Metabolic resistance¶
Confidently known¶
- human mouse Immune-suppressive metabolism in the TME is real and multi-layered. The established frameworks cover adenosine (CD39/CD73), tryptophan/kynurenine (IDO/TDO), lactate/Warburg, arginine depletion (arginase I), and lipid mediators (PGE2 from COX1/2). Each has preclinical support and some level of clinical translation. However, pharmacologic targeting has been a persistent disappointment: IDO inhibitors (epacadostat) failed in combination with pembrolizumab (ECHO-301) despite compelling biology, and adenosine-axis combinations have produced modest-at-best clinical effects so far.
Human-study evidence¶
No human-study citations in this section.
Contradictions / surprises¶
- human IDO clinical translation failed despite strong preclinical rationale. ECHO-301 (epacadostat + pembrolizumab in melanoma) did not improve PFS/OS over pembrolizumab monotherapy. The TME metabolic target class is biologically compelling but clinically harder than single-agent ICI.
Human-study evidence¶
No human-study citations in this section.
Suspected but unconfirmed¶
- mouse Hypoalbuminemia causally drives ICI resistance via macrophage arginine biosynthesis impairment (LLC mouse model)1. TAM depletion or dietary arginine supplementation rescues anti-PD-1 efficacy. Would recast a routine prognostic lab as a reversible driver. Single preclinical model; no patient-level intervention data.
- mouse Lactate → AARS1 → PD-L1 K280 lactylation is also a metabolic input (see tumor-intrinsic)2. The paradoxical finding that exogenous lactate enhances anti-PD-L1 efficacy in preclinical models needs reconciliation.
- mouse HILPDA-driven lipogenesis → PD-L1 palmitoylation connects lipid metabolism directly to PD-L1 stability3. See tumor-intrinsic.
Human-study evidence¶
No human-study citations in this section.
Emerging¶
- mouse M2 TAM-PGE2 → TIGIT axis in MSS CRC (see t-cell-exhaustion)4 — a concrete combinatorial target set (COX2i, PGE2R antagonist, TIGIT blockade) in an indication where PD-1 has consistently failed.
Human-study evidence¶
No human-study citations in this section.
Theme¶
The recent literature in this period places increasing emphasis on metabolic inputs operating on multiple different substrates at once: CD8 killing capacity (ferroptosis via AGPAT3), PD-L1 protein stability (palmitoylation, lactylation), macrophage polarization (arginine), and TAM-driven TIGIT induction. These are not a single "metabolic resistance" mechanism but a heterogeneous set of plausible druggable inputs.
Practical takeaways¶
- Hypoalbuminemia remains an informative prognostic marker; whether arginine supplementation should become part of ICI practice is not yet evidence-based.
- Do not expect the next-generation metabolic combo (CD39/CD73, IDO2, arginase inhibitors) to be a uniform benefit; learn from the IDO experience that preclinical rationale alone does not guarantee clinical benefit.