Pan-Cancer Analysis Revealed SRSF9 as a New Biomarker for Prognosis and Immunotherapy

Abstract

Background: Serine/arginine-rich splicing factor 9 (SRSF9) is one of the members of SRSF gene family and related to the tumorigenesis and the progression of tumor. However, whether SRSF9 has a crucial role across pan-cancer is still unknown.

Methods: In this study, we used public databases, such as The Cancer Genome Atlas (TCGA), Cancer Cell Line Encyclopedia (CCLE), and Genotype-Tissue Expression (GTEx), to analyze SRSF9 expression level among tumor and normal cells. Survival analysis, K-M plotter, and PrognoScan were used to analyze the prognosis value of SRSF9, regarding to overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFI). Moreover, we performed the correlation between SRSF9 and clinical characteristics (including the outcome of prognosis), as well as molecular events of tumor mutation burden (TMB), microsatellite instability (MSI), immune checkpoint gene, tumor microenvironment (TME), immune infiltrating cells, mismatch repair (MMR) genes, m6A genes, DNA methyltransferases, and neoantigen with bioinformatics methods and TISIDB, TIMER, and Sangerbox websites.

Results: In general, SRSF9 expression was upregulated in most cancers, such as BLCA, CHOL, and UCEC, which SRSF9 was associated with short survival and severe progression. In COAD, STAD, and UCEC, SRSF9 expression was positively related to both TMB and MSI. In BRCA, BLCA, ESCA, GBM, HNSC, LUSC, LUAD, OV, PRAD, TGCT, THCA, and UCEC, both immune score and stomal score showed a negative relationship with SRSF9 expression. Immune score showed a positive relationship with SRSF9 expression in LGG. SRSF9 expression had a significant and positive correlation with six types of immune infiltration cells in LGG, KIRC, LIHC, PCPG, PRAD, SKCM, THCA, and THYM, except in LUSC. In LIHC, SRSF9 was highly significant correlated with most immune checkpoint genes. For neoantigens, correlation between SRSF9 and the quantity of neoantigens was significantly positive in some cancer types. SRSF9 was also correlated with MMR genes, m6A genes, and DNA methyltransferases. In the 33 cancer types, gene set enrichment analysis (GSEA) demonstrated that SRSF9 was correlated with multiple functions and signaling pathways.

Conclusion: These findings demonstrated that SRSF9 may be a new biomarker for the prognosis and immunotherapy in various cancers. As a result, it will be beneficial to provide new therapies for cancer patients, thereby improving the treatment and prognosis of cancer patients.

Figure 1

The transcription levels of SRSF9 in various tissues and tumors. (a) The expression levels of SRSF9 in different normal tissues from GTEx database. (b) The expression difference of SRSF9 in tumor and adjacent normal samples in pan-cancer from TCGA database and (c) combining data from TCGA and GTEx database. (d) The expression difference of SRSF9 in tumor and normal tissues from TCGA database by TIMER. The red color represents the expression of SRSF9 in tumors, and other colors represent the normal tissue.

Figure 2

Significant OS difference between high-expression and low-expression groups of SRSF9 in ACC, KIRP, LGG, LIHC, LUAD, THYM, and UVM and the association between SRSF9 expression levels with OS in pan-cancer.

Figure 3

Validation of the relationship between SRSF9 expression and prognosis in Kaplan–Meier database.

Figure 4

The correlation between SRSF9 expression and pan-cancer clinical stage using TISIDB: (a) in ACC, (b) in KICH, (c) in KIRC, (d) in LUAD, (e) in TGCT, and (f) in LUSC. The correlation between SRSF9 expression and pan-cancer tumor grade using TISIDB: (g) in CESC, (h) in KIRC, (i) in LGG, (j) in LIHC, (k) in OV, and (l) in UCEC.

Figure 5

The correlation between SRSF9 expression and pan-cancer immune subtypes and molecular subtypes using TISIDB: (a) in ACC, (b) in BRCA, (c) in COAD, (d) in KIRC, (e) in LGG, (f) in LUAD, (g) in LUSC, (h) in KIRP, (i) in PRAD, (j) in LIHC, and (k) in STAD. C1 (wound healing); C2 (IFN-gamma dominant); C3 (inflammatory); C4 (lymphocyte depleted); C5 (immunologically quiet); C6 (TGF-b dominant). The correlation between SRSF9 expression and pan-cancer molecular subtypes using TISIDB: (l) in ACC, (m) in LGG, (n) in LUSC, (o) in OV, and (p) in STAD.

Figure 6

(a) Radar plot showing the correlation between SRSF9 expression and TMB in pan-cancer. The blue number represents Spearman's correlation coefficient. (b) Radar plot showing the correlation between SRSF9 expression and MSI in pan-cancer. The green number represents Spearman's correlation coefficient. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001.

Figure 7

The correlation between SRSF9 expression and ImmuneScore and StromalScore in pan-cancer.

Figure 8

The correlation between SRSF9 expression and immune infiltrating cells in (a) KIRC, (b) LGG, (c) LIHC, (d) LUSC, (e) PCPG, (f) PRAD, (g) SKCM, (h) THCA, and (i) THYM using TIMER.

Figure 9

(a) The correlation between SRSF9 expression and immune checkpoint genes in pan-cancer. Each square corresponded to the correlation between SRSF9 expression and the expression of one immune checkpoint gene in a particular tumor. The upper triangle of each square represented the magnitude of the p value of the correlation test, and the lower triangle represented the magnitude of correlation coefficient (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001). (b) The correlation between SRSF9 expression and the number of neoantigens in pan-cancer.

Figure 10

(a) Gene set enrichment analysis of SRSF9 in the KEGG database in ACC, BLCA, BRCA, CESC, CHOL, COAD, DLBC, ESCA, GBM, HNSC, KICH, KIRC, KIRP, LAML, LGG, LIHC, LUAD, LUSC, MESO, OV, PAAD, PCPG, PRAD, READ, SARC, SKCM, STAD, TGCT, THCA, THYM, UCEC, UCS, and UVM. (b) GSEA enrichment analysis of SRSF9 with signaling pathways in the KEGG database in ACC, BLCA, GBM, HNSC, LGG, MESO, OV, PCPG, READ, SARC, STAD, and TGCT.