IDH1-mutant astrocytoma produces D-2-hydroxyglutarate (2-HG) at millimolar concentrations in tumor tissue, an oncometabolite essentially absent from healthy brain parenchyma. This extraordinary signal-to-noise differential (100-10,000-fold versus healthy tissue) makes IDH1-mutant glioma uniquely tractable for living therapeutic approaches. Here I describe a design concept for an engineered bacterium that (1) senses 2-HG through a D2HGDH-coupled NADH biosensing circuit, (2) delivers a cytotoxic and immune-recruiting payload selectively within the tumor microenvironment, and (3) self-limits through redundant biocontainment mechanisms keyed to 2-HG availability. The proposed design addresses the fundamental unsolved problem in IDH1-mutant glioma: diffuse infiltration along white matter tracts that surgery cannot follow. I review the current research landscape, covering bacterial tumor colonization, synthetic genetic circuits, AI-accelerated protein design, and clinical delivery infrastructure, and identify that every major subsystem required has been independently validated, though no program integrating them for IDH1-mutant glioma exists. I also present computational results including: codon-optimized D2HGDH expression constructs for E. coli Nissle 1917; de novo protein structures designed by RFDiffusion3 for a 2-HG binding domain (ESMFold pLDDT scores 80-90); a reaction-diffusion simulation confirming extracellular 2-HG gradients exceed the D2HGDH activation threshold at all clinically relevant distances including under vorasidenib suppression; an agent-based model demonstrating that engineered bacteria colonize the tumor periphery first, targeting infiltrating cells along white matter tracts before the tumor core; and a genetic circuit ODE model confirming kill circuit activation at 15-23 minutes with correct kill switch behavior upon 2-HG depletion. A key finding from the agent-based model: computational modeling predicts bacteria will follow the steepest 2-HG gradient toward the tumor-parenchyma interface, targeting infiltrating cells that surgery cannot reach — a prediction that requires biological validation and is subject to real white matter tract geometry effects modeled in Section 8.1. This paper functions simultaneously as a design proposal and as a defensive publication establishing prior art for the described concepts in the public domain.
Tzvika Besor (Thu,) studied this question.