COPD-Associated PANoptosis Genes Predict Prognosis and Chemotherapy Sensitivity in Lung Squamous Cell Carcinoma

Summary

This study derives a COPD-associated PANoptosis gene set from public transcriptomic data and applies it to lung squamous cell carcinoma cohorts to develop a prognostic signature retrospectively evaluated. The signature was associated with immune infiltration and computationally predicted drug sensitivity, providing hypothesis-generating candidate biomarkers for future risk-stratification studies.

Abstract

Chronic obstructive pulmonary disease (COPD) is a progressive inflammatory disease that increases the risk of lung squamous cell carcinoma (LUSC). PANoptosis integrates pyroptosis, apoptosis, and necroptosis, but the relationship between COPD-associated PANoptosis genes and LUSC prognosis remains unclear. In this study, we first derived COPD-associated differentially expressed genes from GSE57148 by comparing COPD lung tissues with normal lung tissues and intersected these genes with a curated PANoptosis-associated gene list. The resulting gene set was then applied to TCGA-LUSC, GSE30219, and GSE37745, which were analyzed as LUSC cohorts without confirmed patient-level COPD comorbidity annotation. We evaluated PANoptosis gene expression patterns, performed unsupervised clustering, characterized immune infiltration differences between clusters, and constructed a prognostic signature with external validation. We identified 38 COPD-associated PANoptosis genes with differential expression across tissue types. Two molecular subtypes displayed distinct immune landscapes, immune checkpoint expression, and overall survival. A 12-gene risk model selected by univariate Cox and LASSO Cox regression stratified patients into high- and low-risk groups and showed modest-to-moderate predictive performance in the TCGA-LUSC, GSE30219, and GSE37745 cohorts. High-risk patients also showed higher computationally predicted IC50 values for multiple agents, suggesting lower predicted drug sensitivity rather than experimentally confirmed chemotherapy resistance. These findings indicate that COPD-associated PANoptosis genes are associated with prognosis, immune microenvironment remodeling, and predicted drug sensitivity in LUSC and may provide hypothesis-generating biomarkers for future validation.

Introduction

Lung squamous cell carcinoma (LUSC), a subtype of non-small cell lung cancer (NSCLC), is characterized by the abnormal proliferation of squamous cells within the lung. Despite advances in surgery, platinum-based chemotherapy, immunotherapy, radiotherapy, and selected molecularly guided strategies, the overall cure rate for LUSC remains low, particularly among patients with advanced-stage disease^1,2. Optimizing treatment efficacy remains challenging because of marked tumor heterogeneity and multiple resistance mechanisms^3,4.

....

Protocol

This study used only publicly available, de-identified datasets and did not involve direct human or animal experimentation; therefore, additional ethics committee approval and informed consent were not required.

Data download and processing
RNA-sequencing data and corresponding clinical information for lung squamous cell carcinoma (LUSC) were obtained from The Cancer Genome Atlas (TCGA) database through the Genomic Data Commons data portal under the TCGA-LUSC project. The TCGA-LUSC expression matrix used in this study was based on FPKM values. Gene expression values were transformed and normalized before downstream an....

Results

Expression patterns and functional annotation of PANoptosis-associated genes in LUSC
Transcriptomic data from GSE57148, including 91 normal lung tissues and 98 COPD lung tissues, were analyzed to identify COPD-associated differentially expressed genes. A total of 2,182 COPD-associated genes were identified, and 38 genes overlapped with the curated PANoptosis-associated gene set (Figure 1A and 1B). The chromosomal distribution of these 38 genes is shown i.......

Discussion

COPD and LUSC are clinically and biologically connected through shared inflammatory, smoking-related, and airway-injury-associated pathogenic processes. Although immune checkpoint inhibitors and other systemic therapies have improved treatment options for LUSC, durable benefit remains limited for many patients because of tumor heterogeneity, incomplete biomarker stratification, and treatment resistance^31,32,33.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
AnnotationDbi R package	Bioconductor	N/A	Used for gene annotation and identifier conversion. Source: https://bioconductor.org/packages/AnnotationDbi/
caret R package	CRAN	N/A	Used for reproducible stratified random splitting of the TCGA-LUSC cohort into training and testing subsets. Source: https://cran.r-project.org/package=caret
CIBERSORT algorithm	CIBERSORT resource	N/A	Used for immune-cell deconvolution with the LM22 signature matrix. Access and licensing are subject to the original resource provider. Source: https://cibersortx.stanford.edu/
clusterProfiler R package	Bioconductor	N/A	Used for GO and KEGG enrichment analyses. Source: https://bioconductor.org/packages/clusterProfiler/
ConsensusClusterPlus R package	Bioconductor	N/A	Used for consensus clustering of PANoptosis-associated gene expression patterns. Source: https://bioconductor.org/packages/ConsensusClusterPlus/
e1071 R package	CRAN	N/A	Used as a dependency for CIBERSORT-based deconvolution analysis. Source: https://cran.r-project.org/package=e1071
edgeR R package	Bioconductor	N/A	Used for RNA-seq count processing when count-based normalization was required. Source: https://bioconductor.org/packages/edgeR/
ESTIMATE R package	MD Anderson Cancer Center / SourceForge	N/A	Used to calculate stromal score, immune score, ESTIMATE score, and tumor purity. Source: https://sourceforge.net/projects/estimateproject/
Genomic Data Commons Data Portal	National Cancer Institute	TCGA-LUSC	Source of TCGA-LUSC RNA-sequencing data and clinical annotations. Source: https://portal.gdc.cancer.gov/
Genomics of Drug Sensitivity in Cancer database	GDSC project	N/A	Drug-response reference dataset used through the pRRophetic-compatible reference data. Source: https://www.cancerrxgene.org/
GEOquery R package	Bioconductor	N/A	Used to download and parse GEO Series Matrix files. Source: https://bioconductor.org/packages/GEOquery/
ggplot2 R package	CRAN	N/A	Used for visualization and figure generation. Source: https://cran.r-project.org/package=ggplot2
ggpubr R package	CRAN	N/A	Used for publication-style plotting and group comparisons. Source: https://cran.r-project.org/package=ggpubr
glmnet R package	CRAN	N/A	Used for LASSO Cox regression and prognostic model construction. Source: https://cran.r-project.org/package=glmnet
GSE30219 dataset	NCBI Gene Expression Omnibus	GSE30219	External LUSC validation cohort with expression and survival information. Source: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE30219
GSE37745 dataset	NCBI Gene Expression Omnibus	GSE37745	External LUSC validation cohort with expression and survival information. Source: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE37745
GSE57148 dataset	NCBI Gene Expression Omnibus	GSE57148	COPD versus normal lung transcriptomic dataset used to derive COPD-associated differentially expressed genes. Source: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE57148
limma R package	Bioconductor	N/A	Used for differential-expression analysis. Source: https://bioconductor.org/packages/limma/
org.Hs.eg.db annotation package	Bioconductor	N/A	Used for human gene annotation in enrichment analysis. Source: https://bioconductor.org/packages/org.Hs.eg.db/
preprocessCore R package	Bioconductor	N/A	Used for expression preprocessing and CIBERSORT-compatible workflows. Source: https://bioconductor.org/packages/preprocessCore/
pRRophetic R package	CRAN	N/A	Used for exploratory prediction of drug IC50 values from gene-expression data. Source: https://cran.r-project.org/package=pRRophetic
R software	R Foundation for Statistical Computing	Version 4.1.0	Used for all statistical analyses and figure generation. Source: https://www.r-project.org/
STRINGdb R package	Bioconductor	N/A	Used for protein-protein interaction-related analysis when applicable. Source: https://bioconductor.org/packages/STRINGdb/
SummarizedExperiment R package	Bioconductor	N/A	Used to handle TCGA expression data objects downloaded through TCGAbiolinks. Source: https://bioconductor.org/packages/SummarizedExperiment/
survival R package	CRAN	N/A	Used for Cox regression and Kaplan-Meier survival analysis. Source: https://cran.r-project.org/package=survival
survminer R package	CRAN	N/A	Used to visualize Kaplan-Meier survival curves and risk tables. Source: https://cran.r-project.org/package=survminer
TCGAbiolinks R package	Bioconductor	N/A	Used to download and prepare TCGA-LUSC expression and clinical data. Source: https://bioconductor.org/packages/TCGAbiolinks/
timeROC R package	CRAN	N/A	Used for time-dependent ROC analysis of the prognostic model. Source: https://cran.r-project.org/package=timeROC