Spatial Multi-Omics and Machine Learning Reveal Lactate-Driven Targets in LUAD
Overview
This study integrates single-cell transcriptomics, spatial transcriptomics, spatial metabolomics, and machine learning to characterize lactate-driven metabolic heterogeneity in lung adenocarcinoma (LUAD). High-lactate tumors show increased epithelial and fibroblast populations, angiogenic endothelial signatures, and immune suppression, highlighting lactate-centered pathways as therapeutic targets.
Background
Lung adenocarcinoma (LUAD) is a common and heterogeneous subtype of non-small cell lung cancer with complex metabolic and cellular microenvironments. Lactate, a key metabolite produced by tumor glycolysis, acts as a signaling molecule that promotes immune suppression, angiogenesis, and tumor progression. Despite its importance, the spatial distribution and cell-type-specific effects of lactate within LUAD tumors remain poorly understood. Integrating multi-omics spatial profiling with machine learning can elucidate how lactate shapes the tumor microenvironment and identify prognostic and therapeutic targets.
Data Highlights
Cell Type
High-Lactate Tumors
Low-Lactate Tumors
Epithelial Cells
Increased abundance
Lower abundance
Fibroblasts
Increased abundance
Lower abundance
T/NK Cells
Decreased abundance
Enriched
Monocytes/Macrophages
Decreased abundance
Enriched
Endothelial Cells
Minimal lactate accumulation but angiogenic/stress signatures
Less angiogenic/stress signatures
Key Findings
High-lactate LUAD tumors exhibit increased epithelial and fibroblast cell populations, while T/NK cells and monocytes/macrophages are enriched in low-lactate tumors.
Spatial metabolomics reveals cell-type-restricted lactate and pyruvate distributions, with endothelial cells showing minimal lactate accumulation.
Endothelial subclusters in high-lactate tumors display angiogenic and stress-response gene signatures associated with poor prognosis.
Machine learning models consistently identify endothelial and fibroblast transcriptional programs as key determinants of high-lactate states and adverse clinical outcomes.
Lactate reshapes the tumor microenvironment by promoting angiogenesis and immune suppression, contributing to prognostic stratification in LUAD.
Clinical Implications
These findings suggest that targeting lactate-driven pathways, particularly those involving endothelial and fibroblast cell programs, may improve therapeutic outcomes in LUAD. Spatial multi-omics combined with machine learning can stratify patients by metabolic tumor profiles, guiding precision interventions to overcome immune suppression and angiogenesis associated with high lactate levels.
Conclusion
Integrative spatial multi-omics and machine learning reveal that lactate accumulation drives metabolic and cellular reprogramming in LUAD, promoting angiogenesis and immune evasion. Targeting lactate-centered pathways offers promising avenues for therapeutic development and prognostic assessment.
References
Original Study 2024 -- Spatial Multi-Omics Enhanced by Machine Learning Reveals Lactate-Driven Therapeutic Targets and Reprogramming of the Tumor Microenvironment in Cancer
