The “Gene_DE” module of tumor immune estimation resource, version 2 (TIMER2) web server (http://timer.cistrome.org/) was explored with input of “SMARCA4.” The differences of SMARCA4 gene expression between tumor and normal tissues of TCGA datasets were explored. For some tumors without normal tissues (e.g., TCGA-diffuse large B cell lymphomas (DLBC), TCGA-glioblastoma multiforme (GBM), TCGA-low-grade glioma (LGG), etc.), “Expression Analysis-Box Plots” module of the gene expression profiling interactive analysis, version 2 (GEPIA2) web server (http://gepia2.cancer-pku.cn/#analysis) was used to obtain expression difference between tumor and normal tissues of GTEx database. p-Value cutoff = 0.01, log2 fold change (FC) cutoff = 1, and “Match TCGA normal and GTEx data” were set as criteria. Violin plots of the SMARCA4 expression in different pathological stages of TCGA tumors through the “pathological stage plot” module of GEPIA2 were obtained. The log2 (transcripts per million (TPM) +1) transformed expression data were applied for the box or violin plots.

UALCAN portal (http://ualcan.path.uab.edu/analysis-prot.html) was used to perform protein expression analysis of the clinical proteomic tumor analysis consortium (CPTAC) dataset (4). Expression level of total protein or phosphoprotein of SMARCA4 between primary tumor and normal tissues were explored. Datasets of six tumors were selected, including breast cancer, ovarian cancer, colon cancer, clear cell renal cell carcinoma (RCC), uterine corpus endometrial carcinoma (UCEC), and lung adenocarcinoma (LUAD).

Survival map of SMARCA4 across TCGA tumors were generated from the “Survival Map” module of GEPIA2, in terms of overall survival (OS) and disease-free survival (DFS). Cutoff value of 50% was set as expression threshold for separating high- and low-expression cohorts. The log-rank test was used in the hypothesis test, and the survival plots were generated via “survival analysis” module of GEPIA2.

The cBioPortal website (https://www.cbioportal.org/) was explored using “quick selection” section to investigate “TCGA Pan Cancer Atlas Studies”. “SMARCA4” was entered for queries of the genetic alteration. Alteration frequency, mutation type, and copy number alteration (CNA) results of all TCGA tumors were obtained from “cancer types summary” module. The “mutations” module was used to explore the mutated site of SMARCA4, which is displayed in the schematic diagram of the protein structure or the three-dimensional (3D) structure. The “comparison” module was used to generate the data on the overall, disease-free, progression-free, and disease-free survival of TCGA cases with or without SMARCA4 alteration. Kaplan-Meier plots with log-rank p-value were displayed.

We first evaluated the relationship between the level of SMARCA4 expression and the abundance of six types of TIICs, including CD4+ T cells, CD8+ T cells, B cells, neutrophils, dendritic cells (DCs), and macrophages. Results were exhibited in the form of these three kinds of scores: ImmuneScore, StromalScore, and ESTIMATEScore. The higher score estimated in ImmuneScore or StromalScore positively correlated with the ratio of immune or stromal, and it referred to the higher the respective score and the larger the ratio of the corresponding component in TME. ESTIMATEScore was the sum of both, denoting the integrated proportion of both components in TME. The ImmuneScore and StromalScore of multiple cancers were obtained via the “estimate” R package.

The purity of tumors was also quantified, and the differences of 22 immune cell subtypes were further explored using the “Immune-Gene” module of TIMER2 web server. The algorithms of TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISEQ, XCELL, MCPCOUNTER, and EPIC were used to estimate immune infiltration. The p-values and partial correlation values (cor) were obtained via the purity-adjusted Spearman’s rank correlation test. The ratio of immune-stromal component in tumor microenvironment (TME) to obtained explores the association of the estimated proportion of immune and stromal with SMARCA4 expression using Spearman’s correlation analysis.

Spearman’s correlation analysis was also used to evaluate the relationships between SMARCA4 expression and expression levels of the immune checkpoint markers. Gene markers of TIICs were analyzed including T cells (general), CD8+ T cells, B cells, monocytes, TAMs, M1 macrophages, M2 macrophages, DCs, neutrophils, natural killer (NK) cells, follicular helper T (Tfh) cells, T-helper 1 (Th1) cells, T-helper 2 (Th2) cells, T-helper 17 (Th17) cells, exhausted T cells, Tregs, and mast cells (5). These gene markers include BLTA, CD200, TNFRSF14, NRP1, LAIR1, TNFSF4, CD244, LAG3, ICOS, CD40LG, CTLA4, CD48, CD28, CD200R1, HAVCR2, ADORA2A, CD276, KIR3DL1, CD80, PDCD1, LGALS9, CD160, TNFSF14, IDO2, ICOSLG, TMIGD2, VTCN1, IDO1, PDCD1LG2, HHLA2, TNFSF18, BTNL2, CD70, TNFSF9, TNFRSF8, CD27, TNFRSF25, VSIR, TNFRSF4, CD40, TNFRSF18, TNFSF15, TIGIT, CD274, CD86, CD44, and TNFRSF9. All of the gene expression levels were log2 transformed.

Relationship between SMARCA4 and TMB or MSI was analyzed. The “maftools” R package was applied to analyze the somatic data (MAF data) in human pan-cancer from the TCGA database. The number of mutations of exons to identify the TMB was analyzed in each cancer. MSI score was obtained from the TCGA database. The analysis of the association between SMARCA4 expression and TMB or MSI was based on the Spearman’s method.

MLH1, MSH2, MSH6, EPCAM, and PMS2 are five MMR genes, and their expression levels in multiple cancers were obtained from the TCGA database. The correlation of expression levels of these MMR genes with the expression levels of SMARCA4 was analyzed by the Spearman’s correlation method. In the present study, we also evaluated the expression levels of DNMT1, DNMT2, DNMT3A, and DNMT3B, and Spearman’s correlation was used to evaluate the correlation of the four methyltransferases with SMARCA4 expression.

The STRING website (https://string-db.org/) was used to query protein name “BRG1” and organism (“Homo sapiens”). The following main parameters were set as: minimum required interaction score [“low confidence (0.150)”], meaning of network edges (“evidence”), max number of interactors to show (“no more than 50 interactors” in 1st shell) and active interaction sources (“experiments”). Then, the available experimentally determined SMARCA4-binding proteins were obtained. The “similar gene detection” module of GEPIA2 was used to obtain the top 100 SMARCA4-correlated targeting genes of all TCGA tumor and normal tissues. The “correlation analysis” module of GEPIA2 was used to perform a pairwise gene Pearson’s correlation analysis of SMARCA4 and selected genes. The log2 TPM was applied for the dot plot. The p-value and the correlation coefficient (R) were indicated. The “Gene_Corr” module of TIMER2 was used to generate heatmap data of the selected genes, which contains the partial correlation (cor) and p-value in the purity-adjusted Spearman’s rank correlation test. Venn diagram was used to conduct an intersection analysis to compare the SMARCA4-binding and interacted genes.

Two sets of data were combined to perform Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. The gene lists were uploaded to Database for Annotation, Visualization, and Integrated Discovery (DAVID) with the settings of selected identifier (“OFFICIAL_GENE_SYMBOL”) and species (“Homo sapiens”) and obtained the data of the functional annotation chart. Enriched pathways were visualized with R packages “tidyr” and “ggplot2.” Gene Ontology (GO) enrichment analysis was conducted using R package “clusterProfiler”. The data for biological process (BP), cellular component (CC), and molecular function (MF) were visualized as cnet plots, using the cnetplot function.

GSEA was performed in the high- and the low-expression groups to explore the biological signaling pathway. The top 5 terms of KEGG and HALLMARK analyses were exhibited. KEGG pathways with significant enrichment results were demonstrated on the basis of net enrichment score (NES), gene ratio, and p-value. Gene sets with |NES| >, NOM p < 0.05, and FDR q < 0.25 were considered to be enrichment significant (6).

R language software (version 4.0.3) (https://www.r-project.org/) was used in this analysis. Differences between the two groups and among multiple groups were analyzed using the default Wilcoxon’s test and one-way analysis of variance (ANOVA), respectively. The differences in overall survival between groups were determined by Kaplan-Meier analysis and a log-rank test. The subtypes, clinicopathological features, risk scores, expression of immune checkpoints, and levels of immune infiltration were determined by a Pearson’s correlation test. Results were considered statistically significant when the p < 0.05.

The expression level of SMARCA4 in tumor tissues was higher than the corresponding control tissues, including bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), stomach adenocarcinoma (STAD), thyroid carcinoma (THCA), UCEC (all p < 0.001) and cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC, p < 0.01), as shown in Figure 1A . The expression level of SMARCA4 in tumor tissues was significantly lower than control tissues of kidney renal clear cell carcinoma (KIRC, p < 0.001) and kidney chromophobe (KICH, p < 0.05).

The expression differences of SMARCA4 between the tumor and normal tissues of lymphoid neoplasm diffuse large B-cell lymphoma (DLBC), glioblastoma multiforme (GBM), brain lower grade glioma (LGG), skin cutaneous melanoma (SKCM), testicular germ cell tumors (TGCT), and thymoma (THYM) were also analyzed in GTEx dataset ( Figure 1B , p < 0.001). The results of the CPTAC dataset showed higher expression of SMARCA4 total protein in the primary tissues of breast cancer, ovarian cancer, colon cancer, UCEC and LUAD ( Figure 1C , p < 0.001) than in normal tissues.

We also used the “pathological stage plot” module of GEPIA2 to observe the correlation between SMARCA4 expression and the pathological stages of cancers, including adrenocortical carcinoma (ACC), BLCA, CESC, COAD, KICH, LUAD, pancreatic adenocarcinoma (PAAD), and THCA ( Figure 1D , all p < 0.05).

As shown in Figure 2A , highly expressed SMARCA4 is linked to poor prognosis for cancers including ACC (p = 0.00034), mesothelioma (MESO, p = 0.00017), sarcoma (SARC, p = 0.011), and SKCM (p = 0.037) of TCGA datasets. DFS analysis ( Figure 2B ) shows high SMARCA4 expression is correlated with poor prognosis for the TCGA cases of ACC (p = 0.0023), BRCA (p = 0.034), and uterine carcinosarcoma (UCS, p = 0.017). The above data indicate that SMARCA4 expression is associated with the prognosis of cases with different tumors.

The genetic alteration status of SMARCA4 in different tumor samples of the TCGA cohorts were analyzed. As shown in Figure 3A , the highest alteration frequency of SMARCA4 appears for patients with uterine tumors with “mutation” as the primary type.

The “amplification” type of CNA is the primary type in the ovarian cancer cases, which shows an alteration frequency of ~9% ( Figure 3A ). It is worth noting that all kidney cases (clear cell and nonclear cell carcinoma) with genetic alteration have SMARCA4 mutations. The types, sites, and case number of the SMARCA4 genetic alterations are further presented in Figure 3B . Missense mutation of SMARCA4 is the main type of genetic alteration. The R1192C/H alteration in the helicase_C domain is detected in 16 cases ( Figure 3B ). The 3D structure of SMARCA4 protein is shown in Figure 3C . The potential association between SMARCA4 alteration and the clinical survival prognosis of cases with different types of cancer was analyzed. The data in Figure 3D indicate that UCEC cases with altered SMARCA4 have better prognosis in PFS (p = 0.0387), but not OS (p = 0.158), DFS (p = 0.0762), and disease-specific survival (DSS) (p = 0.147), compared with patients without SMARCA4 alterations. Interestingly, the data of Figure 3E indicate that LUAD cases with altered SMARCA4 show worse prognosis in OS (p = 0.0348), PFS (p = 0.0387), DFS (p = 0.0762), and DSS (p = 0.147), compared with patients without SMARCA4 alterations.

The SMARCA4/BRG1 phosphorylation levels between tumor versus normal tissues of five types of tumors (breast cancer, clear cell RCC, LUAD, ovarian cancer, and UCEC) were analyzed using CPTAC dataset. SMARCA4 phosphorylation sites and the significant differences are summarized in Figure 4A .

The S613 locus of SMARCA4 exhibits a higher phosphorylation level in several tumor tissues compared with normal tissues, including breast cancer, colon cancer and UCEC ( Figure 4B , all p < 0.05), followed by an increased phosphorylation level of the S695 locus for breast cancer (p = 3.65e−06), UCEC (p = 2.34e−09), and LUAD (p = 5.72e−33). Higher phosphorylation level in some tumor tissues were also observed in S699 and S1417 locus.

ImmuneScore and StromalScore were integrated to evaluate the relationship between SMARCA4 expression and immune infiltration across cancers. According to the results, SMARCA4 expression was negatively correlated with the ImmuneScore in ACC, BLCA, BRCA, etc. ( Figure 5A ). Also, SMARCA4 expression positively correlates with the StromalScore in UVM, while negatively in ACC, BLCA, BRCA, COAD, etc. ( Figure 5B ). SMARCA4 expression is negatively correlated with ESTIMATEScore in most cancer types ( Figure 5C ). The top 3 tumors most significantly correlated with expression of SMARCA4 are BRCA, GBM, and PRAD (StromalScore); GBM, SKCM, and SARC (ImmuneScore); and GBM, SARC, and SKCM (ESTIMATEScore) ( Figure 5D ).

Algorithms of TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISEQ, XCELL, MCPCOUNTER, and EPIC were further used to investigate the potential relationship between the infiltration level of different immune cells and SMARCA4 gene expression in diverse cancer types of TCGA. A statistically negative correlation was observed between the immune infiltration of CD8⁺ T cells and SMARCA4 expression in the tumors of ESCA, PAAD, SKCM, and SKCM-metastasis ( Figure 6A ) based on most algorithms. For example, SMARCA4 expression level in PRAD is negatively correlated with the infiltration level of CD8+ T cells ( Figure 6B , cor = −0.152, p = 1.14e−03) based on the EPIC algorithm. A statistically positive correlation of SMARCA4 expression and the estimated infiltration value of cancer-associated fibroblasts is observed for the TCGA tumors of CESC and HNSC-HPV− but noted a negative correlation for TGCT and THYM ( Figure 6C ). For example, SMARCA4 expression level in HNSC-HPV− is correlated with the infiltration level of cancer-associated fibroblasts ( Figure 6D , cor = 0.2, p = 5.46e−05) based on the MCPCOUNTER algorithm.

The correlation between SMARCA4 and immune checkpoint gene expression was analyzed. In LIHC, SMARCA4 expression is positively correlated with expression of CD200, NRP1, LAIR1, TNFSF4, LAG3, ICOS, CD48, CD200R1, HAVCR2, CD276, CD80, PDCD1, LGALS9, ICOSLG, TIGIT, etc. ( Figure 7A ).

Association between SMARCA4 expression and TMB varies markedly among cancer types. SMARCA4 is positively correlated with TMB in BLCA, KICH, LIHC, LUAD, MESO, etc. ( Figure 7B ). On the contrary, SMARCA4 expression is negatively associated with TMB in UVM, THYM, CHOL, etc. SMARCA4 is positively correlated with MSI in UVM, KICH, SARC, STAD, LUSC, LUAD, etc. ( Figure 7C ). In contrast, SMARCA4 expression is negatively related to MSI in CHOL, READ, SKCM, UCS, etc. All these data together indicate that high SMARCA4 expression is widely associated with immunity in cancers.

In order to determine the potential role of SMARCA4 in tumor progression, we evaluated the association of the expression level of SMARCA4 with mutation levels of five MMR genes. The results shown in Figure 8A revealed that SMARCA4 is highly related to MMR genes in 31 cancers, except for READ and UCS. The relationships between SMARCA4 and four DNA methyltransferases were also evaluated. SMARCA4 expression is highly associated with these four DNA methyltransferases in multiple cancers, such as THYM, BLCA, COAD, etc. ( Figure 8B ). These results indicate that SMARCA4 may regulate the tumor progression by mediating repairment of DNA and DNA methylation across cancers.

A total of 50 SMARCA4-binding proteins were obtained which were supported by experimental evidence via STRING tool. Interaction network of these proteins is shown in Figure 9A . The top 100 genes that correlate with SMARCA4 expression were obtained via GEPIA2 tool to combine pan-cancer expression of TCGA. As shown in Figure 9B , SMARCA4 expression level is positively correlated with that of interleukin enhancer-binding factor 3 (IL-F3) (R = 0.68), Krev interaction trapped/cerebral cavernous malformation 1 (KRI1) (R = 0.61), C19orf52 (TIMM29) (R = 0.6), Peter Pan Homolog (PPAN) (R = 0.56), and phenylalanyl-tRNA synthetase alpha (FARSA) (R = 0.56) genes (all p < 0.001). The corresponding heatmap data also show a positive correlation between SMARCA4 and the above five genes in the majority of detailed cancer types ( Figure 9C ). Intersection analysis of the above two groups show three common members, namely, SMARCB1, SMARCD1, and HDAC2 ( Figure 9D ).

We combined the two datasets to perform KEGG and GO enrichment analyses. KEGG data suggest that “SWI/SNF superfamily-type complex” and “ATPase complex” might be involved in the effect of SMARCA4 on tumor pathogenesis, which was shown in Figure 9E . GO enrichment analysis indicate that most of these genes are linked to the pathways or cellular biology of chromatin, such as chromatin remodeling, chromatin DNA binding, nuclear chromatin, and others ( Figure 9F ).

GSEA was performed to identify the functional enrichment of high and low SMARCA4 expression ( Figure 10 ). KEGG enrichment term exhibits that high expression of SMARCA4 is mainly associated with tyrosine kinase signaling pathways, including neurotrophin signaling pathway, ERBB signaling pathway, and GnRH pathway. HALLMARK terms indicated that high expression of SMARCA4 was associated with mitotic spindle, apical junction, and PI-3K/AKT/mTOR signaling pathways.