Integrative Single-Cell and Spatial Transcriptomics Reveal Organelle Stress-Associated Heterogeneity and Immune Microenvironment Remodeling in Lung Adenocarcinoma.

Background: Lung adenocarcinoma (LUAD) exhibits marked heterogeneity. Organelle stress-adaptive programs that tumor cells develop under hypoxia, nutrient limitation, and proteostasis pressure may drive functional reprogramming of tumor biology, remodel the immune microenvironment, and ultimately influence the benefits of immunotherapy. Therefore, it is necessary to systematically characterize the coordinated changes across organelle stress-related pathways in LUAD, establish a subtyping and prognostic stratification framework, and identify key molecules as well as potential cell-cell communication axes. Methods: Transcriptomic profiles and clinical follow-up data from The Cancer Genome Atlas LUAD (TCGA-LUAD) cohort and public cohorts were collected. Pathway activities were quantified using organelle stress-related gene sets, and core stress programs associated with overall survival (OS) and progression-free survival (PFS) were screened by Cox regression and Kaplan-Meier analyses. Nonnegative matrix factorization (NMF) was used for unsupervised subtyping and stability evaluation. Functional enrichment, genomic features, and immune landscapes were compared between subtypes, and potential benefit from immune checkpoint blockade (ICB) was inferred using tumor immune dysfunction and exclusion (TIDE) and immunophenoscore (IPS). Single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics (ST) were integrated to characterize malignant cell states, spatial niches, and cell-cell communication networks. In LUAD cell lines, the candidate gene was silenced by small interfering RNA (siRNA), and phenotypic assays were performed to validate its effects. Results: The organelle stress activity-based system robustly classified LUAD into two biologically distinct subtypes (the mitochondrial-ribosome biogenesis, MRB, subtype and the lysosomal catabolism, LC, subtype), which showed systematic differences in prognosis and functional programs. The MRB subtype exhibited enhanced stress and metabolic adaptation accompanied by immune exclusion features, whereas the LC subtype showed a relatively immune-active tumor microenvironment. Immunotherapy-related analyses suggested divergent trends in potential ICB benefit between the two subtypes. Multiscale evidence highlighted SLC16A14 (MCT14) as a key node linking stress heterogeneity to malignant progression. At the single-cell level, SLC16A14 was mainly expressed in malignant cells, and communication analyses suggested that CALCR-related signaling may mediate tumor-endothelial interactions and contribute to an immune-excluded microenvironment. In vitro, SLC16A14 silencing suppressed tumor cell proliferation, invasion, and migration, supporting its role as a key molecule connecting stress adaptation and tumor progression. Conclusion: We established an organelle stress program-based subtyping and prognostic framework for LUAD, revealed the coupling between stress adaptation and TME remodeling, and proposed SLC16A14 and its associated communication network as potential intervention targets, providing multiomics evidence for interpreting LUAD heterogeneity and for stratifying immunotherapy and combination strategies.