DNA Methylation-Mediated Overexpression of CXCL1 in Helicobacter pylori-Induced Gastric Cancer: In Silico- and In Vitro-Based Identification of a Potential Biomarker for Carcinogenesis

Abstract

Given the high global prevalence and mortality associated with gastric cancer, and its known causal link with Helicobacter pylori infection, it is important to have a biomarker to identify malignant transformation at early stages. Previously, we, and others, have reported that H. pylori-induced epigenetic changes could mediate carcinogenic transformation of the gastric cells. Also, CXCL1 secreted by gastric cancer cells was reported as a key diagnostic and prognostic biomarker for the pathogenic progression of gastric cancer. In this study, for the first time, we aimed to investigate the role of H. pylori-induced DNA methylation-based epigenetic regulation of CXCL1. In silico analysis of publicly available datasets and in vitro experiments were performed. Our results showed that CXCL1 is highly expressed in both gastric cancer tissues and gastric cancer cells infected with H. pylori. Further, we showed and confirmed that H. pylori-mediated overexpression of CXCL1 is due to hypomethylation of its promoter region. Since epigenetic events such as DNA methylation happen early in the sequence; H. pylori-induced CXCL1 hypomethylation could likely be detected at an early stage of gastric cancer development. Epigenetic modifications, such as CXCL1 hypomethylation, are reversible and could potentially be a therapeutic target using demethylation drugs.

Keywords: CXCL1; DNA methylation; Helicobacter pylori; biomarker; epigenetics; gastric cancer.

Figure 1

Differentially expressed genes (DEGs) detected from gastric cancer and H. pylori-infected cells compared to healthy tissues or control cells. (A) Filtering out DEGs by comparing three datasets, the cancer genome atlas (TCGA) for human gastric tissues and two datasets of H. pylori-infected AGS cells from Gene Expression Omnibus (GEO); (B) Heatmap derived from GTEx showing expression of six candidate genes in the normal human stomach; (C) Heatmap showing the expression patterns of the six candidate genes in normal versus primary gastric adenocarcinoma (TPM: Transcripts per million); (D) Relative mRNA expression of the six candidate genes in H. pylori-infected compared versus the expression data of the same gene from uninfected gastric epithelial cells. The black bars represent mRNA expressions from H. pylori-infected AGS cells compared to the untreated cells from GSE108305, and the white bars represent H. pylori-infected AGS cells compared to the untreated cells from GSE70394. All the gene expressions reported are statistically significant in comparison to the healthy controls or uninfected cells (p < 0.001).

Figure 2

CXCL1 protein–protein interactions and functional pathway enrichment analysis. (A) Interactions of CXCL1 with protein molecules transcribed by the other 5 candidate genes; (B) Functional pathways enrichment output obtained when protein molecules transcribed by the six candidate genes and other closely interacting proteins were used as input; (C) Co-expression analysis of CXCL1 with CXCR1 (Spearman correlation coefficient r² = 0.41), CXCR2 (r² = 0.42), IL6 (r² = 0.38) and IL1B (r² = 0.49) in human gastric cancer tissue samples (n = 375). All the values are statistically significant (p < 0.001).

Figure 3

Detailed analysis of CXCL1 expression in TCGA stomach adenocarcinoma. (A) Expression in normal primary tumor samples; (B) Gender-wise differences in expression; (C) Expression in normal compared to individual cancer stages; (D) Expression based on normal versus tumor grade; (E) Expression based on metastasis in the lymph nodes; (F) Expression in gastric cancer tissue with or without H. pylori infection. All the gene expressions reported are statistically significant (** p < 0.001). (TPM: Transcripts per million; Hp: H. pylori).

Figure 4

Detailed analysis of CXCL1 promoter methylation in TCGA STAD samples. (A) Location of CpG island in the promoter region of the CXCL1 gene. (B) CXCL1 promoter methylation based on sample types; (C) CXCL1 promoter methylation based on patient’s gender; (D) CXCL1 promoter methylation based on individual cancer stages; (E) CXCL1 promoter methylation based on tumor grade; (F) CXCL1 promoter methylation based on nodal metastasis. All the gene expressions reported are statistically significant (** p < 0.001).

Figure 5

In vitro CXCL1 gene expression, promoter methylation, protein expression, and protein secretion in H. pylori-infected, decitabine (DAC)-treated, or TNFα-treated gastric cells. AGS and MKN45 cells were left untreated (Control) or treated with DAC (0.5 and 1.0 μM), H. pylori (MOI 50:1 and 100:1), and TNFα (10 ng/mL). (A) CXCL1 relative gene expression (fold change) (B) Percentage of CXCL1 promoter methylation levels; (C) CXCL1 protein expression; (D) Human CXCL1 protein secretion in the cell culture supernatant. Three independent experiments were performed and a t-test was used to compare treated samples versus control samples (* p < 0.01; NS: not significant).

Figure 6

Tumor microenvironment in correlation with CXCL1 expression and overall patient survival in gastric cancer patients. (A) Purity of infiltrating cells; (B) T cell CD8+ infiltration; (C) T cell CD4+ infiltration; (D) Regulatory T cells infiltration; (E) NK cell infiltration; (F) Macrophages infiltration; (G) Monocytes infiltration; (H) Kaplan–Meier plot for overall patient probability at risk for over 120 months. (p < 0.001).