| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Gene (GSTP1) CpG Island Is Present in a Subset of Proliferative Inflammatory Atrophy Lesions but Not in Normal or Hyperplastic Epithelium of the Prostate
From the Department of Pathology,*the Brady Urological Institute,the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins,and the Bloomberg School of Public Health,the Johns Hopkins Medical Institutions, Baltimore, Maryland
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
gene (GSTP1) via CpG island hypermethylation occurs early during prostate carcinogenesis, present in 
70% of high-grade prostatic intraepithelial neoplasia (high-grade PIN) lesions and more than 90% of adenocarcinomas. Recently, there has been a resurgence of the concept that foci of prostatic atrophy (referred to as proliferative inflammatory atrophy or PIA) may be precursor lesions for the development of prostate cancer and/or high-grade PIN. Many of the cells within PIA lesions contain elevated levels of GSTP1, glutathione S-transferase-
(GSTA1), and cyclooxygenase-II proteins, suggesting a stress response. Because not all PIA cells are positive for GSTP1 protein, we hypothesized that some of the cells within these regions acquire GSTP1 CpG island hypermethylation, increasing the chance of progression to high-grade PIN and/or adenocarcinoma. Separate regions (n =199) from 27 formalin-fixed paraffin-embedded prostates were microdissected by laser-capture microdissection (Arcturus PixCell II). These regions included normal epithelium (n = 48), hyperplasticepithelium from benign prostatic hyperplasia nodules (n = 22), PIA (n = 64), high-grade PIN (n = 32), and adenocarcinoma (n = 33). Genomic DNA was isolated and assessed for GSTP1 CpG island hypermethylation by methylation-specific polymerase chain reaction. GSTP1 CpG island hypermethylation was not detected in normal epithelium (0 of 48) or in hyperplastic epithelium (0 of 22), but was found in 4 of 64 (6.3%) PIA lesions. The difference in the frequency of GSTP1 CpG island hypermethylation between normal or hyperplastic epithelium and PIA was statistically significant (P = 0.049). Similar to studies using nonmicrodissected cases, hypermethylation was found in 22 of 32 (68.8%) high-grade PIN lesions and in 30 of 33 (90.9%) adenocarcinoma lesions. Unlike normal or hyperplastic epithelium, GSTP1 CpG island hypermethylation can be detected in some PIA lesions. These data support the hypothesis that atrophic epithelium in a subset of PIA lesions may lead to high-grade PIN and/or adenocarcinoma. Because these atrophic lesions are so prevalent and extensive, even though only a small subset contains this somatic DNA alteration, the clinical impact may be substantial.
(GSTP1) and (GSTA), enzymes typically induced in response to oxidative or electrophilic stresses, are elevated in PIA.4,6
Also, cyclooxygenase II, another stress-inducible enzyme, is elevated in PIA.7
Several investigators in the early part of the last century suggested that focal prostatic atrophy may be a precursor to prostate cancer by virtue of the predilection for both lesions to occur in the peripheral zone and their frequent occurrence near carcinoma lesions.8,9
Franks10
and Liavag11
postulated that at least some forms of atrophy might directly give rise to cancer in the prostate. Recent morphological, immunohistochemical, and molecular-genetic evidence have provided preliminary support for this concept.4,12-15
For example, atrophic epithelium in PIA is frequently found to merge directly with, and display morphological transitions to, high-grade PIN12
and is often found near small carcinoma lesions.10-12
In addition, some atrophic lesions have been shown to acquire somatic chromosome 8 abnormalities13,15
and p53 mutations14
—genetic changes that are associated with prostate cancer and high-grade PIN.
GSTP1, encoding glutathione S-transferase-, has been proposed to be a caretaker gene, protecting cells against genome damage mediated by oxidants and electrophiles from inflammation or dietary exposures.16-18
In the human prostate, the vast majority of high-grade PIN and prostatic adenocarcinoma lesions fail to express GSTP1, whose loss is associated with hypermethylation of the CpG island encompassing the GSTP1 promoter. Methylation changes at this site have been detected in up to 100% of prostate cancer DNA specimens16,19-32
and in 70% of high-grade PIN,24,28,30,32,33
but are generally not found in normal prostate tissue. Although many of the luminal epithelial cells in PIA lesions express GSTP1, some do not.4
We hypothesize that some PIA cells may acquire GSTP1 CpG island hypermethylation leaving these cells vulnerable to progress to high-grade PIN and/or adenocarcinoma. Because atrophic cells expressing GSTP1 would not be expected to harbor GSTP1 promoter hypermethylation, and because the majority of the tissue within areas of PIA is stromal, we needed a method to enrich the epithelial cells in these regions to perform molecular analysis. For this, we isolated selected cell populations using laser-capture microdissection (LCM). Here, we report the results of a large survey of human clinical prostate tissues that examined the methylation status of the GSTP1 CpG island in matched samples of normal prostate, epithelial benign prostatic hyperplasia (BPH) tissue, PIA, high-grade PIN, and prostatic adenocarcinoma.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Twenty-seven formalin-fixed paraffin-embedded radical prostatectomy specimens were randomly selected from a larger series of patients who underwent radical retropubic prostatectomy for clinically localized prostate adenocarcinoma at The Johns Hopkins Hospital between 2000 and 2001. All patients provided informed consent for use of tissues, and the use of tissues was approved by the Johns Hopkins University School of Medicine Institutional Review Board. The median patient age was 59-years-old and ranged from 47 to 67 years. The median preoperative serum prostate specific antigen (PSA) was 5.5 (range, 4.0 to 11.5). The median Gleason score was 6 (range, 5 to 8) and the pathological stage ranged from pT2N0Mx to pT3aN0Mx. A series of a 5-µm sections and two 10-µm sections were cut from each tissue block. A 5-µm section was hematoxylin and eosin (H&E)-stained for pathological evaluation to identify each region. One or two 10-µm sections were used for LCM. Adjacent sections in some cases were stained by immunohistochemistry for 34ßE12 and/or GSTP1.
Histological Classification of Normal and Hyperplastic Tissues
Epithelium was classified as histologically normal when glands contained two epithelial cell layers lined by luminal cells that were tall and columnar. These luminal cells contained pale-to-clear cytoplasm and nuclei that were generally round, mostly basally located, and nearly even in size and shape and showed either no evidence of nucleoli or only very small nucleoli (<1 µm). These normal appearing glands contained abundant papillary infoldings and were generally lined by a continuous or mostly continuous basal cell layer. Epithelium was classified as hyperplastic when a similar histological appearance to that of normal epithelium, albeit at times with more pronounced papillae, was seen and the region occurred within an enlarged nodule in the transitional zone (nodular hyperplasia or benign prostatic hyperplasia).3
Histological Classification of PIA
We classified PIA lesions into the following types: simple atrophy, postatrophic hyperplasia, or mixed simple atrophy/postatrophic hyperplasia.2
Simple atrophy consists of atrophic cells lining acini with relatively normal caliber that lack papillary fronds, where the number of glands per unit area do not appear to be increased relative to normal tissue. Postatrophic hyperplasia consists of acini that are smaller and round and appear in a lobular distribution, often surrounding a somewhat dilated duct with an apparent increase in the number of small glands compared to normal tissue.2
Often, these two patterns are found in the same region, appearing to merge together (mixed lesions). The epithelium in both types of atrophy is composed of two layers consisting of flat basal cells and cubical to low-columnar luminal cells. The majority of luminal cells in atrophic regions contain scant basophilic cytoplasm. The spatial relation between PIA and high-grade PIN or carcinoma was examined as previously described.12
We recorded whether the atrophic lesion was merging directly with high-grade PIN, merging with carcinoma, adjacent to high-grade PIN and/or carcinoma (
100 µm), near to high-grade PIN and/or carcinoma (>100 µm to 1 mm), and distant from high-grade PIN and/or carcinoma (>1 mm). We also evaluated the nuclear features of the PIA luminal cells and classified individual lesions as containing no nuclear atypia, or having focal atypia (<1% of atrophic cells) consisting of nucleolar and nuclear enlargement beyond that of normal epithelium, but less than that found in high-grade PIN. These regions containing nuclear atypia were designated as focal low-grade PIN, as indicated previously.12
Nucleolar enlargement in atrophic cells directly abutting acute or chronic inflammatory cells was considered reactive atypia and not PIN.
Immunohistochemistry
Immunohistochemistry for high-molecular weight cytokeratin (HMWCK) and GSTP1 was performed with 34ßE12 (primary monoclonal, dilution 1:50; Enzo Biochem, Farmingdale, NY) and anti-GSTP1 (rabbit polyclonal, 1:4000; DAKO, Carpinteria, CA) using the EnVision+ detection system (DAKO). Paraffin sections were deparaffinized and rehydrated through a graded alcohol series. Steam heating in citrate buffer (Vector Laboratories Inc., Burlingame, CA) was performed for 20 minutes for antigen retrieval. Primary antibodies were incubated either overnight at 4°C (GSTP1) or for 45 minutes at room temperature (34ßE12). Slides were counterstained with hematoxylin.
Cell Culture
The established human prostate cancer cell line LNCaP (clone FGC, American Type Culture Collection catalog no. CRL-1740) with known GSTP1 promoter hypermethylation,27 obtained from the American Type Culture Collection (Rockville, MD), was grown in RPMI 1640 (Life Technologies, Inc., Gaithersburg, MD) and 10% fetal calf serum.
Laser Capture Microdissection
Two serial 10-µm sections were deparaffinized in two changes of xylene solutions for 5 minutes and rehydrated in 100%, 100%, and 95% ethanol for 3 minutes, 3 minutes, and 1 minute each, followed by staining with Hematoxylin 7211 (Richard-Allan Scientific, Kalamazoo, MI) for 45 seconds and Eosin-Y (Richard-Allan Scientific) for 1 second after rinsing with distilled water. The sections were dehydrated in 95%, 100%, and 100% ethanol for 1 minute each, followed by two changes of xylene for 5 minutes each and air-dried. Regions of normal epithelium, hyperplastic epithelium from transition zone nodules (BPH), PIA, high-grade PIN, and prostatic adenocarcinoma were obtained by LCM using the Pixcell II LCM system (Arcturus Engineering Inc., Mountain View, CA) from one or two sections using the CapSure HS LCM Caps (Arcturus Engineering Inc.). All prostates contained multiple regions of normal appearing epithelium and PIA, and many contained high-grade PIN and areas of nodular hyperplasia (BPH) in the transition zone. Regions of these histologies that were considered large enough to microdissect were selected at random. Carcinoma lesions were also multifocal in the prostates and these were also selected at random. To augment the resolution of the LCM in cases in which PIA and regions of high-grade PIN were in close proximity or merging, in many cases, slides were temporarily coverslipped and scanned using the BLISS System (Bacus Laboratory Inc., Lombard, IL). This produces a high-resolution map of the entire region of interest that can be used as a guide during LCM. After coverslips were removed, LCM was performed in conjunction with simultaneous visualization on a separate computer screen using the BLISS images. This allowed unambiguous isolation by LCM of atrophic cells from lesions where atrophy was merging with high-grade PIN or adenocarcinoma cells without isolation of high-grade PIN cells or adenocarcinoma cells. That the neoplastic cells were not procured in these cases was verified by microscopic examination and permanently recorded by rescanning the slides using the BLISS system after microdissection.
Isolation of Genomic DNA
DNA was isolated from the LNCaP cell line (positive control for GSTP1 promoter hypermethylation) and hyperplastic tonsil gland (negative control) by use of a QIAamp DNA Mini Kit (Qiagen Inc., Valencia, CA) following the manufacturer’s protocol. The concentrations of extracted DNA were measured with spectrometry. The cells captured with LCM were digested in 20 µl of proteinase K buffer (2 mg/ml proteinase K, 10 mmol/L Tris-HCl, pH 8.0, 0.1 mmol/L ethylenediaminetetraacetic acid, and 1% Tween 20) at 55°C overnight. After overnight incubation, the samples were incubated at 95°C for 8 minutes to inactivate the proteinase K. The DNA was isolated with standard phenol extraction and ethanol precipitation by adding 1.5 µl of glycogen (20 mg/ml; Roche Diagnostics Corp., Indianapolis, IN) as a carrier. The isolated DNA was eluted in 50 µl of H2O.
To determine the starting amount of DNA for bisulfite modification, DNA extracted with LCM was quantified by use of real-time quantitative polymerase chain reaction (PCR) with SYBR Green I. Real-time quantitative PCR analyses were performed with the i-Cycler IQ Real-Time Detection System (Bio-Rad Laboratories, Inc., Hercules, CA). The concentration of DNA was measured using the absolute standard curve method following the i-Cycler IQ Real-Time Detection System Resource guide. An absolute standard curve was constructed with 10-fold serial dilutions of LNCaP DNA ranging from 50 ng to 5 pg using a human ß-globin primer set. PCR primers of ß-globin were 5'-GCAACCTCAAACAGACACCA-3' (forward) and 5'-CCTCACCACCAACTTCATCC-3' (reverse). The quantitative PCR reaction was performed with 1.0 µl of DNA template in 25 µl of reaction mixture containing 2.5 µl of 10x PCR Buffer (Roche Applied Biosystems, Branchburg, NJ), 200 µmol/L each of the four dNTPs, 3.0 mmol/L MgCl2, 0.4 µmol/L each forward and reverse primer, 0.5 µl of 1:2000 dilution SYBR Green I (10,000x stock) (Molecular Probes Inc., Eugene, OR) and 1.25 U of Ampli Taq Gold (Roche Applied Biosystems). PCR reactions were performed in triplicate. The PCR reaction was subjected to hot start at 95°C for 10 minutes followed by 45 cycles of denaturation at 95°C for 15 seconds, annealing, and extension at 65°C for 30 minutes. DNA amplifications were performed in a 96-well reaction plate.
Methylation-Specific PCR (MSP)
In the majority of isolated specimens, 10 ng of genomic DNA, measured with real-time PCR, was modified with bisulfite using the CpGenome DNA modification kit (Intergen, Norcross, GA) following the manufacturer’s protocol. The bisulfite-modified DNA was eluted in 25 µl of H2O. In some LCM samples where yields were lower, between 1.9 and 10 ng of DNA was treated with bisulfite. To enhance the sensitivity of detection of methylated alleles, two different primer sets were used for MSP. The first set34
utilizes two sets of primers for each sample; one that detects densely methylated alleles (a) and the other that detects unmethylated alleles (b). The second set28
utilizes two sets of primers as well; one that detects densely methylated alleles (c) and the other that detects an unrelated gene, MYOD1, that controls for efficacy of bisulfite conversion (d). a, GSTP1-methylated specific primer 1, 5'-TTCGGGGTGTAGCGGTCGTC-3' (forward), 5'-GCCCCAATACTAAATCACGACG-3' (reverse); b, GSTP1 unmethylated specific primer, 5'-GATGTTTGGGGTGTAGTGGTTGTT-3' (forward), 5'-CCACCCCAATACTAAATCACAACA-3' (reverse); c, GSTP1 methylated specific primer 2, 5'-AGTTGCGCGGCGATTTC-3' (forward), 5'-GCCCCAATACTAAATCACGACG-3' (reverse); d, MYOD1 primer, 5'-CCAACTCCAAATCCCCTCTCTAT-3' (forward), 5'-TGATTAATTTAGATTGGGTTTAGAGAAGGA-3' (reverse). The two primer sets to detect densely methylated alleles in the human GSPT1 promoter are depicted in Figure 1
. The PCR reaction was performed with 2.5 µl of bisulfite-modified DNA template in 25 µl of reaction mixture containing 2.5 µl 10x PCR buffer, 200 µmol/L of each dNTP, 3.0 mmol/L MgCl2 (a, b), 3.5 mmol/L MgCl2 (c), 4.5 mmol/L MgCl2 (d), 0.25 µmol/L each primer (a, b), 0.6 µmol/L each primer (c, d), and 1.25 U of Ampli Taq Gold. The PCR reaction was subjected to hot start at 95°C for 10 minutes followed by 50 cycles of denaturation at 95°C for 1 minute, annealing at 59°C for 30 seconds, and extension at 72°C for 1 minute and a final extension at 72°C for 7 minutes (a, b), or 50 cycles of denaturation at 95°C for 15 seconds, annealing, and extension at 65°C for 1 minute and a final extension at 72°C for 7 minutes (c, d). Eight µl of each PCR reaction were loaded onto a 3% agarose gel and stained with Gel Star nucleic acid gel stain (BMA Bio Whittaker Molecular Applications, Rockland, ME). DNA from the prostate cancer cell line LNCaP, which is known to contain only densely methylated GSPT1 alleles,16
was used as a positive control for GSTP1 methylation, normal human tonsil DNA was used as a negative control, and dH20 used as a control for contamination in PCR reactions.
|
Although MSP is a very sensitive method for detecting methylation changes, bisulfite treatment can result in decreased yields of DNA. We found that to consistently obtain the correct PCR products, 1 ng of LNCaP DNA was needed for bisulfite modification. The results of real-time quantitative PCR with the ß-globin primers showed 5 to 10 pg of genomic DNA was obtained from one cell, such that 1000 to 2000 cells were needed to obtain 10 ng of genomic DNA. In an experiment to investigate the sensitivity of MSP, LNCaP DNA was mixed with tonsil DNA in different proportions, and the limit of detection was 1% of LNCaP DNA (100 pg) in a total of 10 ng of bisulfite-treated DNA for both primer sets (data not shown).
Statistical Analysis
Differences in frequency of GSTP1 hypermethylation between normal and hyperplastic epithelium, PIA, high-grade PIN, and adenocarcinoma were assessed with the Fisher’s exact or the chi-square test using Stata 6.0 for Microsoft Windows software.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The histological features of PIA have been described (see Materials and Methods).2,4
The histological classifications of the microdissected PIA lesions are shown in Table 1
. An example of PIA is shown in Figure 2
, a stereotypic case in which a large fraction of the peripheral zone was occupied by atrophic epithelium classified as simple atrophy. Of the 64 PIA lesions microdissected, 44 (68.8%) were classified as simple atrophy, 5 (7.8%) as postatrophic hyperplasia (PAH), and 15 (23.4%) as mixed simple atrophy and postatrophic hyperplasia.
|
|
shows an example of this staining in another typical simple atrophy lesion. Another characteristic immunohistochemical feature of PIA is that GSTP1 protein is also elevated in many, but not all, of the luminal cells (Figure 3)
.
|
, the latter being characterized by cells with numerous prominent nucleoli (between 1.5 and 2.5 µm as measured by the BLISS system). Most of these cases also contained atrophic appearing cells containing somewhat more cytoplasm with nuclear and nucleolar enlargement that was intermediate between normal epithelium and high-grade PIN epithelium, suggestive of transitions between PIA and high-grade PIN.12
PIA was near to high-grade PIN in 9 lesions (14.1%), and distant from high-grade PIN in 39 lesions (60.9%). Of the 48 PIA lesions where there was no apparent merging with high-grade PIN, 9 contained low-grade PIN (Figure 4)
. Careful LCM permitted the selective isolation of atrophic cells, even atrophic cells merging directly with high-grade PIN or adenocarcinoma, without contamination by PIN or adenocarcinoma cells (Figure 5)
. To further avoid the possibility of contamination, during microdissection we also carefully avoided cells in PIA containing any nucleolar enlargement when the lesion was merging with high-grade PIN.
|
|
. PIA was adjacent to adenocarcinoma in 4 lesions (6.3%), near to adenocarcinoma in 10 lesions (15.6%), and distant from adenocarcinoma in 48 lesions (75.0%). Inflammatory infiltrates were found in all microdissected areas. The extent of inflammation varied from mild chronic to moderate acute inflammation, although the majority of cases contained mild to moderate amounts of chronic inflammation and mild or no acute inflammation.
|
To assess the amount of isolated genomic DNA templates obtained by LCM, we used an aliquot of each sample obtained by LCM for real-time PCR. In a typical LCM session, we obtained 1000 to 3000 cells from each of two adjacent slides. An example of a lesion before and after LCM is shown in Figure 5
. The yield of amplifiable template genomic DNA ranged from 1.95 ng to 54.0 ng. One hundred ninety-nine of the 200 (99.5%) samples showed readily amplifiable DNA using either set of primers after bisulfite treatment. Bisulfite-modified DNA was successfully amplified in 178 of the 200 samples (89%) using the first primer set (GSPT1 methylated primer 1 and unmethylated primer) and in 177 of the 200 samples (88.5%) using the second primer set (GSTP1 methylated primer 2 and MYOD primer). One hundred fifty-six of the 200 lesions (78%) amplified for both primer sets. A summary of the results of MSP is shown in Table 2
. Two examples of MSP analysis are shown in Figure 7
.
|
|
. GSTP1 promoter methylation was found in 4 of 64 PIA lesions (6.3%) from 4 of 27 patients (14.8%). One lesion demonstrated methylated alleles using primer set 1, two lesions contained methylated alleles using primer set 2, and one lesion had detectable methylation using both primer sets 1 and 2. All four methylated PIA lesions were considered simple atrophy. Two of four microdissected lesions with methylated alleles contained focal markedly atypical cells (low-grade PIN). Although one of the four methylated atrophy lesions merged with high-grade PIN, the other three lesions did not and none of the positive lesions either merged with cancer or were directly adjacent to cancer.
For statistical analysis, we pooled the results from the normal epithelium (n = 48) with that of hyperplastic epithelium (BPH) (n = 22) to obtain a total of 70 benign nonatrophic epithelial regions. Comparing the benign and hyperplastic epithelium (0 of 70) to that of PIA (4 of 64), there was a statistically significant difference in the frequency of GSTP1 promoter hypermethylation (P = 0.049, Fisher’s exact test) (Table 2)
.
Hypermethylation of the GSTP1 promoter was detected in 13 of 27 high-grade PIN lesions (48.1%) using the first primer set and 16 of 29 lesions (55.2%) using the second primer set. Combined, 22 of 32 high-grade PIN lesions (68.8%) contained methylated alleles from 17 of 23 patients (73.9%). A statistically significant difference in the frequency of GSTP1 promoter hypermethylation was found between PIA foci and high-grade PIN foci (P < 0.001, chi-square) (Table 2)
.
Hypermethylation of the GSTP1 promoter was detected in 23 of 30 adenocarcinoma lesions (76.7%) using the first primer set, and 25 of 31 lesions (80.6%) using second primer set. Combined, therefore, 30 of 33 carcinoma lesions (90.9%) contained methylated alleles. All 27 patients (100%) showed methylation in either of the two different methylated primer sites or both sites. There was a significant difference from high-grade PIN lesions to adenocarcinoma lesions (P = 0.03, chi-square) (Table 2)
.
The ability of a given DNA sample to be amplified by both primer sets, as opposed to a single primer set, may be a measure of the density of methylation of the GSTP1 promoter. The ratio of lesions that contained methylation changes in both primer sites to lesions that showed methylation change in either set of primers was 1 of 4 PIA lesions (25.0%), 7of 18 high-grade PIN lesions (38.9%), and 18 of 27 adenocarcinoma lesions (66.7%).
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
In a previous study examining the topographic relation of PIA to high-grade PIN and adenocarcinoma, we selected 13 radical prostatectomy specimens that contained on average 5 separate foci of carcinoma and 50 separate foci of high-grade PIN.12 To study the earliest lesions, these prostates all had tumor volumes of less than 0.5 cm3. Although 42% of high-grade PIN lesions were found to merge directly with atrophic epithelium, no carcinomas were found to merge directly with atrophic epithelium.12 However, 88% of carcinoma lesions were within 1 mm of PIA.12 Previous studies by others using H&E-stained slides did report areas of atrophy that appeared to be merging directly with carcinoma,10,11 and in the present study two PIA lesions appeared to be directly merging with small foci of adenocarcinoma. Twenty-five percent of the PIA lesions were found to merge directly with high-grade PIN and 31.3% of the PIA lesions contained at least some cells with prominent nucleoli (low-grade PIN). Thus, morphological evidence indicates that atrophic epithelial cells can develop nuclear atypia and architectural changes that proceed either to PIN or to carcinoma directly. Morphological data alone, however, cannot be used to directly implicate a cancer precursor. Thus, several groups have begun to subject prostate atrophy to molecular analysis. Previous studies have indicated that chromosome 8 abnormalities, which are common in prostate adenocarcinoma, were found in areas of simple atrophy and postatrophic hyperplasia. Gain of chromosome 8 centromere sequences was noted in 1.3% of normal, 2.1% of simple atrophy, 2.8% of high-grade PIN, 4.0% of postatrophic hyperplasia, and 6.0% of adenocarcinoma.13 In a study of postatrophic hyperplasia from Japan, Tsujimoto and colleagues14 reported that 5.3% of atrophy lesions, 4.2% of high-grade PIN, and 25.0% of adenocarcinoma contained at least some cells harboring p53 mutations. Thus together with our results, there is accumulating evidence that a subset of focal atrophy lesions in the prostate may contain somatic molecular alterations characteristic of prostate cancer.
GSTP1 is a phase II detoxification enzyme that defends cells against DNA damage mediated by oxidants or electrophiles.36 GSTP1 may function as a caretaker gene, which when inactivated leads to additional somatic genome alterations that promote tumor growth.37 For example, mice with disrupted GSTP1/2 alleles showed an increase in skin tumorigenesis in response to topical carcinogen treatment.38 Rats fed 2-amino-1-methyl-6-phenylimidazo [4,5-ß] pyridine (PhIP), which is the most abundant of the mutagenic heterocyclic amines found in well-done or charred meats,39 showed mutations in prostate DNA and developed prostate carcinoma.40,41 Loss of GSTP1 function was found to increase PhIP induced DNA damage in human prostate cells exposed to metabolically activated PhIP in vitro.42
In human prostates, the majority of high-grade PIN and almost all adenocarcinoma cells fail to express GSTP1. This lack of protein expression is associated with hypermethylation of the GSTP1 CpG island encompassing its 5' upstream promoter region.27,33
Although the normal luminal cells generally do not express GSTP1, strong staining is seen in underlying basal cells. We have proposed that the consistent expression of GSTP1 may protect basal cells from DNA damage.17
In contrast, in addition to the basal cells, many of the luminal cells in PIA lesions stain positively for GSTP1. We hypothesized that the cells in PIA that lack GSTP1 expression may be a consequence of GSTP1 CpG island hypermethylation. The data presented here demonstrate that GSTP1 CpG island hypermethylation changes occur in some PIA lesions. These results provide further molecular evidence that at least some PIA lesions may harbor cells already initiated to progress to high-grade PIN and/or prostatic adenocarcinoma. The fact that only 6.3% of focal atrophy lesions contain methylated GSTP1 alleles does not preclude PIA from being a major precursor to prostate cancer because atrophy can occupy a very large percentage of the total peripheral zone of the prostate (Figure 2)
.10
Therefore, if 6.3% of prostate atrophy regions can potentially lead to carcinoma, then the very high prevalence and extent of atrophy may help explain the high frequency of multifocal prostate cancer and high-grade PIN. It would be of interest to repeat this study in different patient populations from different geographic locations to determine how widespread these findings are.
Our results differ somewhat from that of Jeronimo and colleagues.28 In their study, hypermethylation of the GSTP1 CpG island was identified in 29% of BPH samples whereas none of the BPH samples in our study showed GSTP1 hypermethylation. There are several potential explanations for this discrepancy. First, Jeronimo and colleagues28 used a much larger amount of DNA for MSP. Thus, the chance of finding rare methylated cells in BPH tissue was increased in their study. Another possible explanation is that BPH samples in patients treated for BPH symptoms, such as those used by Jeronimo and colleagues,28 may be biologically different from those samples obtained from an enlarged transition zone containing histological nodular hyperplasia, as used in the present study. Finally, Jeronimo and colleagues28 did not use microdissected samples, and it is possible that some of their specimens contained occult carcinoma or PIN. Consistent with our results, however, they did find quantitatively much less GSTP1 hypermethylation in BPH than in cancer specimens.28
The results of MSP using two different primer sets support the hypothesis that the methylation of individual CpG sites in the island may gradually increase during progression of the neoplastic process. If so, inhibition of de novo methylation at any stage could be important for prevention of cancer development.
Although these studies provide new additional evidence that some foci of prostate atrophy may be at risk for developing into true neoplastic lesions, many new questions are raised by these results. What cells in PIA have GSTP1 hypermethylation? We presume that cells that express GSTP1 in PIA would not have dense CpG island methylation. This question may best be addressed in the future by in situ methods for detection of CpG island hypermethylation43 combined with immunostaining for GSTP1 protein. Another possibility is that a higher percentage of PIA lesions contain a small number of cells with methylated GSTP1 alleles but that this small fraction may not be routinely detected by our method. Another question is: does the extent of methylation changes at the GSTP1 CpG island increase with neoplastic transformation in the prostate? Finally, does immune-mediated oxidant stress lead to aberrant methylation changes? Further studies are needed to address these issues.
In summary, GSTP1 CpG island hypermethylation appeared in a subset of PIA lesions, though no GSTP1 hypermethylation was identified in normal appearing epithelium or epithelium of BPH tissue. High-grade PIN and adenocarcinoma very frequently showed this somatic molecular alteration. Our data support the hypothesis that atrophic epithelium in a subset of PIA lesions may lead to high-grade PIN and/or adenocarcinoma. Because these lesions are so prevalent and extensive, even though only a small subset contain this key somatic DNA alteration, the clinical impact may be highly significant.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Public Health Service (grants NIH/NCI R01CA084997, NIH R01CA70196, and NIH/NCI Specialized Program in Research Excellence in Prostate Cancer P50CA58236 to The Johns Hopkins University).
Accepted for publication May 14, 2003.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. Uemura, H. Ishiguro, Y. Ishiguro, K. Hoshino, S. Takahashi, and Y. Kubota Angiotensin II Induces Oxidative Stress in Prostate Cancer Mol. Cancer Res., February 1, 2008; 6(2): 250 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R. Bethel, D. Faith, X. Li, B. Guan, J. L. Hicks, F. Lan, R. B. Jenkins, C. J. Bieberich, and A. M. De Marzo Decreased NKX3.1 Protein Expression in Focal Prostatic Atrophy, Prostatic Intraepithelial Neoplasia, and Adenocarcinoma: Association with Gleason Score and Chromosome 8p Deletion. Cancer Res., November 15, 2006; 66(22): 10683 - 10690. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S Perry, R. Foley, K. Woodson, and M. Lawler The emerging roles of DNA methylation in the clinical management of prostate cancer. Endocr. Relat. Cancer, June 1, 2006; 13(2): 357 - 377. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Umetani, M. F.G. de Maat, E. Sunami, S. Hiramatsu, S. Martinez, and D. S.B. Hoon Methylation of p16 and Ras Association Domain Family Protein 1a during Colorectal Malignant Transformation Mol. Cancer Res., May 1, 2006; 4(5): 303 - 309. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Yegnasubramanian, X. Lin, M. C. Haffner, A. M. DeMarzo, and W. G. Nelson Combination of methylated-DNA precipitation and methylation-sensitive restriction enzymes (COMPARE-MS) for the rapid, sensitive and quantitative detection of DNA methylation Nucleic Acids Res., February 9, 2006; 34(3): e19 - e19. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Henrique, C. Jeronimo, M. R. Teixeira, M. O. Hoque, A. L. Carvalho, I. Pais, F. R. Ribeiro, J. Oliveira, C. Lopes, and D. Sidransky Epigenetic Heterogeneity of High-Grade Prostatic Intraepithelial Neoplasia: Clues for Clonal Progression in Prostate Carcinogenesis Mol. Cancer Res., January 1, 2006; 4(1): 1 - 8. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. H. Lee, S. Yegnasubramanian, X. Lin, and W. G. Nelson Procainamide Is a Specific Inhibitor of DNA Methyltransferase 1 J. Biol. Chem., December 9, 2005; 280(49): 40749 - 40756. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kim, J. Lapointe, G. Kaygusuz, D. E. Ong, C. Li, M. van de Rijn, J. D. Brooks, and J. R. Pollack The Retinoic Acid Synthesis Gene ALDH1a2 Is a Candidate Tumor Suppressor in Prostate Cancer Cancer Res., September 15, 2005; 65(18): 8118 - 8124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Hughes, A Murphy, C Martin, O Sheils, and J O'Leary Molecular pathology of prostate cancer J. Clin. Pathol., July 1, 2005; 58(7): 673 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. S. Palapattu, S. Sutcliffe, P. J. Bastian, E. A. Platz, A. M. De Marzo, W. B. Isaacs, and W. G. Nelson Prostate carcinogenesis and inflammation: emerging insights Carcinogenesis, July 1, 2005; 26(7): 1170 - 1181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J. Bastian, G. S. Palapattu, X. Lin, S. Yegnasubramanian, L. A. Mangold, B. Trock, M. A. Eisenberger, A. W. Partin, and W. G. Nelson Preoperative Serum DNA GSTP1 CpG Island Hypermethylation and the Risk of Early Prostate-Specific Antigen Recurrence Following Radical Prostatectomy Clin. Cancer Res., June 1, 2005; 11(11): 4037 - 4043. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. K.Y. Chan, U.-S. Khoo, K. Y.K. Chan, H. Y.S. Ngan, S.-S. Li, P.-M. Chiu, L.-S. Man, P. P.C. Ip, W.-C. Xue, and A. N.Y. Cheung Promoter Methylation and Differential Expression of {pi}-Class Glutathione S-Transferase in Endometrial Carcinoma J. Mol. Diagn., February 1, 2005; 7(1): 8 - 16. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Jeronimo, R. Henrique, M. O. Hoque, E. Mambo, F. R. Ribeiro, G. Varzim, J. Oliveira, M. R. Teixeira, C. Lopes, and D. Sidransky A Quantitative Promoter Methylation Profile of Prostate Cancer Clin. Cancer Res., December 15, 2004; 10(24): 8472 - 8478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Jeronimo, R. Henrique, M. O. Hoque, F. R. Ribeiro, J. Oliveira, D. Fonseca, M. R. Teixeira, C. Lopes, and D. Sidransky Quantitative RAR{beta}2 Hypermethylation: A Promising Prostate Cancer Marker Clin. Cancer Res., June 15, 2004; 10(12): 4010 - 4014. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Yegnasubramanian, J. Kowalski, M. L. Gonzalgo, M. Zahurak, S. Piantadosi, P. C. Walsh, G. S. Bova, A. M. De Marzo, W. B. Isaacs, and W. G. Nelson Hypermethylation of CpG Islands in Primary and Metastatic Human Prostate Cancer Cancer Res., March 15, 2004; 64(6): 1975 - 1986. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
