SPDL1 Is an Independent Predictor of Patient Outcome in Colorectal Cancer

Abstract

Spindle Apparatus Coiled-Coil Protein 1 (SPDL1) is a relatively recently identified coiled-coil domain containing protein and an important determinant of DNA fidelity by ensuring faithful mitosis. Hence, SPDL1 is suspected to underlie genomic (in-)stability in human cancers, yet its exact roles in these diseases remain largely underexplored. Given that genomic instability (GIN) is a crucial feature in colorectal cancer (CRC), we primarily asked whether the expression of this protein may account for differences in clinicopathological features and survival rates of CRC patients. Protein expression was evaluated by immunohistochemistry in the institutional tissue microarray (TMA), and gene expression by the analysis of publicly available datasets. To place the prognostic relevance in a predicted biological context, gene co-expression set around SPDL1 identified by public data mining was annotated and assessed for enrichment in gene ontology (GO) categories, BRITE hierarchies, and Reactome pathways. The comparison with adjacent normal tissue revealed a high expression of SPDL1 protein in a subset of tumor cases (48.84%), and these had better prognosis than the SPDL1-low expression counterpart even after adjustment for multiple confounders. SPDL1-high expression within tumors was associated with a median 56-month survival advantage, but not with any clinicopathological characteristics of our cohort. In the TCGA cohort, SPDL1 was overexpressed in tumor tissue and positively associated with improved survival, chromosome instability phenotype, and various GIN markers. In addition to the genes critically involved in the cell cycle and mitosis, a gene set co-expressed with SPDL1 contained checkpoint members of both chromosome segregation and DNA replication, as well as those associated with defective DNA repair, and retrograde vesicle-mediated transport. In conclusion, SPDL1 is an independent predictor of CRC patient survival in a possible connection with chromosomal instability.

Keywords: SPDL1; colorectal cancer; genomic instability; prognostic factor.

Figure 1

(A) Immunohistochemical expression of SPDL1 in colorectal cancer compared to non-cancerous adjacent tissues. The error bars present the range from minimum to maximum values of data. Asterisk (*) indicates statistically significant difference (p < 0.05). (B) The optimal cut-off value established for SPDL1 based on IRS scores and Evaluate Cutpoints software.

Figure 2

Kaplan-Meier survival curves and log-rank test for overall survival of CRC patients based on (A) SPDL1 protein expression and (B) SPDL1 mRNA expression.

Figure 3

(A) The expression level of SPDL1 mRNA in colorectal cancer compared to normal tissues. Analysis was performed based on gene expression data sourced from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) databases. The error bars present the range from minimum to maximum values of data. Asterisks (*) indicate statistically significant differences (p < 0.05). (B) The optimal cut-off value for SPDL1 mRNA expression level was established using Evaluate Cutpoints software.

Figure 4

Functional enrichment analyses for the top 50 genes positively correlated with SPDL1 in CRC. (A) These genes are depicted with the corresponding Pearson correlation coefficient (Pearson CC) values. (B) The Reactome Pathway hierarchy panel. (C) Bar chart of the top 30 Reactome pathways based on the enrichment scores [−log10 (p-value)]. (D) Heatmap of the top 30 Reactome pathways based on the target genes. (E) Bar chart for all terms in BRITE functional hierarchies. The horizontal axis represents the number of genes enriched in KEGG Brite terms represented on the vertical axis. (F) Heatmap for all terms in BRITE functional hierarchies based on the target genes.

Figure 5

Gene ontology (GO) enrichment analysis for the top 50 genes positively correlated with SPDL1 in CRC. The three GO categories [cellular component (CC), biological process (BP) and molecular function MF)] were detected using DAVID Bioinformatics Tool. The top 15 enriched terms for BP (A) and CC (C), as well as 10 terms for MF (E) are displayed. (A,C,E) The horizontal axis of bar chart represents the –log10 (p-value) for the GO terms represented on the vertical axis. (B,D,F) Heatmaps based on the target genes.