The Oncogenic Role and Immune Infiltration for CARM1 Identified by Pancancer Analysis

Abstract

Chromatin-modifying enzymes, especially protein arginine methyltransferases (PRMTs), have been identified as candidate targets for cancer. Cellular or animal-based evidence has suggested an association between coactivator-linked arginine methyltransferase 1 (CARM1) and cancer progression. However, the relationship between CARM1 and patient prognosis and immune infiltration in pancancer patients is unknown. On the basis of the GEO and TCGA databases, we first investigated the possible oncogenic functions of CARM1 in thirty-three tumor types. CARM1 expression was elevated in many types of tumors. In addition, there was a significant association between CARM1 expression and the survival rate of tumor patients. Uterine corpus endometrial carcinoma (UCES) samples had the highest CARM1 mutation frequency of all cancer types. In head and neck squamous cell carcinoma (HNSC) and lung squamous cell carcinoma (LUSC), CARM1 expression was associated with the level of CD8+ T cell infiltration, and cancer-associated fibroblast infiltration was also observed in other tumors including kidney renal papillary cell carcinoma (KIRC) and prostate adenocarcinoma (PRAD). CARM1 was involved in immune modulation and played an important role in the tumor microenvironment (TME). Furthermore, activities associated with RNA transport and its metabolism were included in the possible mechanisms of CARM1. Herein, our first pancancer research explores the oncogenic role of CARM1 in various tumors. CARM1 is associated with immune infiltrates and can be employed as a predictive biomarker in pancancer.

Figure 1

The expression of CARM1 in normal as well as cancer tissues of a human in the HPA database. (a) The expression of CARM1 in healthy tissues (human). (b) The expression of CARM1 protein in healthy tissues (human). (c–f) Characteristic IHC images of CARM1 expression in tissues, such as testis normal tissue, cerebral cortex normal tissue, colorectal cancer, and breast cancer tissues.

Figure 2

The mRNA levels of CARM1 in human cancers. (a) Tissues from TCGA and GTEx used for evaluating the mRNA expression of CARM1 between healthy and tumor tissues. (b) The link between the expression of CARM1 mRNA and various pathological phases in patients with various cancers from TCGA. (c) The CARM1 expression in various molecular subtypes of cancers via TISIDB (^∗∗∗P value, ^∗∗P value, ^∗P value, and P value indicate ≤0.001, ≤0.01, ≤0.05, and ≤0.1, accordingly).

Figure 3

Survival analysis was employed to compare the elevated and lower expression of CARM1 in various kinds of cancer in the TCGA and GEO datasets. (a, j) The link between the expression of CARM1 and prognosis in patients (DSS and OS) of various cancers in the TCGA database (^∗∗∗P < 0.001, ^∗∗P < 0.01, and ^∗P < 0.05). (b–i) OS curves having statistical significance in TCGA for eight cancer types (KIRP, LUAD, ACC, LGG, BLCA, MESO, GBMLGG, and SKCM). (k–r) Survival curves in eight cohorts GSE4412, GSE9893, GSE1456, GSE7378, GSE31210, GSE26712, GSE3141, and GSE1923) with significance.

Figure 4

CARM1 mutation features in various TCGA tumors. cBioPortal tool was used to evaluate CARM1 mutation features for the TCGA tumors. The frequency of mutations is shown along with the mutation type (a) and mutation site (b). The link between CARM1 copy number and expression of mRNA indicated in the dot plot (c) and correlation plot (d) by cBioPortal (e). Using the cBioPortal method, we determined the considerable association between CARM1 mutation status and progression-free survival in various tumors.

Figure 5

Correlation investigation of CARM1 expression and immune infiltration of fibroblasts associated with cancer. (a) The possible association between CARM1 expression and the infiltration level of fibroblasts (cancer-linked) investigated using various algorithms across all cancer types in the TCGA. (b) Significant correlations of CARM1 expression with the infiltration level of fibroblasts (cancer-linked) in KIRC, THYM, BRCALumA, HNSC, KIRP,. PRAD, DLBC, and LUSC by using a single algorithm.

Figure 6

CARM1-associated gene enrichment analysis. (a) The STRING database used to obtain CARM1-binding proteins that were evaluated experimentally. (b) The GEPIA2 approach used to find the top 100 CARM1-related genes in the TCGA data, and the expression correlation between CARM1 and chosen target genes including RAVER1, WIZ, GATAD2A, BRD4, MRPL4, and CHERP was examined. (c) In the detailed cancer types, the equivalent heatmap data are displayed. (d) GO/KEGG pathway analysis conducted according to the CARM1-binding and interacted genes.