DNA methylation paradigm shift: 15-Lipoxygenase-1 upregulation in prostatic intraepithelial neoplasia and prostate cancer by atypical promoter hypermethylation

Prostaglandins & other Lipid Mediators 82 (2007) 185–197

DNA methylation paradigm shift: 15-Lipoxygenase-1 upregulation in prostatic intraepithelial neoplasia and prostate cancer by atypical promoter hypermethylation U.P. Kelavkar a,∗ , N.S. Harya a , J. Hutzley a , D.J. Bacich a , F.A. Monzon b , U. Chandran b , R. Dhir b , D.S. O’Keefe a a b

Department of Urology, University of Pittsburgh and Cancer Institute, PA, USA Department of Pathology, University of Pittsburgh and Cancer Institute, PA, USA Received 24 March 2006; accepted 19 May 2006 Available online 11 July 2006

Abstract Fifteen (15)-lipoxygenase type 1 (15-LO-1, ALOX15), a highly regulated, tissue- and cell-type-specific lipid-peroxidating enzyme has several functions ranging from physiological membrane remodeling, pathogenesis of atherosclerosis, inflammation and carcinogenesis. Several of our findings support a possible role for 15-LO-1 in prostate cancer (PCa) tumorigenesis. In the present study, we identified a CpG island in the 15-LO-1 promoter and demonstrate that the methylation status of a specific CpG within this island region is associated with transcriptional activation or repression of the 15-LO-1 gene. High levels of 15-LO-1 expression was exclusively correlated with one of the CpG dinucleotides within the 15-LO-1 promoter in all examined PCa cell-lines expressing 15-LO-1 mRNA. We examined the methylation status of this specific CpG in microdissected high grade prostatic intraepithelial neoplasia (HGPIN), PCa, metastatic human prostate tissues, normal prostate cell lines and human donor (normal) prostates. Methylation of this CpG correlated with HGPIN, PCa and metastatic human prostate tissues, while this CpG was unmethylated in all of the normal prostate cell lines and human donor (normal) prostates that either did not display or had minimal basal 15-LO-1 expression. Immunohistochemistry for 15-LO-1 was performed in prostates from PCa patients with Gleason scores 6, 7 [(4 + 3) and (3 + 4)], >7 with metastasis, (8–10) and 5 normal (donor) individual males. Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) was used to detect 15-LO-1 in PrEC, RWPE-1, BPH-1, DU-145, LAPC-4, LNCaP, MDAPCa2b and PC-3 cell lines. The specific methylated CpG dinucleotide within the CpG island of the 15-LO-1 promoter was identified by bisulfite sequencing from these cell lines. The methylation status was determined by COBRA analyses of one specific CpG dinucleotide within the 15-LO-1 promoter in these cell lines and in prostates from patients and normal individuals. Fifteen-LO-1, GSTPi and beta-actin mRNA expression in BPH-1, LNCaP and MDAPCa2b cell lines with or without 5-aza-2 -deoxycytidine (5-aza-dC) and trichostatin-A (TSA) treatment were investigated by qRT-PCR. Complete or partial methylation of 15-LO-1 promoter was observed in all PCa patients but the normal donor prostates showed significantly less or no methylation. Exposure of LNCAP and MDAPCa2b cell lines to 5-aza-dC and TSA resulted in the downregulation of 15-LO-1 gene expression. Our results demonstrate that 15-LO-1

Abbreviations: 5-aza-dC, 5-aza-2 -deoxycytidine; COBRA, combined bisulfite restriction analysis; HDACs, histone deacetylases; 15-LO-1, 15-lipoxygenase-type 1; TSA, trichostatin-A ∗ Correspondence to: Urological Research Laboratories, Shadyside Medical Center G-37, 5200 Centre Avenue, Pittsburgh, PA 15232, USA. Tel.: +1 41 26 233 914; fax: +1 41 26 233 907. E-mail address: [email protected] (U.P. Kelavkar). 1098-8823/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.prostaglandins.2006.05.015

186

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

promoter methylation is frequently present in PCa patients and identify a new role for epigenetic phenomenon in PCa wherein hypermethylation of the 15-LO-1 promoter leads to the upregulation of 15-LO-1 expression and enzyme activity contributes to PCa initiation and progression. © 2006 Elsevier Inc. All rights reserved. Keywords: 15-Lipoxygenase-1 (15-LO-1); Promoter; Methylation; Prostate cancer; High-grade PIN; Normal; Overexpression

1. Introduction Although the estimated US cases of diagnosed prostate cancer (PCa) in 2006 has slightly increased, as compared with previous years, a reduction in the number of deaths is projected. Nevertheless, it remains as one of the leading causes of cancer deaths among men in the United States [American Cancer Society, Facts and Figures 2006] in part because neither early onset intervention strategies for improving quality of life nor curative therapy for advanced disease exists [1]. The “baby boomer” generation will significantly exacerbate this major health problem. The age specific incidence of PCa increases after age 60 and in 2006–2007, 80 million “baby boomers” in the USA will approach this milestone. Other projections estimate that the numbers of new cases of PCa during the next 20 years will more than double and the number of men dying of the disease could double or triple [2]. These estimates underscore the urgent need for new biomarkers and therapeutic strategies that target the early stages of the disease. Alterations in epigenetic activity can have a direct and profound effect on carcinogenesis. For example, aberrant methylation of CpG islands in the promoter region of tumor suppressor genes can lead to their transcriptional silencing and concomitant loss of tumor suppressor activity [3,4]. There is also evidence that epigenetic alterations can also lead to up regulation of transcription [5]. Considering that epigenetic aberrations rather than genetic mutations play a key role in early PCa development, the reversibility of these changes with histone deacetylase and methylation inhibitors represents an attractive therapeutic option [6–8]. DNA methylation is an epigenetic alteration carried out by DNA methyltransferases (DNMT). These enzymes catalyze the covalent addition of a methyl group from a donor S-adenosylmethionine to the 5 position of cytosine, predominantly within CpG dinucleotides. A major gap in understanding methylation dysregulation and neoplastic transformation is the lack of knowledge about the mechanisms underlying sequential changes in methylation patterns during the preneoplastic period in vivo. Current knowledge is based primarily on comparisons of methylation profiles between normal cells and tumor cells and it is not clear whether methylation instability in tumor cells stems arises from an isolated determining event, or from progressive alterations in heritable methylation patterns. Studies from our laboratory [9–13] as well as from others [14–17] have reported a key role for 15-LO-1 in several diseases including high grade prostatic intraepithelial neoplasia (HGPIN) and prostate tumors. Fifteen-LO-1 is a highly regulated, tissue- and cell-type-specific polyunsaturated fatty acid-peroxidating enzyme that has several functions, ranging from physiological membrane remodeling to the pathogenesis of atherosclerosis, inflammation and carcinogenesis. Expression of the 15-LO-1 protein is upregulated in PCa, which in combination with a diet high in omega-6 polyunsaturated fatty acids (PUFAs), leads to the formation of tumorigenic metabolite 13-hydroxyoctodecadenoic acid (13-S-HODE) [18,19]. Due to the probable impact of 15-LO-1 on PCa development, targeting overexpression of 15-LO-1 expression may delay disease progression. However, what causes aberrant expression of 15-LO-1 in cancer cells remains unknown. Human 15-LO-1 is located on chromosome 17p13.3 [13], a region which is frequently methylated in PCa [20]. Based on previously published studies on 5-LO and 15-LO-1 promoter methylation patterns and expression [21–23], we hypothesized that 15-LO-1 may also undergo methylation of its promoter region in PCa. Therefore, in this study, we began by examining the methylation and expression status of 15-LO-1 in normal, benign prostatic hyperplasic (BPH) cells and malignant PCa cell lines. We then examined the effect of DNA methyltransferase and histone deacetylase inhibitors on 15-LO-1 expression. Furthermore, we characterized the methylation profile of a specific CpG residue in the 15-LO-1 promoter, that when methylated correlated with 15-LO-1 expression in the cell lines, in cancer-associated HGPIN, PCa specimens, normal tissue adjacent to tumor and normal prostate tissue from cancer-free organ donors. Our findings demonstrated that the hypermethylation of an HpyCH4IV site in the 5 -flanking region −217 to −474 nt of the 15-LO-1 gene is associated with the transcriptional activation of the 15-LO-1 gene in PCa.

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

187

2. Materials and methods 2.1. Cell lines and tissues All cell lines were grown in a 5% CO2 incubator at 37 ◦ C and 85% humidity. Primary prostate epithelial cells (PrEC) were maintained in PrEGM prostate epithelial cell medium (Clonetics, San Diego, CA). RWPE1, an immortalized normal prostate epithelial cell line and the PCa cell lines, LNCaP, PC-3, DU145 and MDAPCa2b were obtained from the American type culture collection (ATCC; Manassas, VA) and were maintained in the recommended medium. Los Angeles Prostate Cancer-4 (LAPC-4) PCa cells were kindly provided by Dr. Robert Reiter (University of California-Los Angeles, Los Angeles, CA) and maintained in phenol red-free Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA) containing 5% heat-inactivated fetal calf serum (Sigma, St. Louis, MO) with streptomycin-penicillin antibiotics. The benign prostatic hyperplasia cell line BPH-1 was kindly provided by Dr. Simon Hayward (Vanderbilt University, USA) and was maintained in RPMI-1640 medium supplemented with 5% FBS, HEPES buffer and penicillin-streptomycin antibiotics. The PC3-15LOS (15-LO-1-overexpressing) PCa cells were grown in RPMI/FBS medium containing 50 ␮g/ml Zeocin (Invitrogen) [11]. Sections from 43 surgically resected primary prostate tumor tissues (age range 41–>71 years) and 5 normal prostate specimens from cancer-free organ donors (age range 20–>71 years) were obtained from the Western Pennsylvania Genitourinary Tissue Bank. Full details of tissue pathology are listed in Table 1. Follow-up information was available from all but 9 patients. Follow-up information ranged from 2 weeks to 166 months, with a mean follow-up time at 91 months and a median of 90 months. Patients with disease recurrence are defined as those with a PSA value of ≥0.2 ng/ml post-prostatectomy or three consecutive increasing PSA values. All patients gave informed consent in accordance with the Institutional Review Board guidelines. 2.2. Combination 5-aza-2 -deoxycytidine and trichostatin-A treatment LNCaP, MDAPCa2b and BPH-1 (control) cells were plated onto several T75 cm2 flasks (a total of 2 × 107 cells each) and treated with a combination of 0.1 mM 5-aza-2 -deoxycytidine (5-aza-dC) and 0.1 mM trichostatin-A (TSA) Table 1 Clinical information of patients with prostate cancer used in this study. Clinical parameter

Frequency (%total)

Age range 41–50 51–60 61–70 >71

2(4.65) 9(20.9) 25(58.1) 7(16.2)

Preoperative serum PSA (ng/ml) Mean Median Range

13.8 9 1.8–90

Pathological stage Organ confined (T1–T2) Capsular penetration (T3a and T3b) Seminal vesicle involvement (T3c) Lymph node/distant metastases (N, M) Unknown Total

16(37.2) 8(18.6) 5(11.6) 5(11.6) 9(20.9) 43(100)

Cumulative Gleason score 6 7(3 + 4) 7(4 + 3) 8–10 Total

11(25.6) 12(27.9) 12(27.9) 8(18.6) 43(100)

188

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

for a total of five passages. Untreated cells served as controls. Medium containing fresh drugs was replaced every 72 h for the duration of the experiment. RNA was harvested after five passages for quantitative reverse transcription (qRT-PCR) analyses. 2.3. DNA and RNA extraction Genomic DNA was isolated from prostate cell lines with a Wizard DNA Purification System (Promega, Madison, WI) according to the manufacturer’s instructions. The quality and integrity of the DNA was determined by the A260/280 ratio. RNA was extracted from drug treated cells using RNeasy kit (Qiagen, 2 × 105 TRIzol reagent (Invitrogen, Carlsbad, CA). All RNA samples were treated with DNase1 (Invitrogen) for 15 min at room temperature before use. 2.4. Tissue microdissection Because of the prostate tissue heterogeneity, each tumor sample was sectioned in two consecutive 5- and 10␮M thick sections. The 5-␮M slide was stained with hematoxylin and eosin (H and E) and regions of cancer, HGPIN, hyperplasia and normal prostate were marked by a pathologist. The 10-␮M section was deparaffinized, stained in Evan’s Blue solution (0.5%, w/v) for 10 min, followed by microdissection of marked regions under a dissecting microscope using Pinpoint resin (Zymo Research, Orange, CA). Briefly, the resin was applied to the marked regions, allowed to dry for 45 min and then carefully lifted off with a disposable scalpel and fine forceps to prevent cross-contamination between samples. The tissue was then incubated in proteinase K at 55 ◦ C for 4 h and genomic DNA purified using a DNA-binding silica column (Zymo Research). 2.5. Bisulfite conversion, cloning, sequencing from cell lines Denatured DNA (0.5 ␮g) from cell lines PrEC, RWPE-1, BPH-1, DU-145, LAPC-4, LNCaP, MDAPCa2b and PC-3 was bisulfite converted using the EZ DNA Methylation Kit (Zymo Research) according to the manufacturer’s directions. Briefly, the denatured DNA was incubated for 16 h in sodium bisulfite and then desalted using DNA-binding columns. Desulphonation by incubation in sodium hydroxide was carried out within the column. The modified DNA was then eluted in 20 ␮l elution buffer and 2 ␮l of the recovered DNA was used for PCR analysis. Bisulfite treatment converts non-methylated cytosines into uracil via deamination, which is replicated as thymidine during PCR. In contrast, 5methyl cytosines are protected and thus identified as cytosines in the resultant PCR product. Bisulfite modified DNA was amplified by PCR with the following reaction conditions: 1X buffer (BD Biosciences; Mountain View, CA), 0.5 mM of each primer, 0.2 mM of each dNTP, 0.5 unit of Titanium Taq (BD Biosciences, Franklin Lakes, NJ) and 2 ␮l of bisulfitemodified DNA in a final volume of 25 ␮l. Primers were: F15LO1COBRA5 -TTTGTAATTTAATTTGTGAGGTTTG-3 and R15LO1COBRA5 -CAAAAAATAAAAACCACTATCTTAAC-3 (a 258 bp PCR fragment). They are designed such that will amplify only the bisulfite converted DNA. The PCR conditions were as follows: 95 ◦ C for 10 min for denaturation, initial 5 cycles of amplification (95 ◦ C, 30 s; 60 ◦ C, 30 s; 72 ◦ C, 45 s) and then 40 cycles of amplification (95 ◦ C, 30 s; 57 ◦ C, 30 s; 72 ◦ C, 45 s) with a final elongation step of 10 min at 72 ◦ C. PCR products were cloned in into pCR2.1-TOPO vector using the Invitrogen TOPO TA cloning kit (Invitrogen Life Technologies, Carlsbad, USA) and were transformed in Escherichia coli Mach1TM -T1R cells. Blue/white screening was performed after 12 h and plasmids extracted and purified from ten white colonies for each cell line (except 9 colonies for DU145 cells) were sequenced by dye-terminator cycle-sequencing in both directions (both strands) on a ABI 3100 16-capillary instrument (Applied Biosystems, Foster City, CA) to identify CpG methylation. 2.6. Bisulfite conversion and combined bisulfite restriction analysis (COBRA) Briefly, 0.5 ␮g of denatured DNA from cell lines or 45 ␮l eluted DNA from microdissected sections were subjected to bisulfite conversion as described above. The methylation status of CpG 10 within the15-LO-1 promoter can be determined as a percentage by subjecting the PCR product of the bisulfite modified DNA by COBRA analysis [24] utilizing restriction digestion with the enzyme HpyCH4IV (New England Biolabs, Ipswich, MA), which cuts specifically at 5 . . ..ACGT. . .. 3 at CpG 10 (Fig. 2) followed by agarose gel electrophoresis. This gives a 258 bp (uncut indicating unmethylated DNA fragment) and two 129 bp (cut indicating unmethylated DNA fragment) fragments. Fifteen-LO-1

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

189

promoter fragment was PCR amplified from the bisulfite modified DNA as described above. Universally methylated DNA was bisulfite converted and included in the analysis as a control for methylation in the COBRA analyses. Normal male lymphocyte DNA, which is usually unmethylated at most promoter sites, was bisulfite converted and included as a control to represent unmethylated DNA. 2.7. Real-time reverse transcriptase polymerase chain reaction (RT-PCR) analysis Total RNA from cell lines isolated by RNeasy kits was treated with DNase I (Qiagen, Valencia, CA), following the manufacturers protocol. RNA was quantified by spectrophotometry (Eppendorf, Germany) and its integrity was assured by analysis using BioRad Experion RNA analyzer chip (BioRad Inc., Hercules, CA). Real-time quantitative PCR (qRT-PCR) reactions were performed in a 25 ␮l mixture containing first-strand cDNA synthesized using 1 ␮g of total RNA (DNase-treated) and reverse transcriptase reaction mixture. A 120-bp region of ␤-actin using primers 5 -CCTGGCACCCAGCACAAT-3 and 5 -GCCGATCCACACGGAGTACT-3 was amplified at 95 ◦ C/10 min, [95 ◦ C/30 s, 59 ◦ C/60 s] × 40 cycles using 1X SYBR Green and buffer (PE Applied Biosystems, Foster City, CA, USA), 4 mM MgCl2 , 0.2 ␮M of each primers (␤-actin and 15-LO-1), 0.2 mM dNTPs mix and 0.025 unit of AmpliTaq Gold® thermostable DNA polymerase (Applied Biosystems, Foster City, CA, USA). A 192 bp region of human 15-lipoxygenase-1 (15-LO-1) using primers 15-LO-1 5 -GACCGAGGGTTTCCTGTCTC-3 and 5 TGTCTCCAGCGTTGCATCC-3 was similarly amplified (but without SYBR Green) at 95 ◦ C/3 min, [95 ◦ C/30 s, 58 ◦ C/60 s] × 40 cycles and quantified by a TaqMan probe◦ 5 -5HEX-CAGGCTCGGGACCAGGTTTGCCAGBHQ2a∼5HEX-3 . Real-time quantitation was performed using the iQ5 Real-Time PCR Detection System (BioRad, Hercules, CA, USA). The fluorescence threshold value was calculated using the system software. Optimization experiments showed that PCR for ␤-actin in triplicate were highly reproducible with a low intra-assay coefficient of variation (0.5%). Relative expression values were represented as 15-LO-1 relative fluorescence units (RFU’s) normalized to ␤-actin from the same sample. The PCR products were subjected to 2% agarose gel electrophoresis; stained with ethidium bromide, visualized under UV illumination and quantified by scanning densitometry using the Gel Doc video camera and Quantity One 4.1.1 software (Bio-Rad). The images were divided into their respective lanes, their background subtracted and 258 bp (uncut) and 129 bp (cut) bands were analyzed. A 100 bp ladder and the positive and negative digested controls were used for gel alignment. The normal peak density was then exported to Microsoft excel and the density values normalized for intercalation fluorescence bias by taking the natural log of the value and dividing by the number of base pairs of the fragment. The percentage of methylation was calculated by dividing the density value of the methylated band (128 bp) by the sum values of the lower and upper band densities × 100. 2.8. Immunohistochemical and image cytometric analysis of prostate tissues Sections of formalin-fixed, paraffin-embedded tissue (5 ␮m) were tested for the presence of 15-LO-1 [1:1600] using an avidin biotin-complex technique and steam heat-induced antigen retrieval and quantitatively examined by image cytometry as described previously [9]. 2.9. Statistical analysis Statistical analysis was performed using the statistical package Splus 6.2 (Insightful Corporation, Seattle, WA) either by students t-test with unequal variance and two tails or by Fisher’s Exact Test with a significance threshold of 0.01. 3. Results 3.1. Relationship between 15-LO-1 mRNA expression and promoter methylation in human prostate cell lines To examine the methylation and expression status of 15-LO-1, we used prostate cell lines representative of normal tissue as well as benign and malignant prostate disease. Based on real-time polymerase chain reaction (RT-PCR) analysis, 15-LO-1 mRNA was undetectable in the normal prostate cell lines PrEC (Fig. 1, lane 7) and RWPE1 (Fig. 1, lane 8). However, low levels were detected in BPH1, DU145, LAPC4 (Fig. 1, lanes 1–3) and PC3 (Fig. 1, lane 6) cells

190

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

Fig. 1. Reverse-transcriptase (RT)-PCR of 15-lipoxygenase-1 (15-LO-1) and beta (␤)-actin expression in prostate cell lines. Lane 1: BPH-1; lane 2: DU145; lane 3: LAPC4; lane 4: LNCaP; lane 5: MDAPCa2b; lane 6: PC3; lane 7: PrEC; lane 8: RWPE1 and lane 9: PC3-15LOS.

while high levels of 15-LO-1 mRNA were detected in LNCaP and MDAPCa2b cell (Fig. 1, lanes 4 and 5). A stably transfected PC3-15LOS cell line served as a positive control (Fig. 1, lane 9). To study whether epigenetic mechanisms trigger aberrant 15-LO-1 expression in PCa, we examined the CpG islands in the 15-LO-1 promoter as previously studied by Liu et al.[25]. The 15-LO-1 promoter contains two CpG islands; one starting from −1 to −216 nt and the other from −217 to −474 nt relative to the ATG [26] (Fig. 2). There were no differences observed in the 15-LO-1 mRNA expression levels and the methylation status within the first CpG island spanning from −1 to −216 nt CpG island examined in normal as well as PCa cells (data not shown). We therefore focused our efforts on the second CpG island, (−217 to −474 nt). We PCR amplified the −217 to −474 nt region from bisulfite converted genomic DNA extracted from PrEC, RWPE-1, BPH-1, DU-145, LAPC-4, LNCaP, MDAPCa2b and PC-3 cells. The amplified DNA was cloned into TA vectors and 10 plasmids from each cell line were sequenced in both directions. Comparing the RT-PCR expression data (Fig. 1) and the sequences obtained for individual cell lines, we found that from a total of 15 CpG dinucleotides in the second CpG island, methylation of the 10th CpG dinucleotide (Fig. 2 shown by *) correlates with 15-LO-1 mRNA expression in LNCaP and MDAPCa2b cells, as shown in the methylation map (Figs. 2 and 3).

Fig. 2. Schematic illustration of the 15-LO-1 promoter sequence. The numbering of promoter sequence is indicated by its position (reverse strand) on chromosome 17p starting at 44,991,559 nt as well as in noted in parenthesis based upon the translation start point (ATG) indicated by 3 asterisks (***). The HpyCH4IV site used for the COBRA assay is indicated by an arrow. The CpG sites are indicated by an asterisk (*). Note that the boxed primer binding sites are designed for converted DNA and the CpG site 10 (boxed) must be methylated to be digested by HpyCH4IV (. . ..ACGT. . ..3 ) in COBRA analysis.

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

191

Fig. 3. Sequencing analysis and methylation map of the CpG island spanning the region −217 to −474 nt within the 15-LO-1 promoter from prostate cell lines. Bisulfite converted genomic DNA from PrEC, RWPE-1, BPH-1, DU-145, LAPC-4, LNCaP, MDAPCa2b and PC-3 cell lines was PCR amplified, TA-cloned and transformed in competent E. coli. A total of 9–10 clones were analyzed by plasmid sequencing from every cell line. Closed circle: methylated cytosine residue; open circle: unmethylated cytosine residue and boxes represent CpG dinucleotide 10. The numbers shown above the circles depict the cytidine residue.

It is interesting to note that the methylation of one CpG site among the 15 examined (Figs. 2 and 3) was correlated with 15-LO-1 expression in LNCaP and MDAPCa2b cells. Thus, methylation at other CpG sites within the 15-LO-1 CpG island promoter region does not seem to interfere with transcription in PCa. (Fig. 3). 3.2. Methylation status of CpG 10 in the second CpG island within 15-LO-1 promoter from normal and cancer cell lines of prostate To further evaluate the methylation status of the 15-LO-1 promoter, we used combined bisulfite restriction analysis (COBRA). Bisulfite-treated genomic DNA from all the prostate cell lines was PCR amplified using F15LO1COBRA and R15LO1COBRA primers, yielding a 258 bp amplified DNA fragment. This amplified fragment was then restriction digested with HpyCH4IV. All the normal cell lines (i.e., BPH1, RWPE1 and PrEC) showed only one uncut 258 bp band (Fig. 4A, lanes 1–3), indicating that the CpG 10 was unmethylated in normal prostate epithelial cells. However, cell lines LNCaP and MDAPCa2b showed 1 major band corresponding to the 129 bp (cut, methylated) band and a faint band at 258 bp (uncut, unmethylated) (Fig. 4A, lanes 4–8). This observation confirms our results shown in the methylation map (Fig. 3). In summary, our data show that partial promoter hypermethylation of 15-LO-1 exists in all PCa cell lines examined. The extent of methylation of the CpG dinucleotide 10 in the second CpG island of 15-LO-1 promoter corresponded with increased expression in LNCaP and MDAPCa2b cell lines. Notably, this aberrant methylation was absent in normal prostate cell lines. 3.3. Methylation status of CpG 10 in the 15-LO-1 promoter from microdissected human prostate tissues To demonstrate that our observations in cell lines also occur in vivo, 15-LO-1 promoter methylation was assessed in 43 microdissected primary PCa specimens, 37 corresponding adjacent normal sections (taken between 2 and 17 mm

192

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

Fig. 4. (A): DNA methylation status and combined bisulfite restriction analysis (COBRA) analysis of the 15-LO-1 promoter in prostate cell lines. Bisulfite-treated genomic DNA from PrEC (lane 1), RWPE-1 (lane 2), BPH-1 (lane 3), DU-145 (lane 4), LAPC-4 (lane 5), LNCaP (lane 6), MDAPCa2b (lane 7) and PC-3 (lane 8) cell lines was PCR amplified and restriction digested with HpyCH4IV. Lane 9 is methylation positive control, lane 10 is unmethylation positive control and lane 11 is unconverted (not bisulfite treated) negative control. The amplicons contain the CpG island spanning the region −217 to −474 nt. The digested DNA was separated in a 2.5% agarose gel and visualized by ethidium bromide (EtBr) staining. UM: unmethylated DNA derived fragment, M: methylated DNA derived fragment; (B) tissue microdissection from paraffin embedded sections. Sections from surgically resected primary prostate were obtained. Tumor, HGPIN and adjacent normal regions were marked by a pathologist, microdissected from the sections, subjected to DNA extraction, bisulfite converted, PCR amplified and analyzed by COBRA as described in Section 2. Representative slides from patients PR253, PR254, PR288 and PR293 containing marked sections (N or AN, normal surrounding or adjacent normal; P, HGPIN and T, tumor) before (shown on a hematoxylin and eosin, H and E slide) and after (shown on a parallel cut and Evan’s blue stained slide) microdissection. X and Y: COBRA analyses of microdissected AN and T. UM: unmethylated DNA derived fragment, M: methylated DNA derived fragment; (C) DNA methylation status and combined bisulfite restriction analysis (COBRA) analysis of the 15-LO-1 promoter in prostate cancer specimens. COBRA analyses on representative samples of tumor tissue (T), adjacent normal tissue (AN) and HGPIN (P) samples. The amplified DNA was digested with HpyCH4IV restriction enzyme and separated in a 2.5% agarose gel, visualized by ethidium bromide (EtBr) staining, quantified by scanning densitometry using the BioRad Gel Doc video camera and Quantity One 4.1.1 software as described in Section 2; (D) Methylation status of 15-LO-1 CpG dinucleotide 10 in prostate tissues. Densitometry analyses of amplified DNA from tumor (cancer, n = 43), adjacent normal (n = 37) and HGPIN (PIN, n = 10) from prostate cancer patients and normal donors (n = 5) digested with HpyCH4IV. The images of 258 bp (uncut) and 129 bp (cut) bands obtained by Gel Doc video camera and Quantity One 4.1.1 software were analyzed as described in Section 2. The percentage of methylation at CpG dinucleotide 10 in prostate samples was calculated by dividing the value for the methylated band (129 bp) by the sum values of the lower and upper band densities. Cancer vs. donor, P < 0.05; cancer vs. normal, P < 0.05 and donor vs. HGPIN, P < 0.1.

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

193

from the tumor, median 10 mm), 10 adjacent HG-PIN sections and 5 normal donor prostates (Table 1). Fig. 4B shows representative slides from patients PR253, PR254, PR288 and PR293 containing marked sections (N or AN, normal surrounding or adjacent normal; P, HGPIN and T, tumor) before (shown on a hematoxylin and eosin, H and E slide) and after (shown on a parallel cut Evan’s blue stained slide) microdissection (Fig. 4B). Representative COBRA analyses of tumor, HGPIN and adjacent normal regions are shown in Fig. 4BX, BY and C. Overall, COBRA analyses revealed that approximately 35% of cancer specimens (P < 0.05 versus normal donor tissue), 20% of adjacent normal specimens, 36% adjacent HG-PIN samples (P < 0.1 versus donor) and 0% normal donor specimens were positive for 15-LO-1 promoter methylation at CpG 10 (Fig. 4D). Given that there was 20% methylation observed in adjacent normal specimens, we examined this DNA for GSTPi methylation and confirmed that there was no crosscontamination between tumor samples and adjacent normal tissue that could yield a false positive result. We have previously found that our technique avoided the potential contamination pitfalls associated with manual microdissection [27]. We statistically compared the methylation data to the 15-LO-1 immunohistochemistry (IHC) analyses (data not shown) performed on all the tissue samples to assess whether a correlation existed between the protein expression and methylation. There was no significant correlation between methylation and protein expression in the tumor, Gleason score or comparing methylation status in tumors from different pathological stages. Similarly, no significant difference in methylation is observed when cancers are compared to normal surrounding tissue, normal surrounding tissue are compared to donors or donors are compared to HGPIN. However, interestingly, when donors were compared to cancer samples, there is a statistically significant difference (P < 0.001) in methylation as well as in 15-LO-1 protein expression (P < 0.01). This observation suggests that there is a correlation between mRNA expression levels and actual protein expression in prostate cell lines which perhaps may be as well in the tissues. 3.4. Requirement of DNA methylation for induction of 15-LO-1 mRNA Normally CpG hypermethylation can result in recruitment of methylation binding proteins (MBPs) and histone deacetylation, which leads to transcriptional silencing. As a recent study had indicated that the 15-LO-1 promoter is regulated by DNA methylation in normal T cells [25], we asked whether demethylation of the 15-LO-1 promoter in prostate cancer cells might contribute to transcriptional upregulation of 15-LO-1. To our surprise, we found that methylation of a specific CpG in prostate cancer cells resulted in transcriptional upregulation of 15-LO-1, which was the opposite of our expectations. To examine the phenomenon further, we took high 15-LO-1-expressing LNCaP and MDAPCa2b cells and low 15-LO-1-expressing BPH-1 cells and treated them in vitro with a combination (0.1 mM of each) of the DNA methylation inhibitor 5-aza-2 -deoxycytidine and the histone deacetylase inhibitor trichostatin-A (TSA) for five passages. Treatment of these cell lines significantly reduced the levels of 15-LO-1 mRNA. The mRNA expression was significantly reduced by two-fold in both the cell lines with P = 0.001 (LNCaP cells treated versus the untreated condition) and P = 0.006 (MDAPCa2b cells treated versus the untreated condition) (Fig. 5). This strongly suggested that hypomethylation of the CpG 10 in the second CpG island of the15-LO-1 promoter is a prerequisite for the 15-LO-1 gene inactivation in prostate cells. 4. Discussion Epigenetic DNA methylation profoundly impacts carcinogenesis; thus, to fully comprehend the mechanisms underlying the etiology and progression of cancer will require a thorough exploration of the relationship between epigenetics and cancer. In addition, a more complete understanding of epigenetics has the potential to lead to the development of novel biomarkers that may significantly improve clinical cancer screening, risk assessment and treatment. Epigenetic processes, such as methylation, have several molecular effects [28]. Methylation can directly interfere with the binding of transcription factors to inhibit transcription. Similarly, methylated CpG binding proteins (MBPs) can bind to methylated DNA as well as to regulatory proteins, again inhibiting transcription. In addition, both DNA methyltransferase (DNMT)-1 and MBPs, such as methyl-CpG-binding protein 2 (MeCP2), can recruit histone deacetylase. Deacetylation of core histone tails by this enzyme results in tighter packing of the DNA into chromatin, reducing the access of transcription factors and thus impeding transcription. Besides an essential role in X-chromosome inactivation and imprinting, CpG methylation in mammalian DNA has been shown to be important in controlling transcription of many other genes, including tumor suppressors [29,30].

194

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

Fig. 5. Real-time PCR analysis of 15-lipoxygenase-1 (15-LO-1) expression in BPH1 (control), LNCaP and MDAPCa2b cell lines treated with (+) and without (−) DNA methyltransferase inhibitor (AZA i.e., 5-aza-dC) and histone deacetylase inhibitor (TSA). Fifteen-LO-1 expression is expressed as the relative expression examined by relative fluorescence units (RFU’s) normalized to ␤-actin. P = 0.001 in comparison of LNCaP cells treated vs. untreated condition. P = 0.006 in comparison of MDAPCa2b cells treated vs. untreated condition.

Because 15-LO-1 gene is highly regulated and specifically overexpressed in PCa, we attempted to define a potential relationship between DNA methylation and 15-LO-1 expression. We demonstrate a novel process of abnormal hypermethylation of a 15-LO-1 CpG island that potentially occurs commonly in human HGPIN and in cancer. Our study is intriguing since one copy of 15-LO-1 gene becomes methylated in nearly all HGPIN and PCa tissues and yet the island remains essentially unmethylated in normal tissues. In addition, our study identified a CpG dinucleotide within the 15-LO-1 promoter whose methylation status is directly associated with 15-LO-1 mRNA expression in human cell lines and quite likely to be expressed in tissues as well. Most interestingly, we observed methylation of this CpG dinucleotide in the normal adjacent cells to the tumor but not in donors. Changes that occur prior to the histological changes that indicate the presence of cancer have been previously defined as “field effects” and have been shown to occur in PCa [27,31,32]. Therefore, changes in 15-LO-1 methylation levels may represent a field effect. These “subtle” changes, not observed by immunohistochemistry in “normal appearing tissue”, are early precursor changes associated with cancer progression and therefore differences in methylation of the CpG dinucleotide 10 in the 15-LO-1 promoter but not 15-LO-1 protein overexpression in normal appearing adjacent cells may occur very early in cancer progression. Aberrant 15-LO-1 overexpression may also result from the mutator response (MR) effect that becomes activated in cellular environments with persistent inflammation as well as sustained proliferative and survival signals. This sustained stress environment (SSE) [33] could give rise to primed cells carrying epigenetic alterations of specific genes, such as 15-LO-1, leading to initiation and progression of cancer. This effect could also account for the observation that 15-LO-1 expression decreases in colorectal and esophageal cancers and that restoration of 15-LO-1 expression by non-steroidal anti-inflammatory drugs (NSAIDs) leads to increased apoptosis [23,34–36]. If the cellular environment in colorectal and esophageal cancer cannot quickly adapt to SSE after NSAID treatment, it can either lead to cell growth arrest or apoptosis. As the authors observed in these NSAID-challenged cancer cells [23,34–36], it is highly likely that the two events, 15-LO-1 upregulation and apoptosis, are mutually exclusive events. This also supports a probable role for 15-LO-1 as a stress survival protein [11], as it is also expressed in cancer cells when challenged with stress-associated cell growth inhibition. Histone deacetylation may be involved in the transcriptional repression mediated by DNA methylation [29]. Methylated CpGs bind to a methylcytosine binding protein2 (MeCP2) that in turn recruits corepressors-HDACs-containing complexes. The HDACs recruited to the specific promoter lead to accumulation of deacetylated histones, which then makes nucleosomal structures inaccessible to the transcriptional machinery. We observed that combined treatment of the cells with both the methylation and deacetylation inhibitors resulted in silencing the 15-LO-1 expression in PCa cell lines overexpressing 15-LO-1 mRNA. Taken together, our findings suggest that disruption of DNA methylation facil-

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

195

itates histone acetylation at the 15-LO-1 promoter by TSA, leading to significant 15-LO-1 gene silencing. Similarly, in one other form of lipoxygenase both DNA methylation and histone acetylation have been implicated in controlling expression of the gene for the lipoxygenase, 5-LO [21,37]. Reversibility of DNA methylation, in contrast to structural modifications of the genome, provides the possibility of treating tumor cells with agents that modify DNA methylation. Indeed, inhibitors of DNA methyltransferase-1 (antisense oligonucleotides and nucleoside cytidine analogs) demonstrate anti-tumor effect on laboratory models in some cases and are in the first stages of clinical trial [38,39]. However, recent studies suggest that, in addition to local hypermethylation, global demethylation also contributes to carcinogenesis. Global demethylation can activate expression of proto-oncogenes as well as genes associated with metastases. These data indicate that the carcinogenic effect of DNA demethylation should be taken into account for design and trial of agents for anti-tumor therapy [38,39]. Since 15-LO-1 plays roles in inflammation and cancer for several types of tumor cells, inactivation of the 15-LO-1 gene by DNA methylation inhibitors as well as HDAC inhibitors may represent an attractive chemoprevention therapeutic approach for these diseases [40]. In addition to acting as a therapeutic target, 15-LO-1 hypermethylation in combination with markers found in PCa [41] may also represent an ideal tumor biomarker for detecting early onset of PCa progression. The 15-LO-1 promoter methylation and enzyme activity is readily detectable in clinical specimens obtained through minimally invasive procedures such as biopsies. Also, methylation occurs in a defined region (i.e., the second CpG island) and can be detected using techniques with high sensitivity, such as COBRA analysis and with high resolution, such as pyrosequencing [42]. Lastly, it is plausible to utilize hypermethylation of 15-LO-1 DNA in combination with different types of tumor apparently having their own signature of methylated genes, such as the methylation of GSTPi in PCa [43]], the mismatch repair gene MLH1 in colon cancer [44] and APC in esophageal cancer [45]. Given the long latency period of prostate disease progression coupled with the close association between aging and disease incidence, any treatment that impedes PCa progression may prove beneficial. For e.g., if 15-LO-1 methylation alone could be detected in prostates of 20% of patients with biopsies that were pathologically negative for PCa and if an early intervention is administered, it will have a significant impact on the quality of life of these individuals. Thus, the clinical utility of 15-LO-1 as an early biomarker and then as a therapeutic target may represent a viable treatment strategy, especially for the baby boomers approaching the age which is associated with increased incidence of PCa. In summary, this study has demonstrated, for the first time, that methylation of 15-LO-1 occurs in PCa, cancerassociated HGPIN and normal prostate that is adjacent to cancer, suggesting that 15-LO-1 methylation represents an early event in PCa initiation and development and may serve as a valuable marker for disease initiation and progression. Fifteen-LO-1 can also be a potential therapeutic target for preventing the initiation of prostate disease by utilizing a combination of DNA methyltransferase and histone deacetylase inhibitors. Acknowledgements We sincerely thank Dr. Moira Hitchens for critical reading and editing the manuscript. This work was fully supported by the Hillman Foundation Award to UPK and by the U.S. Army Department of Defense grant (W81XWH) and Edwin Beer program of the New York Academy of Medicine to DSOK. This work in no way reflects the opinion of the U.S. government. References [1] Marr PL, Elkin EP, Arredondo SA, Broering JM, DuChane J, Carroll PR. Comorbidity and primary treatment for localized prostate cancer: data from CaPSURET. J Urol 2006;175:1326–31. [2] Chan JM, Jou RM, Carroll PR. The relative impact and future burden of prostate cancer in the United States. J Urol 2004;172:S13–6 [discussion S17]. [3] Hanson JA, Gillespie JW, Grover A, et al. Gene promoter methylation in prostate tumor-associated stromal cells. J Natl Cancer Inst 2006;98:255–61. [4] DeMarzo AM, Nelson WG, Isaacs WB, Epstein JI. Pathological and molecular aspects of prostate cancer. Lancet 2003;361:955–64. [5] De Larco JE, Wuertz BRK, Yee D, Rickert BL, Furcht LT. Atypical methylation of the interleukin-8 gene correlates strongly with the metastatic potential of breast carcinoma cells. PNAS 2003;100:13988–93. [6] Steele VE, Holmes CA, Hawk ET, et al. Potential use of lipoxygenase inhibitors for cancer chemoprevention. Expert Opin Investig Drugs 2000;9:2121–38.

196

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

[7] Murgo. Innovative approaches to the clinical development of dna methylation inhibitors as epigenetic remodeling drugs. Sem Oncol 2005;32:458–64. [8] Stern N, Natarajan R, Tuck ML, Laird E, Nadler JL. Selective inhibition of angiotensin-II-mediated aldosterone secretion by 5hydroxyeicosatetraenoic acid. Endocrinology 1989;125:3090–5. [9] Kelavkar U, Cohen C, Kamitani H, Eling TE, Badr K. Concordant induction of 15-lipoxygenase-1 and mutant p53 expression in human prostate adenocarcinoma: correlation with Gleason staging. Carcinogenesis 2000;21:1777–87. [10] Kelavkar U, Glasgow W, Eling TE. The effect of 15-lipoxygenase-1 expression on cancer cells. Curr Urol Rep 2002;3:207–14. [11] Kelavkar UP, Nixon JB, Cohen C, Dillehay D, Eling TE, Badr KF. Overexpression of 15-lipoxygenase-1 in PC-3 human prostate cancer cells increases tumorigenesis. Carcinogenesis 2001;22:1765–73. [12] Kamitani H, Taniura S, Ikawa H, Watanabe T, Kelavkar UP, Eling TE. Expression of 15-lipoxygenase-1 is regulated by histone acetylation in human colorectal carcinoma. Carcinogenesis 2001;22:187–91. [13] Kelavkar UP, Badr KF. Effects of mutant p53 expression on human 15-lipoxygenase-promoter activity and murine 12/15-lipoxygenase gene expression: evidence that 15-lipoxygenase is a mutator gene. Proc Natl Acad Sci USA 1999;96:4378–83. [14] Blask DE, Brainard GC, Dauchy RT, et al. Melatonin-depleted blood from premenopausal women exposed to light at night stimulates growth of human breast cancer xenografts in nude rats. Cancer Res 2005;65:11174–84. [15] Blask DE, Dauchy RT, Sauer LA, Krause JA. Melatonin uptake and growth prevention in rat hepatoma 7288CTC in response to dietary melatonin: melatonin receptor-mediated inhibition of tumor linoleic acid metabolism to the growth signaling molecule 13-hydroxyoctadecadienoic acid and the potential role of phytomelatonin. Carcinogenesis 2004;25:951–60. [16] Sauer LA, Dauchy RT, Blask DE. Polyunsaturated fatty acids, melatonin, and cancer prevention. Biochem Pharmacol 2001;61:1455–62. [17] Blask DE, Sauer LA, Dauchy R, Holowachuk EW, Ruhoff MS. New actions of melatonin on tumor metabolism and growth. Biol Signal Recept 1999;8:49–55. [18] Kelavkar UP, Lin Y, Landsittel D, Chandran U, Dhir R. The yin and yang of 15-lipoxygenase-1 and delta-5-desaturase: Dietary omega-6 linoleic acid metabolic pathway in prostate carcinogenesis. J Carcinogenesis 2006;5:9. [19] Kelavkar U, Hutzley J, Dhir R, Kim P, Allen KGD, McHugh K. Prostate tumor growth and recurrence can be modulated by the omega-6:omega-3 ratio in diet: athymic mouse xenograft model simulating radial prostatectomy. Neoplasia 2006;8:112–24. [20] Morton Jr RA, Watkins JJ, Bova GS, Wales MM, Baylin SB, Isaacs WB. Hypermethylation of chromosome 17P Locust D17S5 in human prostate tissue. J Urol 1996;156:512–6. [21] Uhl J, Klan N, Rose M, Entian K-D, Werz O, Steinhilber D. The 5-lipoxygenase promoter is regulated by DNA methylation. J Biol Chem 2002;277:4374–9. [22] Li C, Xu D, Sjoberg J, Forsell P, Bjorkholm M, Claesson HE. Transcriptional regulation of 15-lipoxygenase expression by promoter methylation. Exp Cell Res 2004;297:61–7. [23] Liu LC, Xi D, Sjoberg J, Forsell P, Bjorkholm M, Claesson HE. The methyltransferase inhibitor 5-aza-2-deoxycytidine induces apoptosis via induction of 15-lipoxygenase-1 in colorectal cancer cells. Mol Cancer Ther 2005;4:1740–6. [24] Xiong Z, Laird PW. COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res 1997;25:2532–4. [25] Liu C, Xu D, Sjoberg J, Forsell P, Bjorkholm M, Claesson H-E. Transcriptional regulation of 15-lipoxygenase expression by promoter methylation. Exp Cell Res 2004;297:61–7. [26] Kelavkar U, Wang S, Montero A, Murtagh J, Shah K, Badr K. Human 15-lipoxygenase gene promoter: analysis and identification of DNA binding sites for IL-13-induced regulatory factors in monocytes. Mol Biol Rep 1998;25:173–82. [27] Collard RL, Harya NS, Monzon FA, Maier CE, O’Keefe DS. Methylation of the ASC gene promoter is associated with aggressive prostate cancer. The Prostate 2006 [epub, n/a.]. [28] Bastian PJ, Yegnasubramanian S, Palapattu GS, Rogers CG, Lin X, De Marzo AM, Nelson WG. Molecular biomarker in prostate cancer: the role of CpG island hypermethylation. Eur Urol 2004;46:698–708. [29] Razin A. CpG methylation, chromatin structure and gene silencing-a three-way connection. EMBO J 1998;17:4905–8. [30] Masumori N, Thomas TZ, Chaurand P, et al. A probasin-large T antigen transgenic mouse line develops prostate adenocarcinoma and neuroendocrine carcinoma with metastatic potential. Cancer Res 2001;61:2239–49. [31] Yu YP, Landsittel D, Jing L, et al. Gene expression alterations in prostate cancer predicting tumor aggression and preceding development of malignancy. J Clin Oncol 2004;22:2790–9. [32] Chandran U, Dhir R, Ma C, Michalopoulos G, Becich M, Gilbertson J. Differences in gene expression in prostate cancer, normal appearing prostate tissue adjacent to cancer and prostate tissue from cancer free organ donors. BMC Cancer 2005;5:45. [33] Karpinets TV, Foy BD. Tumorigenesis: the adaptation of mammalian cells to sustained stress environment by epigenetic alterations and succeeding matched mutations. Carcinogenesis 2005;26:1323–34. [34] Shureiqi I, Lippman SM. Lipoxygenase modulation to reverse carcinogenesis. Cancer Res 2001;61:6307–12. [35] Shureiqi I, Chen D, Lee JJ, et al. 15-LOX-1: a novel molecular target of nonsteroidal anti-inflammatory drug-induced apoptosis in colorectal cancer cells. J Natl Cancer Inst 2000;92:1136–42. [36] Shureiqi I, Xu X, Chen D, et al. Nonsteroidal anti-inflammatory drugs induce apoptosis in esophageal cancer cells by restoring 15-lipoxygenase-1 expression. Cancer Res 2001;61:4879–84. [37] Klan N, Schnur SS, Jung N, Steinhilber DM. Trichostatin A and structurally related histone deacetylase inhibitors induce 5-lipoxygenase promoter activity. Biol Chem 2003;384:777–85. [38] Goffin J, Eisenhauer E. DNA methyltransferase inhibitors—state of the art. Ann Oncol 2002;13:1699–716. [39] Garcia-Manero GIJ. Histone deacetylase inhibitors: a review of their clinical status as antineoplastic agents. Cancer Invest 2005;23:635–42. [40] Dashwood RH, Myzak MC, Ho E. Dietary HDAC inhibitors: time to rethink weak ligands in cancer chemoprevention? Carcinogenesis 2006;27:344–9.

U.P. Kelavkar et al. / Prostaglandins & other Lipid Mediators 82 (2007) 185–197

197

[41] Bismar TADF, Riva A, Kim R, et al. Defining aggressive prostate cancer using a 12-gene model. Neoplasia 2006;8:59–68. [42] Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res 1994;22:2990–7. [43] Santourlidis S, Florl A, Ackermann R, Wirtz HC, Schulz WA. High frequency of alterations in DNA methylation in adenocarcinoma of the prostate. Prostate 1999;39:166–74. [44] Ricciardiello L, Goel A, Mantovani V, et al. Frequent loss of hMLH1 by promoter hypermethylation leads to microsatellite instability in adenomatous polyps of patients with a single first-degree member affected by colon cancer. Cancer Res 2003;63:787–92. [45] Kawakami K, Brabender J, Lord RV, et al. Hypermethylated APC DNA in plasma and prognosis of patients with esophageal adenocarcinoma. J Natl Cancer Inst 2000;92:1805–11.