ITGB2 related to immune cell infiltration as a potential therapeutic target of inflammatory bowel disease using bioinformatics and functional research

Abstract

Inflammatory bowel disease (IBD) is a chronic systemic inflammatory condition regarded as a major risk factor for colitis-associated cancer. However, the underlying mechanisms of IBD remain unclear. First, five GSE data sets available in GEO were used to perform 'batch correction' and Robust Rank Aggregation (RRA) to identify differentially expressed genes (DEGs). Candidate molecules were identified using CytoHubba, and their diagnostic effectiveness was predicted. The CIBERSORT algorithm evaluated the immune cell infiltration in the intestinal epithelial tissues of patients with IBD and controls. Immune cell infiltration in the IBD and control groups was determined using the least absolute shrinkage selection operator algorithm and Cox regression analysis. Finally, a total of 51 DEGs were screened, and nine hub genes were identified using CytoHubba and Cytoscape. GSE87466 and GSE193677 were used as extra data set to validate the expression of the nine hub genes. CD4-naïve T cells, gamma-delta T cells, M1 macrophages and resting dendritic cells (DCs) are the main immune cell infiltrates in patients with IBD. Signal transducer and activator of transcription 1, CCR5 and integrin subunit beta 2 (ITGB2) were significantly upregulated in the IBD mouse model, and suppression of ITGB2 expression alleviated IBD inflammation in mice. Additionally, the expression of ITGB2 was upregulated in IBD-associated colorectal cancer (CRC). The silence of ITGB2 suppressed cell proliferation and tumour growth in vitro and in vivo. ITGB2 resting DCs may provide a therapeutic strategy for IBD, and ITGB2 may be a potential diagnostic marker for IBD-associated CRC.

Keywords: STAT1;ITGB2; bioinformatics analysis; immune cell infiltration; inflammatory bowel disease.

FIGURE 1

Flow chart of the study process for this research. Five GSE data sets were used for bioinformatics analysis; RRA and Batch were used to screen differential genes; Then, GO and KEGG were used to identify the potential function of these differential genes; PPI was to identify the hub genes, and hub genes verification were performed by in vitro and in vivo experiments. Furthermore, we applied CIBERSORT algorithm to evaluate the distribution of infiltration of immune cells in IBD and healthy control tissues. GO, Gene Ontology; PPI, protein–protein interaction; RRA, Robust Rank Aggregation.

FIGURE 2

Identification of DEGs. (A) 78 differential DEGs were visualized in a heatmap determined by ‘Batch.’ (B) Venn diagram was used for DEGs intersection screened from the two methods. DEGs, differentially expressed genes.

FIGURE 3

Functional analysis. (A)The results of KEGG as shown by cneplot. (B) The results of GO as shown in cneplot. GO, Gene Ontology.

FIGURE 4

Construction of PPI network for differentially expressed genes. (A) Visualization of PPI network using Cytoscape software. Nodes in red represent upregulated in IBD, nodes in green indicate no significantly difference between IBD and control, and nodes in orange represent genes downregulated in IBD. (B) Nine hub genes were screened through ten algorithms. (C, D) The expression of nine hub genes shown in a heatmap and volcano map. DEGs, differentially expressed genes; IBD, inflammatory bowel disease; PPI, protein–protein interaction.

FIGURE 5

Validation of hub genes (A) The expression of nine hub genes in in GSE207022. (B) GSE87466 was used for validating the expression of nine hub genes. (C, D) GSE193677 was used to analysis the expression of nine hub genes. (E) ROC analysis was used to evaluate the diagnostic effectiveness of hub genes for IBD in validation GSE87466. (F) Combined AUC of STAT1 and ITGB2 in validation GSE87466. AUC, area under the curve; IBD, inflammatory bowel disease; ROC, receiver operating characteristic.

FIGURE 6

Immune cell infiltration in IBD and control tissues. (A) The 22 types of immune cells shown in a histogram. (B) Wilcoxon test was performed to identify the different immune cells in IBD and control. IBD, inflammatory bowel disease.

FIGURE 7

LASSO regression (A) and cvfit (B) were used to analyse the immune cell infiltration in IBD and control. IBD, inflammatory bowel disease; LASSO, least absolute shrinkage selection operator.

FIGURE 8

The expression of STAT1, CCR5 and ITGB2 in a mouse model of IBD. (A) HE assay was used to analyse the pathological changes in intestinal tissues in each group. (B) ELISA was used to evaluate the inflammatory cytokines in serum. (C, D) The protein level was detected by western blotting. ***p < 0.001, data represented as Mean ± SD. IBD, inflammatory bowel disease.

FIGURE 9

The effect of ITGB2 silence on DSS induced IBD mice. (A) qPCR was used to detect the expression level of ITGB2. (B, C) Western blotting was used to detect the protein level of ITGB2. (D)The protein level of ITGB2 was gradually increased over time under LPS treatment. (E, F) The protein level of ITGB2 in Caco2 cells. (G) The pathological changes in intestinal tissues in each group. (H) Inflammatory cytokines level was measured in each group. (I, J) Interference by ITGB2 significantly decreased the ITGB2 protein level in DSS‐induced IBD mice. **p < 0.01, ***p < 0.001, data represented as Mean ± SD. DSS, Dextran sulphate sodium salt; IBD, inflammatory bowel disease.

FIGURE 10

The silence ITGB2 could suppress cell growth in vitro and in vivo. (A) qPCR was used to detect the expression of ITGB2. (B, C)CCK8 assay was used to detect the cell viability. (D, E) Clone formation assay was used to detect the cell colony formation ability. (F) Pictures of xenograft tumours were shown in the ITGB2 silence group and control group (n = 4). (G) The growth curves of each group of xenograft tumours were displayed. (H) The xenograft tumour weight was measured and analysed. (I) HE assay was used to analysis the morphological changes in tumour. (J) immunofluorescence assay was used to detect the expression of ki67. **p < 0.01, ***p < 0.001, data represented as Mean ± SD.