Quantitation of GSTP1 Methylation in Non-neoplastic Prostatic Tissue and Organ-Confined Prostate Adenocarcinoma

Transcription

1 Quantitation of GSTP1 Methylation in Non-neoplastic Prostatic Tissue and Organ-Confined Prostate Adenocarcinoma Carmen Jerónimo, Henning Usadel, Rui Henrique, Jorge Oliveira, Carlos Lopes, William G. Nelson, David Sidransky Background: Methylation of regulatory sequences near GSTP1,which encodes the class glutathione S-transferase,is the most common epigenetic alteration associated with prostate cancer. We determined whether the quantitation of GSTP1 methylation in histopathologically distinct prostate tissue samples could improve prostate cancer detection. Methods: We used a fluorogenic real-time methylation-specific polymerase chain reaction (MSP) assay to analyze cytidine methylation in the GSTP1 promoter in prostate tissue samples from 69 patients with earlystage prostatic adenocarcinoma (28 of whom also had prostatic intraepithelial neoplasia lesions) and 31 patients with benign prostatic hyperplasia. The relative level of methylated GSTP1 DNA in each sample was determined as the ratio of MSP-amplified GSTP1 to MYOD1,a reference gene. We also performed a prospective, blinded investigation to quantitate GSTP1 promoter methylation in sextant prostate biopsy specimens from 21 additional patients with elevated serum prostate-specific antigen levels,11 of whom had histologically identified adenocarcinoma and 10 of whom had no morphologic evidence of adenocarcinoma. All data were analyzed by using nonparametric two-sided statistical tests. Results: The median ratios (and interquartile ranges) of MSP-amplified GSTP1 to MYOD1 in resected benign hyperplastic prostatic tissue,intraepithelial neoplasia,and adenocarcinoma were 0 (range,0 0.1),1.4 (range,0 45.9),and (range, ), respectively; all of these values were statistically significantly different (P<.001). The median ratios of MSPamplified GSTP1 to MYOD1 in the prospectively collected sextant biopsy samples were for the patients with adenocarcinoma and 0.0 for the patients with no evidence of adenocarcinoma (P<.001). Conclusion: Quantitation of GSTP1 methylation accurately discriminates between normal hyperplastic tissue and prostatic carcinoma in small samples of prostate tissue and may augment the standard pathologic/histologic assessment of the prostate. [J Natl Cancer Inst 2001;93: ] Prostate adenocarcinoma is the most commonly diagnosed cancer among men in Western countries and is the second leading cause of cancer-related deaths in the United States (1). Treatment of the advanced stages of this disease has met with limited success; hence, the development of reliable methods for early detection of tumors while they are still confined to the prostate gland increases the likelihood that radical therapy will effect a cure (2). Although the measurement of serum levels of prostate-specific antigen (PSA) is currently the most powerful screening test available for prostate cancer, the large proportion of false positives ascribed to elevations in serum PSA levels limits this approach. Newer molecular tests for prostate cancer that detect genetic or epigenetic alterations may be able to identify more precisely the presence of neoplastic cells in biopsy specimens and bodily fluids. Mutations in several genes are associated with prostate adenocarcinoma. Nevertheless, some mutations, such as those that activate the ras oncogene or inactivate the tumor suppressor TP53 (3), occur in only a small number of prostate cancer cases, while other mutations, such as those that inactivate the tumor suppressor gene PTEN, are found mainly in advanced disease (4). Thus, the identification of genetic or epigenetic alterations that occur more frequently and in earlier stages of the disease as well as in premalignant lesions, such as prostate intraepithelial neoplasia (PIN) may allow DNAbased detection of prostate cancer (5). Methylation of cytosines at CpG islands within the 5 region of genes has been recognized as an important epigenetic alteration that plays a decisive role in the control of gene expression (6). Methylation of cytidine nucleotides in the 5 -regulatory region of GSTP1, the gene that encodes the drug detoxification enzyme glutathione S-transferase (GST- ), is associated with the loss of GST- expression and is the most common epigenetic alteration described in prostate adenocarcinoma to date (7 9). Methylation of GSTP1 has been detected even in precursor lesions, such as PIN, but is rare in benign prostatic hyperplasia (BPH) (10). Thus, this epigenetic alteration represents a new and potentially powerful molecular marker for the detection of tumor cells in an early stage of prostate cancer. Methylation of the promoters of several other genes has already been successfully used to detect tumor DNA in sputum, serum, and bronchoalveolar lavage from lung cancer patients and in serum from head and neck cancer patients (11, 12). In these studies, a highly sensitive and reproducible methylation-specific polymerase chain reaction (MSP) method was introduced (13). This method, however, does not permit quantification of the level of gene methylation, which may be critical for distinguishing neoplastic from non-neoplastic tissue. A specific real-time quantitative MSP method based on detection of a fluorescent signal produced proportionally during polymerase chain reaction (PCR) amplification was developed in the 1990s that allows the rapid and highly accurate analysis of methylation levels in tissue samples (14,15). This method was used to evaluate the methylation status of the p16 gene in bone marrow aspirates from patients with multiple myeloma, and results obtained by using real-time quantitative MSP were completely concordant with those obtained by using a conventional Affiliations of authors: C. Jerónimo, H. Usadel, Department of Otolaryngology Head and Neck Surgery, Head and Neck Cancer Research Division, The Johns Hopkins University School of Medicine, Baltimore, MD; R. Henrique (Unit of Molecular Pathology, Department of Pathology), Portugal; J. Oliveira, (Department of Urology), C. Lopes (Unit of Molecular Pathology-Department of Pathology), Instituto Português de Oncologia de Francisco Gentil-Centro Regional do Porto, Portugal; W. G. Nelson, Department of Urology; D. Sidransky, Department of Otolaryngology-Head and Neck Surgery, Head and Neck Cancer Research Division and Department of Urology, The Johns Hopkins University School of Medicine. Correspondence to: David Sidransky, M.D., The Johns Hopkins University School of Medicine, Division of Cancer Research, 818 Ross Research Bldg., 720 Rutland Ave., Baltimore, MD , ( dsidrans@jhmi.edu). See Notes following References. Oxford University Press Journal of the National Cancer Institute, Vol. 93, No. 22, November 21, 2001 REPORTS 1747

2 MSP analysis (15). In this same study, real-time quantitative MSP had the sensitivity to detect as few as 10 genome equivalents of a methylated p16 sequence. Here, we investigated whether quantitation of GSTP1 methylation levels by real-time quantitative MSP could be used to augment prostate cancer detection in tissue biopsies. We performed this study by using prostate tissue samples collected from patients who harbored clinically localized prostate cancer and from a control group of patients with BPH. We also measured GSTP1 methylation levels in prostate sextant biopsy specimens collected from 21 patients who had high ( 4.0 ng/ml) PSA values. MATERIALS AND METHODS Cell Culture LNCaP (positive for GSTP1 methylation) and Du145 (negative for GSTP1 methylation) human prostate cancer cell lines, derived from prostate metastatic adenocarcinomas, were purchased from the American Type Culture Collection (Manassas, VA) and maintained in RPMI-1640 medium with 10% fetal bovine serum (7). Patients and Sample Collection We obtained prostate tissue samples from 69 randomly selected patients with clinically localized prostate adenocarcinoma [stage T1c, according to the TNM staging system (16)] who were consecutively diagnosed and treated with radical prostatectomy at the Portuguese Cancer Institute (Porto, Portugal) from September 1999 through April In 28 of these 69 radical prostatectomy specimens with adenocarcinoma, foci of high-grade PIN lesions were identified and further analyzed. In addition, prostate tissue samples were obtained from 31 randomly selected patients with BPH who had undergone transurethral resection of the prostate; these samples were used as control samples. We also obtained prostate sextant biopsy specimens from 21 patients who had elevated levels of serum PSA. Ten of those patients had no morphologic evidence of adenocarcinoma in the biopsy fragments, and 11 patients had histologically proven adenocarcinoma within the biopsy specimen. These 21 patients were prospectively selected, and their tissue samples were tested for GSTP1 methylation in a blinded fashion. Two pathologists (R. Henrique and C. Lopes) reviewed all of the histologic slides containing formalin-fixed, paraffin-embedded tissue fragments obtained from the surgical specimens and prostate biopsy specimens and graded each prostate tumor according to the Gleason grading system (17). We used surgical tissue specimens obtained during radical prostatectomy or transurethral resection that were either snap frozen in isopentane and stored at 80 C or formalin fixed and paraffin embedded for MSP analysis. Sections, 5- m-thick, were cut from frozen-tissue fragments or paraffin blocks to identify areas of high-grade PIN and adenocarcinoma in the radical prostatectomy specimens and BPH in the control-tissue samples obtained from transurethral resections of the prostate. These tissues were then carefully microdissected from 12- m-thick sections to enrich for areas that contained PIN, adenocarcinoma, or hyperplasia. DNA was extracted from an average of fifty 12- m-thick sections, and only from sections that contained at least 70% neoplastic cells (either PIN or adenocarcinoma). Paraffin-embedded tissue was similarly microdissected and was subsequently placed in xylene for 3 hours at 48 C to remove the paraffin. Paraffin blocks from the prostate sextant biopsy specimens were exhaustively cut (30 50 sections per biopsy specimen) without microdissection and were likewise placed in xylene. Genomic DNA was extracted from the tissue sections by using the method described by Ahrendt et al. (18). We used the same method to extract genomic DNA from LNCaP cells and Du145 cells, using one T-75 tissue culture flask of each cell type grown to near 100% cell confluence. Briefly, tissue sections and tissue culture cells were digested overnight at 48 C in 1% sodium dodecyl sulfate and 0.5 mg/ml Proteinase K, extracted with phenolchloroform (2 vol/1 vol), and ethanol precipitated. Bisulfite Treatment Sodium bisulfite conversion of unmethylated cytosine residues to uracil in samples of genomic DNA obtained from patient tissue samples and cell lines was performed by using a modification of a previously described method (19). Briefly, 2 g of genomic DNA was denatured with 0.2 M NaOH in a total volume of 22 L for 20 minutes at 50 C. A volume of 500 L freshly made bisulfite solution (2.5 M sodium metabisulfite and 125 mm hydroquinone [ph 5.0]) was added to each denaturation reaction, and the mixture was incubated at 50 C for 3 hours in the dark. The resulting bisulfite-converted DNA was then purified by using Wizard DNA purification resin (Promega Corp., Madison, WI) according to the manufacturer s instructions and eluted in 45 L of water preheated to 80 C. The eluted DNA was denatured in 0.3 M NaOH for 10 minutes at 37 C. Ammonium acetate (final concentration, 1mM) was added, and the mixture was incubated for 5 minutes at room temperature. Finally, the bisulfiteconverted and denatured genomic DNA was precipitated by adding 2.5 volumes of 100% ethanol and glycogen (final concentration, 20 g/ml), followed by centrifugation at rpm at room temperature for 15 minutes. Each resulting DNA pellet was washed with 70% ethanol, dried, and resuspended in 30 L5mM Tris-HCl (ph 8.0) and stored at 20 C. Real-Time Quantitative MSP Sodium bisulfite-treated genomic DNA was amplified by fluorescence-based real-time MSP by using TaqMan technology (PerkinElmer Corp., Foster City, CA) as described previously (20). We performed MSP in 96-well plates using a PE Applied Biosystems 7700 Sequence Detector (PerkinElmer Corp.). Real-time quantitative MSP is based on the continuous monitoring of a progressive fluorogenic PCR by an optical system. This PCR system uses two amplification primers and an additional, amplicon-specific, fluorogenic hybridization probe whose target sequence is located within the amplicon. The probe is labeled with two fluorescent dyes: 1) a 6-carboxy-fluorescein (FAM), located at the 5 -end, which serves as reporter, and 2) a 6-carboxytetramethyl-rhodamine (TAMRA), located at the 3 -end, which serves as a quencher. When amplification occurs, the 5 3 exonuclease activity of the Taq DNA polymerase cleaves the reporter from the probe during the extension phase, thus releasing it from the quencher. The resulting increase in fluorescence emission of the reporter dye is monitored during the PCR process (21) and represents the number of DNA fragments generated. In brief, oligonucleotide primers were designed to specifically amplify bisulfite-converted DNA within the 3 -end of the promoter of the GSTP1 gene (gene of interest), and a probe was designed to anneal specifically within the amplicon during extension. For the internal reference gene, MYOD1, the primers and probe were designed to amplify and detect a region of the gene that is devoid of CpG nucleotides. Thus, amplification of MYOD1 by MSP occurs independent of its methylation status, whereas the amplification of GSTP1 is proportional to the degree of cytosine methylation within the GSTP1 promoter. The methylation ratio was defined as the ratio of the fluorescence emission intensity values for the GSTP1 PCR products to those of the MYOD1 PCR products obtained by TaqMan analysis, multiplied by This ratio was used as a measure for the relative level of methylated GSTP1 DNA in the particular sample. The fluorogenic probes were custom synthesized by PE Applied Biosystems (PerkinElmer Corp.); the oligonucleotide primers were synthesized by Life Technologies, Inc. (GIBCO BRL, Rockville, MD). The sequences of the primers and probe used to amplify and detect methylated GSTP1 were 5 -AGTTGCGCG- GCGATTTC-3 (sense primer), 6FAM-5 -CGGTCGA- CGTTCGGGGTGTAGCG-3 -TAMRA (probe), and 5 -GCCCCAATACTAAATCACGACG-3 (antisense primer). The sequences of the primers and probe used to amplify and detect MYOD1 were 5 -CCAACTCCAA- ATCCCCTCTCTAT- 3 (sense primer), 6FAM-5 -TCC- CTTCCTATTCCTAAATCCAACCTAAATAC- CTCC-3 -TAMRA (probe), and 5 -TGATTAATTTA- GATTGGGTTTAGAGAAGGA- 3 (antisense primer). Fluorogenic quantitative real-time MSP assays were performed in a reaction volume of 25 L by using components supplied in a TaqMan PCR Core Reagent Kit (Perkin-Elmer). Separate amplification assays were performed for GSTP1 and MYOD1; each assay was performed in duplicate. The final reaction mixtures contained the sense and antisense primers at 600 nm each, the probe at 200 nm, 200 M each of deoxyadenosine triphosphate, deoxycytidine triphosphate, and deoxyguanosine triphosphate, 400 M deoxyuridine triphosphate, 5.5 mm MgCl 2, 1 TaqMan Buffer A, 1 unit of Amplitaq Gold DNA polymerase (PerkinElmer), and 3 L bisulfiteconverted genomic DNA. PCR was performed under the following conditions: 50 C for 2 minutes and, 95 C for 10 minutes, followed by 50 cycles of 95 C for 15 seconds and 60 C for 1 minute. To ensure the specificity of the MSP analysis, each 96-well PCR plate had wells that contained bisulfite-converted DNA isolated from patient tissue samples and wells that contained the following controls: DNA from LNCaP cells, in which GSTP1 is methylated (positive control), DNA from Du145 cells, in which GSTP1 is not methylated (7) (nega REPORTS Journal of the National Cancer Institute, Vol. 93, No. 22, November 21, 2001

3 tive control), and multiple wells that contained water instead of DNA (control for PCR specificity). We used serial dilutions of the positive control DNA to create a standard curve. Using a conversion factor of 6.6 pg of DNA per diploid cell (20), we were able to determine that we could detect a minimum of four genome equivalents of methylated GSTP1 DNA using the MSP assay. Statistical Analysis We determined the median and interquartile range of the methylation ratios for each group of tissue samples. These values were analyzed by using the Kruskal Wallis one-way analysis of variance, followed by the Bonferroni-adjusted Mann Whitney U test. For this comparison test among the three groups of tissue samples, the nonadjusted statistical level of significance of P<.05 corresponds to a Bonferroni adjusted statistical significance of P< The same procedure was used to compare age and PSA levels among those patients with BPH, clinically localized prostate adenocarcinoma and histologically proven prostate adenocarcinoma on biopsy and those whose biopsy specimens showed no morphologic evidence of adenocarcinoma. For this multiple comparison test, which comprised four groups of patients, the nonadjusted statistical level of significance of P<.05 corresponds to a Bonferroni adjusted statistical significance of P< The correlations between the tumor methylation ratios and PSA levels or prostate tumor Gleason scores were determined by calculating a Spearman s correlation coefficient. All statistical analyses were carried out by using a computer-assisted program (Statistica for Windows, version 6.0; StatSoft, Tulsa, OK). After analyzing the GSTP1 methylation levels in the initial samples of hyperplastic tissue, PIN lesions, and adenocarcinoma, we chose a GSTP1/MYOD1 methylation ratio of 10.0 as the cutoff level of GSTP1 methylation to distinguish benign from malignant tissue in the prospectively collected sextant biopsy specimens. All statistical tests were two-sided. RESULTS We initially measured GSTP1 promoter methylation in samples from 69 patients with clinically localized prostate adenocarcinoma who underwent radical prostatectomy and 31 patients with histologically documented BPH who had a transurethral resection of the prostate. Using quantitative real-time MSP, we detected at least four genome equivalents of methylated GSTP1 in 63 (91.3%) of the 69 adenocarcinoma cases and in 15 (53.6%) of the 28 high-grade PIN lesions obtained from the same radical prostatectomy specimens. Moreover, tissue samples from nine (29.0%) of 31 patients with BPH also displayed some GSTP1 methylation. The distributions of the ratios of methylated GSTP1/MYOD1 in tissue samples from patients with BPH, PIN, and clinically localized prostate adenocarcinoma differed markedly when they were plotted on a log scale (Fig. 1). The median ratios (and interquartile ranges) of methylated GSTP1/ MYOD1 were 0 (range, 0 0.1) for BPH, 1.4 (range, ) for PIN, and (range, ) for clinically localized prostate adenocarcinoma. The Kruskal Wallis test revealed a statistically significant difference in the GSTP1/MYOD1 methylation ratios among these three groups of tissue samples (P.00001). Using the Bonferroni-corrected Mann Whitney U test, we also found statistically significant differences in GSTP1/MYOD1 methylation ratios between BPH and PIN (P.014), between BPH and adenocarcinoma (P<.001), and between PIN and adenocarcinoma (P<.001) (Fig. 1). After reviewing these initial data, we set a cutoff level of 10.0 for GSTP1/MYOD1 ratios to distinguish benign (i.e., hyperplastic) from malignant tissue because it represents an optimal balance between the sensitivity and the specificity of the test. With the use of this cutoff value, the sensitivity of MSP quantitation of GSTP1 methylation in the detection of prostate adenocarcinoma was 85.5% and the positive predictive value was 100%. We could not determine the true specificity of quantitative MSP because all of the prostatectomy specimens that we analyzed harbored adenocarcinoma. By using this same cutoff value, however, in the 31 tissue samples harboring BPH, the specificity of this analysis reached 96.8%. On the basis of these initial observations, we investigated whether quantitation of GSTP1 methylation could be used to detect prostate cancer in small biopsy specimens. We therefore used the MSP assay to quantitate, in a blinded fashion, GSTP1 methylation in genomic DNA extracted from prostate sextant biopsy specimens collected prospectively from 21 patients with elevated levels of serum PSA. Eleven of these patients (median PSA level, 21.4 ng/ml; range, ng/ml) were subsequently found to have histologically proven prostatic adenocarcinoma, while 10 (median PSA level, 10.8 ng/ml; range, ng/ml) had no evidence of malignant disease within their biopsy specimens. We detected at least four genome equivalents of methylated Fig. 1. Distribution of GSTP1 methylation levels in prostate tissues displaying benign prostatic hyperplasia (BPH), prostate intraepithelial neoplasia (PIN), and clinically localized prostate adenocarcinoma (T). Nine (29.0%) of 31 patients with BPH, 63 (91.3%) of 69 patients with T, and 15 (53.6%) of the 28 paired PIN lesions harbored methylated GSTP1 promoter DNA as determined by real-time methylation-specific polymerase chain reaction. Each circle represents a different tissue sample. The solid horizontal bar indicates the median ratio of methylated GSTP1/MYOD1 ( 1000) within a group of patients. The median ratio of methylated GSTP1/MYOD1 differed statistically significantly between BHP and PIN (P.014), between BPH and T (P<.001), and between PIN and T (P ). Asterisks indicate the samples (n 22 for BPH; n 13 for PIN; n 6 for T) that had a median ratio of methylated GSTP1/MYOD1 equal to 0, which cannot be plotted on a log scale. Journal of the National Cancer Institute, Vol. 93, No. 22, November 21, 2001 REPORTS 1749

4 GSTP1 in the prostate biopsy specimens from 10 (90.9%) of the 11 patients with histologically proven prostatic adenocarcinoma and from four (40%) of the 10 patients with no evidence of neoplasia. Using the cut-off value of 10.0 for the methylation ratio, we correctly predicted the histologic diagnosis of prostate cancer in 10 (90.9%) of the 11 sextant biopsies from patients with prostate cancer and excluded a diagnosis of malignancy in all 10 patients whose biopsy specimens showed no evidence of malignancy (Fig. 2). The median ratio of methylated GSTP1/MYOD1 in tissue samples from the patients with no malignant disease (median ratio, 0.0) was statistically significantly different (P ) from that of patients with histologically proven adenocarcinomas (median ratio, 410.6) (Fig. 2) and from that of patients with clinically localized adenocarcinoma (P ). Using the same cutoff value of 10.0, we calculated that MSP quantitation of GSTP1 methylation had a sensitivity of 90.9%, a specificity of 100%, and a positive predictive value of 100% when applied to this group of prospectively collected biopsy specimens. There were no statistically significant differences in serum PSA levels between patients with no evidence of adenocarcinoma and those with histologically proven adenocarcinoma (P.014). In contrast, statistically significant differences in serum PSA levels were observed between patients with BPH and those with either histologically proven adenocarcinoma (P ) or clinically localized prostate adenocarcinoma (P<.001). This result could not be attributed to a difference in the age distribution because no statistically significant differences for that parameter were found among these groups of patients (P.087). In patients with clinically localized prostate adenocarcinoma, we found no correlation between the ratio of methylated GSTP1/MYOD1 in the tumor sample and the Gleason score of the tumor (r.13, P.36). In patients with either clinically localized or histologically Fig. 2. Distribution of the levels of serum prostate-specific antigen (PSA) and GSTP1 methylation (log scale) in prostate sextant biopsy samples from patients without (NB normal biopsy; n 10) and with a histologic diagnosis of prostate cancer (TB tumor biopsy; n 11). The y-axis represents the GSTP1/MYOD1 ratio 1000 (Me) or PSA values. Four (40.0%) of 10 patients with methylated NB had a GSTP1/MYOD1 ratio ( 1000) less than 10, and 10 (90.9%) of 11 patients with TB had a GSTP1/MYOD1 ratio ( 1000) greater than 10 in their biopsy specimens (Me methylation level of GSTP1). Asterisks indicatethe samples (n 6 for NB and n 1 for TB) that had a median ratio of methylated GSTP1/MYOD1 equal to 0, which cannot be plotted on a log scale. The solid horizontal bar indicates the median ratio of either serum PSA or methylated GSTP1/MYOD within a group of patients. The median serum PSA levels differed statistically significantly between NB and TB (P.014). The difference between the medians of methylated GSTP1/MYOD1 in NB and TB was also statistically significant (P.0007). proven adenocarcinomas, there were no correlations between the ratios of methylated GSTP1/MYOD1 in the tumor sample and PSA levels (r.04 [P.74] and r.41 [P.20], respectively). DISCUSSION We detected GSTP1 promoter methylation in the majority of the study patients who had either clinically localized tumors (91.3%) or histologically proven (90.9%) prostate adenocarcinoma in prostate sextant biopsy specimens. Other studies that have used conventional (i.e., nonquantitative) MSP assays (7,8) have also reported GSTP1 methylation in more than 90% of the prostate cancer cases analyzed. Conventional MSP for the detection of GSTP1 promoter methylation, however, is severely limited as a technique to detect cancer because many benign prostatic hyperplastic lesions as well as actual cases of prostate cancer score as positive for GSTP1 methylation. By using a robust quantitative assay, we demonstrated a clear difference in GSTP1 methylation levels between benign (both hyperplastic and non-neoplastic) and neoplastic prostate tissues. High-grade PIN lesions are regarded as possible precursors of prostate adenocarcinoma (22). In agreement with a previous study (10), we found that 15 (53.6%) of 28 premalignant PIN lesions displayed cytosine methylation at the GSTP1 promoter. The adenocarcinomas present in the same radical prostatectomy specimens from which these 15 methylated PIN lesions originated also displayed methylation, but the ratio of methylated GSTP1/ MYOD1 was statistically significantly lower in PIN lesions than in the paired tumor sample. These findings add further support to the notion that at least some PIN lesions may play a precursor role in prostate adenocarcinoma. Since PIN lesions are frequently methylated at the GSTP1 promoter region, impaired expression of the protein would be anticipated. Indeed, a loss or decreased expression of GST in PIN lesions has been reported, although the basal cells present in these lesions retained the normal pattern of GSTP1 expression (22). Thus, the lower levels of GSTP1 methylation detected in PIN lesions could also be related to the presence of nonmethylated alleles in basal cells. We found, however, that a statistically significant number of these PIN lesions were negative for GSTP1 methylation. This result could reflect the genetic 1750 REPORTS Journal of the National Cancer Institute, Vol. 93, No. 22, November 21, 2001

5 diversity of multifocal PIN lesions (23). Future studies will be required to determine if the level of methylation in PIN lesions predicts progression to invasive cancer. A major goal of this study was to determine whether GSTP1 methylation detection could distinguish neoplastic from non-neoplastic prostate tissue. We determined the levels of GSTP1 promoter methylation for a total of 41 patients with no clinical or pathologic evidence of prostate adenocarcinoma. Four of these patients displayed methylation of the GSTP1 promoter in the normal (i.e., nonneoplastic) prostatic tissue, and nine patients displayed methylation of the GSTP1 promoter in hyperplastic prostatic tissue. The former group of patients are good clinical controls because they have the same clinical/histologic characteristics as the bulk of patients who present with high PSA values and a need to rule out the possibility of prostate cancer. It is important that the median ratio of methylated GSTP1/MYOD1 was statistically significantly lower in these non-neoplastic tissue samples than in tissue samples displaying PIN and either histologically proven or clinically localized adenocarcinoma. Indeed, since GSTP1 methylation appears to be an early genetic alteration in prostate tumorigenesis (10, 22), it could also occur in morphologically normal premalignant tissue. Although recent evidence suggests that methylation at CpG islands located in the promoter regions of certain genes in normal-appearing tissues may be associated with aging (24,25), we did not see agerelated methylation differences in the tissue samples analyzed in this study. The distinctly different levels of GSTP1 methylation in non-neoplastic tissues and in those with evidence of prostate cancer suggest that quantitation of GSTP1 methylation could be more useful than measurements of serum PSA levels in distinguishing men with a very low risk of harboring prostate cancer from those who carry a clinically silent prostate adenocarcinoma. Using a cutoff value of 10.0, we found that the accuracy of GSTP1 methylation quantitation is excellent, with a positive predictive value of 100%. Moreover, because there was no correlation between PSA levels and GSTP1 methylation levels in prostate cancer patients, the latter potentially represents an independent marker for this disease. Indeed, while the serum levels of PSA were not statistically significantly different between patients with no morphologic evidence of prostate adenocarcinoma and those with prostate cancer, the GSTP1/MYOD1 methylation ratios confirmed that the biopsy samples from the former patients were non-neoplastic. The results of our study could prove particularly useful in the evaluation of negative sextant prostate biopsy specimens (i.e., those with no malignant features). A previous study (26) demonstrated that 24% of men who had a prostate biopsy because of abnormal (increased) serologic PSA, ultrasonographic, or clinical findings were found to harbor prostate cancer in subsequent biopsies. Moreover, only a few neoplastic cells are usually harvested in a core prostate biopsy despite clinically significant disease within the gland (27). Thus, such foci of neoplasia could eventually be missed in routine diagnostic testing because histologic sampling usually only evaluates a small amount of prostate tissue in a given section. Hence, the quantitation of GSTP1 methylation levels in prostate tissue from sextant biopsy specimens that show no histopathologic evidence of malignancy would be able to select a group of patients who could benefit from an earlier diagnosis of their disease. Although our preliminary blinded analysis of small biopsy samples from 21 individuals with elevated serum PSA levels suggests that the quantitation of GSTP1 methylation may have important clinical value, further clinical, serological, pathologic, and molecular evaluation of these patients is required to confirm this. This study demonstrates that quantitation of GSTP1 methylation may be a useful marker for prostate cancer in patients with clinically localized disease. The use of real-time PCR technology further enhances this approach as a powerful ancillary tool in molecular detection of prostate cancer. It is intriguing that methylated DNA was recently detected by using quantitative real-time MSP in urine and plasma samples from patients with prostate cancer (28,29). Thus, quantitative real-time MSP could be useful for monitoring patients for the presence of minimal residual disease after radical treatment once the GSTP1 methylation status of the primary tumor is established. Because so many patients at risk for prostate cancer present with a high serum PSA, quantitation of GSTP1 methylation in tissue biopsy specimens could augment current diagnostic histology and facilitate the triage of patients into appropriate risk categories for further intervention (30). REFERENCES (1) Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics. CA Cancer J Clin 1999;49: (2) Andriole GL, Catalona WJ. The case for aggressive diagnosis and therapy of localized prostate cancer. In: Raghavan D, Scher HI, Leibel SA, Lange PH, editors. Principles and practice of genitourinary oncology. 1 st ed. Philadelphia (PA): Lippincott-Raven; p (3) Isaacs WB, Isaacs JT. Molecular genetics of prostate cancer progression. In: Raghavan D, Scher HI, Leibel SA, Lange PH, editors. Principles and practice of genitourinary oncology. 1 st ed. Philadelphia (PA): Lippincott-Raven; p (4) Cairns P, Okami K, Halachami S, Halachami N, Esteller M, Herman JG, et al. Frequent inactivation of PTEN/MMAC1 in primary prostate cancer. Cancer Res 1997;57: (5) Sidransky D. Nucleic acid-based methods for the detection of cancer. Science 1997;278: (6) Baylin SB, Herman JG, Graff JR, Vertino P, Issa JP. Alterations in DNA methylation: a fundamental aspect of neoplasia. Adv Cancer Res 1998;72: (7) Lee W, Morton RA, Epstein JI, Brooks JD, Campbell PA, Bova GS, et al. Cytidine methylation of regulatory sequences near the piclass glutathione S-transferase gene accompanies human prostatic carcinogenesis. Proc Natl Acad Sci U S A 1994;91: (8) Lee WH, Isaacs WB, Bova GS, Nelson WG. CG island methylation changes near the GSTP1 gene in prostatic carcinoma cells detected using the polymerase chain reaction: a new prostate cancer biomarker. Cancer Epidemiol Biomark Prev 1997;6: (9) Esteller M, Corn PG, Urena JM, Gabrielson E, Baylin SB, Herman JG. Inactivation of glutathione S-transferase P1 gene by promoter hypermethylation in human neoplasia. Cancer Res 1999;58: (10) Brooks JD, Weinstein M, Lin X, Sun Y, Pin SS, Bova GS, et al. CG island methylation changes near the GSTP1 gene in prostatic intraepithelial neoplasia. Cancer Epidemiol Biomark Prev 1998;7: (11) Esteller M, Sanchez-Cespedes M, Rosell R, Sidransky D, Baylin SB, Herman JG. Detection of aberrant promoter hypermethylation of tumor suppressor genes in serum DNA from non-small cell lung cancer patients. Cancer Res 1999;59: (12) Sanchez-Cespedes M, Esteller M, Wu L, Nawroz-Danish H, Yoo GH, Koch WM, et al. Gene promoter hypermethylation in tumors and serum of head and neck patients. Cancer Res 2000;3: (13) Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A 1996;93: Journal of the National Cancer Institute, Vol. 93, No. 22, November 21, 2001 REPORTS 1751

6 (14) Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome Res 1996;6: (15) Lo YM, Wong IH, Zhang J, Tein M, Ng MH, Hjelm NM. Quantitative analysis of aberrant p16 methylation using real-time quantitative methylation-specific polymerase chain reaction. Cancer Res 1999;59: (16) Hermanek P, Hutter RVP, Sobin LH, Wagner G, Wittekind C. Prostate. In: Hermanek P, Hutter RVP, Sobin LH, Wagner G, Wittekind C, editors. Illustrated guide to the TNM/pTNM classification of malignant tumors. 4 th ed. Heidelberg (Germany): Springer-Verlag; p (17) Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histologic grading and clinical staging. J Urol 1974;111: (18) Ahrendt SA, Chow JT, Xu LH, Yang SC, Eisenberger CF, Esteller M, et al. Molecular detection of tumor cells in bronchoalveolar lavage fluid from patients with early stage lung cancer. J Natl Cancer Inst 1999;91: (19) Olek A, Oswald J, Walter J. A modified and improved method of bisulfite based cytosine methylation analysis. Nucleic Acids Res 1996; 24: (20) Eads CA, Danenberg KD, Kawakami K, Saltz LB, Danenberg PV, Laird PW. CpG island hypermethylation in human colorectal tumors is not associated with DNA methyltransferase overexpression. Cancer Res 1999;59: (21) Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 1988;239: (22) Montironi R, Mazzucchelli R, Stramazzotti D, Pomante R, Thompson D, Bartels PH. Expression of pi-class glutathione S-transferase: two populations of high grade prostatic intraepithelial neoplasia with different relations to carcinoma. Mol Pathol 2000;53: (23) Bostwick DG, Shan A, Qian J, Darson M, Maihle NJ, Jenkins RB, et al. Independent origin of multiple foci of prostatic intraepithelial neoplasia. Cancer 1998;83: (24) Toyota M, Issa JP. CpG island methylator phenotypes in aging and cancer. Semin Cancer Biol 1999;9: (25) Ahuja N, Li Q, Mohan AL, Baylin SB, Issa JP. Aging, DNA methylation in colorectal mucosa and cancer. Cancer Res 1998;58: (26) Keetch DW, Catalona WJ, Smith DS. Serial prostatic biopsies in men with persistently elevated serum prostate specific antigen values. J Urol 1994;151: (27) Epstein JI, Walsh PC, Carmichael M, Brendler CB. Pathologic and clinical findings to predict tumor extent of nonpalpable (stage T1c) prostate cancer. JAMA 1994;271: (28) Goessl C, Krause H, Muller M, Heicappell R, Schrader M, Sachsinger J, et al. Fluorescent methylation-specific polymerase chain reaction for DNA-based detection of prostate cancer in bodily fluids. Cancer Res 2000;60: (29) Cairns P, Esteller M, Herman JG, Schoenberg M, Jeronimo C, Sanchez-Cespedes M, et al. Detection of prostate cancer in urine by GSTP1 hypermethylation. Clin Cancer Res 2001;7: (30) Sharifi R, Shaw M, Ray V, Rhee H, Nagubadi S, Guinan P. Evaluation of cytologic techniques for diagnosis of prostate cancer. Urology 1983;21: NOTES C. Jerónimo and H. Usadel contributed equally to this work. Editor s note: Funding for the study described in this article was provided by Virco, Inc. Under a licensing agreement between The Johns Hopkins University and Virco, Dr. Sidransky is entitled to a share of royalty received by the university on sales or products described in this article. The university and Dr. Sidransky own Virco stock, which is subject to certain restrictions under university policy. Dr. Sidransky is a paid consultant to Virco. The terms of this arrangement are being managed by The Johns Hopkins University in accordance with its conflict of interest policies. W. G. Nelson has a patent (U.S. Patent 5,552,277) entitled Genetic Diagnosis of Prostate Cancer. Supported by Public Health Service grants RO1DE (National Institute of Dental Research and Craniofacial Research), RO1CA77664 (National Cancer Institute), PO1CA58184 (National Cancer Institute), and UO1CA84986 (National Cancer Institute), National Institutes of Health, Department of Health and Human Services. C. Jerónimo and H. Usadel are supported by grants from the Fundação para a Ciência e Tecnologia, Portugal (Program PRAXIS XXI-BD 13398/97), and the Dr. Mildred Scheel-Stiftung für Krebsforschung, Deutsche Krebshilfe, respectively. Manuscript received April 30, 2001; revised August 31, 2001; accepted September 7, REPORTS Journal of the National Cancer Institute, Vol. 93, No. 22, November 21, 2001