Whole-Exome Sequencing Analyses of Inflammatory Bowel Disease-Associated Colorectal Cancers

Abstract

Background & aims: A long duration of inflammatory bowel disease (IBD) increases the risk for colorectal cancer. Mutation analysis of limited numbers of genes has indicated that colorectal tumors that develop in patients with IBD differ from those of patients without IBD. We performed whole-exome sequencing analyses to characterize the genetic landscape of these tumors.

Methods: We collected colorectal tumor and non-neoplastic tissues from 31 patients with IBD and colorectal cancer (15 with ulcerative colitis, 14 with Crohn's disease, and 2 with indeterminate colitis) and performed whole-exome sequencing analyses of the microdissected tumor and matched nontumor tissues. We identified somatic alterations by comparing matched specimens. The prevalence of mutations in sporadic colorectal tumors was obtained from previously published exome-sequencing studies.

Results: Two specimens had somatic mutations in the DNA proofreading or mismatch repair genes POLE, MLH1, and MSH6 and the tumor cells had a hypermutable phenotype. The remaining tumors had, on average, 71 alterations per sample. TP53 was the most commonly mutated gene, with prevalence similar to that of sporadic colorectal tumors (63% of cases). However, tumors from the patients with IBD had a different mutation spectrum. APC and KRAS were mutated at significantly lower rates in tumors from patients with IBD than in sporadic colorectal tumors (13% and 20% of cases, respectively). Several genes were mutated more frequently or uniquely in tumors from patients with IBD, including SOX9 and EP300 (which encode proteins in the WNT pathway), NRG1 (which encodes an ERBB ligand), and IL16 (which encodes a cytokine). Our study also revealed recurrent mutations in components of the Rho and Rac GTPase network, indicating a role for noncanonical WNT signaling in development of colorectal tumors in patients with IBD.

Conclusions: Colorectal tumors that develop in patients with IBD have distinct genetic features from sporadic colorectal tumors. These findings could be used to develop disease-specific markers for diagnosis and treatment of patients with IBD and colorectal cancer.

Keywords: Crohn’s Disease; Exome; Sequencing; Ulcerative Colitis.

Figure 1

A. Number of genetic alterations (mutations, small insertions/deletions, focal amplifications) detected through sequencing and copy number analyses in each of 32 IBD-associated tumors. Samples are organized by descending total number of alterations and type of IBD. B. Median number of somatic mutations per tumor in non-hypermutated IBD-Colorectal (this study, red frame) and in other tumors of the GI tract. Horizontal bars indicate the 25 and 75% quartiles. MSI, microsatellite instability; hypermut, hypermutable phenotype; EAC, esophageal adenocarcinoma; MSS, microsatellite stable; ESCC, esophageal squamous cell carcinoma. Published data on which this figure is based are provided in Supplementary Table 4.

Figure 2

A. Mutation sequence context of non-hypermutated IBD-associated tumors. Base substitutions were collated into categories representing the 6 possible base changes (represented by colors in the upper right) and further subdivided into the 16 possible combinations that take into account the identity of nucleotides flanking the mutated base, or “trinucleotide context” (for more detail see Supplementary Figure 3). The fractional breakdown of mutation counts is shown in the pie chart on the upper left. B. Mutation sequence context of non-hypermutated tumors from UC cases. C. Mutation sequence context of non-hypermutated tumors from CD cases.

Figure 3

Genes recurrently mutated in IBD-associated CRC and their prevalence in sporadic CRC (TCGA). Each column denotes an individual tumor and each row represents a gene. Depicted are genes with mutation rates > 10 mutations/Mb, colored by the type of coding mutation, and marked by * if present in COSMIC. Right, percent of mutated cases in this study (black bars) and in sporadic CRC from TCGA (gray bars). Genes marked by * show mutation prevalence different from that of CRC (two-tailed Fisher’s exact test with Benjamini-Hochberg correction for multiple testing, P < .05). Bottom, recurrently-amplified genes and private hotspot mutations.

Figure 4

TP53 mutation spectrum in IBD-associated CRC. A. Mutation signature of 19 IBD-associated tumors with TP53 mutations compared to that of all somatic TP53 mutations in human sporadic CRC obtained from the IARC TP53 database (http://p53.iarc.fr/). B. Distribution of TP53 single nucleotide substitutions in IBD-associated CRC. C. Distribution of TP53 single nucleotide substitutions in sporadic CRC obtained from the IARC TP53 database. D. Distribution of TP53 single nucleotide substitutions in sporadic CRC identified by exome sequencing in the TCGA study, obtained from cBioPortal (http://www.cbioportal.org/).

Figure 5

Signaling pathways affected in CD and UC. Genes recurrently mutated or amplified were assigned to manually curated pathways based on KEGG, Gene Ontology, Pathway Commons, and Ingenuity Pathway Analysis (IPA Ingenuity Systems, http://www.ingenuity.com). Pathways for discussion were selected based on prior association with CRC or the presence of multiple recurrently mutated genes in the pathway.