Thibodeau, Stephen N.

Abstract Several platforms for genome sequencing and methods for exon enrichment are now available and continue to evolve quickly. In this study, we compared the performance of whole exome sequencing using two instruments, Illumina's GAIIx and HiSeq 2000, comparing single versus duplicate runs. Eight runs of eight samples each were performed with genomic DNA derived from whole blood and sample libraries prepared with the Agilent SureSelect capture array. One run was conducted in duplicate on the GAIIx, and, for some analyses, the combined data were used. We used TREAT (Targeted RE-sequencing and Annotation Tool), developed in-house for data analysis, including sequence alignment (MAQ, BWA), local re-alignment (GATK), variant calling (utilizing MAQ), annotation (SIFT and Seattle Seq), and visualization. Overall, duplicate runs on the GAIIx platform (compared to a single run) increased the number of total reads by approximately two-fold: from ∼66 M to ∼130 M. For both single and duplicate runs, ∼90% of reads mapped to the reference sequence, and ∼55% of reads mapped on-target. Of importance, the percent coverage of the target region increased substantially with duplicate runs; ∼72% of the target region was covered at 40-fold or more when run in duplicate compared to only ∼52% when run once. At 30x, 20x, and 10x, there was a 1.3- (62% v. 78%), 1.2- (72% v. 84%), and 1.1-fold (84% v. 91%) increase respectively, in the percent coverage of the target region for samples run twice. A similar number of filtered, on-target SNPs per sample (∼24 K) was found for both single and duplicate run analyses. However, a 1.3-fold increase in the number of on-target indels was seen in the duplicate run (∼1,200) compared to the single run (∼900). When the combined data for samples run twice on a GAIIx were compared to the samples run using a HiSeq, the total number of reads were similar (∼130 M), although several of the samples had substantially more reads on the HiSeq platform. Samples run on a HiSeq had an increase in the percent coverage of the target region at 40x (72% v. 81%), 30x (78% v. 86%), 20x (84% v. 91%) and 10x (91% v. 95%) compared to the two-run approach on a GAIIx. A similar number of filtered on-target SNPs (∼24 K/sample) and indels (∼1200) was found with the HiSeq compared to two runs on a GAIIx. In summary, our results demonstrate increases in the total number of reads and in overall coverage with the HiSeq 2000 instrument. The total number of SNPs and indel that mapped on target for the version 1 Agilent SureSelect capture array is ∼25 K and 1,200 per sample, respectively. As capture is of variable efficiency for individual runs, higher average coverage of the target region is necessary for sufficient coverage of poorly captured regions. Additional analyses comparing Agilent SureSelect versions 1 and 2 arrays are underway. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 2742. doi:10.1158/1538-7445.AM2011-2742

Abstract Introduction: Genome-wide association studies of colorectal cancer have identified 18 risk variants that explain ∼25% of the hereditable risk of colorectal cancer. These variants are located in intronic, inter-genic or gene-desert regions. To uncover whether some of these variants may have a regulatory function on the expression of near-by genes, we conducted expression quantitative trait loci (eQTL) analyses of these 18 regions. Methods: DNA and RNA were extracted from 40 fresh-frozen paired colon tumor (microsatellite stable and CpG island methylator phenotype-negative) and adjacent normal tissue samples collected by the Colon Cancer Family Registry. Whole genome gene expression assessment was performed in the tumors and adjacent normal tissue with the Affymetrix GeneChip Human Exon 1.0 ST array. Genome-wide SNP data was generated with DNA extracted from the same normal tissue samples using the Affymetrix Genome-Wide Human SNP 6.0 array. Partek software was used to compare the expression levels by genotype for each gene spanning a 2 Mb region up- and down- stream of each of the 18 established risk variants for colorectal cancer. Results: Three of the 18 risk variants (rs10795668, rs4444235, rs9929218) demonstrated significant cis-eQTLs (FDR q-values<0.05). Specifically, differential gene expression was observed in genes neighboring rs10795668 at 10p14 (ATP5C1, q=0.024), rs4444235 at 14q22.2 (DLGAP5, q=0.041), and rs9929218 at 16q22.1 (NOL3, q=0.017; DDX28, q=0.046). In particular, DLGAP5 (discs, large (Drosophila) homolog-associated protein 5) at 14q22.2 and NOL3 (nucleolar protein 3) at 16q22 are interesting candidate genes for colorectal cancer. DLGAP5 encodes for a microtubule-associated protein and functions as a mitotic regulator that has been reported to be overexpressed in colon tumors. NOL3 (nucleolar protein 3) encodes for an anti-apoptotic protein that down-regulates the expression of caspase 2 and 8, and p53. NOL3 also appears to be overexpressed in colorectal tumors. Conclusion: In summary, these findings suggest putative functional activities for three established colorectal risk variants and highlight potentially interesting candidate genes for this malignancy. If confirmed, these results may help explain the total genetic risk of the disease, as well as further our understanding of the underlying biology of colorectal cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 4732. doi:10.1158/1538-7445.AM2011-4732

Abstract Short constitutional telomere length has been associated with aging and cancer, but how telomere length is affected by other epidemiologic or environmental variables in health has not been thoroughly studied. Understanding these associations with telomere length in health is important for studies to evaluate the relationship of telomere length and diseases such as cancer. We measured peripheral blood leukocyte DNA telomere length with quantitative PCR in 1977 non-related healthy controls collected through the Colon Cancer Family Registry (an NCI cooperative group) and through the Mayo Biobank and Biobank for Gastrointestinal Health Research, two IRB approved prospective biospecimen collections based at Mayo Clinic, Rochester, MN. The controls were a median age of 50.6 years old, and 49.6% of the controls were male. Ninety-six percent of the controls were white; 1.8% were Asian, and <1% were either black, Pacific Islander, or Native American. Less than 1% of controls were Hispanic. On univariate analysis, as age increased, telomere length decreased (Pearson correlation coefficient = −0.07, 95% CI: −0.11, −0.03). Telomeres were longer in people who consumed more fruits (p<0.01) and vegetables (p<0.01); took aspirin on a regular basis (p<0.01); or had a family history of cancers of any type (p=0.05). Women had longer telomeres than men (p=0.03). Telomeres were shorter in people who consumed alcohol (p<0.01), took folate on a regular basis (p=0.01), or who consumed red meat (p<0.01). Statistical significance was not found in the following variables: tobacco use, diagnosis of diabetes, intake of calcium, use of hormone replacement therapy in women, or body mass index. On multivariate analysis, age, vegetable consumption, family history of colorectal cancer or any type of cancer, and folate intake were identified as being significantly associated with telomere length. Telomere length may be associated with multiple environmental and demographic variables in healthy individuals, and adjusting for these may be important for studies that test for an association between telomere length and diseases such as cancer. Citation Information: Cancer Prev Res 2011;4(10 Suppl):A82.

Abstract Background: Colorectal cancer (CRC) is the third most common incident cancer among women in the United States. Alcohol use has been proposed as a CRC risk factor, but existing observational data remain inconsistent. To date, relatively few studies have examined associations between alcohol use and CRC risk based on molecularly-defined subtypes. Aim: To evaluate associations between alcohol use and incident CRC overall, as well as by molecularly-defined subtypes including microsatellite instability (MSI), CpG island methylator phenotype (CIMP) and BRAF mutation status in the prospective, population-based, Iowa Women's Health Study (IWHS). Methods: The IWHS recruited 41,836 randomly selected Iowa women, ages 55–69 years at enrollment, in 1986. Women with prevalent cancers or no follow-up were excluded, leaving 38,001 eligible for this study. Alcohol use was obtained from a self-administered food frequency questionnaire at study baseline. Incident CRCs were identified through annual linkage with the State Health Registry of Iowa. Archived, paraffin-embedded tissue specimens were recently requested for incident CRC cases diagnosed through December 31, 2002. In the present study, useable tissue specimens were obtained and molecularly characterized for 563/1,255 cases (45%), whose baseline age, body mass index, physical activity level, total energy intake and alcohol use were similar to non-retrieved cases (p > 0.05 for each comparison). Alcohol use was categorized as never (n=21,464), < 3.4 g/day (n=8,313) and > 3.4 g/day (n=8,224), based on the median split among users. Molecularly-defined CRC subtypes were categorized as MSI-high (MSIH; n=148), MSI-low or microsatellite stable (MSI-L/MSS; n=400); CIMP-positive (n=167) or CIMP-negative (n=368); and BRAF-mutation (n=154) or BRAF-wildtype (n=391). Relative risks (RRs) and 95% confidence intervals (95% CIs) were estimated for alcohol use compared to never drinkers, using multivariable Cox regression models adjusted for body mass index, waist-to-hip ratio, smoking status, physical activity level, exogenous estrogen use, and daily intakes of total energy, total fat, red meat, sucrose, calcium, folate, methionine and vitamin E. Results: There was no statistically significant association between alcohol use and overall CRC risk (RR=1.00; 95% CI = 0.86–1.15 for < 3.4 g/day versus none; RR=1.06; 95% CI=0.91–1.24 for >3.4 g/day versus none). Further, there were no associations with alcohol use for specific, molecularly-defined CRC subtypes (RRs shown for > 3.4 g/day versus none): MSI-H (RR=1.07; 95% CI=0.70–1.64); MSI-L or MSS (RR=1.12; 95% CI=0.85–1.47); CIMP-positive (RR=0.97; 95% CI=0.64–1.47); CIMP-negative (RR=1.13; 95% CI=0.85–1.50); BRAF-mutation (RR=0.94; 95% CI=0.61–1.45); and BRAF-wildtype (RR=1.20; 95% CI=0.91–1.57). Conclusion: In this cohort of older women, alcohol use did not appear to be a risk factor for incident CRC, overall or by the defined molecular subtypes. Citation Information: Cancer Prev Res 2010;3(1 Suppl):A115.

Abstract Background: Postmenopausal hormone (PMH) therapy represents a controversial colorectal cancer (CRC) preventive intervention. Because colorectal carcinogenesis is a heterogeneous process, we evaluated associations between PMH therapy and incident CRC in relation to KRAS mutation status in a population-based cohort of older women [Iowa Women's Health Study (IWHS)]. Methods: The IWHS enrolled 41,836 randomly selected women, ages 55 to 69 years, in 1986. PMH therapy and other exposure data were recorded at baseline. Tissue samples from prospectively identified CRC cases (n = 507) were analyzed for somatic KRAS mutations (exon 2, codons 12 and 13). Multivariable Cox regression models were fit to estimate relative risks (RR) and 95% confidence intervals (CI). Results: PMH therapy (ever vs. never) was inversely associated with KRAS mutation–negative (RR = 0.83; 95% CI, 0.66–1.06; P = 0.14) and KRAS mutation–positive (RR = 0.82; 95% CI, 0.58–1.16; P = 0.27) tumors, although the observed risk estimates were not statistically significant. When anatomic subsite was additionally considered, the strongest association was found for KRAS mutation–negative, distal colorectal tumors (RR = 0.64; 95% CI, 0.43–0.96; P = 0.03). Conclusions: To our knowledge, we provide the first report of KRAS-defined CRC risks associated with PMH therapy. These data suggest that PMH therapy may reduce CRC risk through mechanisms beyond KRAS mutation status but might provide greater benefits for KRAS mutation–negative than mutation-positive tumors (at least in the distal colorectum). Impact: Findings from this prospective cohort study provide novel insights about the molecular biology of PMH therapy–related CRC risk reduction. Cancer Epidemiol Biomarkers Prev; 21(4); 681–4. ©2012 AACR.

Background Lynch syndrome is caused by germline mismatch repair (MMR) gene mutations. The PREMM 1,2,6 model predicts the likelihood of a MMR gene mutation based on personal and family cancer history. Objective To compare strategies using PREMM 1,2,6 and tumour testing (microsatellite instability (MSI) and/or immunohistochemistry (IHC) staining) to identify mutation carriers. Design Data from population-based or clinic-based patients with colorectal cancers enrolled through the Colon Cancer Family Registry were analysed. Evaluation included MSI, IHC and germline mutation analysis for MLH1 , MSH2 , MSH6 and PMS2 . Personal and family cancer histories were used to calculate PREMM 1,2,6 predictions. Discriminative ability to identify carriers from non-carriers using the area under the receiver operating characteristic curve (AUC) was assessed. Predictions were based on logistic regression models for (1) cancer assessment using PREMM 1,2,6 , (2) MSI, (3) IHC for loss of any MMR protein expression, (4) MSI+IHC, (5) PREMM 1,2,6 +MSI, (6) PREMM 1,2,6 +IHC, (7) PREMM 1,2,6 +IHC+MSI. Results Among 1651 subjects, 239 (14%) had mutations (90 MLH1 , 125 MSH2 , 24 MSH6 ). PREMM 1,2,6 discriminated well with AUC 0.90 (95% CI 0.88 to 0.92). MSI alone, IHC alone, or MSI+IHC each had lower AUCs: 0.77, 0.82 and 0.82, respectively. The added value of IHC+PREMM 1,2,6 was slightly greater than PREMM 1,2,6 +MSI (AUC 0.94 vs 0.93). Adding MSI to PREMM 1,2,6 +IHC did not improve discrimination. Conclusion PREMM 1,2,6 and IHC showed excellent performance in distinguishing mutation carriers from non-carriers and performed best when combined. MSI may have a greater role in distinguishing Lynch syndrome from other familial colorectal cancer subtypes among cases with high PREMM 1,2,6 scores where genetic evaluation does not disclose a MMR mutation.

Background Postmenopausal hormone (PMH) therapy may reduce colorectal cancer (CRC) risk, but existing data are inconclusive. Objectives To evaluate associations between PMH therapy and incident CRC, overall and by molecularly defined subtypes, in the population-based Iowa Women's Health Study of older women. Methods Exposure data were collected from Iowa Women's Health Study participants (55–69 years) at baseline (1986). Archived, paraffin-embedded tissue specimens for 553 CRC cases were collected and analysed to determine microsatellite instability (MSI-L/MSS or MSI-H), CpG island methylator phenotype (CIMP-negative or CIMP-positive) and BRAF mutation ( BRAF -wildtype or BRAF -mutated) status. Multivariable Cox regression models were fit to estimate RRs and 95% CIs. Results PMH therapy (ever vs never use) was inversely associated with incident CRC overall (RR=0.82; 95% CI 0.72 to 0.93), with a significantly lower risk for MSI-L/MSS tumours (RR=0.75; 95% CI 0.60 to 0.94), and borderline significantly lower risks for CIMP-negative (RR=0.79; 95% CI 0.63 to 1.01) and BRAF -wildtype (RR=0.83; 95% CI 0.66 to 1.04) tumours. For PMH therapy >5 years, the subtype-specific risk estimates for MSI-L/MSS, CIMP-negative and BRAF -wildtype tumours were: RR=0.60, 95% CI 0.40 to 0.91; RR=0.68, 95% CI 0.45 to 1.03; and RR=0.70, 95% CI 0.47 to 1.05, respectively. PMH therapy was not significantly associated with the MSI-H, CIMP-positive or BRAF -mutated CRC subtypes. Conclusions In this prospective cohort study, PMH therapy was inversely associated with distinct molecularly defined CRC subtypes, which may be related to differential effects from oestrogen and/or progestin on heterogeneous pathways of colorectal carcinogenesis.

Abstract Increased alcohol consumption is a putative colorectal cancer (CRC) risk factor. However, existing data are less conclusive for women than men. Also, to date, relatively few studies have reported alcohol-related CRC risks based on molecularly defined tumor subtypes. We evaluated associations between alcohol intake and incident CRC, overall and by microsatellite instability [MSI high (MSI-H) or MSI low/microsatellite stable (MSI-L/MSS)], CpG island methylator phenotype (CIMP positive or CIMP negative), and BRAF mutation (mutated or wild-type) status in the prospective, population-based Iowa Women's Health Study (IWHS; n = 41,836). Subjects were 55 to 69 years at baseline (1986), and exposure data were obtained by self-report. Incident CRCs were prospectively identified and archived, paraffin-embedded tissue specimens were collected from 732 representative cases, diagnosed through December 31, 2002. Multivariate Cox regression models were fit to estimate relative risks (RR) and 95% confidence intervals (CI). Among alcohol consumers, the median intake (range) was 3.4 (0.9–292.8) g/d. Compared with nonconsumers, alcohol intake levels of 3.4 g/d or less (RR = 1.00; 95% CI, 0.86–1.15) and more than 3.4 g/d (RR = 1.06; 95% CI, 0.91–1.24) were not significantly associated with overall CRC risk. Analyses based on alcohol intake levels of 30 g/d or less and more than 30 g/d or quartile distributions yielded similar risk estimates. Null associations were also observed between each alcohol intake level and the MSI-, CIMP- or, BRAF-defined CRC subtypes (P > 0.05 for each comparison). These data do not support an adverse effect from alcohol intake on CRC risk, overall or by specific molecularly defined subtypes, among older women. Cancer Prev Res; 4(12); 2035–43. ©2011 AACR.

Purpose With three available chemotherapy drugs for advanced colorectal cancer (CRC), response rate (RR) and survival outcomes have improved with associated morbidity, accentuating the need for tools to select optimal individualized treatment. Pharmacogenetics identifies the likelihood of adverse events or response based on variants in genes involved in drug transport, metabolism, and cellular targets. Patients and Methods Germline DNA was extracted from 520 patients on the North American Gastrointestinal Intergroup N9741 study. Three study arms were evaluated: IFL (fluorouracil [FU] + irinotecan [IRN]), FOLFOX (FU + oxaliplatin), and IROX (IRN + oxaliplatin). Information on adverse events, response, and disease-free survival was available. Thirty-four variants in 15 candidate genes for analysis based on previous associations with adverse events or outcome were assessed. Genotyping was performed using pyrosequencing. Results All variants were polymorphic. The homozygous UGT1A1*28 allele observed in 9% of patients was associated with risk of grade 4 neutropenia in patients on IROX (55% v 15%; P = .002). Deletion in GSTM1 was associated with grade 4 neutropenia after FOLFOX (28% v 16%; P = .02). Patients with a homozygous variant genotype for GSTP1 were more likely to discontinue FOLFOX because of neurotoxicity (24% v 10%; P = .01). The presence of a CYP3A5 variant was significantly associated with RR on IFL (29% v 60%; P = .0074). Most previously published genotype-toxicity or -efficacy relationships were not validated in this study. Conclusion This study provides a platform to evaluate pharmacogenetic predictors of response or severe adverse events in advanced CRC. Pharmacogenetic studies can be conducted in multicenter trials, and our findings demonstrate that with continued research, clinical application is practical.