Prognostic DNA methylation biomarkers in ovarian cancer.

PURPOSE: Aberrant DNA methylation, now recognized as a contributing factor to neoplasia, often shows definitive gene/sequence preferences unique to specific cancer types. Correspondingly, distinct combinations of methylated loci can function as biomarkers for numerous clinical correlates of ovarian and other cancers. EXPERIMENTAL DESIGN: We used a microarray approach to identify methylated loci prognostic for reduced progression-free survival (PFS) in advanced ovarian cancer patients. Two data set classification algorithms, Significance Analysis of Microarray and Prediction Analysis of Microarray, successfully identified 220 candidate PFS-discriminatory methylated loci. Of those, 112 were found capable of predicting PFS with 95% accuracy, by Prediction Analysis of Microarray, using an independent set of 40 advanced ovarian tumors (from 20 short-PFS and 20 long-PFS patients, respectively). Additionally, we showed the use of these predictive loci using two bioinformatics machine-learning algorithms, Support Vector Machine and Multilayer Perceptron. CONCLUSION: In this report, we show that highly prognostic DNA methylation biomarkers can be successfully identified and characterized, using previously unused, rigorous classifying algorithms. Such ovarian cancer biomarkers represent a promising approach for the assessment and management of this devastating disease.

The fragile histidine triad gene: a molecular link between cigarette smoking and cervical cancer.

PURPOSE: Smoking is an epidemiologic risk factor for cervical cancer. The fragile histidine triad (FHIT) gene is a tumor suppressor gene that is altered in 80% of tobacco-associated lung cancers. We hypothesized that reduced FHIT protein expression, homozygous deletions (HD) or hemizygous deletions (HemiD) and microsatellite alterations (MA) at the FHIT/FRA3B locus occur more commonly in cervical cancers of smokers than nonsmokers. EXPERIMENTAL DESIGN: Archival tissues of 58 patients with stage IA1 to IB2 squamous cell carcinoma of the cervix were identified. FHIT protein expression was studied with immunohistochemistry. Laser capture microdissection was used to isolate tumor and normal DNA. HD/HemiD of FHIT exons 4 and 5 were analyzed by monoplex real-time PCR. MA at FHIT/FRA3B were studied with multiplex nested PCR with three fluorescently labeled microsatellite markers (D3S1300, D3S1312, and D3S1480). RESULTS: Eighteen of 26 tumors from smokers (69%) and 13 of 32 nonsmokers (41%; P < 0.05) showed loss of FHIT protein expression. Thirty-seven stage IB tumors yielded sufficient DNA for analyses. HD or HemiD of both exons tested occurred in 8 of 17 smokers (47%) and 2 of 20 nonsmokers (10%; P < 0.05). MA at more than two sites were found in 11 of 17 tumors of smokers (65%) and 6 of 20 nonsmokers (30%; P < 0.05). Mean composite genomic FHIT alteration scores were significantly higher for tumors of smokers versus nonsmokers (0.67 versus 0.40; P < 0.02). CONCLUSION: Loss of FHIT expression, HD, HemiD, and MA at the FHIT/FRA3B locus occur significantly more commonly in cervical cancers of smokers. These findings suggest that the tumor suppressor gene FHIT may represent a molecular target in cigarette smoking-associated cervical carcinogenesis.

Loss of c-myc repression coincides with ovarian cancer resistance to transforming growth factor beta growth arrest independent of transforming growth factor beta/Smad signaling.

Many epithelial carcinomas, including ovarian, are refractory to the antiproliferative effects of transforming growth factor (TGF) beta. In some cancers, TGF-beta resistance has been linked to TGF-beta receptor II (TbetaR-II) and Smad4 mutations; however, in ovarian cancer, the mechanism of resistance remains unclear. Primary ovarian epithelial cell cultures were used as a model system to determine the mechanisms of TGF-beta resistance. To simulate in vivo responses to TGF-beta, primary cultures derived from normal human ovarian surface epithelium (HOSE) and from ovarian carcinomas (CSOC) were grown on collagen I gel, the predominant matrix molecule in the ovarian tumor milieu. When treated with 5 ng/ml TGF-beta for 72 h, HOSE (n = 11) proliferation was inhibited by 20 +/- 21% on average. In contrast, CSOC (n = 10) proliferation was stimulated 5 +/- 10% in response to TGF-beta (a statistically significant difference in response when compared with HOSE; P = 0.001). To dissect the TGF-beta/Smad signaling pathway we used a quantitative RNase protection assay (RPA) for measuring mRNA levels of TGF-beta pathway components in 20 HOSE and 20 CSOC cultures. Basal mRNA levels of TGF-beta receptors I and II, downstream signaling components Smad2, 3, 4, 6, 7, and the transcriptional corepressors Ski and SnoN did not show a statistically significant difference between HOSE and CSOC, and cannot explain their differential susceptibility to TGF-beta-induced cell cycle arrest. To assess functional differences of the TGF-beta pathway in TGF-beta-sensitive HOSE and TGF-beta-resistant CSOC, we measured Smad2/4 and 3/4 complex induction after TGF-beta treatment. HOSE and CSOC showed equivalent Smad2/4 and 3/4 complex induction after TGF-beta exposure for 0, 0.5, 2, and 4 h. It has been proposed that SnoN and Ski are corepressors of the TGF-beta/Smad pathway and undergo TGF-beta-induced degradation followed by reinduction of SnoN mRNA. However, our data show equivalent SnoN degradation in HOSE and CSOC, and equivalent SnoN mRNA induction after TGF-beta treatment. Surprising, TGF-beta-induced Ski degradation was not observed in HOSE or CSOC, suggesting that Ski may not function as a TGF-beta/Smad corepressor in ovarian epithelial cells. These data implied that the TGF-beta/Smad pathway remains functional in CSOC, although CSOC cells are resistant to antimitogenic TGF-beta effects. CSOC resistance to TGF-beta coincided with the loss of c-myc down-regulation. These data suggest that TGF-beta/Smad signaling is blocked downstream of Smad complex formation or that an alternate signaling pathway other than TGF-beta/Smad may transmit TGF-beta-induced cell cycle arrest in the ovarian epithelium.

Gene expression in epithelial ovarian carcinoma.

We analysed the mRNA levels corresponding to 12,600 transcripts in primary cultures of ovarian epithelial cells derived from nine normal ovaries and 21 epithelial ovarian carcinoma. The class distinction and hierarchical clustering of expression data revealed a clear distinction in gene expression between normal and carcinoma-derived ovarian epithelial cells. Comparison of expression levels revealed 111 genes with mean expression values of >2.5-fold higher in carcinoma cells. Similarly, 62 genes were expressed at >2.5-fold higher levels in normal ovarian epithelial cells. For a few selected genes, we demonstrate that the pattern of differential expression observed in cultured epithelial cells is present in the normal ovaries and epithelial ovarian carcinoma. Use of cultured epithelial cells represents a novel strategy to study gene expression in a cell-type specific manner.

Cancer incidence in a population of Jewish women at risk of ovarian cancer.

PURPOSE: To evaluate the incidence and clinical characteristics of ovarian and other cancers in a cohort of women at risk of developing ovarian cancer. PATIENTS AND METHODS: The Gilda Radner Ovarian Cancer Detection Program in Los Angeles, CA, was established in 1991 to study the efficacy of screening in the early detection of ovarian cancer. We present findings from a historical cohort of 290 Jewish women who were offered BRCA testing for three common founder mutations (BRCA1 185delAG and 5382insC and BRCA2 6174delT). RESULTS: In 10 years, 17 cancers were observed (1,111 per 100,000 per year), including six breast and eight ovarian or related cancers. A high proportion of cancers of peritoneal origin was observed. The majority (86%) of women with incident breast or ovarian/peritoneal cancer carried a mutation in the BRCA1 gene. The overall cancer incidence among carriers of mutations in the BRCA1 gene was estimated to be 5,450 per 100,000 per year, corresponding to a cumulative incidence of 47.5% at 10 years. In contrast, the cumulative incidence of cancer among noncarriers was 2.5% (P < 10(-8)). After adjustment for sampling, the risks to BRCA1 mutation carriers at 10 years were estimated to be 21% for ovarian/peritoneal/tubal cancer, 16% for breast cancer, and 36% for all cancers. CONCLUSION: The excess risk of breast and ovarian cancer in Jewish women with a family history of ovarian cancer is largely attributable to mutations in BRCA1. Intensive surveillance by use of CA-125 and ultrasound does not seem to be an effective means of diagnosing early-stage ovarian cancer in this high-risk cohort.

Short androgen receptor allele length is a poor prognostic factor in epithelial ovarian carcinoma.

PURPOSE: Epidemiological evidence implicates a heightened androgenic state in women with epithelial ovarian cancer. Androgen activity may be modulated by altered expression or activity of the androgen receptor (AR) or AR polymorphisms. Exon 1 of the AR gene contains a polymorphic (CAG)(n) sequence whose length is inversely correlated with transcriptional activity. EXPERIMENTAL DESIGN: Differential expression of AR mRNA and protein was examined in 46 primary cultures of normal human ovarian surface epithelium (HOSE) and malignant Cedars-Sinai ovarian cancer (CSOC) ovarian epithelial cells. AR allele length was characterized by genotyping in 77 ovarian cancer specimens. RESULTS: AR mRNA expression was higher in CSOC primary cultures (1.58 +/- 0.17) when compared with HOSE (1 +/- 0.09, P = 0.005), but protein expression was not statistically different. CAG repeat lengths were shorter in CSOC (20.6 +/- 1.2) than in HOSE (23.4 +/- 0.9, P = 0.04). Patients with an AR allele containing < or =19 CAG repeats had a shorter time to recurrence (5.5 versus 19.4 months, P < 0.0001) and overall survival (9 versus 32.6 months, P = 0.0007). There was no correlation between AR allelotype and age of diagnosis, stage, or grade; however, a short CAG length < or =19 repeats was associated with decreased surgical cytoreducibility (44.4 versus 10.3%, P = 0.035). Multivariate analyses confirmed a short AR allele as an independent prognostic factor (P = 0.02). CONCLUSIONS: These data support epidemiological evidence linking heightened androgenicity to the pathogenesis and tumor biology of epithelial ovarian cancer.

BRCA1 promoter methylation predicts adverse ovarian cancer prognosis.

OBJECTIVE: To compare the clinical outcome of ovarian cancer patients whose tumors contain BRCA1 genes silenced by promoter hypermethylation to patients with germline BRCA1 mutations and to patients with wild-type BRCA genes. METHODS: Ovarian cancers from a hospital-based tumor bank were characterized as having a BRCA1 mutation; or a methylated BRCA1, BRCA1 pseudogene or MLH1 promotor; or a wild-type BRCA gene. Survival of patients with methylated BRCA1 promoters (N = 11) was compared to that of patients with wild-type BRCA genes (N = 30) and BRCA1 mutations (N = 22). A methylator phenotype was defined to include tumors with hypermethylation of BRCA1, hMLH1 and/or dBRCA1 pseudogene promoters (N = 23). RESULTS: All cohorts had comparable clinical factors except for age at diagnosis. Median age of methylated BRCA1 and wild-type BRCA patients was older than BRCA1 mutation carriers (60 and 63 versus 48 years; P = 0.04). The median disease-free interval was significantly shorter for patients with a methylated BRCA1 promoter (9.8 months) than for BRCA1 mutation carriers (39.5 months; P = 0.04). Median overall survival was also significantly shorter for patients with a methylated BRCA1 promoter (35.6 months) than BRCA1 mutation carriers (78.6 months; P = 0.02). The combined methylator phenotype cohort had significantly shorter survival (36.1 months) compared to wild-type BRCA patients (63.3 months; P = 0.02). CONCLUSION: These data suggest that methylation of the BRCA1 promoter is associated with poor patient outcome. BRCA1 may be part of a global panel of methylated genes associated with aggressive disease.

Estrogen and progesterone receptor subtype expression in normal and malignant ovarian epithelial cell cultures.

OBJECTIVE: Epidemiologic data suggest that the malignant transformation of ovarian epithelium may be linked to altered steroid hormone homeostasis. STUDY DESIGN: Estrogen receptor-alpha, estrogen receptor-beta, progesterone receptor A, and progesterone receptor B messenger RNA and protein expression were evaluated by reverse transcriptase-polymerase chain reaction and Western blot analysis in primary cell cultures of human ovarian surface epithelium (n = 23 cultures) and Cedars-Sinai ovarian cancer (n = 23 cultures). RESULTS: The ratio of estrogen receptor-alpha/estrogen receptor-beta messenger RNA expression was 10 times higher in primary ovarian cancer cultures (9.94 +/- 3.90) than in normal ovarian surface epithelium cultures (1.00 +/- 0.16, P =.04). Estrogen receptor-alpha/estrogen receptor-beta protein ratio in primary ovarian cancer cultures (2.13 +/- 0.43) was twice that of normal human ovarian surface epithelium cultures (1.00 +/- 0.13, P =.05). Individual estrogen receptor-alpha and estrogen receptor-beta messenger RNA and protein expression were not significantly different. Progesterone receptor B protein levels in primary ovarian cancer cultures (2.08 +/- 0.42) were twice that of normal surface ovarian epithelium cultures (1.00 +/- 0.10, P =.04), although differences in progesterone receptor B messenger RNA and progesterone receptor A protein expression were not observed. CONCLUSION: Malignant ovarian epithelial cells demonstrated multiple alterations in the expression of sex steroid hormone receptors.

Improved survival in women with BRCA-associated ovarian carcinoma.

BACKGROUND: The objective of this study was to determine the clinical characteristics, treatment response, and frequency of p53 overexpression in Ashkenazi Jewish women with hereditary ovarian carcinoma. METHODS: Seventy-one Jewish women with epithelial ovarian carcinoma (EOC) were tested for the three BRCA founder mutations using single-strand conformation polymorphism analysis, heteroduplex analysis, and protein truncation testing. Clinical and histopathologic data were reviewed retrospectively. In vitro chemoresistance was analyzed in 32 patients. Mutations of p53 were studied using immunohistochemical detection of p53 overexpression. RESULTS: Thirty-four of 71 Jewish patients with EOC (48%) had germline BRCA mutations (BRCA heterozygotes), including 22 BRCA1 mutations and 12 BRCA2 mutations. BRCA heterozygotes were younger compared with Jewish patients who had EOC without mutations (sporadic carcinoma; 50 years vs. 59 years, respectively; P = 0.01). BRCA1 heterozygotes were younger compared with BRCA2 heterozygotes (48 years vs. 57 years, respectively; P = 0.01). Histopathologic tumor features were similar; however, tumors with low malignant potential were seen only in women with sporadic carcinoma. Both groups had equivalent rates of surgical cytoreduction and similar median follow-up (72 months). BRCA heterozygotes had higher response rates to primary therapy compared with patients who had sporadic disease (P = 0.01). In vitro chemoresistance predicted tumor response to platinum chemotherapy correctly in BRCA heterozygotes (P = 0.0096). BRCA heterozygotes with advance-stage disease had improved survival compared with patients who had advanced stage sporadic carcinoma (91 months vs. 54 months, respectively; P = 0.046) and had a longer disease free interval (49 months vs. 19 months, respectively; P = 0.16). p53 overexpression was common in BRCA heterozygotes (80%). CONCLUSIONS: BRCA1 heterozygotes developed EOC at a younger age compared with BRCA2 heterozygotes and women who had sporadic ovarian carcinoma. BRCA heterozygotes had a better response to platinum chemotherapy compared with women who had sporadic disease, which may have contributed to their improved prognosis.

A human ovarian carcinoma murine xenograft model useful for preclinical trials.

OBJECTIVE: To establish a murine xenograft model of human ovarian carcinoma. METHODS: A slurry of fresh human tumor from patients with intraperitoneal malignancies was heterotransplanted intraperitoneally into nude (nu/nu) and severely combined immunodeficient mice (CB-17, SCID). Xenograft growth was assessed by serial examination and necropsy. The xenografts were passaged to new animals when tumors were palpably greater than 1 cm(3). Histopathologic analysis of the xenografts was performed at each passage as well as immunohistochemical staining for p53 mutations. Persistent expression of human genes by the xenografts at higher passages was assessed by RT-PCR amplification of the human beta-globin gene. This xenograft model was used in the preclinical evaluation of an adenoviral vector containing a beta-galactosidase reporter gene and a wild-type p53 gene. RESULTS: Tumor growth was not established in any of the nude mice heterotransplanted with tissue from six different ovarian cancer patients. Eleven of 13 specimens established xenograft growth when injected in SCID mice. Nine xenografts have been subsequently passaged between 6 and 24 animal generations to date. All xenografts retained histopathologic similarities to their original human tumors and the p53 expression patterns remained stable through higher passages. Within 24 h after intraperitoneal administration of an adenoviral vector, transduction of the reporter gene was evident in the xenografts. In addition, administration of an adenoviral vector containing a wild-type p53 gene significantly decreased the tumor burden compared to controls (P < 0.04). CONCLUSIONS: This murine xenograft model of human ovarian carcinoma appears to be reliable and reproducible and has utility for the study of novel therapeutics.