An immune-related prognostic gene ULBP2 is correlated with immunosuppressive tumor microenvironment and immunotherapy in breast cancer

Abstract

Breast cancer (BC) is one of the major dangerous tumors threatening women's lives. We here aimed to sort out prognostic immune-related genes by univariate Cox regression analysis and build a model of immune-related genes for forecasting the prognosis of BC patients. We identified UL16 binding protein 2 (ULBP2) as a valuable gene for study in the model using related databases and algorithms analysis. We found the stromal and immune cells scores were higher in ULBP2 high expression group and ULBP2 was related to kinds of immune cells, most importantly had negative correlation with CD8⁺ T cell. Notably, ULBP2 was positively correlated with tumor mutational burden (TMB) and had relationship with many immune checkpoints. Correlation analysis revealed that ULBP2 expression was closely linked to the clinicopathological characters and negatively related to BC patient survival. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed the functional enrichment of differential genes related to ULBP2. Gene Set Enrichment Analysis (GSEA) indicated pathway enrichment in ULBP2 high and low expression groups. In short, this study comprehensively investigated the potential function of ULBP2 in BC, which might make ULBP2 to be an important therapeutic target for BC.

Keywords: Breast cancer; Immune genes; Immunotherapy effect prediction; Prognostic model; Tumor microenvironment status; ULBP2.

Fig. 1

Screening of differential genes of BC patients. (a, b) the volcano map and heatmap showed total differential genes, the color bar means logFC. (c, d) the volcano map and heatmap showed differential immune-related genes, the color bar means logFC. (e, f) the volcano map and heatmap showed differential TFs. DEGs were detected by Wilcoxon rank sum test (|log2FC| > 1 & fdr <0.05), the color bar means logFC. (g) In the interaction network, the red and green ellipsoidal nodes represent IRGs with high- and low-hazard ratios respectively. The red and green edges indicate positive- and negative-regulatory relationships respectively. (h) Univariate Cox analysis of IRGs and prognostic information of BC patients. Five prognosis-associated IRGs were screened out by drawing the forest map (p < 0.001).

Fig. 2

The risk score of the model is an important indicator of the survival of BC patients. (a) Kaplan–Meier curve of OS of BC patients. The differences between the two curves were evaluated by the two-side log-rank test. (b) The ROC curve analysis was used to evaluate the accuracy of the model. (c) Risk score distribution of the BC patients. (d) Presenting of survival status of BC patients. (e) Expression profile of five IRGs, the color bar means logFC. (f) Prognostic factors including age, stage, TNM stage and risk score were detected by univariate independent prognostic analysis (p < 0.05). (g) Prognostic factors including age, NM stage and risk score were detected by multivariate independent prognostic analysis (p < 0.05).

Fig. 3

ULBP2 was related to TME score, immune cell infiltration, checkpoints and TMB. (a) The association between the CTL level and OS of BC patients with different ULBP2 levels. (b) stromal cell score and immune cell score in ULBP2 high and low expression group. (c) The content of different kinds of immune cells in the high and low expression group of ULBP2. Spearman correlation analysis was the statistical methods. (d) the relationship of various kinds of immune cells and ULBP2. (e) ULBP2 had significant correlation with different checkpoint. Pearson analysis evaluates the significance of correlation. (f) ULBP2 was positively related to TMB. P value are based on spearman analysis.

Fig. 4

The expression of ULBP2 and association with survival of BC patients and clinicopathological characterastics. (a) Differential expression of ULBP2 in BC and normal tissues. Statistical method was Wilcoxon-Mann-Whitney test. (b) ULBP2 expression levels in the paired BC tissues were evaluated. P-value was detected by the Wilcoxon Test. (c–d) The IHC of ULBP2 expression in BC and normal tissue in the HPA dataset. (e–i) The relationship of ULBP2 expression and OS, DMFS, RFS, PPS, PFS of BC patients by KM survival analysis. (j–l) The correlation of ULBP2 expression with age, gender and T-stage of BC patients by Wilcoxon rank sum test. (m) Enrichment of clinical features in different expression groups of ULBP2. (n) Nomogram model was used to predict the probability of 1-, 3-, and 5-year OS of BC patients. Points are assigned for clinical features. The total points axis shows the sum of all these points. The total points on the bottom scales correspond to the forecast of 1-, 3-, and 5-year survival probability. (o) Calibration curve of the nomogram at 1-, 3-, and 5-year. The X-axis is nomogram-predicted OS. The Y-axis is observed OS. (p–q) independent prognostic analysis of BC patients with ULBP2, age, gender, stage.

Fig. 5

Analysis of co-expressed genes of ULBP2. (a) the correlation of ENO1, SPEF2, SFXN2, TTC39A, PCP2 and P2RX4 with ULBP2. Statistical methods is pearson correlation analysis. (b) circle graph showed 5 genes with the smallest correlation coefficient and the 5 genes with the largest correlation coefficient. (c) 100 genes differentially expressed in different expression groups of ULBP2. (d–f) GO and KEGG enrichment analyses of co-expressed genes of ULBP2. (g) Gene sets enriched in ULBP2 high and low expression group. NOM p < 0.05, and |NES| > 1 are the significance threshold.