A genome-wide search identifies epigenetic silencing of somatostatin, tachykinin-1, and 5 other genes in colon cancer.

BACKGROUND & AIMS: Gene silencing via promoter hypermethylation is a central event in the pathogenesis of cancers. To identify novel methylation targets in colon cancer, we conducted a genome-wide, microarray-based, in silico, and epigenetic search. METHODS: Complementary DNA microarray experiments were first performed to identify genes down-regulated in primary colon cancers and up-regulated in colon cancer cell lines after global DNA demethylation by 5-aza-2'-deoxycitidine. Candidate methylation targets were then identified by combining these microarray data with in silico genetic and functional searches. Candidate genes recognized by these searches were further investigated for promoter hypermethylation in colon cancer using methylation-specific polymerase chain reaction. RESULTS: We identified 51 novel and 3 known candidate methylation targets. Subsequent epigenetic analysis revealed that primary colon cancers demonstrated frequent methylation of somatostatin (SST, 30 of 34 cases, 88%) and the substance P precursor gene tachykinin-1 (TAC1; 16 of 34 cases, 47%). TAC1 methylation intensity was significantly higher in Dukes A/B than in Dukes C/D cancers (P = .01). SST methylation intensity was significantly higher in low-level microsatellite instability (MSI-L) than in non-MSI-L cancers (P = .02). Methylation was associated with messenger RNA down-regulation for both SST and TAC1. Furthermore, we isolated 5 additional novel promoter methylation targets: NELL1, AKAP12, caveolin-1, endoglin, and MAL. CONCLUSIONS: These data strongly suggest that SST and TAC1 are involved in colon carcinogenesis. Further studies are now indicated to elucidate mechanisms underlying their involvement in colon cancer and their values as clinical biomarkers. NELL1, AKAP12, caveolin-1, endoglin, and MAL are also promising tumor suppressor gene candidates deserving of further study.

Mutational and LOH analyses of the chromosome 4q region in esophageal adenocarcinoma.

OBJECTIVE: Mortality due to esophageal adenocarcinoma has risen markedly, but the molecular mechanisms underlying this carcinogenesis are still incompletely understood. Findings from loss of heterozygosity (LOH) studies have suggested that the long arm of chromosome 4 might harbor tumor suppressor genes relevant to esophageal adenocarcinoma. METHODS: We performed LOH analysis of 4q in esophageal adenocarcinomas. Regions of LOH were further evaluated by studying two candidate tumor suppressor genes, hCDC4 and CARF, located within them. RESULTS: 54% of the adenocarcinomas examined showed allelic deletion. LOH was observed in 53, 40, 32, 38, and 27% of tumors at positions D4S1554 (the locus of CARF), D4S1572, D4S1548, D4S2934, and D4S3021, respectively. An area of allelic deletion (spanning 3 million bases) was identified at 4q31.1-3 in 37% of tumors. This region harbors a candidate tumor suppressor gene: hCDC4. However, sequencing of the coding regions of CARF and hCDC4 at 4q35 and 4q31, respectively, did not identify mutations. CONCLUSIONS: Our findings demonstrate frequent LOH in esophageal adenocarcinoma at several loci including a novel area of allelic deletion at 4q31.1-3. The results imply that mutational or other alterations at these loci may be involved in the pathogenesis of esophageal adenocarcinoma. Candidate tumor suppressor genes located within these regions merit further study.

Promoter methylation and response to chemotherapy and radiation in esophageal cancer.

BACKGROUND & AIMS: Multiple studies have shown that promoter methylation of tumor suppressor genes underlies esophageal carcinogenesis. Hypothetically, methylation resulting in tumor suppressor gene inactivation might result in tumors that are unresponsive to chemotherapy and radiation. Accordingly, our aim was to find methylation markers that could be used to predict response to chemoradiation. METHODS: Tumor specimens were obtained before treatment from 35 patients enrolled in a uniform chemoradiation treatment protocol. Methylation-specific quantitative polymerase chain reaction was performed on all samples. Pathology reports from esophagectomy specimens were used to define response to treatment. RESULTS: Thirteen (37%) of 35 patients were responders, and 22 (63%) of 35 patients were nonresponders. The number of methylated genes per patient was significantly lower in responders than in nonresponders (1.4 vs 2.4 genes per patient; Student t test, P = .026). The combined mean level of promoter methylation of p16, Reprimo, p57, p73, RUNX-3, CHFR, MGMT, TIMP-3, and HPP1 was also lower in responders than in nonresponders (Student t test, P = .003; Mann-Whitney test, P = .001). The frequency (15% of responders vs 64% of nonresponders; Fisher exact test, P = .01) and level (0.078 in responders vs 0.313 in nonresponders; Mann-Whitney test, P = .037) of Reprimo methylation was significantly lower in responders than in nonresponders. CONCLUSIONS: Reprimo methylation occurred at significantly lower levels and less frequently in chemoradioresponsive than in nonresponsive esophageal cancer patients, suggesting potential clinical application of this single-gene biomarker in defining prognosis and management. In addition, increased methylation of a 9-gene panel correlated significantly with poor responsiveness to chemoradiation.

Polo-like kinase and survivin are esophageal tumor-specific promoters.

For developing successful cancer gene therapy strategies, tumor-specific gene delivery is essential. In this study, we used esophageal cancer (EC) cells to identify and evaluate esophageal tumor-specific gene promoters. Four genes (polo-like kinase-1/PLK, survivin/BIRC5, karyopherin alpha 2/KPNA2, and pituitary tumor transforming gene protein 1/PTTG1) were identified by a microarray analysis as highly expressed in EC cell lines vs. five normal organ tissues (liver, lung, kidney, brain, and heart). By quantitative RT-PCR, the average mRNA expression levels of these four genes in 20 primary ECs were 2.7-fold (PLK), 6.1-fold (survivin), 2.6-fold (KPNA2), and 2.4-fold (PTTG1) higher than that of each gene in 24 different normal organs. By dual luciferase assay, the promoter activity of PLK and survivin in EC cell lines was 18.9-fold and 28.5-fold higher, respectively, than in normal lung and renal cells. The promoters of PLK and survivin could be useful tools for developing EC-specific gene therapy vectors.

Molecular phenotype of inflammatory bowel disease-associated neoplasms with microsatellite instability.

BACKGROUND & AIMS: Patients with inflammatory bowel disease (IBD) are at increased risk of developing colorectal cancer (CRC). We sought to determine the frequency of high-level microsatellite instability (MSI-H) and the mutational and methylation profile of MSI-H IBD-related neoplasms (IBDNs). METHODS: A total of 124 IBDNs (81 cancers, 43 dysplasias) from 78 patients were studied for the frequency of MSI-H and hypermethylation of 3 target genes: MLH1 , HPP1 , and RAB-32 . Fifteen MSI-H IBDNs were characterized according to their profile of frameshift mutations in 28 mononucleotide repeats and compared with 46 sporadic MSI-H CRCs. RESULTS: Nineteen of 124 IBDNs were MSI-H. The frequency of frameshift mutations in coding mononucleotide repeats was significantly lower in MSI-H IBDNs than in sporadic MSI-H CRCs for TGFBR2 (7 of 14 vs 34 of 43 samples; P = .047) and ACVR2 (3 of 14 vs 25 of 43 samples; P = .029). In contrast, ICA1 was mutated in 3 of 9 MSI-H IBDNs vs 2 of 54 sporadic MSI-H CRCs ( P = .028). HPP1 and RAB32 methylation was independent of MSI status and was observed in 4 of 59 and 0 of 64 nondysplastic mucosae, 20 of 38 and 1 of 25 dysplasias, and 28 of 61 and 20 of 60 carcinomas, respectively. CONCLUSIONS: The profiles of coding microsatellite mutations (instabilotypes) differ significantly between MSI-H IBDNs and MSI-H sporadic CRCs. Specifically, TGFBR2 and ACVR2 mutations are significantly rarer in MSI-H IBDNs than in MSI-H sporadic CRCs. Furthermore, HPP1 methylation occurs early, in 7% of nondysplastic and approximately half of dysplastic mucosae, whereas RAB32 methylation occurs at the transition to invasive growth, being rarer in dysplasias.