Single-cell RNA sequencing reveals palmitoylation-driven cellular heterogeneity and prognostic biomarkers in lung adenocarcinoma

Abstract

Background

Lung adenocarcinoma (LUAD) is marked by significant variation within tumor cells and continues to be a major global cause of cancer deaths. Palmitoylation is a dynamic lipid-based modification that occurs after protein synthesis and influences the behavior and lifespan of various cancer-related proteins. However, its role in shaping cellular complexity and predicting outcomes in LUAD patients is not yet fully clarified.

Methods

We examined single-cell RNA sequencing datasets from LUAD samples to identify distinct malignant cell groups. Palmitoylation-related gene activity was estimated using GSVA and ssGSEA techniques. To further define cellular characteristics, we applied copy number variation mapping, pseudotime progression modeling, transcription factor regulatory scoring, and cell–cell interaction analyses. A 12-gene risk model was developed using marker genes from the cluster (C1) with the most prominent palmitoylation pattern. This model was trained on The Cancer Genome Atlas (TCGA) dataset and confirmed using separate GEO datasets. To evaluate tumor immune context, we analyzed immune cell presence and tumor mutational burden across different risk levels. Laboratory experiments involving both upregulation and silencing of aspartate beta-hydroxylase (ASPH) in LUAD cell cultures were conducted to validate its biological significance.

Results

We identified six tumor cell subsets (C0 to C5), with cluster C1 showing peak palmitoylation levels, distinct genomic alterations, and stronger communication with stromal and immune cells. The 12-gene model effectively categorized LUAD patients into high- and low-risk profiles, showing marked survival differences (p < 0.001) and strong performance in time-dependent ROC analysis. Patients in the high-risk group had increased tumor mutational burden and a more immunosuppressive tumor environment. Laboratory findings revealed that raising ASPH expression promoted cell growth, motility, and epithelial–mesenchymal transition. In contrast, reducing ASPH levels triggered cell death and decreased invasiveness.

Conclusions

Our single-cell analysis focused on palmitoylation reveals new dimensions of tumor diversity in LUAD and establishes a validated 12-gene risk signature. Functional studies highlight ASPH as a promising candidate for therapeutic targeting. These results deepen our understanding of palmitoylation-associated pathways and present a foundation for both outcome prediction and precision-based treatment strategies in LUAD.