH specimens.
Further, we found that TβRII suppression was reversible. After 5-aza-dC treatment, the expression level of TβRII mRNA transcripts was increased in three PC cancer cell lines (DuPro, ND-1, and PC-3), an increase in mRNA that paralleled the changes observed in MSP and USP before and after 5-aza-dC treatment.
Both the specificity of TβRII methylation in prostate cancer and its reversibility have important clinical implications for prostate cancer screening and treatment. Future work can determine if TβRII promoter methylation occurs in cancers at other sites, or, alternatively, if it is prostate cancer-specific.
Inactivation of TβRII by CpG methylation at the −140 site is commonly observed in PC leading to functional loss of the TβRII gene. Our study provides a valuable insight into the mechanism underlying the pathogenesis of prostate cancer, its relation to functional loss of the TGF-β signaling pathway, and provides an exciting new target for prostate cancer therapy.