Immune- and Stemness-Related Genes Revealed by Comprehensive Analysis and Validation for Cancer Immunity and Prognosis and Its Nomogram in Lung Adenocarcinoma

Abstract

Objective: Lung adenocarcinoma (LUAD) is a familiar lung cancer with a very poor prognosis. This study investigated the immune- and stemness-related genes to develop model related with cancer immunity and prognosis in LUAD.

Method: The Cancer Genome Atlas (TCGA) was utilized for obtaining original transcriptome data and clinical information. Differential expression, prognostic value, and correlation with clinic parameter of mRNA stemness index (mRNAsi) were conducted in LUAD. Significant mRNAsi-related module and hub genes were screened using weighted gene coexpression network analysis (WGCNA). Meanwhile, immune-related differential genes (IRGs) were screened in LUAD. Stem cell index and immune-related differential genes (SC-IRGs) were screened and further developed to construct prognosis-related model and nomogram. Comprehensive analysis of hub genes and subgroups, involving enrichment in the subgroup [gene set enrichment analysis (GSEA)], gene mutation, genetic correlation, gene expression, immune, tumor mutation burden (TMB), and drug sensitivity, used bioinformatics and reverse transcription polymerase chain reaction (RT-PCR) for verification.

Results: Through difference analysis, mRNAsi of LUAD group was markedly higher than that of normal group. Clinical parameters (age, gender, and T staging) were ascertained to be highly relevant to mRNAsi. MEturquoise and MEblue were found to be the most significant modules (including positive and negative correlations) related to mRNAsi via WGCNA. The functions and pathways of the two mRNAsi-related modules were mainly enriched in tumorigenesis, development, and metastasis. Combining stem cell index-related differential genes and immune-related differential genes, 30 prognosis-related SC-IRGs were screened via Cox regression analysis. Then, 16 prognosis-related SC-IRGs were screened to construct a LASSO regression model at last. In addition, the model was successfully validated by using TCGA-LUAD and GSE68465, whereas c-index and the calibration curves were utilized to demonstrate the clinical value of our nomogram. Following the validation of the model, GSEA, immune cell correlation, TMB, clinical relevance, etc., have found significant difference in high- and low-risk groups, and 16-gene expression of the SC-IRG model also was tested by RT-PCR. ADRB2, ANGPTL4, BDNF, CBLC, CX3CR1, and IL3RA were found markedly different expression between the tumor and normal group.

Conclusion: The SC-IRG model and the prognostic nomogram could accurately predict LUAD survival. Our study used mRNAsi combined with immunity that may lay a foundation for the future research studies in LUAD.

Keywords: RT-PCR; cancer stem cell; immune; lung adenocarcinoma; muti-omics analysis; nomogram; stem cell index.

Figure 1

Flow diagram of the study.

Figure 2

(A) Kaplan–Meier displays no significant difference between the high- and low-mRNAsi groups. (B–D) The correlation of global mRNAsi profiles with LUAD clinical subtypes: (B) age, (C) sex, and (D) T staging. (E) Different analysis of the mRNAsi level between normal and LUAD tissues. (F) The heat map of DEGs in LUAD.

Figure 3

WGCNA of LUAD and enrichment analysis of the significant modules. (A) Correlation of the gene module with mRNAsi and EREG-mRNAsi. (B, C) Scatter graph of the blue module (module membership vs. gene significance). Scatter graph of the turquoise module (module membership vs. gene significance). (D, E). GO enrichment analysis of the blue and turquoise modules. (F, G) KEGG pathway enrichment analysis of the blue and turquoise modules.

Figure 4

(A) The heat map of immune-related DEGs in LUAD. (B) Volcano map of immune-related DEGs in LUAD. Green, downregulated genes; red, upregulated genes. (C) Venn diagram of the intersection genes related to both mRNAsi and immunity. (D) Univariate COX regression analysis of prognosis-related stem cell and immune-related differential genes (SCIRGs) in LUAD. (E, F) Kaplan–Meier curves show a considerable difference between the high- and the low-risk groups. (G) Heat maps of the hub genes’ expression pattern, where the red to green means changes from high to low expression in TCGA. (H) Distribution of multi-genes signature risk score in TCGA datasets. (I) The survival status and interval of TCGA-LUAD patients.

Figure 5

(A, B) Univariate Cox regression analyses of overall survival in TCGA and GEO dataset. (C, D) Multivariate Cox regression analyses of overall survival in TCGA and GEO dataset. (E, F) Comparing the AUCs of the risk scores with other clinical parameters in TCGA and GEO dataset.

Figure 6

(A) Nomogram was assembled by clinical parameters and risk signature for predicting survival of patients with LUAD. (B) One-year nomogram calibration curves. (C) Two-year nomogram calibration curves. (D) Three-year nomogram calibration curves.

Figure 7

(A) Bar plot presents the distribution of 22 kinds of TICs in LUAD tumor samples. Column names represent sample ID. (B) Bar plot presents the difference of TICs between the high- and low-risk groups. (C) Bar plot presents the difference of immune-related function between the high- and low-risk groups. (D, E) Association between tumor immune microenvironment (TIME) and risk score. (D) ImmuneScore; (E) StromalScore. (F, I) TIMER: Immune correlation analysis of SCIRGs in the model based on immune infiltration, (F) ADRB2, (G) CX3CR1, (H) GPER (GPER1), and (I) IL3RA. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 8

(A) The difference of TMB between the high- and low-risk groups. (B) Association between TMB and risk score. (C) The proportion of clinical characteristics of every sample in relative risk group was presented in heat map. (D) Proportion of patients in different stages of high- and low-risk groups. (E) Proportion of patients in different immune sub-typing of high- and low-risk groups. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 9

The expression levels of SCIRGs in the model between Beas-2B, HCC827 cell lines, HCC827 cancer stem cell, and results of the RT-PCR to determine gene expression. *p < 0.05, **p < 0.01, ***p < 0.001; ns, no significance.