Background Leukocyte immunoglobulin-like receptor B2 (LILRB2; ILT4) is an immunoinhibitory protein expressed on the surface of myeloid cells that has been increasingly recognized as a therapeutic target of interest in immuno-oncology (IO). Upon binding its ligands, MHC I molecules (e.g. HLA-G/HLA-A), LILRB2 inhibits myeloid cell activation and promotes an M2-like (anti-inflammatory) state. LILRB2 was the first target prioritized from a macrophage discovery effort leading to the development of JTX-8064, a humanized monoclonal antibody that specifically binds to and antagonizes LILRB2. JTX-8064 has been shown to induce an M1-like (pro-inflammatory; anti-tumor) functional state in macrophages. Rodents do not express LILRB proteins limiting their usefulness as a model for preclinical study of JTX-8064. To overcome this limitation, we conducted an ex vivo human tumor histoculture study to assess the pharmacodynamic effects of LILRB2 antagonism. Protein and/or gene expression analysis of matched tumor samples enabled the discovery of predictive biomarkers associated with the induction of specific pharmacodynamic signatures in ex vivo-cultured human tumors in response to JTX-8064. Finally, tumor types were identified that had a high prevalence of these predictive biomarkers suggesting they may be priority indications for JTX-8064 therapy.

Methods More than 100 fresh treatment-naïve human tumor samples obtained post-surgery from kidney, lung, and head and neck cancer were treated with JTX-8064 or isotype control antibody for 24 hrs in the histoculture system. RNA was isolated from tumors prior to any treatment as well as from JTX-8064 and isotype control treated samples. Gene expression was analyzed using the NanoString nCounter® and qPCR assays. Additional IHC analyses were performed on baseline untreated tumor samples.

Results JTX-8064 was shown to induce pharmacodynamic responses to treatment significantly above isotype control indicative of macrophage polarization, IFNg-signaling, and T cell inflammation. To identify predictive biomarkers of pharmacodynamic response to JTX-8064, matched untreated samples were characterized by gene expression analysis and by IHC (CD8, CD163, and HLA-G proteins). Numerous LILRB2 pathway-related molecules (e.g. HLA-A, HLA-B, CD163, LILRB2) and gene signatures were found to be statistically significantly higher in the untreated kidney, head and neck, and lung cancer samples of matched pharmacodynamic responders compared to non-responders. Further bioinformatics analysis revealed additional cancer subtypes where these biomarkers are enriched.

Conclusions These data will inform indication selection and combination strategies for JTX-8064 to maximize potential therapeutic benefit for patients with solid tumor malignancies.