Genomic instability genes in lung and colon adenocarcinoma indicate organ specificity of transcriptomic impact on Copy Number Alterations

Abstract

Genomic instability (GI) in cancer facilitates cancer evolution and is an exploitable target for therapy purposes. However, specific genes involved in cancer GI remain elusive. Causal genes for GI via expressions have not been comprehensively identified in colorectal cancers (CRCs). To fill the gap in knowledge, we developed a data mining strategy (Gene Expression to Copy Number Alterations; "GE-CNA"). Here we applied the GE-CNA approach to 592 TCGA CRC datasets, and identified 500 genes whose expression levels associate with CNA. Among these, 18 were survival-critical (i.e., expression levels correlate with significant differences in patients' survival). Comparison with previous results indicated striking differences between lung adenocarcinoma and CRC: (a) less involvement of overexpression of mitotic genes in generating genomic instability in the colon and (b) the presence of CNA-suppressing pathways, including immune-surveillance, was only partly similar to those in the lung. Following 13 genes (TIGD6, TMED6, APOBEC3D, EP400NL, B3GNT4, ZNF683, FOXD4, FOXD4L1, PKIB, DDB2, MT1G, CLCN3, CAPS) were evaluated as potential drug development targets (hazard ratio [> 1.3 or < 0.5]). Identification of specific CRC genomic instability genes enables researchers to develop GI targeting approach. The new results suggest that the "targeting genomic instability and/or aneuploidy" approach must be tailored for specific organs.

Figure 1

Identifying genes associated with Copy Number Alterations in colon adenocarcinoma with the “Gene Expression to Copy Number Alterations” (“GE-CNA”) approach. For all genes, we recorded CNA for high expressor tumors (N = 10) and for low expressor tumors (N = 10). The CNA from the “high expressor” and “low expressor” groups were compared using unpaired t-test for each gene, testing the correlation between gene expression and numbers of CNA (q-value < 0.05). Genes whose high expression was associated with high CNA were annotated as CNA suppressors, while genes whose high expression was associated with low CNA were annotated as CNA suppressors. Genes specifically associated with a type of CNA ([a] amplification/insertion [amp/ins] CNA, often associated with Microsatellite Instability [MIN], and [b] deletion CNA, often associated with mitotic error-mediated Chromosome Instability [CIN]), were identified. Figure was generated with cBioportal (https://www.cbioportal.org/datasets).

Figure 2

Pathway analysis of CNA suppressors. The 247 CNA facilitator genes in Supplementary Table 1 did not show significant enrichment in a pathway. The 253 CNA suppressor genes in Supplementary Table 2 were further subcategorized to amplification/insertion CNA suppressors (Supplementary Table 5) and deletion CNA suppressors (Supplementary Table 6). Amp/ins CNA suppressors include only 23 genes, while deletion CNA suppressors include 253 genes, suggesting that CRC cells with amplification/insertion CNA and deletion CNA are suppressed through different modalities. Deletion CNA suppressor genes show enrichment in the (A) Antigen Presentation Pathway, (B) Interferon signaling pathway, and (C) JAK-STAT signaling pathway, suggesting that CRC cells carrying CIN-associated deletion CNA are targeted by these immune-associated pathways and that they represent an immunosurveillance mechanism of CIN cells in CRC. Purple highlighting indicates particular genes with significant GE-CNA correlations and/or a cluster of such genes in the IPA pathways. Figures were generated with IPA (Ingenuity Pathway Analysis, QIAGEN, Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis).

Figure 3

Lung adenocarcinomas carry higher CNA than do CRCs at all stages and in both types of CNA (amp/ins CNA and deletion CNA). (A) At all stages, lung adenocarcinomas carry higher numbers of CNA (all types of CNA) than do CRCs (green: lung adenocarcinoma, orange: CRC). The difference is particularly notable at earlier stages. For stages 1–3, the difference was statistically significant (Bonferroni corrected p-value < 0.05). The trend is the same in both (B) Amp/ins CNA and (C) deletion CNA. Figures were generated from R v4.0.3 (https://www.R-project.org/).

Figure 4

CNA facilitator/suppressor genes affecting patients’ survival (“survival-critical”). For 18 genes, expression levels correlate with both CNA and patients’ survival in CRC (i.e., “survival-critical” genes). The genes represent potential targets for drug development. There are four categories, as follows. (A) Lower altered expression with improved survival. For TMED6 and TIDG6, lower expression was associated with improved survival; thus, they are potential inhibitor development targets. Hazard ratio (HR) < 1 (i.e., decreased risk). “Altered” (red), “Not Altered” (green). (B) Higher altered expression with improved survival. For DDB2, WARS, and KLHDC7B, higher expression was associated with improved survival; thus, they are potential enhancer development targets. (C) Lower altered expression with decreased survival. For MT1G, CLCN3, and CAPS, lower expression was associated with decreased patients’ survival. For HR > 1, expression alterations increase risk. For estimating magnitude of HR, small, medium, and large HRs comparing two groups would be approximately 1.3, 1.9, and 2.8, respectively. (D) Higher altered expression with decreased survival. For APOBEC3D, EP400NL, B3GNT4, ZNF683, FOXD4, FOXD4L1, and PKIB, higher expression was associated with decreased survival; thus, they are potential targets for inhibitors. (E) ROBO3 is consistently shown to be over-expressed in CRCs. This finding is corroborated by the present study. However, the impact of ROBO3 expression on patients’ survival in CRCs is small (not trivial, but possibly inconsequential) with HR1.058. Figures were generated with cBioportal and with R v4.0.3.