Tissue-Resident NK Cells Engineered for Solid Tumors: Preclinical Promise Meets Limits of Mouse Models

Stanford-led work published in Science Translational Medicine demonstrates that engineered tissue-resident natural killer (trNK) cells infiltrate and suppress multiple solid tumor types in mice far better than circulating NK counterparts, with additive effects when paired with tumor-targeting antibodies. This builds directly on the MedicalXpress coverage yet reveals what was under-emphasized: the study remains strictly preclinical (n=6-12 mice per arm across models), with no human safety or efficacy data. Prior observational work in human tumor microenvironments (e.g., a 2022 Nature Reviews Immunology synthesis of 14 patient cohorts) showed trNK subsets can be immunosuppressive in some tissues, underscoring the microenvironmental plasticity Sunwoo’s team exploits via their “Goldilocks” differentiation protocol. A 2024 Cancer Cell paper on off-the-shelf CAR-NK products (phase 1, n=42) already flagged manufacturing scalability advantages, yet reported cytokine-release variability absent in the current mouse data. The original coverage missed how this approach could bypass autologous T-cell manufacturing bottlenecks while inheriting NK cells’ MHC-unrestricted killing—yet it overstates immediate clinical translation given absent large-animal toxicology and potential exhaustion in chronic human tumors. Conflicts of interest are unreported in the press summary; Stanford-affiliated authors hold related IP filings. Overall evidence grade: strong mechanistic reproducibility in murine models, weak for human prediction.