Kelemen, Linda E.

ScopeWe reevaluated previously reported associations between variants in pathways of one‐carbon (1‐C) (folate) transfer genes and ovarian carcinoma (OC) risk, and in related pathways of purine and pyrimidine metabolism, and assessed interactions with folate intake.Methods and resultsOdds ratios (OR) for 446 genetic variants were estimated among 13 410 OC cases and 22 635 controls, and among 2281 cases and 3444 controls with folate information. Following multiple testing correction, the most significant main effect associations were for dihydropyrimidine dehydrogenase (DPYD) variants rs11587873 (OR = 0.92; p = 6 × 10−5) and rs828054 (OR = 1.06; p = 1 × 10−4). Thirteen variants in the pyrimidine metabolism genes, DPYD, DPYS, PPAT, and TYMS, also interacted significantly with folate in a multivariant analysis (corrected p = 9.9 × 10−6) but collectively explained only 0.2% of OC risk. Although no other associations were significant after multiple testing correction, variants in SHMT1 in 1‐C transfer, previously reported with OC, suggested lower risk at higher folate (pinteraction = 0.03–0.006).ConclusionVariation in pyrimidine metabolism genes, particularly DPYD, which was previously reported to be associated with OC, may influence risk; however, stratification by folate intake is unlikely to modify disease risk appreciably in these women. SHMT1 SNP‐by‐folate interactions are plausible but require further validation. Polymorphisms in selected genes in purine metabolism were not associated with OC.

Abstract Background: Ovarian carcinoma is composed of five major histologic types, which associate with outcome and predict therapeutic response. Our aim was to evaluate histologic type assessments across the centers participating in the Ovarian Tumor Tissue Analysis (OTTA) consortium using an immunohistochemical (IHC) prediction model. Methods: Tissue microarrays (TMA) and clinical data were available for 524 pathologically confirmed ovarian carcinomas. Centralized IHC was conducted for ARID1A, CDKN2A, DKK1, HNF1B, MDM2, PGR, TP53, TFF3, VIM, and WT1, and three histologic type assessments were compared: the original pathologic type, an IHC-based calculated type (termed TB_COSPv2), and a WT1-assisted TMA core review. Results: The concordance between TB_COSPv2 type and original type was 73%. Applying WT1-assisted core review, the remaining 27% discordant cases subdivided into unclassifiable (6%), TB_COSPv2 error (6%), and original type error (15%). The largest discordant subgroup was classified as endometrioid carcinoma by original type and as high-grade serous carcinoma (HGSC) by TB_COSPv2. When TB_COSPv2 classification was used, the difference in overall survival of endometrioid carcinoma compared with HGSC became significant [RR 0.60; 95% confidence interval (CI), 0.37–0.93; P = 0.021], consistent with previous reports. In addition, 71 cases with unclear original type could be histologically classified by TB_COSPv2. Conclusions: Research cohorts, particularly those across different centers within consortia, show significant variability in original histologic type diagnosis. Our IHC-based reclassification produced more homogeneous types with respect to outcome than original type. Impact: Biomarker-based classification of ovarian carcinomas is feasible, improves comparability of results across research studies, and can reclassify cases which lack reliable original pathology. Cancer Epidemiol Biomarkers Prev; 22(10); 1677–86. ©2013 AACR.

Abstract Background: Invasive ovarian cancer is a significant cause of gynecologic cancer mortality. Methods: We examined whether this mortality was associated with inherited variation in approximately 170 candidate genes/regions [993 single-nucleotide polymorphisms (SNPs)] in a multistage analysis based initially on 312 Mayo Clinic cases (172 deaths). Additional analyses used The Cancer Genome Atlas (TCGA; 127 cases, 62 deaths). For the most compelling gene, we immunostained Mayo Clinic tissue microarrays (TMA, 326 cases) and conducted consortium-based SNP replication analysis (2,560 cases, 1,046 deaths). Results: The strongest initial mortality association was in HGF (hepatocyte growth factor) at rs1800793 (HR = 1.7, 95% CI = 1.3–2.2, P = 2.0 × 10−5) and with overall variation in HGF (gene-level test, P = 3.7 × 10−4). Analysis of TCGA data revealed consistent associations [e.g., rs5745709 (r2 = 0.96 with rs1800793): TCGA HR = 2.4, CI = 1.4–4.1, P = 2.2 × 10−3; Mayo Clinic + TCGA HR = 1.6, CI = 1.3–1.9, P = 7.0 × 10−5] and suggested genotype correlation with reduced HGF mRNA levels (P = 0.01). In Mayo Clinic TMAs, protein levels of HGF, its receptor MET (C-MET), and phospho-MET were not associated with genotype and did not serve as an intermediate phenotype; however, phospho-MET was associated with reduced mortality (P = 0.01) likely due to higher expression in early-stage disease. In eight additional ovarian cancer case series, HGF rs5745709 was not associated with mortality (HR = 1.0, CI = 0.9–1.1, P = 0.87). Conclusions: We conclude that although HGF signaling is critical to migration, invasion, and apoptosis, it is unlikely that HGF genetic variation plays a major role in ovarian cancer mortality. Furthermore, any minor role is not related to genetically-determined expression. Impact: Our study shows the utility of multiple data types and multiple data sets in observational studies. Cancer Epidemiol Biomarkers Prev; 20(8); 1638–48. ©2011 AACR.

Abstract Introduction: Alterations in components of the stromal tissue that underlie the epithelium can initiate, promote or inhibit epithelial tumorigenesis. Decorin (DCN) and lumican (LUM) are stromal genes with reduced expression in serous ovarian cancer. We hypothesized that single nucleotide polymorphisms (SNPs) in these genes may influence susceptibility to serous ovarian cancer. Methods: To test this hypothesis, we compared genotype frequencies of 11 tagSNPs in DCN and LUM between 397 Caucasian cases with primary serous ovarian cancer and 920 Caucasian controls in two studies conducted at Mayo Clinic and Duke University (discovery set). Associations were evaluated further in a third Caucasian population of 436 cases and 1,098 controls in Australia (replication set). TagSNPs were selected from unrelated Caucasian samples within HapMap based on the following criteria: (i) minor allele frequency ≥ 0.05, (ii) pairwise linkage disequilibrium (LD) of r2 ≥ 0.8, and (iii) location within, and 5kb upstream and downstream of, each gene region. Genotyping was performed using the Illumina GoldenGate™ assay and associations between SNPs and serous cancer were evaluated using unconditional logistic regression to estimate per allele odds ratios (OR) and 95% confidence intervals (95%CI). Untyped genotypes for DCN rs13312816 in the discovery set, and for DCN rs3138165 and LUM rs17018765 in the replication set, were imputed using the MACH software. Results: In the discovery set, we observed inverse associations between DCN rs3138165 (OR=0.7; 95%CI=0.5-1.0), DCN rs13312816 (OR=0.7; 95%CI=0.5-1.0), DCN rs516115 (OR=0.8; 95%CI=0.7-1.0) and LUM rs17018765 (OR=0.7; 95%CI=0.5-1.0) and risk of serous ovarian cancers (all Ptrend=0.06). Associations were confirmed in the replication set for DCN rs3138165 (OR=0.6; 95%CI=0.4-0.9; Ptrend=0.009), DCN rs13312816 (OR=0.7; 95%CI=0.5-0.9; Ptrend=0.01) and LUM rs17018765 (OR=0.6; 95%CI=0.4-0.9; Ptrend=0.008), but not for DCN rs516115 (Ptrend=0.20). To increase statistical power, data from the discovery and replication sets were combined (833 cases and 2,013 controls) following confirmation of no statistical heterogeneity in the ORs between study sites. In the combined sample, associations were strengthened for DCN rs3138165 (OR=0.7; 95%CI=0.5-0.9; Ptrend=0.002), DCN rs13312816 (OR=0.7; 95%CI=0.5-0.9; Ptrend=0.002), DCN rs516115 (OR=0.9; 95%CI=0.8-1.0; Ptrend=0.03) and LUM rs17018765 (OR=0.6; 95%CI=0.5-0.8; Ptrend=0.001). All four SNPs were in the same LD block on chromosome 12, suggesting that serous ovarian cancer susceptibility may be driven by a single, correlated variant in the region. Conclusion: The findings of this study suggest that genetic variation in DCN and LUM may influence susceptibility to serous ovarian cancer and contribute to the mounting evidence that supports the role of the stroma in epithelial ovarian cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 2854.

Abstract Background: Gene amplification leading to overexpression is a common event in breast tumors that is linked to tumor development and progression. The 17q23 region is amplified in >40% of breast tumors and contains several candidate oncogenes. Because common genetic variation in several oncogenes has been associated with cancer risk, we assessed the relevance of common variants in the 17q23 candidate oncogenes to breast cancer. Methods: We investigated 60 polymorphisms in the TUBD1, SEPT4, PRKCA, TBX2, TBX4, TEX14, TLK2, YPEL2, and PPM1E genes from this amplicon for association with breast cancer risk among 798 Caucasian breast cancer cases and 843 unaffected Caucasian controls from the Mayo Clinic. Results: Eight polymorphisms in PRKCA, TBX4, TLK2, and YPEL2 displayed significant dose-response associations with breast cancer risk (Ptrend < 0.05). Of these, PRKCA rs7342847 and TLK2 rs2245092 and rs733025 were also associated with hormone receptor–positive breast cancer: PRKCA rs7342847 (odds ratio, 0.7; 95% confidence interval, 0.6-0.9; Ptrend = 0.002) and TLK2 rs733025 and rs2245092 (both: odds ratio, 0.8; 95% confidence interval, 0.7-1.0; Ptrend = 0.03). Interactions between SEPT4 rs758377 and TEX14 rs302864 (Pinteraction = 0.0003) and between TLK2 rs733025 and YPEL2 rs16943468 (Pinteraction = 0.05) for risk of breast cancer were also observed. Conclusion: These findings suggest that single polymorphisms and combinations of polymorphisms within candidate oncogenes from the 17q23 amplicon may influence risk of breast cancer overall and possibly specific molecular subtypes of breast tumors. The findings are discussed within the context of the results from two independent data sets. (Cancer Epidemiol Biomarkers Prev 2009;18(6):1864–8)

Abstract Whole genome amplification (WGA) offers a means to enrich DNA quantities for epidemiologic studies. We used an ovarian cancer study of 1,536 single nucleotide polymorphisms (SNPs) and 2,368 samples to assess performance of multiple displacement amplification (MDA) WGA using an Illumina GoldenGate BeadArray. Initial screening revealed successful genotyping for 93.4% of WGA samples and 99.3% of genomic samples, and 93.2% of SNPs for WGA samples and 96.3% of SNPs for genomic samples. SNP failure was predicted by Illumina-provided designability rank, %GC (P ≤ 0.002), and for WGA only, distance to telomere and Illumina-provided SNP score (P ≤ 0.002). Distance to telomere and %GC were highly correlated; adjustment for %GC removed the association between distance to telomere and SNP failure. Although universally high, per-SNP call rates were related to designability rank, SNP score, %GC, minor allele frequency, distance to telomere (P ≤ 0.01), and, for WGA only, Illumina-provided validation class (P < 0.001). We found excellent concordance generally (>99.0%) among 124 WGA:genomic replicates, 15 WGA replicates, 88 replicate aliquots of the same WGA preparation, and 25 genomic replicates. Where there was discordance, it was across WGA:genomic replicates but limited to only a few samples among other replicates suggesting the introduction of error. Designability rank and SNP score correlated with WGA:genomic concordance (P < 0.001). In summary, use of MDA WGA DNA is feasible; however, caution is warranted regarding SNP selection and analysis. We recommend that biological SNP characteristics, notably distance to telomere and GC content (<50% GC recommended), as well as Illumina-provided metrics be considered in the creation of GoldenGate assays using MDA WGA DNA. (Cancer Epidemiol Biomarkers Prev 2008;17(7):1781–9)

Abstract Dysfunction in enzymes involved in one-carbon (1-C) metabolism can lead to increased chromosomal strand breaking and abnormal methylation patterns, which are both associated with cancer risk. Availability of 1-C units may modify risk. We investigated the association of single-nucleotide polymorphisms (SNP) in 21 genes in the 1-C transfer pathway among 829 Caucasian cases with primary epithelial ovarian cancer and 941 frequency-matched unaffected controls enrolled at Mayo Clinic (Rochester, MN) and Duke University (Durham, NC) and examined risk modification by multivitamin supplement use. Multivariable-adjusted SNP-specific logistic regression and haplotype analyses were done for 180 SNPs and false positive report probabilities (FPRP) were calculated. Each copy of the minor allele in SHMT1 intron 5 A>G (rs9909104) was associated with epithelial ovarian cancer [odds ratio (OR), 1.2; 95% confidence interval (95% CI), 1.0–1.4; P trend = 0.02; FPRP = 0.16] and a 5-SNP SHMT1 haplotype was associated with decreased risk (P = 0.01; FPRP = 0.09). Three SNPs in DNMT3A were associated with risk among multivitamin supplement users: 3′ untranslated region (UTR) C>G (rs13420827: OR, 0.8; 95% CI, 0.6–1.0; P interaction = 0.006; FPRP = 0.54), intron 6 G>A (rs11887120: OR, 0.8; 95% CI, 0.7–1.0; P interaction = 0.007; FPRP = 0.57), and intron 22 A>T (rs11695471: OR, 1.2; 95% CI, 1.0–1.5; P interaction = 0.01; FPRP = 0.66). These data extend previous findings from other cancers of a role for SHMT1 in ovarian cancer, and provide evidence that SNPs in methylation and DNA synthesis reactions are associated with risk of ovarian cancer. Interventions with modifiable factors such as multivitamin intake may reduce risk. [Cancer Res 2008;68(7):2498–506]

Abstract Studies suggest that underglycosylation of the cell membrane mucin MUC1 may be associated with epithelial ovarian cancer. We identified 26 genes involved in glycosylation and examined 93 single nucleotide polymorphisms (SNP) with a minor allele frequency of ≥0.05 in relation to incident ovarian cancer. Cases were ascertained at the Mayo Clinic, Rochester, MN (n = 396) or a 48-county region in North Carolina (Duke University; n = 534). Ovarian cancer–free controls (n = 1,037) were frequency matched to the cases on age, race, and residence. Subjects were interviewed to obtain data on risk factors and a sample of blood for DNA and genotyped using the Illumina GoldenGate assay. We excluded subjects and individual SNPs with genotype call rates of <90%. Data were analyzed using logistic regression, with adjustment for age and residence. We fitted dominant, log additive, and recessive genetic models. Among Caucasians, nine SNPs in eight genes were associated with risk at P < 0.05 under at least one genetic model before adjusting for multiple testing. A SNP in GALNT1 (rs17647532) was the only one that remained statistically significant after Bonferroni adjustment for multiple testing but was not statistically significant in Hardy-Weinberg equilibrium among controls. Haplotype analyses revealed a global association of GALNT1 with risk (P = 0.038, under a recessive genetic model), which largely reflected a decreased risk of one haplotype (0.10 frequency; odds ratio, 0.07; P = 0.01) compared with the most common haplotype (0.39 frequency). These results suggest that genetic polymorphisms in the glycoslyation process may be novel risk factors for ovarian cancer. (Cancer Epidemiol Biomarkers Prev 2008;17(2):397–404)

Abstract Increased mammographic density (MD), the proportion of dense tissue visible on a mammogram, is a strong risk factor for breast cancer, common in the population and clusters in families. We conducted the first genome-wide linkage scan to identify genes influencing MD. DNA was obtained from 889 relatives (756 women, 133 men) from 89 families. Percent MD was estimated on 618 (82%) female family members using a validated computer-assisted thresholding method. The genome-wide scan included 403 microsatellite DNA markers with an average spacing of 9 cM. Fine mapping of a region of chromosome 5p (5p13.1-5p15.1) was done using 21 additional closely spaced DNA markers. Linkage analyses were conducted to quantify the evidence for a gene responsible for MD across the genome. The maximum log odds for linkage (LOD) score from the genome-wide scan was on chromosome 5p (LOD = 2.9, supporting linkage by a factor of 102.9 or 794 to 1) with a 1-LOD interval spanning 28.6 cM. Two suggestive regions for linkage were also identified on chromosome 12 (LOD = 2.6, 1-LOD interval of 14.8 cM; and LOD = 2.5, 1-LOD interval of 17.2 cM). Finer mapping of the region surrounding the maximum LOD on chromosome 5p resulted in stronger and statistically significant evidence for linkage (LOD = 4.2) and a narrowed 1-LOD interval (13.4 cM). The putative locus on chromosome 5p is likely to account for up to 22% of variation in MD. Hence, 1 or more of the 45 candidate genes in this region could explain a large proportion of MD and, potentially, breast cancer. [Cancer Res 2007;67(17):8412–8]