Integrative analysis of bulk and single-cell RNA sequencing reveals PHLDA3 as a key regulator of Hypoxic TAMs and prognostic biomarker in HNSC

BACKGROUND: Head and neck squamous cell carcinoma (HNSC) is a highly aggressive malignancy characterized by a complex tumor microenvironment (TME). Despite advances in immunotherapy, the immunosuppressive microenvironment driven by metabolic reprogramming remains a critical factor contributing to treatment failure. The prognostic value of PHLDA3, a target gene of the p53 family, in HNSC and its specific mechanisms in shaping TME heterogeneity remain ill-defined. METHODS: This study integrated large-scale transcriptomic data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases, alongside single-cell RNA sequencing (scRNA-seq) data. A prognostic nomogram incorporating PHLDA3 and clinical characteristics was constructed using multivariate Cox regression analysis. Biological functions were assessed utilizing Weighted Gene Co-expression Network Analysis (WGCNA), Gene Set Variation Analysis (GSVA), and the drug sensitivity prediction algorithm, oncoPredict. At the single-cell resolution, algorithms including Seurat, CellChat, scMetabolism, and Monocle 2 were employed to comprehensively characterize cell type annotation, intercellular communication, metabolic activity scoring, and pseudotime differentiation trajectories. RESULTS: PHLDA3 was significantly upregulated in HNSC tissues and strongly correlated with poor overall survival (OS). The PHLDA3-based nomogram demonstrated superior calibration and discrimination capabilities in both the training and validation cohorts. Genomic analysis revealed that high PHLDA3 expression was associated with a lower Tumor Mutation Burden (TMB) but an increased sensitivity to the PLK1 inhibitor, BI-2536. Single-cell analysis further unveiled that PHLDA3 expression was not ubiquitous; rather, it was specifically enriched in tumor-associated macrophage (TAM) subsets characterized by hypoxic and lipid metabolic signatures (Hypoxic TAMs and Lipid TAMs). These subpopulations exhibited transcriptomic signatures indicative of enhanced glycolysis and orchestrated immunosuppressive communication networks within the TME, predominantly via the SPP1-CD44 and MIF-CD74 signaling axes. Ultimately, through its specific enrichment in these metabolically reprogrammed TAM subsets, PHLDA3 intimately participates in remodeling an immunosuppressive TME, thereby facilitating tumor progression and immune evasion. CONCLUSIONS: TAM subpopulations. By potentially sustaining the survival and function of these specific TAMs under hypoxic conditions, PHLDA3 drives the formation of an immunosuppressive microenvironment, highlighting a promising therapeutic target for combined metabolic-immune strategies in HNSC.