Comparative Molecular Analysis of Gastrointestinal Adenocarcinomas

Abstract

We analyzed 921 adenocarcinomas of the esophagus, stomach, colon, and rectum to examine shared and distinguishing molecular characteristics of gastrointestinal tract adenocarcinomas (GIACs). Hypermutated tumors were distinct regardless of cancer type and comprised those enriched for insertions/deletions, representing microsatellite instability cases with epigenetic silencing of MLH1 in the context of CpG island methylator phenotype, plus tumors with elevated single-nucleotide variants associated with mutations in POLE. Tumors with chromosomal instability were diverse, with gastroesophageal adenocarcinomas harboring fragmented genomes associated with genomic doubling and distinct mutational signatures. We identified a group of tumors in the colon and rectum lacking hypermutation and aneuploidy termed genome stable and enriched in DNA hypermethylation and mutations in KRAS, SOX9, and PCBP1.

Keywords: cancer; colon; colorectal; epigenetic; esophagus; genomic; methylation; rectum; stomach; tumor.

Figure 1. Genomic Features of Gastrointestinal Adenocarcinomas

(A) Significantly mutated genes in gastrointestinal adenocarcinomas (GIAC) indicated by green circles, significantly mutated genes identified in other adenocarcinomas (non-GIAC) indicated by red circles, and genes identified as significantly mutated in all adenocarcinomas indicated by white circles. (B) Genes identified as significantly recurrently amplified (left) or deleted (right) in GIAC compared to in non-GIAC. (C) DNA hypermethylation frequency (top), mutation density (middle), and arm-level and focal copy-number events (bottom) in GIAC and non-GI AC. (D) Percent GOF or LOF events in developmental transcription factors by cancer type. See also Figure S1 and Tables S1–S6.

Figure 2. Molecular Subtypes of Gastrointestinal Adenocarcinomas

(A) Flowchart of molecular subtypes: Epstein-Barr virus (EBV)-positive (red); hypermutated-single-nucleotide variant predominant (HM-SNV) (gold); microsatellite instability (MSI) (blue); chromosomal instability (CIN) (purple); and genomically stable (GS) (green). (B) 3D plot of GIAC by SNV density, indel density, and clonal deletion score (CDS). Tumors annoted as Upper GI (crosses) and lower GI (circles) and color-coded by subtypes. (C) IFNγ pathway score (top) and CD8⁺ T-cell score (adjusted for total leukocytes; bottom) by subtypes stratified by upper vs lower GI. Horizontal bars indicate median values, boxes represent interquartile range, and whiskers indicate values within 1.5 times interquartile range. (D) Unsupervised analysis of DNA methylation across GIAC. (E, F) Distribution of subtypes (E) and CIMP subgroups (F) across anatomic regions. (G, H) Distribution of MLH1/CDKN2A silencing (G) and subtypes (H) in CIMP-H tumors by anatomic region. See also Figure S2.

Figure 3. Analysis of MSI and CIMP Tumors

(A) Methylation subtypes with four CIMPs: EBV-CIMP (red), CIMP-high (blue), GEA-CIMP low (yellow), CRC-CIMP low (green) with alterations of indicated genes. (B) Methylation profiles of union of CIMP-high and MSI tumors with MLH1 silencing, KRAS, BRAF, MLH1 and MSH2 mutations. (C) Features of MSI tumors stratified by upper vs. lower GI and by CIMP-high status. Horizontal bars indicate median values, boxes represent interquartile range, and whiskers indicate values within 1.5 times interquartile range. (D) Unique and overlapping epigenetically silenced genes (>25%) in upper GI (top left), upper GI tumors excluding EBV⁺ (top right), lower GI (bottom, left), and MSI (bottom, right). (E) Frequency of silencing (black) and mutation (blue) of select genes in upper GI MSI (vertical axis) vs. lower GI MSI tumors (horizontal axis). See also Figure S3 and Table S7.

Figure 4. Molecular Features of the CIN Subtype in Upper GI

(A) Copy-number heatmap of non-hypermutated GIAC with amplification (red) and deletion (blue) with upper GI CIN tumors (top), CIN CRC (middle) and GS (bottom). (B) Plots of arm-level and focal copy-number events in CIN tumors by upper and lower GI tract. Horizontal bars indicate median values, boxes represent interquartile range, and whiskers indicate values within 1.5 times interquartile range. (C) Distribution of CIN-F (CIN-Focal) score by upper and lower GI CIN tumors. CIN-B denotes CIN-Broad. (D) Distribution of CIN-F score by clinical stage in Upper GI. Horizontal bars indicate median values, boxes represent interquartile range, and whiskers indicate values within 1.5 times interquartile range. (E) Whole genome doubling (WGD) in CIN-F and CIN-B tumors in the upper GI tract; WGD1 indicates one WGD, and WGD2 indicates >WGD (F) Frequency of distinct classes of somatic alterations in RAS and receptor tyrosine kinases (RTK; KRAS, PIK3CA, BRAF, ERBB3, ERBB2, NRAS, EGFR, FGFR1, FGFR2), cell cycle (CC; FBXW7, CCNE1, CDK6, CDKN2A, CDKN1B, CCND1, CCND2) and tumor suppressor genes (TSG) including WNT(APC, RNF43, SOX9, TCF7L2, CTNNB1), TGFβ: TGFBR2, ACVR2A, ACVR1B, SMAD4, SMAD2, SMAD) and TP53 in upper GI CIN-F and CIN-B tumors (G) Schematic model of CIN-F and CIN-B pathogenesis in upper GI. See also Figure S4.

Figure 5. Molecular Features of CIN and GS Colorectal Cancer

(A) Frequency of somatic alterations in indicated genes or pathways in non-CIMP CIN, CIMP-H/L CIN, and GS lower GI tumors. (B) SCNAs (top), mutation density (middle), and CIMP classes (bottom) across subtypes in lower GI tract. Horizontal bars indicate median values, boxes represent interquartile range, and whiskers indicate values within 1.5 times interquartile range. (C) Distribution of somatic mutations in SOX9 and PCBP1 in lower GI GS. (D) Schematic model of pathogenesis of molecular subtypes in lower GI. (E) Frequency of mutations in indicated genes in lower GI CIN/GS stratified anatomically.See also Figure S5.

Figure 6. Gastrointestinal Adenocarcinoma Mutational Signatures

(A) Mutation signatures in non-hypermutated GIAC displayed by substitution class and sequence immediately 3′ and 5′ to the mutated base. (B) Key molecular features of GIAC by anatomical distribution. (C) Intensities of mutational signatures in CIN and GS subtypes by upper and lower GI. (D) BRCA signature in CIN and GS tumors in the upper and lower GI tract. (E) AA>AC signature stratified by CIN-F and CIN-B (top) and TP53 mutation (bottom) in upper GI CIN tumors. (F) CpG>TpG signature in CIN and GS tumors in upper and lower GI stratified by CIMP status. For all boxplots, horizontal bars indicate median values, boxes represent interquartile range, and whiskers indicate values within 1.5 times interquartile range. See also Figure S6.

Figure 7. Integrated Molecular Comparison of Somatic Alterations Across GIAC Molecular Subtypes

(A–C) Alterations in select genes and pathways including RTK/RAS/PI3-K (A), TP53, cell cycle (B), and WNT/TGFβ (C). Deep deletions representing loss of more than half of the gene copies for the given ploidy of the tumor, blue; amplifications, red; missense mutations in the COSMIC repository, green; nonsense or frameshift mutations, black. Percentage of somatic alteration is indicated by numbers to the left of each gene box and divided by upper (U) and lower GI (L).