Gallinger, Steven

Abstract Purpose: Reducing colorectal cancer incidence and mortality through early detection would improve efficacy if targeted. We developed a colorectal cancer risk prediction model incorporating personal, family, genetic, and environmental risk factors to enhance prevention. Methods: A familial risk profile (FRP) was calculated to summarize individuals' risk based on detailed cancer family history (FH), family structure, probabilities of mutation in major colorectal cancer susceptibility genes, and a polygenic component. We developed risk models, including individuals' FRP or binary colorectal cancer FH, and colorectal cancer risk factors collected at enrollment using population-based colorectal cancer cases (N = 4,445) and controls (N = 3,967) recruited by the Colon Cancer Family Registry Cohort (CCFRC). Model validation used CCFRC follow-up data for population-based (N = 12,052) and clinic-based (N = 5,584) relatives with no cancer history at recruitment to assess model calibration [expected/observed rate ratio (E/O)] and discrimination [area under the receiver-operating-characteristic curve (AUC)]. Results: The E/O [95% confidence interval (CI)] for FRP models for population-based relatives were 1.04 (0.74–1.45) for men and 0.86 (0.64–1.20) for women, and for clinic-based relatives were 1.15 (0.87–1.58) for men and 1.04 (0.76–1.45) for women. The age-adjusted AUCs (95% CI) for FRP models for population-based relatives were 0.69 (0.60–0.78) for men and 0.70 (0.62–0.77) for women, and for clinic-based relatives were 0.77 (0.69–0.84) for men and 0.68 (0.60–0.76) for women. The incremental values of AUC for FRP over FH models for population-based relatives were 0.08 (0.01–0.15) for men and 0.10 (0.04–0.16) for women, and for clinic-based relatives were 0.11 (0.05–0.17) for men and 0.11 (0.06–0.17) for women. Conclusions: Both models calibrated well. The FRP-based model provided better risk stratification and risk discrimination than the FH-based model. Impact: Our findings suggest detailed FH may be useful for targeted risk-based screening and clinical management.

Abstract Background: Smoking has been associated with colorectal cancer (CRC) risk; but limited evidence has shown the association between smoking and molecular subtypes of CRC. Methods: We analyzed 9,422 CRC cases and 9,950 controls from 10 observational studies. Multinomial logistic regression analysis was performed to assess the association between sex-study-specific quartiles of pack years of smoking and risk of CRC molecular subtypes, using non-smokers as reference group, adjusting for age, sex, and study. Known oncogenic mutations in four somatic colorectal tumor markers were assessed individually and in combination, including BRAF mutations, KRAS mutations, CpG island methylator phenotype (CIMP), and microsatellite instability (MSI) status. Case-only analysis was also performed to estimate heterogeneity in risk of molecular subtypes of CRC. Results: Compared with controls, higher pack years of smoking were statistically significantly associated with increased CRC risk when stratified, individually, by all four markers, and the associations got stronger with higher quartiles pack years (p-trend<0.001). Associations between smoking and CRC risk also differed significantly among molecular subtypes. Compared to nonsmokers, the risk of BRAFmut CRC was 83% higher for smokers within the highest quartile of pack-years (OR=1.83; 95% CI: 1.50, 2.25), and 29% higher for BRAFwt CRC (OR=1.29; 95% CI: 1.17, 1.43; Ratio of ORs (ROR)=1.45; 95% CI: 1.18, 1.17; p=3.45x10-4). Similarly, heavy pack-years of smoking was associated with almost two times higher risk of CIMP-high CRC (OR=1.93; 95% CI: 1.60, 2.31), but only 33% higher risk of CIMP-low/negative CRC (OR=1.33; 95% CI: 1.19, 1.48; ROR=1.49; 95% CI: 1.24, 1.79; p=1.72x10-5). The association between smoking and CRC was also stronger in MSI-high tumors (ORMSI-H=1.65; 95% CI: 1.36, 2.00; ORMSI-L/MSS=1.37; 95% CI: 1.23, 1.52; ROR=1.22; 95% CI: 1.00, 1.48; p=0.046). In contrast, the association between smoking and CRC risk was stronger for KRASwt (OR=1.43; 95% CI: 1.27, 1.60), than KRASmut tumors (OR=1.18; 95% CI: 1.02, 1.37; ROR=0.83; 95% CI: 0.71, 0.97; p =0.016). When combining tumor markers, smoking was found to be significantly associated with higher risk of colorectal tumors from the serrated pathway. Conclusion: In this largest study with a total of 19,372 subjects, we found that heavier pack years of smoking was associated with increased risk of all CRC molecular subtypes. Smokers with heavier pack-years of smoking had particularly higher risk of CRC subtypes with BRAF mutation and CIMP-high, suggesting smoking may be particularly involved in the development of these subtypes of colorectal tumor. Citation Format: Xiaoliang Wang, Efrat Amitay, Barbara L. Banbury, Hermann Brenner, Daniel D. Buchanan, Peter T. Campbell, Jenny Chang-Claude, Steven J. Gallinger, Graham G. Giles, Tabitha A. Harrison, John L. Hopper, Mark A. Jenkins, Yi Lin, Reiko Nishihara, Polly A. Newcomb, Shuji Ogino, Lori C. Sakoda, Robert E. Schoen, Martha L. Slattery, Steven N. Thibodeau, Bethany Van Guelpen, Michael O. Woods, Li Hsu, Michael Hoffmeister, Ulrike Peters. Smoking is associated with risks of molecular subtypes of colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 630.

Interindividual differences in DNA repair systems may play a role in modulating the individual risk of developing colorectal cancer. To better ascertain the role of DNA repair gene polymorphisms on colon and rectal cancer risk individually, we evaluated 15,419 single nucleotide polymorphisms (SNPs) within 185 DNA repair genes using GWAS data from the Colon Cancer Family Registry (CCFR) and the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO), which included 8,178 colon cancer, 2,936 rectum cancer cases and 14,659 controls. Rs1800734 (in MLH1 gene) was associated with colon cancer risk (p‐value = 3.5 × 10−6) and rs2189517 (in RAD51B) with rectal cancer risk (p‐value = 5.7 × 10−6). The results had statistical significance close to the Bonferroni corrected p‐value of 5.8 × 10−6. Ninety‐four SNPs were significantly associated with colorectal cancer risk after Binomial Sequential Goodness of Fit (BSGoF) procedure and confirmed the relevance of DNA mismatch repair (MMR) and homologous recombination pathways for colon and rectum cancer, respectively. Defects in MMR genes are known to be crucial for familial form of colorectal cancer but our findings suggest that specific genetic variations in MLH1 are important also in the individual predisposition to sporadic colon cancer. Other SNPs associated with the risk of colon cancer (e.g., rs16906252 in MGMT) were found to affect mRNA expression levels in colon transverse and therefore working as possible cis‐eQTL suggesting possible mechanisms of carcinogenesis.

Abstract Background: Pancreatic cancer is the fourth-leading cause of cancer death in both men and women in the United States. The currently identified common susceptibility loci account for a small fraction of estimated heritability. We sought to estimate overall heritability of pancreatic cancer and partition the heritability by variant frequencies and functional annotations. Methods: Analysis using the genome-based restricted maximum likelihood method (GREML) was conducted on Pancreatic Cancer Case-Control Consortium (PanC4) genome-wide association study (GWAS) data from 3,568 pancreatic cancer cases and 3,363 controls of European Ancestry. Results: Applying linkage disequilibrium- and minor allele frequency-stratified GREML (GREML-LDMS) method to imputed GWAS data, we estimated the overall heritability of pancreatic cancer to be 21.2% (SE = 4.8%). Across the functional groups (intronic, intergenic, coding, and regulatory variants), intronic variants account for most of the estimated heritability (12.4%). Previously identified GWAS loci explained 4.1% of the total phenotypic variation of pancreatic cancer. Mutations in hereditary pancreatic cancer susceptibility genes are present in 4% to 10% of patients with pancreatic cancer, yet our GREML-LDMS results suggested these regions explain only 0.4% of total phenotypic variance for pancreatic cancer. Conclusions: Although higher than previous studies, our estimated 21.2% overall heritability may still be downwardly biased due to the inherent limitation that the contribution of rare variants in genes with a substantive overall impact on disease are not captured when applying these commonly used methods to imputed GWAS data. Impact: Our work demonstrated the importance of rare and common variants in pancreatic cancer risk.

Abstract Genome-wide association studies of colorectal cancer (CRC) have identified over 50 susceptibility loci. These variants represent only a small fraction of total heritability for CRC. Gene-environment (GxE) interaction studies may help identify novel loci and biological interactions that give insight to the pathogenesis of CRC. Previous genome-wide GxE studies with dietary factors have identified interactions between loci and processed meat consumption and alcohol; however, limited statistical power remains a primary concern. Set-based SNP testing has the potential to increase statistical power to detect GxE interactions by aggregating functionally relevant SNPs. In this large pooled analysis using 14 case-control and nested case-control studies, we incorporated functional information from the transcriptome prediction tool, PrediXcan, into a novel set-based approach for testing GxE interactions. We used variant weights from the PrediXcan models of tissue-specific gene expression in the transverse colon as a priori variant information for a set-based GxE approach. We restricted our analysis to variants in the PrediXcan transverse colon gene models (n = 4,842). This discovery phase included 10,360 CRC and advanced adenoma cases, and 11,183 controls of European ancestry from the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO) and the Colon Cancer Family Registry. All 14 studies were analyzed together in a pooled data set using the Mixed Effects Score Tests for interactions. We tested for gene interactions with sex- and study- specific quartiles of dietary intake of red meat (servings/day), processed meat (servings/day), vegetables (servings/day), fruits (servings/day), and fiber (g/day). We detected two genes with suggestive interactions (false discovery rate (FDR) < 0.2) with intake of red meat and risk of CRC: Superoxide Dismutase 2 (SOD2) and Ubiquitin Conjugating Enzyme E2 H (UBE2H). No interactions at FDR < 0.2 were observed for processed meat, vegetables, fruits, or fiber. The SOD2 gene, which encodes an enzyme important in apoptotic signaling and clearing of reactive oxygen species, may regulate response to colonic exposure to heme iron or increased bile acid from high-fat content in red meat. UBE2H is part of the ubiquitin-proteasome system that has a role in Wnt signaling, which can be mediated by heme iron in red meat and is commonly found dysregulated in cancer. These findings highlight the efficacy of integrating functional information and set-based testing for novel discovery of genes interacting with known dietary risk factors of CRC. We plan to replicate these findings in additional studies. Citation Format: Paneen S. Petersen, Yu-Ru Su, Sonja I. Berndt, Stephanie A. Bien, Hermann Brenner, Graham Casey, Andrew T. Chan, Jenny Chang-Claude, Jane C. Figueiredo, Steven J. Gallinger, Robert W. Haile, Tabitha A. Harrison, Michael Hoffmeister, Mark A. Jenkins, Amit D. Joshi, Sébastien Küry, Loic Le Marchand, Yi Lin, Noralane M. Lindor, Polly A. Newcomb, John D. Potter, Robert Schoen, Martha L. Slattery, Stephen N. Thibodeau, Emily White, Li Hsu, Ulrike Peters, CCFR, GECCO. Interactions between genetic predictors of gene expression and dietary factors associated with risk of colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5268.

Abstract Background Regular use of aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs) is associated with lower risk of colorectal cancer (CRC). Genome-wide interaction analysis (GxE) has identified a few variants that may modify the effects of NSAIDs on CRC risk. However, limited statistical power remains a concern. Restricting analyses by using functional genomic information to aggregate variants into biologically relevant sets can reduce the number of tests, thereby increasing statistical power. We tested the interactions between variant models of gene expression and NSAIDs use on CRC risk. Methods Functional weights of each variant were estimated using PrediXcan based on jointly measured transcriptomes and genomes data of transverse colon tissues from the Genotype-Tissue Expression (GTEx) Project for all genes with sufficient heritability (≥1%). A mixed-effects model was used to assess the GxE effects in a gene among 9,917 incident CRC cases and 10,533 controls from 17 (nested) case-control studies from the Colon Cancer Family Registry (CCFR) and Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO). GxE analysis was done by modeling the interaction between predicted gene expression and NSAID use (fixed effects), and residual variant-specific GxE effects that are not accounted for (random effects). Regular user of aspirin and/or non-aspirin NSAIDs was compared to non-regular users. False discovery rate (FDR) was used to account for multiple comparison and FDR≤0.2 was considered genome-wide significant. Secondary analysis was performed on significant genes and duration of use among users. Results Regular use of aspirin was higher among controls (30.3%) compared to cases (24.3%). Among the 4,842 genes tested, SORD significantly modified the effect of regular use of aspirin on CRC risk (p-interaction = 1.45×10-5; FDR=0.07). The SORD gene encodes sorbitol dehydrogenase (SORD), which oxidizes sorbitol to fructose in the polyol pathway. Decreased SORD concentrations were previously observed to dramatically increase in colorectal adenoma cells when compared to normal mucosa cells, suggesting the involvement of dysregulated polyol metabolisms in colorectal tumorigenesis. However, the duration of aspirin use was not statistically significantly associated with SORD gene expression on CRC risk (p=0.149) among 5,560 aspirin users. No significant interactions were observed between genetically determined colon gene expression levels and any NSAID use or non-aspirin NSAID use at FDR<0.2. Conclusions Incorporating functional information, we discovered a novel gene that may interact with aspirin use to confer CRC risk. These findings provide preliminary support for new biological insights that could help understand the chemopreventive mechanisms of aspirin on CRC. We aim to replicate these findings in additional studies. Citation Format: Xiaoliang Wang, Yu-Ru Su, Andrew T. Chan, Stephanie Bien, Sonja I. Bernt, Hermann Brenner, Graham Casey, Jenny Chang-Claude, Steven J. Gallinger, Robert W. Haile, Tabitha A. Harrison, Michael Hoffmeister, Mark A. Jenkins, Amit Joshi, Yi Lin, Noralane M. Lindor, Loic Le Marchand, Hongmei Nan, Polly A. Newcomb, John D. Potter, Martha L. Slattery, Steve N. Thibodeau, Emily White, Li Hsu, Ulrike Peters. Functionally informed genome-wide interaction analysis of nonsteroidal anti-inflammatory drugs on colorectal cancer risk [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2965.

Greater physical activity is associated with a decrease in risk of colorectal cancer for the general population; however, little is known about its relationship with colorectal cancer risk in people with Lynch syndrome, carriers of inherited pathogenic mutations in genes affecting DNA mismatch repair (MMR). We studied a cohort of 2,042 MMR gene mutations carriers (n = 807, diagnosed with colorectal cancer), from the Colon Cancer Family Registry. Self‐reported physical activity in three age‐periods (20–29, 30–49 and ≥50 years) was summarized as average metabolic equivalent of task hours per week (MET‐hr/week) during the age‐period of cancer diagnosis or censoring (near‐term exposure) and across all age‐periods preceding cancer diagnosis or censoring (long‐term exposure). Weighted Cox regression was used to estimate the hazard ratio (HR) and 95% confidence intervals (CI) for the association between physical activity and colorectal cancer risk. Near‐term physical activity was associated with a small reduction in the risk of colorectal cancer (HR ≥35 vs. <3.5 MET‐hr/week, 0.71; 95% CI, 0.53–0.96). The strength and direction of associations were similar for long‐term physical activity, although the associations were not nominally significant. Our results suggest that physical activity is inversely associated with the risk of colorectal cancer for people with Lynch syndrome; however, further confirmation is warranted. The potential modifying effect of physical activity on colorectal cancer risk in people with Lynch syndrome could be useful for risk prediction and support counseling advice for lifestyle modification to reduce cancer risk.

Abstract Background Previous genome-wide association studies (GWAS) have identified 42 loci (P < 5 × 10−8) associated with risk of colorectal cancer (CRC). Expanded consortium efforts facilitating the discovery of additional susceptibility loci may capture unexplained familial risk. Methods We conducted a GWAS in European descent CRC cases and control subjects using a discovery–replication design, followed by examination of novel findings in a multiethnic sample (cumulative n = 163 315). In the discovery stage (36 948 case subjects/30 864 control subjects), we identified genetic variants with a minor allele frequency of 1% or greater associated with risk of CRC using logistic regression followed by a fixed-effects inverse variance weighted meta-analysis. All novel independent variants reaching genome-wide statistical significance (two-sided P < 5 × 10−8) were tested for replication in separate European ancestry samples (12 952 case subjects/48 383 control subjects). Next, we examined the generalizability of discovered variants in East Asians, African Americans, and Hispanics (12 085 case subjects/22 083 control subjects). Finally, we examined the contributions of novel risk variants to familial relative risk and examined the prediction capabilities of a polygenic risk score. All statistical tests were two-sided. Results The discovery GWAS identified 11 variants associated with CRC at P < 5 × 10−8, of which nine (at 4q22.2/5p15.33/5p13.1/6p21.31/6p12.1/10q11.23/12q24.21/16q24.1/20q13.13) independently replicated at a P value of less than .05. Multiethnic follow-up supported the generalizability of discovery findings. These results demonstrated a 14.7% increase in familial relative risk explained by common risk alleles from 10.3% (95% confidence interval [CI] = 7.9% to 13.7%; known variants) to 11.9% (95% CI = 9.2% to 15.5%; known and novel variants). A polygenic risk score identified 4.3% of the population at an odds ratio for developing CRC of at least 2.0. Conclusions This study provides insight into the architecture of common genetic variation contributing to CRC etiology and improves risk prediction for individualized screening.