3.1. Expression profile of SRC in different human cancers
In this study, we downloaded the pan-cancer dataset (TCGA, TARGET, GTEx, PANCAN, N=19131, G=60499) from UCSC database (https://xenabrowser.net/), which has been uniformly normalized, to obtain the SRC expression data of 34 types of cancers. Then, we calculated the expression differences of SRC between normal and tumor samples for each cancer type (Figure 1). It was observed that the expression levels of SRC were significantly up-regulated in 21 types of cancers, including glioblastoma multiforme (GBM), glioma (GBMLGG), brain lower grade glioma (LGG), breast invasive carcinoma (BRCA), lung adenocarcinoma (LUAD), esophageal carcinoma (ESCA), stomach and esophageal carcinoma (STES), kidney renal papillary cell carcinoma (KIRP), colon adenocarcinoma (COAD), colon adenocarcinoma/rectum adenocarcinoma/esophageal carcinoma (COADREAD), stomach adenocarcinoma (STAD), and so on. Meanwhile, the expression levels of SRC were significantly down-regulated in 10 types of cancers, including prostate adenocarcinoma (PRAD), kidney renal clear cell carcinoma (KIRC), high-risk Wilms tumor (WT), thyroid carcinoma (THCA), ovarian serous cystadenocarcinoma (OV), testicular germ cell tumors (TGCT), uterine carcinosarcoma (UCS), pheochromocytoma and paraganglioma (PCPG), adrenocortical carcinoma (ACC), and kidney chromophobe (KICH).
Figure 1. Differential expression of SRC between normal and tumor samples in TCGA and GTEx databases
(*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001).
3.2 Pan-cancer analysis of the prognostic value of SRC expression
In order to further elucidate the impact of SRC expression on the prognosis of cancer patients, we downloaded the pan-cancer dataset from UCSC database which has been uniformly normalized: TCGA TARGET GTEx, and analyzed the correlation of SRC expression with cancer OS and DSS. Firstly, we downloaded the pan-cancer dataset which has been uniformly normalized from UCSC, and obtained TARGET follow-up data as a supplement. We excluded samples with follow-up time less than 30 days and cancer types with less than 10 samples in a single cancer. We then performed univariate Cox regression analysis to explore the association between SRC expression and OS in pan-cancer, and observed that SRC expression was significantly associated with OS in 7 types of cancers, including LAML, PRAD, KIRC, LIHC, PAAD, GBMLGG, and BLCA (P<0.05, Figure 2A). Kaplan-Meier (KM) survival analysis showed that high expression of SRC was significantly associated with poorer OS in BLCA, PRAD, LIHC, and PAAD (P<0.05), but was associated with better OS in GBMLGG, LAML, and KIRC (Figure 2B). We further explored the relationship between SRC expression and DSS in cancer patients. Cox regression analysis of DSS showed that SRC expression was negatively associated with DSS in PRAD, LUSC, LIHC, and GBMLGG (Figure 2C). KM analysis of DSS showed that up-regulation of SRC expression was significantly associated with poor DSS in PRAD and LIHC patients, but had a good outcome in LUSC and GBMLGG patients (Figure 2D). These results suggest that SRC expression may have important biological and clinical implications in different types of cancers.
(A) Correlation between SRC expression and OS in various cancers analyzed by Cox regression model. (B) Kaplan-Meier analysis of the association between SRC expression and OS. (C) Forest plot of the correlation between SRC expression and DSS in pan-cancer. (D) Kaplan-Meier analysis of the association between SRC expression and DSS.
Figure 2. Correlation of SRC expression with pan-cancer OS and DSS
3.3 Relationship between SRC expression and different clinical features of tumors
To investigate the relationship between SRC expression and clinical pathological features, we used the Sangerbox database to analyze the correlation between SRC expression levels and various clinical pathological features, such as clinical stage, grade, gender, and age. Our results showed that in terms of gender, SRC expression exhibited significant differences in 7 types of tumors, including STES, SARC, KIPAN, STAD, KIRC, LIHC, and THCA (Figure 3A). Additionally, we analyzed the correlation between SRC expression and age, and found that the expression of SRC was positively correlated with age in cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), LAML, ESCA, STES, and LUSC (P<0.05), but was negatively correlated with age in GBMLGG, LGG, BRCA, HNSC, and LIHC (Figure 3B). In terms of clinical staging, SRC expression was significantly correlated with COAD, COADREAD, KIRC, THCA, PAAD, UVM, and BLCA (P<0.05, Figure 3C). Furthermore, SRC expression was significantly correlated with the tumor grade of STES, KIPAN, STAD, HNSC, KIRC, LIHC, and PAAD (Figure 3D). These findings suggest that SRC expression may be associated with various clinical pathological features of different types of tumors.
(A) Differential expression levels of SRC between genders; (B) Correlation between SRC expression levels and gender; (C) Differential expression of SRC in different clinical stages of tumors; (D) Differential expression of SRC in different tumor grades.
Figure 3. Differential expression of SRC in different clinical features of tumors.
3.4 Correlation analysis between SRC expression and TMB/MSI in certain cancers
TMB and MSI are effective prognostic biomarkers and immune therapy response indicators in various cancers. We analyzed the relationship between SRC expression levels and TMB/MSI status in all tumors from the TCGA database. Our results showed that SRC expression levels were positively correlated with TMB in LUAD patients (P<0.05), but were negatively correlated with TMB in COAD, COADREAD, STES, STAD, and CHOL patients (P<0.05, Figure 4A). Furthermore, SRC expression levels were significantly positively correlated with MSI in GBMLGG, LUAD, and LUSC patients (P<0.05), but were significantly negatively correlated with MSI in COAD, COADREAD, STES, STAD, HNSC, PAAD, and DLBC patients (P<0.05, Figure 4B). These findings suggest that SRC expression may be associated with TMB and MSI status in certain types of cancers.
(A)Correlation between SRC expression and TMB; (B) Correlation between SRC expression and MSI.
Figure 4. Correlation analysis between SRC expression and TMB/MSI in certain cancers
2.5 Correlation between SRC expression and immune system-related genes
2.5.1 Correlation between SRC expression and immune checkpoint (ICP) genes
To determine the role of SRC in immune therapy, we analyzed the correlation between SRC expression levels and immune checkpoint genes (ICP) using the Sangerbox database. We found that SRC expression was correlated with 60 ICP genes, including 24 inhibitory genes and 36 stimulatory genes, in various cancer types such as THYM, DLBC, LAML, UVM, BLCA, SARC, GBMLGG, LGG, CHOL, PAAD, NB, KIPAN, UCS, GBM, LUSC, MESO, ACC, KIRP, SKCM, CESC, LUAD, STAD, STES, ESCA, COAD, COADREAD, TGCT, OV, WT, READ, KICH, KIRC, and LIHC. Specifically, SRC expression was positively correlated with the expression of several immune-related genes in GBM, THYM, DLBC, LAML, UVM, PAAD, KIPAN, OV, WT, READ, KICH, KIRC, and LIHC. However, SRC expression was negatively correlated with the expression of several immune-related genes in BLCA, SARC, and GBMLGG (Figure 5). These data suggest that SRC expression is correlated with immune-related genes in most tumors, indicating that SRC may be a promising therapeutic target for cancer.
The y-axis shows the Pearson correlation coefficient. Red boxes indicate positive correlation, while blue boxes indicate negative correlation. *P<0.1, **P<0.05, ***P<0.01.
Figure 5. Correlation between SRC expression and immune checkpoint genes
2.5.2 Correlation between SRC expression and tumor immune infiltration in pan-cancer dataset
To determine the role of SRC in immune infiltration, we evaluated the correlation between SRC expression levels and immune and stromal components, as well as pan-cancer tumor purity. ESTIMATE scores indirectly reflect tumor purity, while immune and stromal scores indirectly reflect the immune and stromal components of tumor purity. We analyzed the relationship between SRC expression levels and pan-cancer immune infiltration status by estimating immune scores, stromal scores, and ESTIMATE scores in the TCGA database. We found that in most types of cancer, SRC expression levels were negatively correlated with stromal scores, immune scores, and ESTIMATE scores. Specifically, SRC expression levels were negatively correlated with immune scores in UVM, LAML, and DLBC; and negatively correlated with immune scores in GBM, GBMLGG, LGG, UCEC, CESC, LUAD, ESCA, STES, SARC, KIRP, KIPAN, COAD, COADREAD, STAD, HNSC, LUSC, THYM, WT, SKCM, BLCA, MESO, OV, and UCS (Figure 6A). SRC expression levels were positively correlated with stromal scores in LAML, THCA, UVM, TGCT, DLBC, and KICH (Figure 6B); and were significantly negatively correlated with stromal scores in 13 types of tumors, including GBM, GBMLGG, LGG, LUAD, ESCA, STES, SARC, KIRP, KIPAN, STAD, LUSC, BLCA, and PAAD (Figure 6C).
(A) Positive and negative correlation between SRC expression levels and immune scores in various cancers; (B) Positive correlation between SRC expression levels and stromal scores in various cancers; (C) Negative correlation between SRC expression levels and stromal scores in various cancers
Figure 6. Correlation between SRC expression and immune/stromal scores in various cancers
In ESTIMATE scores, SRC expression was significantly negatively correlated with 17 types of cancer, including GBM, GBMLGG, LGG, UCEC, LUAD, ESCA, STES, SARC, KIRP, KIPAN, COAD, STAD, LUSC, THYM, BLCA, OV, and PAAD (Figure 7A); and significantly positively correlated with 4 types of cancer, including UVM, LAML, DLBC, and KICH (Figure 7B).
(A) Negative correlation; (B) Positive correlation.
Figure 7. Correlation between SRC expression and ESTIMATE scores in various cancers.
To explore the relationship between SRC expression and immune cell infiltration levels, we analyzed the relationship between immune cell infiltration levels and SRC expression in different types of tumors in the TCGA database using CIBERSORT, QUANTISEQ, MCPCOUNTER, TIMER, and EPIC algorithms in the Sangerbox 3.0 database. CIBERSORT analysis showed that in most cancers, SRC expression was positively correlated with the tumor infiltration levels of various B cells and CD8+ T cells, and negatively correlated with the infiltration levels of memory B cells (Figure 8).
Figure 8. Correlation between SRC expression and immune cell infiltration levels.
*P<0.05
According to TIMER database analysis, SRC expression levels were positively correlated with the infiltration levels of B cells, CD4+ T cells, CD8+ T cells, neutrophils, macrophages and dendritic cells (DC) in PRAD, while negatively correlated in BLCA. Moreover, in KIRC, KIPAN, LIHC and BRCA cancers, the levels of tumor infiltration of B cells, CD4+ T cells, neutrophils, and DC were significantly positively correlated with SRC expression levels (Figure 9A). In addition, QUANTISEQ database analysis showed that SRC was significantly associated with immune infiltration in 38 cancer types, including GBM, GBML, LGG, UCEC, BRCA, CES, LUAD, ESCA, STES, SARC, KIRP, KIPAN, COAD, COADREAD, PRAD, STAD, HNSC, KIRC, LUSC, THYM, LIHC, SKCM-P, SKCM, BLCA, SKCM-M, THCA, READ, OV, UVM, PAAD, TGCT, UCS, LAML, PCPG, ACC, TCGA, KICH and CHOL. Furthermore, there was a positive correlation between SRC expression levels and the infiltration levels of neutrophils and NK cells in multiple cancer types based on TCGA data (Figure 9B).
MCPCOUNTER database analysis revealed a significant positive correlation between SRC expression levels and the tumor infiltration levels of neutrophils and endothelial cells in multiple cancers. Additionally, significant positive correlations were observed between SRC expression levels and the tumor infiltration levels of monocytic cell DCs, myeloid dendritic cells MDCs, and neutrophils in KIPAN, DLRC, LAML, SKCM-M, HNSC, and TGCT cancers (Figure 10A). EPIC analysis showed that in most cancers, there was a negative correlation between SRC expression levels and the infiltration levels of B cells, especially in CHOL. However, in most cancers, there was a positive correlation between SRC expression levels and the infiltration levels of CD4+ T cells, particularly in WT (Figure 10B).
(A) TIMER database analysis showed a significant correlation between SRC expression levels and the infiltration levels of various immune cells. (B) QUANTISEQ analysis indicated a significant correlation between SRC expression levels and the infiltration levels of various immune cells. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.
Figure 9: Correlation between the expression levels of SRC and the infiltration levels of various immune cells.
(A) MCPCOUNTER analysis showed a significant correlation between SRC expression levels and the infiltration levels of various immune cells. (B) EPIC analysis indicated a significant correlation between SRC expression levels and the infiltration levels of various immune cells. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.
Figure 10: Significant correlation between the expression levels of SRC and the infiltration levels of various immune cells.