Gastroenterology
Volume 137, Issue 5 , Pages 1573-1576, November 2009

Genome-Wide Association Study in Esophageal Squamous Cell Carcinoma

  • Adam J. Bass

      Affiliations

    • Department of Medical Oncology and Center for Cancer Genome Discovery, Dana-Farber Cancer Institute and Department of Medicine, Harvard Medical School, Boston, Massachusetts and Broad Institute, Cambridge, Massachusetts
    • Corresponding Author InformationReprint requests Address requests for reprints to Adam J, Bass, MD, Department of Medical Oncology, Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115
    • ,
  • Matthew Meyerson

      Affiliations

    • Department of Medical Oncology and Center for Cancer Genome Discovery, Dana-Farber Cancer Institute and Department of Pathology, Harvard Medical School, Boston, Massachusetts and Broad Institute, Cambridge, Massachusetts

published online 28 September 2009.

Article Outline

 

See “Functional variants in ADH1B and ALDH2 coupled with alcohol and smoking synergistically enhance esophageal cancer risk,” by Cui R, Kamatani Y, Takahashi A, et al, on page 1768.

It is often stated that cancer is a disease of the genome. In reality, cancer develops in the context of the complex interplay of each individual's germline or inherited genetic inheritance, the somatic alterations that the cells in the body accumulate over time, and one's environmental exposures. For both the germline and somatic genome, new technologies have revolutionized our ability to identify the genetic underpinnings of cancer. Large efforts such as The Cancer Genome Atlas (available: http://cancergenome.nih.gov/) and the International Cancer Genome Consortium (available: http://www.icgc.org/) have taken on the task of systematic characterization of the somatic genomic alterations in cancer. Efforts to identify the germline contribution to cancer risk have instead relied on many smaller scale investigations across different tumor types.

The study of the germline contributions to cancer has focused traditionally on the characterization of uncommon cancer syndromes. These syndromes, such as familial adenomatous polyposis (FAP), are caused by very rare but highly penetrant inherited genetic alterations in critical genes. The genes behind such syndromes are often some of the most frequently somatically mutated tumor suppressors and oncogenes such as APC1 (FAP), PTEN2 (Cowden syndrome), p533 (Li–Fraummeni syndrome), or KRAS4 (Noonan syndrome). As important as these genes are to the somatic cancer genome, these syndromes contribute to only a very small fraction of cancer and even to only a small fraction of the inferred germline contribution to cancer risk. Thus, most of the predisposing alleles that influence cancer risk remain undiscovered. Other efforts to identify the genetic contributors to cancer have focused on candidate gene approaches where variants in genes heavily suspected to influence cancer risk are queried in case-control studies.

The completion of the human genome sequence when coupled with new technologies to profile the genome has transformed the strategies for identification of the genetic determinants of cancer risk. In the prior era of limited but extremely laborious and expensive genotyping of patients, germline studies were limited to individuals in select families with unique pedigrees strongly suggestive of a genetic syndrome. The cost of genotyping germline polymorphisms, namely single nucleotide polymorphisms (SNPs), has dropped by orders of magnitude in recent years. It has thus become feasible to perform genome-wide association studies (GWAS) in large collections of affected individuals. (The genetic technology and associated improved mapping of the human genome that has allowed such studies is reviewed elsewhere.5) In the GWAS era, the focus has shifted from an effort to identify rare cancer syndromes or perform candidate-gene studies toward the goal of agnostic discovery of common genetic variants that contribute to sporadic cancers.

Among gastrointestinal tumors, several GWAS have been performed in colorectal cancers and one study in pancreatic adenocarcinoma leading to the identification of several risk alleles for these diseases (Table 1).6, 7, 8, 9, 10, 11, 12 In these and other GWAS, the identified loci are often not located in candidate genes or annotated coding genes, and even less frequently do they represent nonsynonymous changes that alter the amino acid sequence of proteins. One recent notable example comes from colorectal cancer, where a recently discovered risk allele on chromosome 8q24 was identified and recently shown to influence the binding at an enhancer for the MYC proto-oncogene.13, 14 Unlike for the rare Mendelian syndromes, the biology behind most other risk alleles identified in GWAS has not been elucidated.

Table 1. Risk Alleles Identified via GWAS in Gastrointestinal Cancer
DiseaseSNPLoci/CandidatesOdds RatioReference
Colorectalrs380284211q231.18
rs70143468q241.19
rs493982718q211.2
Colorectalrs69832678q24.211.277
Colorectalrs10505477–rs69832678q241.176
rs7197259p241.08
Colorectalrs69832678q24.211.229
Colorectalrs493982718q21.1/SMAD71.1510
Colorectalrs1079566810p141.2511
rs168927668q23.3/EIF3H0.89
Pancreaticrs5059229q34/ABO1.212
Esophageal SCCrs12299844q21-23/ADH1B1.6618
rs67112q24/ALDH21.85

In esophageal squamous cell carcinoma (SCC), there have been some candidate-based studies to identify risk alleles as has been recently reviewed by Cheung and Liu15 and Hiyama et al.16 Such studies have focused on variants in pathways controlling alcohol metabolism given epidemiologic data linking alcohol consumption to esophageal SCC.17 In this issue of Gastroenterology, Cui et al18 report the first large-scale WGAS performed in esophageal SCC. This study, conducted in a Japanese population, was significant for finding significant associations near the loci of 2 well-characterized, nonsynonymous SNPS within 2 metabolic genes, ADH1B and ALDH2. The authors followed up the identification of risk loci near these genes with striking confirmation of prior reports on the synergies of these genetic risk factors with consumption of alcohol and tobacco toward the increased risk for esophageal SCC.

The authors performed a multistage study. An initial screening phase of 188 cases and 938 controls were genotyped at ∼550,000 SNP loci using an Illumina HumanHap550v3 array. The top 12,000 SNPs from the initial analysis were tested in a further 517 cases and 548 controls. Taken together, these 2 stages of analysis identified 10 SNPs significant after correction for multiple hypothesis testing. Nine of the 10 SNPs were located in a single region spanning 4q21–q23 and the remaining SNP was located on 12q24. The identified locus on 4q contains 7 alcohol dehydrogenase members, including ADH1B. Although the SNP responsible for the previously identified nonsynonymous variant ADH1B was not on the Illumina assay, the authors genotyped samples for this allele and found modestly greater significance than for the SNPs covered in their screen. The authors did not perform fine mapping for other variants in this region. Future elucidation of the germline variants in this region may thus identify other alleles that also contribute to esophageal SCC risk, albeit likely much less than does the functional ADH1B allele.

The SNP identified on 12q24 was a silent variant in the gene BRAP, located within a cluster of genes including the aldehyde dehydrogenase ALDH2 that acts to oxidize the ethanol metabolite aldehyde. A variant allele of ALDH2 with a glutamic acid to lysine substitution, resulting in the inability to properly metabolize acetaldehyde, has been associated with increased esophageal SCC risk in heterozygotes (but not among homozygotes owing to impaired alcohol tolerance in such individuals). As this candidate allele was not assayed on the initial Illumina array, the authors further genotyped cases and controls for this SNP and identified a stronger association than for the BRAP SNP and confirmed the known correlation of this risk allele among heterozygotes, but not those with 2 nonfunctional alleles. The authors then retested the 2 functional SNPs in ALDH2 and ADH1B in an additional set of Japanese cases and controls and confirmed the strong associations.

The authors, using this large database of cases and controls genotyped for the 2 risk alleles, next performed an integrated analysis of the effect of the SNPs together or individually in conjunction with tobacco and alcohol exposure. A synergistic effect was noted in individuals harboring both risk alleles. More striking were the interactions with alcohol and tobacco exposure. The ADH1B1 SNP showed interaction with ethanol exposure and the variant in ALDH2 showed interactions with both alcohol and tobacco exposure (the latter attributed by the authors to the presence of acetaldehyde in tobacco). Individuals with both risk factors who also had significant tobacco and alcohol exposure demonstrated a remarkable odds ratio (OR) for developing esophageal SCC of 189 (95% confidence interval [CI], 95–376). By contrast, those with both risk alleles with no alcohol or tobacco use had OR of 6.8 (95% CI, 3.8–12) and those users of tobacco and alcohol with neither risk allele had an OR of only 3.4 (95% CI, 2.2–5.1).

Earlier case-control studies evaluating polymorphisms in genes involved in alcohol metabolism have come to similar conclusions regarding these risk alleles. Hori et al,19 in a case-control series from only 94 cases, identified increased risk of carcinoma in carriers of both the ALDH2 and ADH1B1 (also called ADH2) alleles and noted synergy in those carrying both alleles. Subsequent larger series with hundreds of cases and controls such as those by Yokoyama et al20 and Lee et al21 also confirmed the synergy of these risk alleles with each other and with alcohol exposure. The Yokoyama et al report notably demonstrated an imprecise but striking OR of 384 for development of esophageal SCC for those with both alleles and heavy alcohol exposure (>30 g/d). The report by Cui et al18 in this issue provides the largest dataset so far of cases and controls with matched lifestyle and genotype data and solidifies the link between alcohol consumption with both risk alleles—while demonstrating that there is a significant contribution of tobacco use with only the ALDH2 allele.

The design of the Cui et al19 report has several shortcomings, however. There is concern about the proper matching of the cases and controls in the initial phases of the study. Unfortunately, the controls in these 2 stages of the analysis were not adequately matched for age or gender as evidenced by the 2nd screen phase, where the controls were 58% male and of average age of 43 years, whereas the cases were 86% male and of average age 68 years. The initial control population of 938 individuals also came from a volunteer population of a local Rotary Club, a group younger and with less alcohol and tobacco use and potentially also differing in other socioeconomic characteristics and exposures relative to the cases with esophageal SCC, although no population substructure was identified. Some in the control population may have harbored true risk alleles but may not have yet developed esophageal SCC, given either younger age or reduced exposure to the environmental insults associated with esophageal carcinogenesis.

Although this represents the largest such study to date, the small size greatly limits its power for discovery. As Altshuler et al5 discuss, even in a study with 1,000 cases and 1,000 controls, there is only a 1% power to detect a variant present at 20% population prevalence associated with a 1.3-fold increased risk of disease. The small size also led the authors to analyze only SNPs present at a ≥10% minor allele frequency, and thus, not testing many potential risk loci. The combination of the limited power and the bias introduced by the mismatch in the cases and controls in this study hampered the potential of this study to discover new loci. It will be important to further analyze these data to determine the population attributable risk of these 2 risk alleles to better estimate what fraction of the genetic risk of esophageal SCC remains undiscovered. Future GWAS in esophageal SCC will likely require consortia to allow the collection of adequately sized collections of samples for initial and validation phases of analysis. Evaluation of cases and controls from multiple ethnic backgrounds will be essential given geographic variations in esophageal SCC incidence.

Despite these shortcomings, the authors have confirmed 2 strong genetic loci for esophageal SCC risk and demonstrated strong synergy with 2 modifiable risk factors. Future studies need to evaluate in large cohorts all germline variants within genes responsible for alcohol and acetaldehyde metabolism and study in detail their interactions with gradations of alcohol and tobacco exposure. While further studies progress, these results underscore the potential for development of a simple assay that by genotyping individuals for only 2 SNPs can provide data to strongly influence recommendations on lifestyle modifications and identify those at high risk who warrant added vigilance in screening for ESCC. Although we are still early in our understanding of the genetic predictors of cancer, results such as provided by Cui et al19 demonstrate the potential benefit of better understanding of contributions of common genetic variants to cancer risk and point to clear and achievable steps that can be taken to reduce the burden of this deadly disease.

Back to Article Outline

Acknowledgments 

The authors thank Craig Mermel for critical reading of this manuscript.

Back to Article Outline

References 

  1. Kinzler KW, Nilbert MC, Su LK, et al. Identification of FAP locus genes from chromosome 5q21. Science. 1991;253:661–665
  2. Liaw D, Marsh DJ, Li J, et al. Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome. Nat Genet. 1997;1:64–67
  3. Malkin D, Li FP, Strong LC, et al. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas and other neoplasms. Science. 1990;250:12333–12338
  4. Schubbert S, Zenker M, Rowe SL, et al. Germline KRAS mutations cause Noonan syndrome. Nat Genet. 2006;3:331–336
  5. Altshuler D, Daly MJ, Lander ES. Genetic mapping in human disease. Science. 2008;322:881–888
  6. Zanke BW, Greenwood CM, Rangrej J, et al. Genome-wide association scan identifies a colorectal cancer susceptibility locus on chromosome 8q24. Nat Genet. 2007;39:989–994
  7. Tomlinson I, Webb E, Carvajal-Carmona L, et al. A genome-wide association scan of tag SNPs identifies a susceptibility variant for colorectal cancer at 8q24.21. Nat Genet. 2007;39:984–988
  8. Tenesa A, Farrington SM, Prendergast JGD, et al. Genome-wide association scan identifies a colon cancer susceptibility locus on 11q23 and replicates loci at 8q24 and 18q21. Nat Genet. 2008;40:631–637
  9. Haiman CA, Marchand LL, Yamamato J, et al. A common genetic risk factor for colorectal and prostate cancer. Nat Genet. 2007;39:954–956
  10. Broderick P, Carvajal-Carmona L, Pittman AM, et al. A genome-wide association study shows that common alleles of SMAD7 influence colorectal cancer risk. Nat Genet. 2007;39:1315–1317
  11. Tomlinson I, Webb E, Carvajal-Carmona , et al. A genome-wide association study identified colorectal cancer susceptibility loci on chromosomes 10p14 and 8q23.3. Nat Genet. 2008;40:623–630
  12. Amundadottir L, Kraft P, Stolzenberg-Soloman RZ, et al. Genome-wide association study identifies variants in the ABO locus associated with susceptibility to pancreatic cancer. Nat Genet. 2009;41:986–990
  13. Pomerantz MM, Ahmadiyeh N, Jia L, et al. The 8q24 cancer risk variant rs6983267 shows long-range interaction with MYC in colorectal cancer. Nat Genet. 2009;41:882–884
  14. Tuupanen S, Turunen M, Lehtonen R, et al. The common colorectal cancer predisposition SNP rs6983267 at chromosome 8q24 confers potential to enhance Wnt signaling. Nat Genet. 2009;41:885–890
  15. Cheung WY, Liu G. Genetic variations in esophageal cancer risk and prognosis. Gastroenterol Clin North Am. 2009;38:75–91
  16. Hiyama T, Toshihara M, Tanaka S, et al. Genetic polymorphisms and esophageal cancer risk. Int J Cancer. 2007;121:1643–1658
  17. Pandeya N, Williams G, Green AC, et al. Alcohol consumption and the risks of adenocarcinoma and squamous cell carcinoma of the esophagus. Gastroenterology. 2009;136:1215–1224
  18. Cui R, Kamatani Y, Takahashi A, et al. Functional variants in ADH1B and ALDH2 coupled with alcohol and smoking synergistically enhance esophageal cancer risk. Gastroenterology. 2009;137:1768–1775
  19. Hori H, Kawano T, Endo M, et al. Genetic polymorphisms of tobacco-and alcohol-related metabolizing enzymes and human esophageal squamous cell carcinoma susceptibility. J Clin Gastro. 1997;25:568–575
  20. Yokoyama A, Kato H, Yokoyama T, et al. Genetic polymorphisms of alcohol and aldehyde dehydrenases and glutathione S-transferase M1 and drinking, smoking, and diet in Japanese men with esophageal squamous cell carcinoma. Carcinogenesis. 2002;23:1851–1859
  21. Lee CH, Lee JM Wu DC, et al. Carcinogenetic impact of ADH1B and ALDH2 genes on squamous cell carcinoma risk of the esophagus with regard to the consumption of alcohol, tobacco and betel quid. Int J Cancer. 2008;122:1347–1356

 Conflicts of interest The authors disclose no conflicts.

PII: S0016-5085(09)01667-9

doi:10.1053/j.gastro.2009.09.026

Refers to article:

  • Editorial AccompaniesAdditional Online Content Available Functional Variants in ADH1B and ALDH2 Coupled With Alcohol and Smoking Synergistically Enhance Esophageal Cancer Risk , 20 August 2009

    Ri Cui, Yoichiro Kamatani, Atsushi Takahashi, Masayuki Usami, Naoya Hosono, Takahisa Kawaguchi, Tatsuhiko Tsunoda, Naoyuki Kamatani, Michiaki Kubo, Yusuke Nakamura, Koichi Matsuda
    Gastroenterology November 2009 (Vol. 137, Issue 5, Pages 1768-1775)

Gastroenterology
Volume 137, Issue 5 , Pages 1573-1576, November 2009

Article Tools