Immune checkpoint molecules (ICMs) make up the immune suppressive network (ISN) that blocks spontaneous or therapy-induced anti-tumoral immune responses. A decade of research into checkpoint inhibitors (CPIs) has identified only few predictive biomarkers for single-agent CPI therapy. Those discovered show mediocre predictive capability. A considerable fraction of predicted non-responders demonstrate response and vice versa. It is unlikely that current biomarker discovery tools based on single antigen immunohistochemistry or whole-tumor genomics/transcriptomics will uncover predictive biomarkers for therapy with CPI combinations. Brain tumors, specifically glioblastoma, are almost uniformly lethal. Preclinical studies have shown that more than one agent is needed to relieve the strong immunosuppression in these tumors. No rational approach to select the right combination of agents is currently available.

We have developed several unique methodologies that delineate the ISN formed between all immune and tumor cell subsets within individual human brain tumors. Our current ISN models incorporate cell-level data, frequency of all intratumoral immune subsets (up to 13 immune subsets), and gene-expression data with each subset's deep-sequenced transcriptome. We produce transcriptomes from all intratumoral immune subsets (as few as 100 cells) and tumor cells that are sorted using elaborate multicolor flow cytometric panels. These transcriptomes reveal each subset's full immune state: expression of all ICMs, all inter-cellular functional and communication molecules (e.g. cytokines, chemokines, receptors, ligands), and its active pathways.

The collected data is integrated using systems-biology tools to generate models of the ISN which delineate the interactions that each cell subset has with all other immune or tumor cell subsets within a tumor. Individual ISN data can be evaluated in respect to each patient's response to CPI treatment to determine which ISN patterns correlate with response, or lack of response, to CPI treatment. Emerging results may retrospectively explain the failure of some CPI clinical trials conducted on glioblastoma, and reveal new targets.

Studying the relationship between ISN patterns and clinical responses to checkpoint inhibition may enable identification of much needed predictive biomarkers. A systematic evaluation of the suppressive components found to be frequently active in many patients' ISNs may guide rational decisions in planning of clinical trials with improved chances to succeed. This methodology may, in the future, guide physicians in selecting reagents to treat an individual patient's tumor, bringing about the age of personalized medicine to the field of immuno-oncology.

Citation Format: Ilan Volovitz, Gil Diamant, Marina Roitman, Nati Shapira, Roni Hagai, Barak Bensimhon, Merav Lustgarten, Rachel Grossman, Zvi Ram. Identification of predictive biomarkers for immunotherapy of brain and solid tumors by studying inter-cellular immune networks [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2272.