Identification of Differentially Expressed Genes of BC Patients in datasets
Firstly, we downloaded the breast tissue genes expression dataset GSE10810 from GEO database to analyze the changes of BC related genes. Then, the expression of related genes in healthy control (HC) and BC groups was compared and analyzed. As shown in Figs. 1A, B, there were a total of 2371 differentially expressed genes with 747 upregulated and 1624 downregulated in BC patients. After analysis, this may be closely related to the progression and prognosis of BC.
Functional Enrichment Analysis for the BC Target Genes
In order to adequately find out the disparity of these target genes, we performed GO and KEGG enrichment analysis using DAVID database. A variety of GO enrichment terms were enriched, including 169 biological processes, 70 cellular components, and 46 molecular functions. The top 10 GO terms are shown in Fig. 1C. We found that the biological processes such as negative regulation of apoptotic process and neuronal migration, cellular components like nucleus and mitochondrion, and molecular function like chromatin binding and ATP binding, were enriched, which may be involved in the biological activity in the BC treatment process. In addition, 10 KEGG pathways were enriched (Fig. 1D), and the important genes were mainly distributed in the cell cycle signaling pathway. Although the JAK/STAT1 pathway did not make the top 10, the data suggest that the JAK/STAT1 signaling pathway is associated with the development of BC.
Expression of PD-L1, STAT1, PD-1, or CTLA-4 is associated with poor OS in BC
We performed OS analysis on 657 BC patients in TCGA database. And the results showed that the higher expression of PD-L1, PD-1, STAT1 and CTLA-4 in BC patients correlated with poor OS in the TCGA database analysis (20-year OS 0% vs 40.76%, 0% vs 35.09%, 0% vs 32.63% and 0% vs 32.69%, respectively, P = 0.262, P = 0.002, P = 0.105,P = 0.446) (Figures. 2A-C and F). But not all data were statistically significant. Expression of IDO1, IDO2, LAG3 and FGL1 in BC patients uncorrelated with poor OS in the TCGA database analysis (20-year OS 18.22% vs 33.55%, 28.16% vs 28.01%, 36.00% vs 17.05% and 18.39% vs 32.77%, respectively, P = 0.496, P = 0.002, P = 0.997,P = 0.140) (Figures. 2 DE and GH). We further analyzed the correlation between PD-1, PD-L1 and STAT1 expression patterns of other important ICs. Then, with Spearman’s correlation analysis, we found that the expression of PD-1 was positively associated with the expression of PD-L1 (r = 0.420, P < 0.0001), lymphocyte activation gene-3 (LAG-3) (r = 0.713, P < 0.0001), STAT1 (r = 0.485, P < 0.0001) and cytotoxic T-lymphocyte associated protein 4 (CTLA-4) (r = 0.789, P < 0.0001) in the TCGA group (Figures. 4B-D and F). Similarly, we found that the expression of PD-L1 was positively associated with the expression of STAT1 (r = 0.624, P < 0.0001) and CTLA-4 (r = 0.553, P < 0.0001) in the TCGA group (Figures. 4A and E). And it was statistically significant. But the expression of FGL1 was no correlation in the expression of LAG3 (r = 0.313, P = 0.424) in the TCGA group (Figures. 4G). In conclusion, the expression of PD-1, PD-L1 and STAT1 is related to the progression and prognosis of BC patients.
Co-Expression of PD-L1 and STAT1 or CTLA-4 is associated with poor OS in BC
Combination of IC inhibitors (ICIs) has the potential to improve responses. We analyzed expression patterns of ICs and found that pairwise combinations of PD-L1 and PD-1, STAT1 or CTLA-4 correlated with poor OS in BC patients (20-year OS: PD-1high and PD-L1high vs PD-1high or PD-L1high vs PD-1lowPD-L1low 0% vs 34.96% vs 39.28%) (20-year OS: PD-L1high and STAT1high vs PD-L1high or STAT1high vs PD-L1lowSTAT1low 0% vs 46.48% vs 39.43%; PD-L1high and CTLA-4high vs PD-L1high or CTLA-4high vs PD-L1lowCTLA4low 0% vs 15.72% vs 42.26%) (P = 0.113, P = 0.021, P = 0.014) in the TCGA (Figures. 3A-C).
Co-Expression of PD-1 and STAT1, CTLA-4 or LAG3 is associated with poor OS in BC
We analyzed expression patterns of ICs and found that pairwise combinations of PD-1 and STAT1, LAG-3 and CTLA-4 correlated with poor OS in BC patients (20-year OS: PD-1high and LAG3high vs PD-1high or LAG3high vs PD-1lowLAG3low 0% vs 49.78% vs 32.59%) (20-year OS: PD-1high and STAT1high vs PD-1high or STAT1high vs PD-1lowSTAT1low 0% vs 50.99% vs 26.73%) (20-year OS: PD-1high and CTLA-4high vs PD-1high or CTLA-4high vs PD-1lowCTLA-4low 0% vs 27.77% vs 34.71%) in the TCGA (P = 0.032; P = 0.005; P = 0.086) (Figures. 3D-F). And finally, what we found expression of FGL1 and LAG3 failed to correlate with OS in BC patients (20-year OS: LAG3high and FGL1high vs LAG3high or FGL1high vs LAG3lowFGL1low23.35% vs 30.25% vs 32.59%) (P = 0.474) in the TCGA (Figures. 3G).
Gene expression levels of immunofunction-related molecules in BC patients and patients with status in different clinical subgroups of BC
A total of 1085 BC patients (include Basal-like / Triple negative: 135, HER2 + non-luminal: 66, Luminal A: 415, Luminal B: 192) and 291 HC from the GTEx database were used analysis. We found that there was no significant difference in the gene expression levels of PD-L1, PD-1, CTLA-4, FGL1 and IDO2 between the BC group and the HC group (P > 0.05), and the Basal-like/Triple negative, HER2 + non-luminal, Luminal A, Luminal B had no significant difference (P > 0.05), it was not related to the clinical stage of BC (Figures. 5A, C, E, F, H). However, the expression levels of STAT1 and IDO1 genes in BC patients and HC patients were significantly different (P ≤ 0.05). The expression of STAT1 in Basal-like/Triple negative, HER2 + non-luminal, Luminal A, Luminal B was statistically significant compared with that in HC (P ≤ 0.05). But the expression of IDO1 in Basal-like / Triple negative, HER2 + non-luminal had significant difference (P ≤ 0.05) (Figures. 5B and D). Besides, the expression levels of LAG3 genes in BC patients and HC patients were no significantly different (P > 0.05), but the expression in Basal-like/Triple negative in BC patients and HC patients were significantly different (P ≤ 0.05) (Figures. 5G). Subsequently, using the difference in clinical stages of BC patients in GTEx database, it was found that the expression of PD-L1 and IDO1 was reduced in stage IV-X compared with stage I-III (P = 0.004; P = 0.006). Although the expression of PD-1, STAT1, CTLA-4, IDO2, FLG1 and LAG3 showed no significant difference in different stages of BC (P = 0.523, P = 0.083, P = 0.124, P = 0.699, P = 0.99, P = 0.058) (Figures. 5). But also confirmed that the key role of immune checkpoints BC progress.
The expression changes of target genes in the BC process
To further validate whether these four genes are related to BC pathology, we performed a serial of RQ-PCR quantification in the MCF-10A, MCF-10AT, MCF-7, MDA-MB-231 cells. As shown in Fig. 6, We found significant differences in PD-L1 gene expression levels in MCF-10A, MCF-10AT, MCF-7 and MDA-MB-231 (Figs. 6A, P = 0.016), and showed MCF-10A < MCF-7 < MCF-10AT < MDA-MB-231 trend. Similarly, STAT1 gene expression levels differed significantly among MCF-10A, MCF-10AT, MCF-7, and MDA-MB-231 (Figs. 6B, P = 0.025), and showed MCF-7 < MCF-10A < MCF-10AT < MDA-MB-231 trend, which were consistent with the GSE10810 dataset results. Moreover, the mRNA levels of IFNγ-R and IDO were no significantly in MCF-10A, MCF-10AT, MCF-7 and MDA-MB-231 (Figs. 6C and D, P = 0.110, P = 0.248). These results support that PD-L1 and STAT1 genes are closely related to the occurrence, development and prognosis of BC.