m7G-related gene NUDT4 as a novel biomarker promoting cancer cell proliferation in lung adenocarcinoma

Abstract

Background: Lung cancer is the leading cause of mortality in cancer patients. N7-methylguanosine (m7G) modification as a translational regulation pattern has been reported to participate in multiple types of cancer progression, but little is known in lung cancer. This study attempts to explore the role of m7G-related proteins in genetic and epigenetic variations in lung adenocarcinoma, and its relationship with clinical prognosis, immune infiltration, and immunotherapy.

Methods: Sequencing data were obtained from the Genomic Data Commons (GDC) Data Portal and Gene Expression Omnibus (GEO) databases. Consensus clustering was utilized to distinguish m7G clusters, and responses to immunotherapy were also evaluated. Moreover, univariate and multivariate Cox and Least absolute shrinkage and selection operator LASSO Cox regression analyses were used to screen independent prognostic factors and generated risk scores for constructing a survival prediction model. Multiple cell types such as epithelial cells and immune cells were identified to verify the bulk RNA results. Short hairpin RNA (shRNA) Tet-on plasmids, Clustered Regularly Interspaced Short Palindromic Repeats CRISPR/Cas9 for knockout plasmids, and nucleoside diphosphate linked to moiety X-type motif 4 (NUDT4) overexpression plasmids were constructed to inhibit or promote tumor cell NUDT4 expression, then RT-qPCR, Cell Counting Kit-8 CCK8 proliferation assay, and Transwell assay were used to observe tumor cell biological functions.

Results: Fifteen m7G-related genes were highly expressed in tumor samples, and 12 genes were associated with poor prognosis. m7G cluster-B had lower immune infiltration level, worse survival, and samples that predicted poor responses to immunotherapy. The multivariate Cox model showed that NUDT4 and WDR4 (WD repeat domain 4) were independent risk factors. Single-cell m7G gene set variation analysis (GSVA) scores also had a negative correlation tendency with immune infiltration level and T-cell Programmed Death-1 PD-1 expression, but the statistics were not significant. Knocking down and knocking out the NUDT4 expression significantly inhibited cell proliferation capability in A549 and H1299 cells. In contrast, overexpressing NUDT4 promoted tumor cell proliferation. However, there was no difference in migration capability in the knockdown, knockout, or overexpression groups.

Conclusions: Our study revealed that m7G modification-related proteins are closely related to the tumor microenvironment, immune cell infiltration, responses to immunotherapy, and patients' prognosis in lung adenocarcinoma and could be useful biomarkers for the identification of patients who could benefit from immunotherapy. The m7G modification protein NUDT4 may be a novel biomarker in promoting the progression of lung cancer.

Keywords: N7-methylguanosine modification; NUDT4; immune infiltration; immunotherapy; lung adenocarcinoma; prognosis.

Figure 1

Genetic alterations in N7-methylguanosine (m7G) modification-related genes. (A) m7G modification-related genes’ chromosome locations. (B) m7G-related gene copy number amplification. (C) Top-rating mutated m7G-related genes. (D) Single-nucleotide polymorphisms in all variant types. (E) Variant allele frequencies of m7G-related genes. (F) Single-nucleotide variants of m7G-related genes. (G) Variant classification of m7G-related genes. (H) Survival analysis between mutant and wild-type patients. CNV: Copy number variation; SNP: Single-nucleotide polymorphism; DEL: Gene deletion; WT: Wild type; SNV: Single nucleotide variant; HR: Hazard ratio.

Figure 2

N7-methylguanosine (m7G)-related gene expression and survival analysis. (A) Boxplot of m7G-related gene mRNA expression between tumor and normal samples. (B) Venn plots for common existing genes in the three groups. (C) Correlation network between m7G-related genes. (D) Survival plot of m7G-related genes. (*P <0.05, **P < 0.01, ***P < 0.001, ns: Not statistically significant).

Figure 3

Unsupervised consensus clustering analysis. (A) Heatmap showed that the samples were distinctly divided into two clusters. (B) Boxplot of N7-methylguanosine (m7G)-related gene mRNA expression between the two clusters. (C) Survival plot for the two clusters. (D) Infiltration level of immune cells in different clusters. (E) Heatmap for KEGG pathway enrichment in the two clusters. (F) Immunophenoscore for predicting the responsiveness to CTLA-4 checkpoint inhibitors and PD-1 checkpoint inhibitors. CTLA-4: Cytotoxic T-lymphocyte associated protein 4; PD-1: Programmed cell death 1. (*P <0.05, **P < 0.01, ***P < 0.001, ns: Not statistically significant).

Figure 4

Cox regression analysis for N7-methylguanosine (m7G)-related genes and clinical characteristics. (A) Univariate Cox regression analysis results of m7G-related genes. (B) Coefficient plot for each factor coefficient. (C) The partial likelihood deviance plot showed partial likelihood deviance of each model. (D) Forest plot showed multivariate Cox regression analysis results of the nine factors. (E) ROC curve showed the survival prediction model performance. (F) The nomogram for survival prediction model. (G) Survival analysis between high-risk and low-risk groups. (H) Boxplot showed Cox risk scores between the two m7G clusters. ROC: Receiver operating characteristic curve; AUC: Area under curve. (*P <0.05, **P < 0.01, ***P < 0.001)

Figure 5

Single-cell analysis and cell type identification for 11 lung adenocarcinoma patient samples. (A) Seven main cell types of all 11 samples were integrated in one Seurat project and visualized in t-SNE plot. (B) The top 10 marker genes of different cell types were visualized in the heatmap. (C) Cell type proportions in different patient samples. (D) Seven main types of cells in different patient samples were visualized in the t-SNE plot. t-SNE: t-Distributed Stochastic Neighbor Embedding.

Figure 6

The relationship between N7-methylguanosine (m7G) GSVA scores and tumor immune infiltration cells. (A) The relationship between m7G mean GSVA scores and tumor immune infiltration levels. (B) m7G mean GSVA scores in different cell types and different samples. (C) The expression level of Cox independent risk factors in different cell types. (D) PD-1 expression of T cells in different samples. (E) The relationship between m7G mean GSVA scores and T cell mean PD-1 expression. (F) PD-L1 expression of epithelial cells in different samples. (G) The relationship between m7G mean GSVA scores and epithelial cell PD-L1 mean expression. GSVA: Gene set variation analysis; PD-1: Programmed cell death 1; PD-L1: Programmed cell death 1 ligand 1.

Figure 7

NUDT4 expression in bulk RNA level and protein level. (A) NUDT4 expression in lung adenocarcinoma and squamous carcinoma in TCGA datasets. (B) NUDT4 expression in lung adenocarcinoma in GEO datasets. (C) The protein level of NUDT4 in lung adenocarcinoma chips using immunochemistry. (D) The protein level of NUDT4 in normal lung tissue chips using immunochemistry. LUAD: Lung adenocarcinoma; LUSC: Lung squamous cell carcinoma; ns: Not statistically significant.

Figure 8

NUDT4 expression in single-cell RNA level. (A) Seven main cell types of tumor and normal samples were integrated in one Seurat project and visualized in t-SNE plot. (B) The top 10 marker genes of seven cell types. (C) Only epithelial cells were selected for NUDT4 analysis. (D) NUDT4 expression in different cell types. (E) NUDT4 expression in different samples. (F) NUDT4 expression in paired tumor and normal derived epithelial cells. (**P< 0.01, ****P < 0.0001, ns: Not statistically significant).

Figure 9

NUDT4 function validation in lung adenocarcinoma cell lines. (A) Sequence map of tet-on system shRNA plasmids. (B) NUDT4 knocking down efficiency in cell lines. (C) CCK8 assay for detecting cell proliferation (n = 3 duplicates). (D) Transwell assay for detecting A549 cell migration (n = 3 duplicates). (E) Transwell assay for H1299 cell migration capability detection (n = 3 duplicates). (F) Knockout efficiency of the NUDT4 in H1299 cell line. (G) CCK8 assay for detecting H1299 cell proliferation (n = 3 duplicates). (H) Transwell assay for detecting H1299 cell migration (n = 3 duplicates). (I) Knockout efficiency of the NUDT4 in A549 cell line. (J) CCK8 assay for detecting A549 cell proliferation (n = 3 duplicates). (K) Transwell assay for detecting A549 cell migration (n = 3 duplicates). (**P < 0.01, ***P < 0.001, ****P < 0.0001, ns: Not statistically significant).

Figure 10

Overexpression of NUDT4 in lung adenocarcinoma cell lines. (A) Overexpression level in A549 cell line. (B) CCK8 assay for detecting cell proliferation (n = 3 duplicates). (C) NUDT4 overexpression level in H1299 cell line. (D) CCK8 assay for observing H1299 cell proliferation (n = 3 duplicates). (E) Transwell assay for A549 cell capability of the migration (n = 3 duplicates). (F) Transwell assay for detecting H1299 cell migration (n = 3 duplicates). (*P < 0.05, ***P < 0.001, ****P < 0.0001, ns: Not statistically significant).