Intratumoral heterogeneity has been increasingly established as a critical phenotype of cancer genomes. Its characterization at the DNA level is based on the identification of somatic mutations consisting of single nucleotide variants (SNVs) and copy number alterations (CNAs), which include deletions, amplifications, and copy neutral loss of heterozygosity. Currently, most methods for the detection of intratumoral heterogeneity use an implicit “infinite sites” model for both SNVs and CNAs. While this may often be appropriate for SNVs, we demonstrate its violation for CNAs in unstable cancer genomes. Here, we propose a novel method to identify CNAs that were created by independent mutational events but alter the same genomic region. Our method identifies regions where the germline heterozygous signals (allelic intensities for DNA arrays or frequencies for next-generation sequencing) shift toward different parental haplotypes between different samples from the same tumor, thus indicative of divergent tumor clones. In this context we define a divergent CNA as one found on multiple samples from the same tumor but with different chromosomal changes giving rise to the CNAs. We applied our method to data from core needle biopsies extracted from the tumors of 31 non-small cell lung cancer (NSCLC) patients and processed using Illumina SNP arrays. We overlapped CNA calls from the same tumor, and then tested whether overlapping segments showed divergent CNAs. We observe instances of divergent CNAs in 23 of the 31 patient tumors comprising 260 in total (median = 5 divergent CNAs per tumor). Strikingly, one tumor had 34. We then assessed whether the level of recurrent mutation correlated with clinical or genomic features. While there was no association with smoking or histology, we did observe a positive association between the rate of divergence and somatic mutations (including loss) in putative genome “gatekeeper” genes, p53 and CDKN2A (P = 0.001). We detected divergent CNAs that spanned shared genomic regions in three or more NSCLC tumors. These included large (> 1Mb) events in chromosome 6 (q13-14, q21-22, q25) and chromosome 21 (q22), as well as smaller events, which included the integrin collagen receptor locus ITGA1-PELO-ITGA2, 8p23.1, 8q24.3, 18q11.2 (ZNF521 gene), and 21q21.3, which has bindings sites for GATA2, GATA3, and STAT3. Our observed divergent genomic alterations represent half of the total number expected since imbalances of the same haplotype will not be observable in such data. In summary, our approach allows for the detection of genomic regions that are divergently altered. This information may support methods to identify CNAs under positive or negative selection in the tumor microenvironment as well as regions of increased genomic instability. This provides an added dimension to intratumoral heterogeneity analysis for a more comprehensive characterization of cancer genomes.

Citation Format: Yasminka A. Jakubek, Smruthy Sivakumar, Louise C. Strong, Humam Kadara, Paul Scheet. Intratumoral divergence of copy number alterations in NSCLC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3587. doi:10.1158/1538-7445.AM2017-3587

©2017 American Association for Cancer Research.
2017
American Association for Cancer Research.