Survival marker genes of colorectal cancer derived from consistent transcriptomic profiling

Abstract

Background: Identification of biomarkers associated with the prognosis of different cancer subtypes is critical to achieve better therapeutic assistance. In colorectal cancer (CRC) the discovery of stable and consistent survival markers remains a challenge due to the high heterogeneity of this class of tumors. In this work, we identified a new set of gene markers for CRC associated to prognosis and risk using a large unified cohort of patients with transcriptomic profiles and survival information.

Results: We built an integrated dataset with 1273 human colorectal samples, which provides a homogeneous robust framework to analyse genome-wide expression and survival data. Using this dataset we identified two sets of genes that are candidate prognostic markers for CRC in stages III and IV, showing either up-regulation correlated with poor prognosis or up-regulation correlated with good prognosis. The top 10 up-regulated genes found as survival markers of poor prognosis (i.e. low survival) were: DCBLD2, PTPN14, LAMP5, TM4SF1, NPR3, LEMD1, LCA5, CSGALNACT2, SLC2A3 and GADD45B. The stability and robustness of the gene survival markers was assessed by cross-validation, and the best-ranked genes were also validated with two external independent cohorts: one of microarrays with 482 samples; another of RNA-seq with 269 samples. Up-regulation of the top genes was also proved in a comparison with normal colorectal tissue samples. Finally, the set of top 100 genes that showed overexpression correlated with low survival was used to build a CRC risk predictor applying a multivariate Cox proportional hazards regression analysis. This risk predictor yielded an optimal separation of the individual patients of the cohort according to their survival, with a p-value of 8.25e-14 and Hazard Ratio 2.14 (95% CI: 1.75-2.61).

Conclusions: The results presented in this work provide a solid rationale for the prognostic utility of a new set of genes in CRC, demonstrating their potential to predict colorectal tumor progression and evolution towards poor survival stages. Our study does not provide a fixed gene signature for prognosis and risk prediction, but instead proposes a robust set of genes ranked according to their predictive power that can be selected for additional tests with other CRC clinical cohorts.

Keywords: Bioinformatics; Cancer; Colon; Colorectal cancer; Gene Expression; Gene marker; Kaplan-Meier analysis; Survival; Transcriptomics.

Fig. 1

Symmetric heatmaps representing the similarity between the overall gene expression signal of the samples compared with each other. Each heatmap is composed of 210 samples (30 × 7, 30 samples random selected from each batch, i.e. from each one of the 7 GSE datasets). The samples of each batch are identified by a color in the top bar below the top dendrograms (following the colors legend). Each heatmap represents a different preprocessing and normalization method performed to merge the datasets in one batch. The methods applied were: a RMA; b RMA plus ComBat; c fRMA; d fRMA plus ComBat; e fRMA plus scaling of the data using mean-centered expression values

Fig. 2

Plots presenting the distribution of the 1273 samples from 7 datasets (GSEs) obtained by Principal Component Analysis (PCA) of the global gene expression profile of each sample; that converts the signal of each sample using an orthogonal transformation in linearly uncorrelated variables called principal components or dimensions. Each plot presents the values of the two main dimensions (dim 1 versus dim 2) and corresponds to the PCA results obtained using the expression data calculated with different preprocessing and normalization methods. The methods applied were: a RMA; b RMA plus ComBat; c fRMA; d fRMA plus ComBat; e fRMA plus scaling of the data using mean-centered expression values. The samples of each batch are identified by color dots following the colors legend

Fig. 3

Kaplan-Meier plots of the survival analysis of the set of 1273 samples from colorectal cancer (CRC) patients. The patients are separated in two groups (high in red and low in green) according to the expression profiles of 4 genes: a DCBLD2, b PTPN14, c EPHB2, d DUS1L. These genes provided the best split between patients of high and low risk based in their expression levels. In the case of genes DCBLD2 and PTPN14 (labelled in red) the over-expression is correlated with poor survival; and in the case of genes EPHB2 and DUS1L (labelled in green) the over-expression is correlated with good survival. In all cases the adjusted p.values of the analyses are very significant (as indicated inside each plot), indicating that the two populations represented by the two curves have a very clear difference in their overall survival

Fig. 4

Risk prediction done for the cohort of 1273 patients of CRC based in the multivariate analysis using the top 100 genes that showed up-regulation correlated with poor prognosis (i.e. overexpressed in low survival cases). a Plot presenting the patients according to their risk score, from Low (blue) to High (red) risk. A recursive algorithm using 10-fold cross-validation finds the value of risk score (marked with a vertical black line) that allows the best splitting of the cohort in two groups. b Kaplan-Meier plot showing the separation of these two groups: a high-risk group including 425 individuals (in red) and a low-risk group including 848 individuals (in blue). The analysis has been done using a multivariate Cox proportional-hazards regression. As shown, the division is very significant (p-value = 8.25e-14) and allows an optimal separation of individuals according to their survival. The analysis of the beta factors assigned by the regression to each of the top 100 genes (i.e. to each variable within the multivariate vector) allows the identification of the genes that are the most influential factors in this risk analysis and therefore it helps in the selection of the best “gene survival markers”