CD247, a Potential T Cell-Derived Disease Severity and Prognostic Biomarker in Patients With Idiopathic Pulmonary Fibrosis

Abstract

Background: Idiopathic pulmonary fibrosis (IPF) has high mortality worldwide. The CD247 molecule (CD247, as known as T-cell surface glycoprotein CD3 zeta chain) has been reported as a susceptibility locus in systemic sclerosis, but its correlation with IPF remains unclear.

Methods: Datasets were acquired by researching the Gene Expression Omnibus (GEO). CD247 was identified as the hub gene associated with percent predicted diffusion capacity of the lung for carbon monoxide (Dlco% predicted) and prognosis according to Pearson correlation, logistic regression, and survival analysis.

Results: CD247 is significantly downregulated in patients with IPF compared with controls in both blood and lung tissue samples. Moreover, CD247 is significantly positively associated with Dlco% predicted in blood and lung tissue samples. Patients with low-expression CD247 had shorter transplant-free survival (TFS) time and more composite end-point events (CEP, death, or decline in FVC >10% over a 6-month period) compared with patients with high-expression CD247 (blood). Moreover, in the follow-up 1st, 3rd, 6th, and 12th months, low expression of CD247 was still the risk factor of CEP in the GSE93606 dataset (blood). Thirteen genes were found to interact with CD247 according to the protein-protein interaction network, and the 14 genes including CD247 were associated with the functions of T cells and natural killer (NK) cells such as PD-L1 expression and PD-1 checkpoint pathway and NK cell-mediated cytotoxicity. Furthermore, we also found that a low expression of CD247 might be associated with a lower activity of TIL (tumor-infiltrating lymphocytes), checkpoint, and cytolytic activity and a higher activity of macrophages and neutrophils.

Conclusion: These results imply that CD247 may be a potential T cell-derived disease severity and prognostic biomarker for IPF.

Keywords: CD247; biomarker; idiopathic pulmonary fibrosis; immune response; inflammation response; prognosis.

Figure 1

Identification of candidate hub genes. (A) The intersection of genes associated with Dlco% predicted in the GSE38958, GSE132607, and GSE93606 datasets. (B) Heatmap plot of the 33 genes associated with Dlco% predicted in the three datasets. (C) The intersection of genes associated with FVC% predicted in the three datasets. (D) Heatmap plot of the 17 genes associated with FVC% predicted in the three datasets. The numbers in the heatmap represent the correlation coefficients, red represents the positive correlation, green represents the negative correlation, the darker shade of red or green represents the higher correlation level, and the pink rectangles showed the genes were selected as the candidate hub genes.

Figure 2

The expression of CD247 and MYL4 and their value for Dlco% predicted after visiting at 0, 4, 8, and 12 months in the GSE132607 dataset. (A) The change of mean CD247 expression. (B) The comparison of mean Dlco% predicted between high-expression and low-expression CD247. (C) The change of mean MYL4 expression. (D) The comparison of mean Dlco% predicted between high-expression and low-expression MYL4. *p < 0.05.

Figure 3

The correlation between CD247 expression and lung function. Blood: GSE38958 dataset (A), GSE93606 dataset (B), visiting at 0, 4, 8, and 12 months in the GSE132607 dataset (C–F). Lung tissue: GSE47460 dataset (G, H) and GSE32537 dataset (I, J).

Figure 4

The expression of CD247 in both blood and lung tissue. Blood: GSE38958 dataset (A), GSE93606 dataset (B), GSE33566 dataset (C). Lung tissue: GSE47460 dataset (D), GSE32537 dataset (E), GSE110147 dataset (F). ns, not significant.

Figure 5

Results of the univariate logistic regression regarding Dlco15 in the GSE132607 dataset (A) and the univariate Cox regression regarding CEP in the GSE93606 dataset and non-TFS in the GSE27957 and GSE28042 datasets (B). Dlco15, Dlco% predicted decline ≥15% over 12 months; TFS, non-transplant-free survival; CEP, composite end point (death or decline in FVC >10% over six months period).

Figure 6

CD247-related gene enrichment analysis. (A) The available experimentally determined CD247-binding proteins using the STRING tool. (B) According to the MEM database, the co-expression correlations between CD247 and 13 selected genes were analyzed. (C) The corresponding Pearson coefficients of the 13 genes correlated with CD247 were visualized by the heatmap in the blood, lung tissue, and BALF samples. The darker shade of red or blue represents the higher correlation level. (D) The top 10 significant terms in the biological processes (BP), cellular component (CC), and molecular function (MF) analysis for the 14 genes. (E) The top 30 significant terms in the KEGG pathways analysis for the 14 genes.

Figure 7

Significant GO terms and KEGG pathway analysis for the DEGs between patients with low-expression CD247 and patients with high-expression CD247. The most significant GO enrichment and KEGG pathways in the GSE38958 dataset (A, B), GSE132607 dataset (C, D), and GSE93606 datasets (E, F) are displayed.

Figure 8

Comparison of the ssGSEA scores between patients with low-expression CD247 and patients with high-expression CD247 in the GSE38958 dataset (A, B), GSE132607 dataset (C, D), and GSE93606 dataset (E, F). The scores of 19 immune cells (A, C, E) and 15 immune-related functions (B, D, F) are displayed in violin plots. DC, dendritic cell; TIL, tumor-infiltrating lymphocytes; CCR, cytokine–cytokine receptor. P values were shown as: ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001.