Antibody-Coated mRNA Nanoparticles Reprogram Tumor Macrophages to Pull Killer T Cells Into Tumors

Many solid tumors resist immunotherapy because their microenvironment is immunosuppressive and poorly infiltrated by cytotoxic T lymphocytes (CTLs, or CD8+ T cells). Tumor-associated macrophages (TAMs) are among the most abundant cells in tumors and often drive this suppression. Systemic cytokine or chemokine mRNA therapies have struggled with poor targeting, limited efficacy, and toxicity. This study asks whether macrophages in the tumor can be targeted directly and converted into a source of T cell–attracting signals.

The team first used human breast cancer data to pick a chemokine. Using the LinkedOmics platform, they found that among CXCL chemokines, CXCL9 correlated most strongly with CD8A expression (coefficient 0.8191). In TCGA data, patients with high CXCL9 expression survived significantly longer than those with low expression. Gene set enrichment analysis linked high CXCL9 to lymphocyte and adaptive immune pathways (P < 0.0001). CXCL9 also correlated with CD8+ T cell activity genes such as granzyme A and interferon-γ. These are correlations in public datasets, not proof of causation. They did, however, motivate restoring CXCL9 in the tumor.

The researchers then engineered an antibody-coated lipid nanoparticle (Ab-LNP). It is coated with an anti-TREM2 antibody and co-loaded with resiquimod, a TLR7/8 agonist meant to shift immunosuppressive TAMs toward a proinflammatory state, and with CXCL9-encoding mRNA. Per the paper's introduction and abstract, TREM2 targeting improved uptake by TAMs relative to 4T1 breast tumor cells in vitro and in vivo. Co-delivery reprogrammed TAMs and increased CXCL9 production. In mice, this reduced TAM immunosuppressive phenotypes and increased CTL infiltration and activity in the tumor. Combining the Ab-LNPs with anti-PD-L1 and anti-CTLA-4 checkpoint inhibitors gave a more CTL-favorable environment. It also increased helper T cells and promoted central memory CTLs in tumor-draining lymph nodes, suggesting durable memory immunity.

The approach is conceptually notable. Instead of depleting TAMs, it uses them as local factories for a T cell–recruiting chemokine, while the TLR agonist removes their suppressive programming. It could in principle make checkpoint blockade work in poorly infiltrated tumors.

Caveats are substantial. The text available for this summary ends early in the Results, so the specific in vivo effect sizes, dosing, tumor-growth and survival data, and safety measurements could not be verified and are described here only as the authors characterize them in the abstract and introduction. The work uses mouse models, notably 4T1 breast cancer, and TREM2 is also expressed on other myeloid cells, so off-target effects and toxicity such as systemic inflammation from TLR agonism need rigorous evaluation. Human translation is far off.