Abstract 1203: Whole-exome sequencing analysis across 23 histotypes of patient-derived tumor xenografts reveals their similarities with TCGA patient tumors

Abstract

Abstract Cell lines insufficiently recapitulate the genomic structure and diversity of its parent tumor. Consequently, results from in-vitro experiments can often not be reproduced in the clinic. In contrast, patient derived xenograft models (PDX) more closely mimic the molecular characteristics of patient tumors. In the present study, we analyzed a large part of our collection of patient derived xenografts by means of whole-exome sequencing (WES) to investigate whether their genomic alterations were consistent with those observed in patient tumors from “The Cancer Genome Atlas” (TCGA). Exonic regions from 268 Oncotest PDX were targeted using Agilent SureSelect Human all exon kits 38MB or 51MB. Enriched genomic DNA was sequenced with Illumina HiSeq2000 in 100bp paired-end (PE) reads with expected mean coverage of 100X. The WES data was analyzed for concordance with a) Sanger and Sequenom OncoCarta panels b) patient tumor mutational signatures described by Alexandrov et al. (Nature 2013) and c) gene mutation frequencies from publically available data of TCGA tumors. In the present study, we show that most of the mutations previously determined by Sequenom or Sanger sequencing were also detected by WES analyses. The mutations in our models were distributed as follows: 86% non-synonymous SNVs, 8% small indels (1 nucleotide), 3% splicing site and 3% start/stop gains or losses. The number and the type of mutations varied significantly across tumor types. Similar to the findings of Alexandrov et al., who analyzed TCGA tumors, most of our PDX were characterized by [C>T] transitions that correlated with the patient age. High mutation rates were observed in tissues typically exposed to mutagens, such as melanomas (UV radiation) or lung PDX (tobacco smoke), associated with specific mutation signatures such as increased rates of [C>T] or [C>A] mutations, respectively. Furthermore, we found an association between the presence of mutations in DNA repair genes such as MLH1 and increased mutation levels in some tumor types including colon, gastric and pancreatic cancer. Interestingly, in PDX with high mutation levels, frameshift mutations were increased two-fold, and represented 14% of all mutations. Analyzing our PDX gene mutation profiles and those of the TCGA dataset revealed comparable gene alteration patterns across the tumor types, with some exceptions possibly attributed to the enrichment of aggressive tumors in our collection. This study demonstrates that our PDX tumor collection shows typical signatures of mutational processes and gene alteration patterns similar to patient tumors of the TCGA database. These results confirm the relevance of using PDX models to evaluate anticancer agents in preclinical settings. Citation Format: Frederic Foucault, Florian Kiefer, Bruno Zeitouni, Jagatheswari Virayah, Thomas Metcalfe, Vincent Vuaroqueaux, Heinz-Herbert Fiebig. Whole-exome sequencing analysis across 23 histotypes of patient-derived tumor xenografts reveals their similarities with TCGA patient tumors. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1203. doi:10.1158/1538-7445.AM2014-1203