Document resource
Background EGFR/ALK-driven and STK11/KEAP1-mutant non-small cell lung cancers (NSCLC) display low T-cell infiltration and poor responses to PD-1 blockade. These ‘cold’ tumors rely on GLUT1-mediated aerobic glycolysis, generating glucose competition and lactate accumulation that suppress effector T cells. We hypothesized that inhibiting GLUT1 would relieve metabolic immunosuppression and potentiate dual checkpoint blockade (DCB: anti-PD-1 + anti-CTLA-4 or anti-LAG-3).Methods Syngeneic murine models representing EGFR-mutant, Lkb1-null, and KEAP1-deficient NSCLC were treated with anti-PD-1 alone, DCB, GLUT1 inhibitors (BAY-876, WZB117, or 2-deoxy-D-glucose), or the combination. Tumor growth and survival were recorded. Flow cytometry, multiplex IHC, and spatial transcriptomics quantified metabolic markers, immune subsets, and cytokine signatures. An AI-assisted pathology pipeline generated a composite ‘immune-metabolic’ score (GLUT1 expression × intratumoral CD8 density) in both murine tumors and patient-derived xenografts (PDXs).Results GLUT1 inhibition decreased intratumoral lactate, increased glucose availability, and down-regulated the anti-apoptotic protein c-FLIP, sensitizing tumor cells to TNF-α-mediated T-cell killing. DCB alone modestly delayed tumor growth, whereas adding GLUT1 blockade produced synergistic regressions: >80% inhibition in Lkb1-null tumors and significant survival extension in EGFR models (p < 0.01). Combination therapy tripled intratumoral CD8 + T-cell numbers, reduced myeloid-derived suppressor cells, and promoted M1 macrophage polarization. Spatial analyses revealed glycolysis-low, T-cell-rich niches and de-novo tertiary lymphoid structures after triple therapy; these features were absent with DCB alone. The AI immune-metabolic score correlated strongly with response and outperformed PD-L1 or tumor mutational burden in predicting benefit. In KEAP1-mutant PDXs, WZB117 restored sensitivity to anti-PD-1 therapy, validating translational relevance. No additive systemic toxicity was observed.Conclusions GLUT1-driven glycolysis is a key, targetable barrier to immune infiltration in genetically ‘cold’ NSCLC. Pharmacologic GLUT1 inhibition reprograms the tumor microenvironment, amplifies dual checkpoint blockade, and achieves durable regressions in models historically resistant to immunotherapy. The clinically available components (nivolumab/pembrolizumab + ipilimumab or relatlimab; investigational GLUT1 inhibitors).