Dataset Information

Establishment and verification of a prognostic tumor microenvironment-based and immune-related gene signature in colon cancer

ABSTRACT:

Background

Gastrointestinal malignant cancers affect many sites in the intestinal tract, including the colon. In this study, we purposed to improve prognostic predictions for colon cancer (CC) patients by establishing a novel biosignature of immune-related genes (IRGs) based on the tumor microenvironment (TME).

Methods

Using the estimation of stromal and immune cells in malignant tumor tissues using expression data (ESTIMATE) algorithm, we calculated the stromal and immune scores of every CC patient extracted from The Cancer Genome Atlas (TCGA). We then identified 4 immune-related messenger RNA (mRNA) biosignatures through a Cox and least absolute shrinkage and selection operator (LASSO) univariate analysis, and a Cox multivariate analysis. Relationships between tumor immune infiltration and the risk score were evaluated through the CIBERSORT algorithm and Tumor Immune Estimation Resource (TIMER) database.

Results

Our studies showed that individuals who had a high immune score (P=0.017) and low stromal score (P=0.041) had a favorable overall survival (OS) rate. By comparing high/low scores cohort, 220 differentially expressed genes (DEGs) were determined. Then an immune-related four-mRNA biosignature, including PDIA2, NAFTC1, VEGFC, and CD1B was identified. Kaplan-Meier, calibration, and receiver operating characteristic (ROC) curves verified the model’s performance. By using univariate and multivariate Cox analyses, we found each biosignature was an independent risk factor for assessing a CC patient’s survival. Three external GEO cohorts validated its good efficiency in estimating OS among individuals with CC. Moreover, the signature was also related to infiltration of several cells of the immune system in the tumor microenvironment.

Conclusions

The resultant model in our study included 4 IRGs associated with the TME. These IRGs can be utilized as an auxiliary variable to estimate and help improve the prognosis of individuals with CC.

SUBMITTER: Guo T

PROVIDER: S-EPMC8576251 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundAccumulating evidence demonstrated that tumor microenvironmental cells played important roles in predicting clinical outcomes and therapeutic efficacy. We aimed to develop a reliable immune-related gene signature for predicting the prognosis of ovarian cancer (OC).MethodsSingle sample gene-set enrichment analysis (ssGSEA) of immune gene-sets was used to quantify the relative abundance of immune cell infiltration and develop high- and low-abundance immune subtypes of 308 OC samples. The presence of infiltrating stromal/immune cells in OC tissues was calculated as an estimate score. We estimated the correlation coefficients among the immune subtype, clinicopathological feature, immune score, distribution of immune cells, and tumor mutation burden (TMB). The differentially expressed immune-related genes between high- and low-abundance immune subtypes were further used to construct a gene signature of a prognostic model in OC with lasso regression analysis.ResultsThe ssGSEA analysis divided OC samples into high- and low-abundance immune subtypes based on the abundance of immune cell infiltration, which was significantly related to the estimate score and clinical characteristics. The distribution of immune cells was also significantly different between high- and low-abundance immune subtypes. The correlation analysis showed the close relationship between TMB and the estimate score. The differentially expressed immune-related genes between high- and low-abundance immune subtypes were enriched in multiple immune-related pathways. Some immune checkpoints (PDL1, PD1, and CTLA-4) were overexpressed in the high-abundance immune subtype. Furthermore, the five-immune-related-gene-signature prognostic model (CCL18, CXCL13, HLA-DOB, HLA-DPB2, and TNFRSF17)-based high-risk and low-risk groups were significantly related to OC overall survival.ConclusionImmune-related genes were the promising predictors of prognosis and survival, and the comprehensive landscape of tumor microenvironmental cells of OC has potential for therapeutic schedule monitoring.

Project description:BackgroundThe tumor microenvironment (TME) has significantly correlation with tumor occurrence and prognosis. Our study aimed to identify the prognostic immune-related genes (IRGs)in the tumor microenvironment of colorectal cancer (CRC).MethodsTranscriptome and clinical data of CRC cases were downloaded from TCGA and GEO databases. Stromal score, immune score, and tumor purity were calculated by the ESTIMATE algorithm. Based on the scores, we divided CRC patients from the TCGA database into low and high groups, and the differentially expressed genes (DEGs) were identified. Immune-related genes (IRGs) were selected by venn plots. To explore underlying pathways, protein-protein interaction (PPI) networks and functional enrichment analysis were used. After utilizing LASSO Cox regression analysis, we finally established a multi-IRGs signature for predicting the prognosis of CRC patients. A nomogram consists of the thirteen-IRGs signature and clinical parameters was developed to predict the overall survival (OS). We investigated the association between prognostic validated IRGs and immune infiltrates by TIMER database.ResultsGene expression profiles and clinical information of 1635 CRC patients were collected from the TCGA and GEO databases. Higher stromal score, immune score and lower tumor purity were observed positive correlation with tumor stage and poor OS. Based on stromal score, immune score and tumor purity, 1517 DEGs, 1296 DEGs, and 1892 DEGs were identified respectively. The 948 IRGs were screened by venn plots. A thirteen-IRGs signature was constructed for predicting survival of CRC patients. Nomogram with a C-index of 0.769 (95%CI, 0.717-0.821) was developed to predict survival of CRC patients by integrating clinical parameters and thirteen-IRGs signature. The AUC for 1-, 3-, and 5-year OS were 0.789, 0.783 and 0.790, respectively. Results from TIMER database revealed that CD1B, GPX3 and IDO1 were significantly related with immune infiltrates.ConclusionsIn this study, we established a novel thirteen immune-related genes signature that may serve as a validated prognostic predictor for CRC patients, thus will be conducive to individualized treatment decisions.

Project description:Neuroblastoma (NB) is the most prevalent malignant solid tumor in children. Tumor metabolism, including lipid, amino acid, and glucose metabolism, is intricately linked to the genesis and progression of tumors. This study aimed to establish a prognostic gene signature for NB patients, based on metabolism-related genes, and to investigate a treatment approach that could enhance the survival rate of high-risk NB patients. From the NB dataset GSE49710, we identified metabolism-related gene markers utilizing the "limma" R package and univariate Cox analysis combined with least absolute shrinkage and selection operator (LASSO) regression analysis. We explored the correlation between these gene markers and the overall survival of NB patients. Gene set enrichment analysis (GSEA) and single-sample GSEA algorithms were used to assess the differences in metabolism and immune status. Furthermore, we examined the association between metabolic subgroups and drug responsiveness. Concurrently, data downloaded from TARGET and MTAB were used for external verification. Using multicolor immunofluorescence and immunohistochemistry, we investigated the relationship between the lipid metabolism-related gene ELOVL6 with both the International Neuroblastoma Staging System classification of NB and survival rate. Finally, we explored the effect of high ELOVL6 expression on the immune microenvironment in NB using flow cytometry. We identified an eight-gene signature comprising metabolism-related genes in NB: ELOVL6, OSBPL9, RPL27A, HSD17B3, ACHE, AKR1C1, PIK3R1, and EPHX2. This panel effectively predicted disease-free survival, and was validated using an internal dataset from GSE49710 and two external datasets from the TARGET and MTAB databases. Moreover, our findings confirmed that ELOVL6 fosters an immunosuppressive microenvironment and contributes to the malignant progression in NB. The eight-gene signature is significant in predicting the prognosis of NB, effectively classifying patients into high- and low-risk groups. This classification may guide the development of innovative treatment strategies for these patients. Notably, the signature gene ELOVL6 markedly encourages an immunosuppressive microenvironment and malignant progression in NB.