BioProject

Using data from high-density genomic profiling arrays, we describe the profiles of somatic copy-number aberrations (SCNAs) in 486 adenocarcinomas across all three major digestive organs, including 296 gastric and esophageal cancers. More...

Using data from high-density genomic profiling arrays, we describe the profiles of somatic copy-number aberrations (SCNAs) in 486 adenocarcinomas across all three major digestive organs, including 296 gastric and esophageal cancers. This analysis revealed that although patterns of broad, chromosome arm-level alterations are similar across the three types of adenocarcinoma, focal genomic amplifications are substantially more prevalent in gastric/esophageal adenocarcinoma. A statistical analysis identified 64 regions of significantly recurrent amplification and deletion, including those shared across these tumors and those uniquely significant in adenocarcinomas from a single organ. Among significantly amplified genes are those encoding therapeutically targetable kinases such as ERBB2, FGFR1, FGFR2, EGFR, and MET, events noted in 14% of colorectal adenocarcinomas and 37% of gastric/esophageal tumors suggesting that analysis of genomic amplification will be a critical source of biomarkers to guide therapies in upper gastrointestinal adenocarcinomas. While many of the other significant loci of amplifications implicate genes recognized to play roles in gastrointestinal and other cancers, other loci point to regions that may harbor novel genes contributing to these cancers. One such event is a recurrent focal deletion present in 15% of esophageal adenocarcinomas, which we narrow to a single likely target, the Runt transcription factor subunit RUNX1. Indeed, reintroduction of RUNX1 into a cell model with this deletion inhibited anchorage-independent growth. Overall, these results demonstrate genomic features common to these tumors and identify key differences that reflect distinctive biology and potential opportunities for therapeutic intervention. Overall design: Affymetrix SNP arrays were performed according to the manufacturer's directions on DNA extracted from 271 cancer DNAs derived from primary tissues, as well as from DNA obtained from 60 normal DNA samples. Signal intensities were normalized to raw copy number estimates using the tangent normalization method, as described in Beroukhim et al., In Press and Mermel et al., In preparation. The SNP 6.0 data from this submission were segmented using CBS. These segmented data were then combined with segmented Affymetrix 250K Sty data for 128 colon, 13 gastric and 74 esophageal adenocarcinomas using common markers to anchor the segments. Data analysis across samples was performed using this GISTIC 2.0 algorithm (Mermel C et al, Genome Biology 2011). Less...