Development of a Novel Immune-Related Gene Prognostic Index for Breast Cancer

Abstract

Objective: To construct an immune-related gene prognostic index (IRGPI) for breast cancer (BC) and investigate its prognostic specificity and the molecular and immune characteristics.

Methods: BC hub genes were identified from The Cancer Genome Atlas and immune-related databases using weighted gene co-expression network analysis (WGCNA). IRGPI was constructed using univariate, LASSO, and multivariate regression analyses, and was validated with GSE58812 and GSE97342 in the Gene Expression Omnibus database (GEO). At the same time, we evaluated the predictive ability of IRGPI for different BC subtypes. Subsequently, the molecular and immune characteristics, clinical relevance, and benefits of immune checkpoint inhibitor treatment were analyzed for different IRGPI subgroups.

Results: IRGPI consisted of six genes: SOCS3, TCF7L2, TSLP NPR3, ANO6, and HMGB3. The IRGPI 1-, 5-, and 10-years area under curve (AUC) values were 0.635, 0.752, and 0.753, respectively, indicating that IRGPI has good potential in predicting the long-term survival of BC patients, consistent with the results in the GEO cohort. IRGPI showed good predictive power in four different breast cancer subtypes: ER positive, PR positive, HER2 positive and triple-negative (P<0.01). Compared with the low-IRGPI group, the high-IRGPI group had a worse prognosis and a lower degree of immune infiltrating cells (p < 0.05). IRGPI showed specificity in distinguishing age, TNM stage, ER, and HER2 statuses, and our study found that the high-IRGPI group had low tumor immune dysfunction and exclusion (TIDE), microsatellite instability (MSI), and T cell dysfunction scores (p < 0.05). In addition, compared with the TIDE and TIS models, showed that the AUCs of IRGPI were better during the 5-year follow-up.

Conclusion: IRGPI can be used as an independent prognostic indicator of breast cancer, providing a method for monitoring the long-term treatment of BC.

Keywords: WGCNA; breast cancer; immune checkpoint inhibitor; immune-related genes; prognostic index.

Figure 1

Weighted co-expression network analysis (WGCNA) of immune-related differential genes. (A) Determining the soft threshold power in WGCNA analysis. (B) Gene modules related to BC obtained by WGCNA.

Figure 2

Functional enrichment of genes in the brown module. (A) Gene Ontology (GO) functional enrichment, include biological process (BP), cellular component (CC), and molecular function (MF). (B) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enrichment analysis (P < 0.05).

Figure 3

Construction of immune-related prognostic index (IRGPI) for BC. (A) LASSO coefficient profiles of the IRGs associated with disease-free survival of BC. (B) Plots of the cross-validation error rates. Each dot represents a lambda value along with error bars that represent the confidence interval for the cross-validated error rate. (C) Forest plots of hazard ratios (HR) of survival-associated IRGs obtained using multivariate Cox regression analysis. (D) K-M survival curves of 6 genes. *P <0.05, **P <0.01

Figure 4

The verification of IRGPI. (A) Survival curves of BC patients with high-IRGPI and low-IRGPI subgroups in TCGA cohort. (B) Survival curves of BC patients with high-IRGPI and low-IRGPI subgroups in GEO cohort. (C) The IRGPI 1, 5and 10 years area under curve (AUC). (D) IRGPI univariate COX regression analysis of independent prognosis. (E) IRGPI multivariate COX regression analysis of independent prognosis.

Figure 5

GSEA analysis of different IRGPI subgroups. (A) GSEA enriched in high-IRGPI subgroup. (B) GSEA enriched in low-IRGPI subgroup. (P < 0.05).

Figure 6

Immune characteristics analysis of different IRGPI subgroups. (A) The proportions of immune infiltration cells in different IRGPI subgroups. (B) Kaplan-meier survival curve of differential immune cells (P < 0.05). (C) The difference immune cell function in different IRGPI subgroups. *P <0.05, **P <0.01, ***P <0.001.

Figure 7

The prognostic value of IRGPI in ICI therapy. (A) TIDE, MSI, and T-cell exclusion and dysfunction score in different IRGPI subgroups. The scores between the two IRGPI subgroups were compared through the Wilcoxon test (ns, not significant, *P <0.05, ***P <0.001). (B) ROC analysis of IRGPI, TIS, and TIDE on OS at 5-year follow-up in breast cancer cohort.