Dataset Information

Development and Validation of an Autophagy-Stroma-Based Microenvironment Gene Signature for Risk Stratification in Colorectal Cancer.

ABSTRACT:

Background

Colorectal cancer is the fourth most common cancer and the second leading cause of cancer-related death in the USA. The aim of this study was to establish a tumor gene signature based on tumor stromal cell and autophagy for predicting the risk of recurrence in patients with colorectal cancer.

Methods

We used "Rtsne" and "xCell" R packages to estimate autophagy and stroma status, respectively. The discovery cohort used microarray gene expression data retrieved from the GSE39582 dataset. The Cox regression model and Least Absolute Shrinkage and Selection Operator (LASSO) were used to identify prognostic genes and to construct an autophagy-stroma-based gene signature. Moreover, external validation was conducted using GSE17538, GSE38832, TCGA database, and patient data obtained from the First Hospital of China Medical University (CMU).

Results

The LASSO model identified three genes (TNS1, TAGLN, and SFRP4) which were used to develop a risk stratification gene signature. The autophagy-stroma-based gene signature was identified as an independent prognostic factor by multivariate analysis (p = 0.0023). The results were validated in GSE17538 (p=0.0062), GSE38832 (p=0.028), TCGA (p=0.046) database, and patient data obtained from the First Hospital of China Medical University (CMU) (p=0.027).

Conclusion

We have established and verified a feasible prognostic model of colorectal cancer based on autophagy and stromal cell characteristics of patients. The model can be used to evaluate recurrence risk of cancer patients, and the hub genes in the model provide potential targets for targeted colorectal cancer treatment.

SUBMITTER: Chen L

PROVIDER: S-EPMC8180295 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ObjectiveStromal cells and immune cells have important clinical significance in the microenvironment of colorectal cancer (CRC). This study is aimed at developing a CRC gene signature on the basis of stromal and immune scores.MethodsA cohort of CRC patients (n = 433) were adopted from The Cancer Genome Atlas (TCGA) database. Stromal/immune scores were calculated by the ESTIMATE algorithm. Correlation between prognosis/clinical characteristics and stromal/immune scores was assessed. Differentially expressed stromal and immune genes were identified. Their potential functions were annotated by functional enrichment analysis. Cox regression analysis was used to develop an eight-gene risk score model. Its predictive efficacies for 3 years, 5 years, overall survival (OS), and progression-free survival interval (PFI) were evaluated using time-dependent receiver operating characteristic (ROC) curves. The correlation between the risk score and the infiltering levels of six immune cells was analyzed using TIMER. The risk score was validated using an independent dataset.ResultsImmune score was in a significant association with prognosis and clinical characteristics of CRC. 736 upregulated and two downregulated stromal and immune genes were identified, which were mainly enriched into immune-related biological processes and pathways. An-eight gene prognostic risk score model was conducted, consisting of CCL22, CD36, CPA3, CPT1C, KCNE4, NFATC1, RASGRP2, and SLC2A3. High risk score indicated a poor prognosis of patients. The area under the ROC curves (AUC) s of the model for 3 years, 5 years, OS, and PFI were 0.71, 0.70, 0.73, and 0.66, respectively. Thus, the model possessed well performance for prediction of patients' prognosis, which was confirmed by an external dataset. Moreover, the risk score was significantly correlated with immune cell infiltration.ConclusionOur study conducted an immune-related prognostic risk score model, which could provide novel targets for immunotherapy of CRC.

Project description:Objective: The present study aimed to establish a prognostic signature based on the autophagy-related long non-coding RNAs (lncRNAs) analysis in patients with hepatocellular carcinoma (HCC). Methods: Patients with HCC from The Cancer Genome Atlas (TCGA) were taken as the training cohort, and patients from the International Cancer Genome Consortium (ICGC) were treated as the validation cohort. Autophagy-related lncRNAs were obtained via a co-expression network analysis. According to univariate and multivariate analyses, a multigene prognostic signature was constructed in the training cohort. The predictive power of the signature was confirmed in both cohorts. The detailed functions were investigated using functional analysis. The single-sample gene set enrichment analysis (ssGSEA) score was used to evaluate the tumor microenvironment. The expression levels of immunotherapy and targeted therapy targets between the two risk groups were compared. Finally, a nomogram was constructed by integrating clinicopathological parameters with independently predictive value and the risk score. Results: Four autophagy-related lncRNAs were identified to establish a prognostic signature, which separated patients into high- and low-risk groups. Survival analysis showed that patients in the high-risk group had a shorter survival time in both cohorts. A time-independent receiver-operating characteristic (ROC) curve and principal component analysis (PCA) confirmed that the prognostic signature had a robust predictive power and reliability in both cohorts. Functional analysis indicated that the expressed genes in the high-risk group are mainly enriched in autophagy- and cancer-related pathways. ssGSEA revealed that the different risk groups were associated with the tumor microenvironment. Moreover, the different risk groups had positive correlations with the expressions of specific mutant genes. Multivariate analysis showed that the risk score also exhibited excellent predictive power irrespective of clinicopathological characteristics in both cohorts. A nomogram was established. The nomogram showed good discrimination, with Harrell's concordance index (C-index) of 0.739 and good calibration. Conclusion: The four autophagy-related lncRNAs could be used as biological biomarkers and therapeutic targets. The prognostic signature and nomogram might aid clinicians in individual treatment optimization and clinical decision-making for patients with HCC.