Lars Beckmann

Gene–environment interactions for complex traits: definitions, methodological requirements and challenges

Genetic and environmental risk factors and their interactions contribute to the development of complex diseases. In this review, we discuss methodological issues involved in investigating gene–environment (G × E) interactions in genetic–epidemiological studies of complex diseases and their potential relevance for clinical application. Although there are some important examples of interactions and applications, the widespread use of the knowledge about G × E interaction for targeted intervention or personalized treatment (pharmacogenetics) is still beyond current means. This is due to the fact that convincing evidence and high predictive or discriminative power are necessary conditions for usefulness in clinical practice. We attempt to clarify conceptual differences of the term ‘interaction’ in the statistical and biological sciences, since precise definitions are important for the interpretation of results. We argue that the investigation of G × E interactions is more rewarding for the detailed characterization of identified disease genes (ie at advanced stages of genetic research) and the stratified analysis of environmental effects by genotype or vice versa. Advantages and disadvantages of different epidemiological study designs are given and sample size requirements are exemplified. These issues as well as a critical appraisal of common methodological concerns are finally discussed.

Matrix Metalloproteinase 1 (MMP1) Is Associated with Early-Onset Lung Cancer

Matrix metalloproteinases (MMP) play a key role in the breakdown of extracellular matrix and in inflammatory processes. MMP1 is the most highly expressed interstitial collagenase degrading fibrillar collagens. Overexpression of MMP1 has been shown in tumor tissues and has been suggested to be associated with tumor invasion and metastasis. Nine haplotype tagging and additional two intronic single nucleotide polymorphisms (SNP) of MMP1 were genotyped in a case control sample, consisting of 635 lung cancer cases with onset of disease below 51 years of age and 1,300 age- and sex-matched cancer-free controls. Two regions of linkage disequilibrium (LD) of MMP1 could be observed: a region of low LD comprising the 5′ region including the promoter and a region of high LD starting from exon 1 to the end of the gene and including the 3′ flanking region. Several SNPs were identified to be individually significantly associated with risk of early-onset lung cancer. The most significant effect was seen for rs1938901 (P = 0.0089), rs193008 (P = 0.0108), and rs996999 (P = 0.0459). For rs996999, significance vanished after correction for multiple testing. For each of these SNPs, the major allele was associated with an increase in risk with an odds ratio between 1.2 and 1.3 (95% confidence interval, 1.0-1.5). The haplotype analysis supported these findings, especially for subgroups with high smoking intensity. In summary, we identified MMP1 to be associated with an increased risk for lung cancer, which was modified by smoking. (Cancer Epidemiol Biomarkers Prev 2008;17(5):1127–35)

Haplotype Sharing Analysis Using Mantel Statistics

Objective: The potential value of haplotypes has attracted widespread interest in the mapping of complex traits. Haplotype sharing methods take the linkage disequilibrium information between multiple markers into account, and may have good power to detect predisposing genes. We present a new approach based on Mantel statistics for spacetime clustering, which is developed in order to improve the power of haplotype sharing analysis for gene mapping in complex disease. Methods: The new statistic correlates genetic similarity and phenotypic similarity across pairs of haplotypes for case-only and case-control studies. The genetic similarity is measured as the shared length between haplotypes around a putative disease locus. The phenotypic similarity is measured as the mean-corrected cross-product based on the respective phenotypes. We analyzed two tests for statistical significance with respect to type I error: (1) assuming asymptotic normality, and (2) using a Monte Carlo permutation procedure. The results were compared to the χ2 test for association based on 3-marker haplotypes. Results: The results of the type I error rates for the Mantel statistics using the permutational procedure yielded pointwise valid tests. The approach based on the assumption of asymptotic normality was seriously liberal. Conclusion:Power comparisons showed that the Mantel statistics were better than or equal to the χ2 test for all simulated disease models.

CYP450 polymorphisms as risk factors for early-onset lung cancer: gender-specific differences

Cytochrome P450 (CYP) enzymes, involved in metabolism of tobacco carcinogens, are also involved in estrogen metabolism and many are regulated by estrogens. These genes may thus be of relevance to gender-specific differences in lung cancer risk, particularly in early-onset lung cancer, where a high proportion of women is observed. We conducted a case-control study to investigate genetic polymorphisms in cytochromes that might modify the risk of developing early-onset lung cancer. In total, 638 Caucasian patients under the age of 51 with primary lung cancer and 1300 cancer-free control individuals, matched by age and sex, were included in this analysis. Thirteen polymorphisms in the CYP1A1, CYP1B1, CYP2A13, CYP3A4 and CYP3A5 genes were analyzed. No significant association was found for any of the analyzed polymorphisms and lung cancer risk overall. However, among women, a significantly increased risk of early-onset lung cancer was observed for carriers of the minor allele of CYP1B1 SNP rs1056836 [odds ratio (OR) 1.97; 95% confidence interval (CI) 1.32-2.94; P < 0.001]. Also, a non-significant increase in lung cancer risk was observed in the group of women carriers of the minor allele of CYP2A13 SNP rs1709084 (OR 1.64; 95% CI 1.00-2.70; P = 0.05). The effect of these two polymorphisms was shown to be modified by smoking. Haplotype analysis was performed for CYP1B1 and CYP2A13. No differences between cases and controls were observed for both genes (P = 0.63 and P = 0.42 for CYP1B1 and CYP2A13, respectively). Our results suggest that the CYP1B1 and the CYP2A13 genotypes may contribute to individual susceptibility to early-onset lung cancer in women.

Polymorphisms in fatty acid metabolism-related genes are associated with colorectal cancer risk

Colorectal cancer (CRC) is the third most common malignant tumor and the fourth leading cause of cancer death worldwide. The crucial role of fatty acids for a number of important biological processes suggests a more in-depth analysis of inter-individual differences in fatty acid metabolizing genes as contributing factor to colon carcinogenesis. We examined the association between genetic variability in 43 fatty acid metabolism-related genes and colorectal risk in 1225 CRC cases and 2032 controls participating in the European Prospective Investigation into Cancer and Nutrition study. Three hundred and ninety two single-nucleotide polymorphisms were selected using pairwise tagging with an r(2) cutoff of 0.8 and a minor allele frequency of >5%. Conditional logistic regression models were used to estimate odds ratios and corresponding 95% confidence intervals. Haplotype analysis was performed using a generalized linear model framework. On the genotype level, hydroxyprostaglandin dehydrogenase 15-(NAD) (HPGD), phospholipase A2 group VI (PLA2G6) and transient receptor potential vanilloid 3 were associated with higher risk for CRC, whereas prostaglandin E receptor 2 (PTGER2) was associated with lower CRC risk. A significant inverse association (P < 0.006) was found for PTGER2 GGG haplotype, whereas HPGD AGGAG and PLA2G3 CT haplotypes were significantly (P < 0.001 and P = 0.003, respectively) associated with higher risk of CRC. Based on these data, we present for the first time the association of HPGD variants with CRC risk. Our results support the key role of prostanoid signaling in colon carcinogenesis and suggest a relevance of genetic variation in fatty acid metabolism-related genes and CRC risk.

Genetic polymorphisms of MPO, GSTT1, GSTM1, GSTP1, EPHX1 and NQO1 as risk factors of early-onset lung cancer

Early‐onset lung cancer diagnosed up to the age of 50 is a very rare disease, with an increasing incidence rate. Differences in aetiology, characteristics and epidemiology of early and older onset lung cancer have been described previously, suggesting the importance of genetic factors in early‐onset lung cancer aetiology. A case‐control study was conducted to investigate the effects of genetic polymorphisms in the MPO, EPHX1, GSTT1, GSTM1, GSTP1 and NQO1 genes on the risk of early‐onset lung cancer development. Six hundred thirty‐eight Caucasian patients under the age of 51 with confirmed primary lung cancer and 1,300 cancer free control individuals, matched by age and sex, were included in this analysis. Seventeen single nucleotide polymorphisms and two deletion polymorphisms were genotyped. No significant association was found for any of the analyzed polymorphisms and overall lung cancer risk. Nonsignificantly decreased risk of lung cancer was observed for carriers of 1 or 2 copies of GSTM1. Subgroup analysis revealed gender‐ and/or smoking‐specific effects of EPHX1 rs2854455 (IV‐1464C > T) and rs2234922 (His139Arg), GSTT1 deletion, GSTP1 rs1695 (Ile105Val), rs947895 (+991C > A) and rs4891 (Ser185Ser) and NQO1 rs1800566 (Pro187Ser) polymorphisms. However, none of the observed effects were confirmed by interaction tests nor were they significant after Bonferroni correction for multiple testing. In summary, our study suggested a modifying effect of polymorphisms in EPHX1, GSTP1, GSTT1, GSTM1 and NQO1 genes on the risk of early‐onset lung cancer. To confirm these observations and to eliminate possible bias in our analyses, larger studies are warranted.

Sample size requirements for indirect association studies of gene–environment interactions (G × E)

Association studies accounting for gene–environment interactions (G × E) may be useful for detecting genetic effects. Although current technology enables very dense marker spacing in genetic association studies, the true disease variants may not be genotyped. Thus, causal genes are searched for by indirect association using genetic markers in linkage disequilibrium (LD) with the true disease variants. Sample sizes needed to detect G × E effects in indirect case–control association studies depend on the true genetic main effects, disease allele frequencies, whether marker and disease allele frequencies match, LD between loci, main effects and prevalence of environmental exposures, and the magnitude of interactions. We explored variables influencing sample sizes needed to detect G × E, compared these sample sizes with those required to detect genetic marginal effects, and provide an algorithm for power and sample size estimations. Required sample sizes may be heavily inflated if LD between marker and disease loci decreases. More than 10,000 case–control pairs may be required to detect G × E. However, given weak true genetic main effects, moderate prevalence of environmental exposures, as well as strong interactions, G × E effects may be detected with smaller sample sizes than those needed for the detection of genetic marginal effects. Moreover, in this scenario, rare disease variants may only be detectable when G × E is included in the analyses. Thus, the analysis of G × E appears to be an attractive option for the detection of weak genetic main effects of rare variants that may not be detectable in the analysis of genetic marginal effects only. Genet. Epidemiol. 2007.

Prediction of breast cancer risk by genetic risk factors, overall and by hormone receptor status.

Objective There is increasing interest in adding common genetic variants identified through genome wide association studies (GWAS) to breast cancer risk prediction models. First results from such models showed modest benefits in terms of risk discrimination. Heterogeneity of breast cancer as defined by hormone-receptor status has not been considered in this context. In this study we investigated the predictive capacity of 32 GWAS-detected common variants for breast cancer risk, alone and in combination with classical risk factors, and for tumours with different hormone receptor status. Material and methods Within the Breast and Prostate Cancer Cohort Consortium, we analysed 6009 invasive breast cancer cases and 7827 matched controls of European ancestry, with data on classical breast cancer risk factors and 32 common gene variants identified through GWAS. Discriminatory ability with respect to breast cancer of specific hormone receptor-status was assessed with the age adjusted and cohort-adjusted concordance statistic (AUROCa). Absolute risk scores were calculated with external reference data. Integrated discrimination improvement was used to measure improvements in risk prediction. Results We found a small but steady increase in discriminatory ability with increasing numbers of genetic variants included in the model (difference in AUROCa going from 2.7% to 4%). Discriminatory ability for all models varied strongly by hormone receptor status. Discussion and conclusions Adding information on common polymorphisms provides small but statistically significant improvements in the quality of breast cancer risk prediction models. We consistently observed better performance for receptor-positive cases, but the gain in discriminatory quality is not sufficient for clinical application.

A Genome-wide Association Study of Early-Onset Breast Cancer Identifies PFKM as a Novel Breast Cancer Gene and Supports a Common Genetic Spectrum for Breast Cancer at Any Age

Early-onset breast cancer (EOBC) causes substantial loss of life and productivity, creating a major burden among women worldwide. We analyzed 1,265,548 Hapmap3 single-nucleotide polymorphisms (SNP) among a discovery set of 3,523 EOBC incident cases and 2,702 population control women ages ≤ 51 years. The SNPs with smallest P values were examined in a replication set of 3,470 EOBC cases and 5,475 control women. We also tested EOBC association with 19,684 genes by annotating each gene with putative functional SNPs, and then combining their P values to obtain a gene-based P value. We examined the gene with smallest P value for replication in 1,145 breast cancer cases and 1,142 control women. The combined discovery and replication sets identified 72 new SNPs associated with EOBC (P < 4 × 10−8) located in six genomic regions previously reported to contain SNPs associated largely with later-onset breast cancer (LOBC). SNP rs2229882 and 10 other SNPs on chromosome 5q11.2 remained associated (P < 6 × 10−4) after adjustment for the strongest published SNPs in the region. Thirty-two of the 82 currently known LOBC SNPs were associated with EOBC (P < 0.05). Low power is likely responsible for the remaining 50 unassociated known LOBC SNPs. The gene-based analysis identified an association between breast cancer and the phosphofructokinase-muscle (PFKM) gene on chromosome 12q13.11 that met the genome-wide gene-based threshold of 2.5 × 10−6. In conclusion, EOBC and LOBC seem to have similar genetic etiologies; the 5q11.2 region may contain multiple distinct breast cancer loci; and the PFKM gene region is worthy of further investigation. These findings should enhance our understanding of the etiology of breast cancer. Cancer Epidemiol Biomarkers Prev; 23(4); 658–69. ©2014 AACR.

Exploring haplotype sharing methods in general and isolated populations to detect gene(s) of a complex genetic trait

We applied a new haplotype sharing method to the simulated Genetic Analysis Workshop 12 data for both isolated and general populations without knowledge of the disease model, using affection status as phenotype and three different sample sizes. The highest peak for the mean sharing of the haplotypes was found in the isolated population for the markers D06G034 and D06G035, which flank the candidate genes located on chromosome 6, with ‐log10(p) values of 2.9 and 7.0 in the moderate and large study samples, respectively. The whole genome screen detected three further loci with ‐log10(p) values of 3.0, which turned out to be false positives. None of the true gene loci were detected in the general population even in the largest sample. The test of linkage disequilibrium based on excess haplotype sharing over the linkage equilibrium expectation revealed z‐values one order of magnitude higher in the isolated than in the general population. This approach appears to be promising for mapping genes of complex diseases depending on population characteristics.