Molecular subtypes of lung adenocarcinoma patients for prognosis and therapeutic response prediction with machine learning on 13 programmed cell death patterns

Abstract

Background: Lung adenocarcinoma (LUAD) seriously threatens people's health worldwide. Programmed cell death (PCD) plays a critical role in regulating LUAD growth and metastasis as well as in therapeutic response. However, currently, there is a lack of integrative analysis of PCD-related signatures of LUAD for accurate prediction of prognosis and therapeutic response.

Methods: The bulk transcriptome and clinical information of LUAD were obtained from TCGA and GEO databases. A total of 1382 genes involved in regulating 13 various PCD patterns (apoptosis, necroptosis, pyroptosis, ferroptosis, cuproptosis, netotic cell death, entotic cell death, lysosome-dependent cell death, parthanatos, autophagy-dependent cell death, oxeiptosis, alkaliptosis and disulfidptosis) were included in the study. Weighted gene co-expression network analysis (WGCNA) and differential expression analysis were performed to identify PCD-associated differential expression genes (DEGs). An unsupervised consensus clustering algorithm was used to explore the potential subtypes of LUAD based on the expression profiles of PCD-associated DEGs. Univariate Cox regression analysis, Least Absolute Shrinkage and Selection Operator (LASSO) regression, Random Forest (RF) analysis and stepwise multivariate Cox analysis were performed to construct a prognostic gene signature. The "oncoPredict" algorithm was utilized for drug-sensitive analysis. GSVA and GSEA were utilized to perform function enrichment analysis. MCPcounter, quanTIseq, Xcell and ssGSEA algorithms were used for tumor immune microenvironment analysis. A nomogram incorporating PCDI and clinicopathological characteristics was established to predict the prognosis of LUAD patients.

Results: Forty PCD-associated DEGs related to LUAD were obtained by WGCNA analysis and differential expression analysis, followed by unsupervised clustering to identify two LUAD molecular subtypes. A programmed cell death index (PCDI) with a five-gene signature was established by machine learning algorithms. LUAD patients were then divided into a high PCDI group and a low PCDI group using the median PCDI as a cutoff. Survival and therapeutic analysis revealed that the high PCDI group had a poor prognosis and was more sensitive to targeted drugs but less sensitive to immunotherapy compared to the low PCDI group. Further enrichment analysis showed that B cell-related pathways were significantly downregulated in the high PCDI group. Accordingly, the decreased tumor immune cell infiltration and the lower tumor tertiary lymphoid structure (TLS) scores were also found in the high PCDI group. Finally, a nomogram with reliable predictive performance PCDI was constructed by incorporating PCDI and clinicopathological characteristics, and a user-friendly online website was established for clinical reference ( https://nomogramiv.shinyapps.io/NomogramPCDI/ ).

Conclusion: We performed the first comprehensive analysis of the clinical relevance of genes regulating 13 PCD patterns in LUAD and identified two LUAD molecular subtypes with distinct PCD-related gene signature which indicated differential prognosis and treatment sensitivity. Our study provided a new index to predict the efficacy of therapeutic interventions and the prognosis of LUAD patients for guiding personalized treatments.

Keywords: Gene signature; Lung adenocarcinoma; Machine learning; Prognosis; Programmed cell death.

Fig. 1

The workflow for comprehensive analysis of thirteen programmed cell death (PCD) patterns in LUAD patients

Fig. 2

Screening of PCD-related DEGs associated with LUAD. a Clustering dendrogram of TCGA-LUAD. b Heatmap of correlation between WGCNA modules and phenotypes. c Volcano plot of PCD-related DEGs between LUAD and normal tissues. d A Venn diagram of the WGCNA module genes and PCD-related DEGs. e An oncoplot showing the condition of variant classification and variant type (left) of 40 PCD-related genes as well as the top 10 PCD-related genes with mutation frequency (right) in the TCGA-LUAD dataset

Fig. 3

Identification of potential molecular subtypes of LUAD based on PCD-related DEGs. a Consensus clustering matrix when k = 2. b Representative cumulative distribution function (CDF) curves. c The score of consensus clustering. d A principal component analysis (PCA) plot visualizing the distribution of two clusters. e Kaplan–Meier survival curve analysis between the two clusters

Fig. 4

Construction of PCD-related gene signature and clinical relevance analysis. a Coefficient path diagram of Lasso regression analysis. b The top 15 genes with variable importance in RF analysis. c A Venn diagram of the genes screened from Lasso regression and RF analysis. d The immunohistochemistry of the model genes in LUAD and normal tissues. e Kaplan–Meier survival curve analysis between the high and low PCDI groups in the TCGA-LUAD cohort. f Distribution of PCDI according to the survival status in the TCGA-LUAD cohort. g A principal component analysis (PCA) plot visualizing the distribution of high and low PCDI groups in the TCGA-LUAD cohort. h–l Violin plots of the relationship between PCDI and survival status (h), clinical stage (i), T stage (j), N stage (k), and M stage (l). m The differentially enriched pathways between high and low PCDI groups

Fig. 5

Therapeutic response analysis. a The correlation between drug sensitivity in the GDSC database and the mRNA expression of model genes. b The correlation between drug sensitivity in the CTRP database and the mRNA expression of model genes. c–f Boxplots of IC50 of Gefitinib (c), Erlotinib (d), Crizotinib (e) and Osimertinib (f) in high and low PCDI groups. g The immunotherapy response of the high and low PCDI groups through TIDE analysis

Fig. 6

Comparison of tumor microenvironment between high and low PCDI groups. a Heatmap of the tumor microenvironment scores calculated by MCPcounter, quanTIseq, and XCell algorithms. b Comparison of immune scores between high and low PCDI groups. c Relationships between infiltration levels of 28 immune cell types and PCDI. d Comparison of the immune infiltration level of 28 immune cell types between high and low PCDI groups

Fig. 7

Correlation analysis between PCDI and tumor tertiary lymphoid structure (TLS). a Pathway enrichment analysis between high and low PCDI groups by GSEA. b Kaplan–Meier survival curve analysis between high and low TLS groups. c Kaplan–Meier survival curves analysis based on B cell infiltration level and Tfh cell infiltration level. d The comparison of TLS scores between high and low PCDI groups

Fig. 8

Construction and assessment of the nomogram survival model. a Univariate analysis of PCDI and the clinicopathologic characteristics. b Multivariate analysis of PCDI and the clinicopathologic characteristics. c A nomogram integrating PCDI and clinicopathologic characteristics to predict the prognosis of LUAD patients. d The calibration curve of the nomogram. e Decision curve analysis (DCA) of the nomogram. f Receiver operator characteristic (ROC) analysis of the nomogram