Association of candidate single nucleotide polymorphisms with somatic mutation of the epidermal growth factor receptor pathway
© Wormald et al.; licensee BioMed Central Ltd. 2013
Received: 12 July 2013
Accepted: 21 October 2013
Published: 23 October 2013
Tumour growth in colorectal cancer and other solid cancers is frequently supported by activating mutations in the epidermal growth factor receptor (EGFR) signaling pathway (Patholog Res Int 2011:932932, 2011). Treatment of metastatic colorectal cancer with targeted anti-EGFR therapeutics such as cetuximab extends survival in only 25% of patients who test wild-type for KRAS, while the majority of patients prove resistant (J Clin Oncol 28(7):1254–1261, 2010).
Prediction of cetuximab responsiveness for KRAS wild-type colorectal cancers is currently not well defined, and prognostic biomarkers would help tailor treatment to individual patients. Somatic mutation of the EGFR signalling pathway is a prevalent mechanism of resistance to cetuximab (Nature 486(7404):532–536, 2012). If the human genome harbours variants that influence susceptibility of the EGFR pathway to oncogenic mutation, such variants could also be prognostic for cetuximab responsiveness.
We assessed whether patient genetic variants may associate with somatic mutation of the EGFR signalling pathway. We combined tumour mutation data from the Cancer Genome Atlas with matched patient genetic data, and tested for germline variants that associate with somatic mutation of the EGFR pathway (including EGFR, KRAS, BRAF, PTEN and PIK3CA).
Two single nucleotide polymorphisms (SNPs) located 90 kb upstream of the TERT oncogene associated with somatic mutation of the EGFR pathway beyond the threshold of genome-wide significance: rs7736074 (P = 4.64 × 10-9) and rs4975596 (P = 5.69 × 10-9). We show that allelic variants of rs7736074 and rs4975596 modulate TERT expression levels in multiple cancer types, and exhibit preliminary prognostic value for response to cetuximab.
We have identified two germline SNPs that associate with somatic mutation of the EGFR pathway, and may be prognostic for cetuximab responsiveness. These variants could potentially contribute to a panel of prognostic biomarkers for assessing whether metastatic colorectal cancer patients are likely to derive benefit from cetuximab treatment. Genotyping of a large cohort of cetuximab-treated colorectal cancer patients is called for to further clarify the association.
The growth of solid tumours is frequently supported by aberrant expression of epidermal growth factor receptor (EGFR) or activating mutations in downstream signalling components . Monoclonal antibodies directed against EGFR, including cetuximab and panitumumab, have shown efficacy both as monotherapies and in combination with chemotherapy for the treatment of colorectal cancer (CRC) . Despite providing new avenues of treatment for solid cancers, effectiveness in the clinic has proved variable. 40% of CRC cases harbor an activating mutation in KRAS and derive no benefit from anti-EGFR therapy, while only 13% of KRAS wild-type cases show an objective response [3, 4].
Regardless of their initial response, patients invariably develop resistance to targeted EGFR therapy [3, 5, 6]. Resistance is likely acquired by the emergence of mutations within EGFR or the EGFR pathway, including KRAS, BRAF, PIK3CA and PTEN. In KRAS wild-type CRC treated with cetuximab, 6 out of 10 cases acquire activating mutations in KRAS , and activating mutations in EGFR occur in 2 out of 10 cases . Likewise, half of all non-small cell lung cancers treated with the EGFR inhibitors gefitinib or erlotinib acquire a second mutation in exon 20 of EGFR that confers resistance . As response durations are typically measured in months, strategies to circumvent acquired drug resistance are needed.
The personalization of cancer care aims to predict effective therapy regimes according to the molecular profiles of individual patients and their cancers . Germline SNPs in two components of the EGFR signalling pathway, EGF and Cyclin D1, are associated with overall survival in advanced CRC patients treated with cetuximab monotherapy , and a SNP in LIFR shows association with response to cetuximab combination therapy . At the tumour level, somatic mutations in EGFR, KRAS, BRAF, PTEN and PIK3CA are associated with poor response to anti-EGFR therapy in CRC [2, 10]. Even the majority of cancers initially negative for these mutations fail to respond , probably because subpopulations harboring drug-resistant mutations have been selected . The identification of germline biomarkers that can predict whether a cancer is predisposed to activating mutations in the EGFR pathway would therefore be an extremely useful therapeutic tool.
Germline SNP data (Affymetrix SNP 6.0) for cancer patients were obtained from The Cancer Genome Atlas (TCGA - level 2 Birdseed output) . Matched somatic mutation data and RNA seq data were obtained from the TCGA exome sequencing pipeline and the TCGA RNA seq pipeline respectively. Where multiple replicate specimens were available from a single patient, one representative specimen was selected at random. For association analysis, patients were only included where both germline SNP data and matched somatic mutation data were available. For RNA-seq analysis, patients were only included where both germline SNP data and matched tumour RNA-seq data were available.
Genome wide association analysis
Genome wide SNP association was performed using the GWASTools package for R . Associations were tested for using logistic regression under an additive model. For quality control, SNPs exhibiting > 5% missing genotype calls or non-Hardy-Weinberg equilibrium (p < 0.001) were excluded. A relatively high minor allele frequency cutoff of 10% was chosen due to the moderate number of patients and the high frequency of the measured outcome within the cancer patient population (meaning that rare SNPs are unlikely to prove informative). Non-autosomal SNPs were also excluded. In total, 580,710 out of 906,600 SNPs on the Affymetrix Human SNP 6.0 array were included in the final analysis. The genome-wide significance p-value cutoff was calculated as 0.05/(580,710 SNPs tested) = 8.61 × 10-8. Measurement of the genomic inflation factor (λ) and adjustment of P values for genomic inflation was performed using the genomic control functionality of the METAL  software package. Alternatively, eigenvectors as determined by EIGENSTRAT  were included as covariates in a linear regression model.
RNA-Seq analysis of TERT expression levels
Raw counts from TCGA RNA-seq data were processed using edgeR . Briefly, counts were normalized within samples, and negative binomial linear models applied, allowing gene-level variance to be quantified using Cox-Reid estimates of common and tagwise dispersions. Differential expression was then tested for using a generalized linear model likelihood ratio test.
Genetic association with EGFR pathway status in cancer
We sought to determine whether a patient’s germline genetic profile influences susceptibility to mutation in EGFR or downstream signaling components. To approach this problem, we made use of The Cancer Genome Atlas (TCGA)  project which collects both somatic mutation data for patient tumours, as well as patients’ germline genetic profiles. Individual cancer types within TCGA comprise too few patients to attempt large scale association analysis, however as somatic mutation of the EGFR pathway is a hallmark of multiple types of solid cancer types, we sought to maximize the power of our study by combining patients across multiple cancer types that exhibit high frequency of mutation in the EGFR pathway. We note this increase in patient numbers comes at the expense of potentially losing signals specific to only single cancer types.
Classification of TCGA patients according to EGFR pathway mutation status
Number of patients
Top 5 SNPs identified by GWAS for EGFR pathway status
An additional quality control measure, principle component analysis, indicated that mutation status of the EGFR pathway is not simply driven by population structure (Additional file 1: Figure S1). Furthermore, different approaches of accounting for population structure did not dramatically alter the p-values for rs7736074 and rs4975596 (Figure 1B and Additional file 1: Figure S2). At the probe level, genotype intensity groups are generally well defined (Additional file 1: Figure S3), however we identified some samples with genotype call p-values above 0.05 (4% for rs7736074, 7% for rs4975596), indicating lower confidence calls (Additional file 1: Figure S4). Both SNPs remain beyond genome-wide significance with these lower-confidence calls excluded from the analysis (rs7736074: 9.46×10-9; rs4975596: 2.67 × 10-8).
Rs7736074 and rs4975596 are located approximately 12 kb upstream of SLC6A19, and 90 kb downstream of the gene encoding telomerase reverse transcriptase (TERT). Genetic variants near TERT are strongly associated with predisposition to eight or more different cancer types , suggesting a potential mechanism by which rs7736074 and rs4975596 could influence the oncogenic potential of the EGFR signaling pathway through modulation of TERT activity.
SNPs rs7736074 and rs4975596 associate with TERT expression levels
TERT expression profiles for rs7736074 and rs4975596 were nearly identical, reflecting the high degree of linkage between these polymorphisms (Figure 3A and 3B). The relationship between genotype and TERT expression was generally consistent between cancer tumour types (Figure 3A and 3B). Heterozygotes typically exhibited heightened expression levels, suggesting a complex relationship between genotype and other factors (such as copy-number or methylation) in determining TERT expression levels. In particular, the substantial differential expression of TERT between genotypes in the two NSCLC subtypes suggests that genotype could play a role in determining copy-number amplification of TERT.
SNPs rs7736074 and rs4975596 associate with in vitrotumor sensitivity to cetuximab
At the molecular level, most human cancers can be classified into one or more subtypes of disease. The germline genetic profile of a patient can influence predisposition to specific cancer subtypes; in breast cancer, for example, FGFR2 variants are strongly associated with ER-positive but not ER-negative breast cancer . In CRC, outgrowth of tumour subpopulations harboring mutations in components of the EGFR pathway is strongly associated with acquired resistance to cetuximab . Cancer heterogeneity may confound the detection of such mutations by biopsy, or they may arise during the course of treatment. This study aimed to determine whether specific germline genetic factors may predispose patients to the acquisition of mutations in the EGFR pathway, and thus to cetuximab resistance. By including multiple components of the EGFR pathway in our association analysis, we aimed to isolate genetic variants that influence the EGFR pathway as a whole, as we reasoned these would likely be most informative.
We identified germline SNPs at 15p5.33 that associate with somatic mutation of the EGFR signaling pathway in TCGA patients. In an attempt to further validate this finding, we examined association of the SNPs with in vitro resistance to cetuximab (which likely reflects to some extent the mutation status of the EGFR pathway) in an independent cohort of CRC patients, and found them to be significant.
15p5.33 is a hotspot of genetic predisposition for multiple cancer types, probably because oncogenesis and cell immortalization are closely linked with the telomere maintenance activities of TERT . We postulate that the SNPs we identified may be in linkage with a regulatory element that modulates TERT expression. Consistent with this hypothesis, we found TERT mRNA expression levels to be associated with genotype at rs4975596/rs7736074 in multiple cancer types. Association was strongest in squamous-cell carcinomas and adenocarcinomas of the lung, where the 15p5.33/TERT locus is amplified at particularly high frequency . The other cancer types we examined exhibited similar regulatory trends albeit at decreased magnitude and significance, possibly due to differences in TERT dependence, tumour heterogeneity, or the action of alternative regulatory pathways at rs4975596/rs7736074 in lung cancer.
Numerous studies have reported regulation of TERT by EGFR-responsive factors including Wnt/B-catenin , Myc , and NFkB . Further evidence for a regulatory link between EGFR and TERT was reported recently in malignant glioma, where 92% of cases harboring EGFR amplification were accompanied by a mutation in the TERT promoter . Polymorphisms that disrupt a regulatory element linking EGFR signaling to TERT expression would thus impede the oncogenic potential of the EGFR pathway, and may reduce the likelihood of the pathway succumbing to somatic mutation.
The EGFR pathway induces pro-proliferative and anti-apoptotic signals, and constitutes a convenient target for somatic mutation in cancer. The occurrence of such a mutation can impede the effectiveness of anti-EGFR therapeutics such as cetuximab. We used TCGA patient data to assess whether genetic variants may predispose to somatic mutation of the EGFR pathway. We identified two SNPs located 90 kb upstream of TERT, rs7736074 and rs4975596, that associate with EGFR pathway mutation (P < = 5.69 × 10-9). We found the same two SNPs were also predictive of in vitro cetuximab resistance using publicly available genetic data from Korean colorectal cancer patients . Our results suggest that genetic variants may predispose to somatic mutation of the EGFR pathway, and consequently to resistance with anti-EGFR therapeutics. Larger studies are called for to further characterize the contribution of patient genetic variation to anti-EGFR therapeutic resistance.
S.W. is supported by an NHMRC Biomedical Postdoctoral Fellowship (519795). This research was supported by the VLSCI's Life Sciences Computation Centre, a collaboration between Melbourne, Monash and La Trobe Universities and an initiative of the Victorian Government, Australia.
- Krasinskas AM: EGFR signaling in colorectal carcinoma. Patholog Res Int. 2011, 2011: 932932.PubMedPubMed CentralGoogle Scholar
- Bardelli A, Siena S: Molecular mechanisms of resistance to cetuximab and panitumumab in colorectal cancer. J Clin Oncol. 2010, 28 (7): 1254-1261. 10.1200/JCO.2009.24.6116.View ArticlePubMedGoogle Scholar
- Misale S, Yaeger R, Hobor S, Scala E, Janakiraman M, Liska D, Valtorta E, Schiavo R, Buscarino M, Siravegna G, et al: Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer. Nature. 2012, 486 (7404): 532-536.PubMedPubMed CentralGoogle Scholar
- Karapetis CS, Khambata-Ford S, Jonker DJ, O'Callaghan CJ, Tu D, Tebbutt NC, Simes RJ, Chalchal H, Shapiro JD, Robitaille S, et al: K-ras mutations and benefit from cetuximab in advanced colorectal cancer. N Engl J Med. 2008, 359 (17): 1757-1765. 10.1056/NEJMoa0804385.View ArticlePubMedGoogle Scholar
- Kobayashi S, Boggon TJ, Dayaram T, Janne PA, Kocher O, Meyerson M, Johnson BE, Eck MJ, Tenen DG, Halmos B: EGFR mutation and resistance of non-small-cell lung cancer to gefitinib. N Engl J Med. 2005, 352 (8): 786-792. 10.1056/NEJMoa044238.View ArticlePubMedGoogle Scholar
- Montagut C, Dalmases A, Bellosillo B, Crespo M, Pairet S, Iglesias M, Salido M, Gallen M, Marsters S, Tsai SP, et al: Identification of a mutation in the extracellular domain of the epidermal growth factor receptor conferring cetuximab resistance in colorectal cancer. Nat Med. 2012, 18 (2): 221-223. 10.1038/nm.2609.View ArticlePubMedGoogle Scholar
- Chan IS, Ginsburg GS: Personalized medicine: progress and promise. Annu Rev Genomics Hum Genet. 2011, 12: 217-244. 10.1146/annurev-genom-082410-101446.View ArticlePubMedGoogle Scholar
- Zhang W, Gordon M, Press OA, Rhodes K, Vallbohmer D, Yang DY, Park D, Fazzone W, Schultheis A, Sherrod AE, et al: Cyclin D1 and epidermal growth factor polymorphisms associated with survival in patients with advanced colorectal cancer treated with Cetuximab. Pharmacogenet Genomics. 2006, 16 (7): 475-483. 10.1097/01.fpc.0000220562.67595.a5.View ArticlePubMedGoogle Scholar
- Kim JC, Kim SY, Cho DH, Ha YJ, Choi EY, Kim CW, Roh SA, Kim TW, Ju H, Kim YS: Novel chemosensitive single-nucleotide polymorphism markers to targeted regimens in metastatic colorectal cancer. Clin Cancer Res. 2011, 17 (5): 1200-1209. 10.1158/1078-0432.CCR-10-1907.View ArticlePubMedGoogle Scholar
- De Roock W, De Vriendt V, Normanno N, Ciardiello F, Tejpar S: KRAS, BRAF, PIK3CA, and PTEN mutations: implications for targeted therapies in metastatic colorectal cancer. Lancet Oncol. 2011, 12 (6): 594-603. 10.1016/S1470-2045(10)70209-6.View ArticlePubMedGoogle Scholar
- The Cancer Genome Atlas. http://cancergenome.nih.gov/.
- Gene Expression Omnibus. http://www.ncbi.nlm.nih.gov/geo/.
- Gogarten SM, Bhangale T, Conomos MP, Laurie CA, McHugh CP, Painter I, Zheng X, Crosslin DR, Levine D, Lumley T, et al: GWASTools: an R/Bioconductor package for quality control and analysis of genome-wide association studies. Bioinformatics. 2012, 28 (24): 3329-3331. 10.1093/bioinformatics/bts610.View ArticlePubMedPubMed CentralGoogle Scholar
- Willer CJ, Li Y, Abecasis GR: METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 2010, 26 (17): 2190-2191. 10.1093/bioinformatics/btq340.View ArticlePubMedPubMed CentralGoogle Scholar
- Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D: Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006, 38 (8): 904-909. 10.1038/ng1847.View ArticlePubMedGoogle Scholar
- Robinson MD, McCarthy DJ, Smyth GK: EdgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010, 26 (1): 139-140. 10.1093/bioinformatics/btp616.View ArticlePubMedGoogle Scholar
- Tam IY, Chung LP, Suen WS, Wang E, Wong MC, Ho KK, Lam WK, Chiu SW, Girard L, Minna JD, et al: Distinct epidermal growth factor receptor and KRAS mutation patterns in non-small cell lung cancer patients with different tobacco exposure and clinicopathologic features. Clin Cancer Res. 2006, 12 (5): 1647-1653. 10.1158/1078-0432.CCR-05-1981.View ArticlePubMedGoogle Scholar
- Johnson AD, Handsaker RE, Pulit SL, Nizzari MM, O'Donnell CJ, de Bakker PI: SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics. 2008, 24 (24): 2938-2939. 10.1093/bioinformatics/btn564.View ArticlePubMedPubMed CentralGoogle Scholar
- Baird DM: Variation at the TERT locus and predisposition for cancer. Expert Rev Mol Med. 2010, 12: e16.View ArticlePubMedGoogle Scholar
- Kang JU, Koo SH, Kwon KC, Park JW, Kim JM: Gain at chromosomal region 5p15.33, containing TERT, is the most frequent genetic event in early stages of non-small cell lung cancer. Cancer Genet Cytogenet. 2008, 182 (1): 1-11. 10.1016/j.cancergencyto.2007.12.004.View ArticlePubMedGoogle Scholar
- Garcia-Closas M, Hall P, Nevanlinna H, Pooley K, Morrison J, Richesson DA, Bojesen SE, Nordestgaard BG, Axelsson CK, Arias JI, et al: Heterogeneity of breast cancer associations with five susceptibility loci by clinical and pathological characteristics. PLoS Genet. 2008, 4 (4): e1000054-10.1371/journal.pgen.1000054.View ArticlePubMedPubMed CentralGoogle Scholar
- Zhang Y, Toh L, Lau P, Wang X: Human telomerase reverse transcriptase (hTERT) is a novel target of the Wnt/beta-catenin pathway in human cancer. J Biol Chem. 2012, 287 (39): 32494-32511. 10.1074/jbc.M112.368282.View ArticlePubMedPubMed CentralGoogle Scholar
- Wu KJ, Grandori C, Amacker M, Simon-Vermot N, Polack A, Lingner J, Dalla-Favera R: Direct activation of TERT transcription by c-MYC. Nat Genet. 1999, 21 (2): 220-224. 10.1038/6010.View ArticlePubMedGoogle Scholar
- Yin L, Hubbard AK, Giardina C: NF-kappa B regulates transcription of the mouse telomerase catalytic subunit. J Biol Chem. 2000, 275 (47): 36671-36675. 10.1074/jbc.M007378200.View ArticlePubMedGoogle Scholar
- Arita H, Narita Y, Fukushima S, Tateishi K, Matsushita Y, Yoshida A, Miyakita Y, Ohno M, Collins VP, Kawahara N, et al: Upregulating mutations in the TERT promoter commonly occur in adult malignant gliomas and are strongly associated with total 1p19q loss. Acta Neuropathol. 2013Google Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1755-8794/6/43/prepub
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.