Protein Disulfide-Isomerase A3 Is a Robust Prognostic Biomarker for Cancers and Predicts the Immunotherapy Response Effectively

Abstract

Background: Protein disulfide isomerase A3 (PDIA3) is a member of the protein disulfide isomerase (PDI) family that participates in protein folding through its protein disulfide isomerase function. It has been reported to regulate the progression of several cancers, but its function in cancer immunotherapy is unknown.

Methods: The RNA-seq data of cancer and normal tissues were downloaded from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) databases. The Cbioportal dataset was used to explore the genomic alteration information of PDIA3 in pan-cancer. Human Protein Atlas (HPA) and ComPPI websites were employed to mine the protein information of PDIA3, and western blot assay was performed to monitor the upregulated PDIA3 expression in clinical GBM samples. The univariate Cox regression and the Kaplan-Meier method were utilized to appraise the prognostic role of PDIA3 in pan-cancer. Gene Set Enrichment Analysis (GSEA) was applied to search the associated cancer hallmarks with PDIA3 expression. TIMER2.0 was the main platform to investigate the immune cell infiltrations related to PDIA3 in pan-cancer. The associations between PDIA3 and immunotherapy biomarkers were performed by Spearman correlation analysis. The immunoblot was used to quantify the PDIA3 expression levels, and the proliferative and invasive ability of glioma cells was determined by colony formation and transwell assays.

Findings: PDIA3 is overexpressed in most cancer types and exhibits prognosis predictive ability in various cancers, and it is especially expressed in the malignant cells and monocytes/macrophages. In addition, PDIA3 is significantly correlated with immune-activated hallmarks, cancer immune cell infiltrations, and immunoregulators, and the most interesting finding is that PDIA3 could significantly predict anti-PDL1 therapy response. Besides, specific inhibitors that correlated with PDIA3 expression in different cancer types were also screened by using Connectivity Map (CMap). Finally, knockdown of PDIA3 significantly weakened the proliferative and invasive ability of glioma cells.

Interpretation: The results revealed that PDIA3 acts as a robust tumor biomarker. Its function in protein disulfide linkage regulation could influence protein synthesis, degradation, and secretion, and then shapes the tumor microenvironment, which might be further applied to develop novel anticancer inhibitors.

Keywords: CMap.; immunotherapy response; pan-cancer; prognostic biomarker; protein disulfide-isomerase A3 (PDIA3).

Figure 1

Basic information of PDIA3. (A) The expression level of PDIA3 between tumor and normal tissues in each cancer based on the integrated data from TCGA and GTEx datasets. (B) The expression level of PDIA3 between GBM and normal brain tissues. (C) Western blot protein detection of the PDIA3 expression levels in paired GBM and adjacent normal tissues. (D) PDIA3 alteration frequency analysis in pan-cancer study according to the cBioPortal database. (E) The PDIA3 expression levels and prognosis of DSS between PDIA3-deletion and diploid DLBC patients. (F) The immunofluorescence images of PDIA3 protein, nucleus, endoplasmic reticulum (ER), microtubules and the merged images in A-431 and U251 cell lines. (G) The protein-protein interaction (PPI) network presents the proteins interacting with PDIA3. The labelled asterisk indicated the statistical p value (ns p > 0.05, *p < 0.05, **p < 0.01 and ***p < 0.001). ns, nonsense.

Figure 2

(A) Summary of PDIA3 expression of 33 cell types in 79 single cell datasets. (B) Scatter plot showed the distributions of 10 different cell types of the GSE120575 SKCM dataset. (C) Scatter plot showed the PDIA3 expression levels of cells in the GSE120575 dataset. (D) Scatter plot showed the distributions of 5 different cell types of the GSE102130 glioma dataset. (E) Scatter plot showed the PDIA3 expression levels of cells in the GSE120575 dataset.

Figure 3

(A) Summary of the correlation between expression of PDIA3 with overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI) and progression-free interval (PFI) based on the univariate Cox regression and Kaplan-Meier models. Red indicates that PDIA3 is a risk factor affecting the prognosis of cancer patients, and green represents a protective factor. Only p values < 0.05 are shown. (B) The forest plot exhibited the prognostic role of PDIA3 in cancers by univariate Cox regression method. The cancer type in red represents the PDIA3 acts as a risky factor with statistical significance. (C–F) Kaplan-Meier overall survival curves of PDIA3 in GBM (C), LGG (D), KICH (E) and UVM (F).

Figure 4

The hallmarks gene set enrichment analysis of PDIA3 in pan-cancer. The size of circle represents the FDR value of enrich term in each cancer, and the color indicates the normalized enrichment score (NES) of each term.

Figure 5

The correlations of PDIA3 expression and the infiltration levels of CD4+ T cells, CAF, progenitor, Endo, Eos, HSC, Tfh, γδT, NKT, regulatory T cells (Tregs), B cells, neutrophils, monocytes, macrophages, dendritic cells, NK cells, Mast cells and CD8+ T cells in cancers. Positive correlation in red and negative correlation in blue.

Figure 6

(A) The spearman correlation heatmap shows the correlations between the PDIA3 expressions and the 47 types of immune regulators in pan-cancer. Red represents positive correlation and blue represent negative correlation. (B) Correlations between PDIA3 expression and tumor mutation burden in pan-cancer. (C) Correlations between PDIA3 expression and microsatellite instability in pan-cancer. (D) Kaplan-Meier curves for low- and high-PDIA3 patient groups in IMvigor210 cohort (anti-PD-L1, urological), and the fraction of urological tumors patients with response to anti-PD-1 therapy in low- and high-PDIA3 subgroups of IMvigor210 cohort. (E) Kaplan-Meier curves for low- and high-PDIA3 patient groups in GSE91061(anti-PD-L1, melanoma), and the fraction of melanoma patients with response to anti-PD-1 therapy in low- and high-PDIA3 subgroups of GSE91061. The labelled asterisk indicated the statistical p value (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 7

Heatmap representing enriched score (positive in blue, negative in red) of each drug from the CMap database for each cancer. Components or drugs are ordered by decreasing number of enriched cancers from right to left.