Platform

PLATFORM TECHNOLOGY

Each cancer consists of a diverse mix of cells. Each person’s immune system generates an individualized response against each specific cancer cell type. This results in extreme variabilities of interactions between people’s immune systems and their cancers. Delivering a cancer immunotherapy that stimulates the immune system to kill cancer cells (and only cancer cells) requires a broad and deep understanding of immunology.

To understand the mechanism behind the immunotherapeutic effect of cancer immunotherapies TREOS scientists investigated the relationship between clinical outcome and complete HLA genotype in treated individuals that participated in clinical trials. They demonstrated that an individual’s HLA class I and class II genotype is the main determinant of clinical response. Specifically, personal epitopes (PEPIs) identified with the PEPI Test are genetic biomarkers that predict peptide-specific T cell responses of individual patients. The PEPI biomarker not only predicts the clinical outcome of peptide recipients but also the clinical trial outcome of cancer immunotherapies in a model population of HLA genotyped individuals.

TREOS findings provided a solution to the long standing puzzle of the variability of individual patient’s clinical responses to cancer immunotherapies. Clarifying the mechanism by which highly variable HLA genes activate T cell responses provides clinicians with a novel personal genetic biomarker that help to select likely responders to cancer immunotherapies and perhaps other immune-based therapies.

DIGITAL IMMUNOTHERAPY DEVELOPMENT

At TREOS we use knowledgebase and algorithm embedded in PASCal (Personal Antigen Selection Calculator) tool to develop personal cancer immunotherapies. PASCal supports the development of immunotherapies for single individuals and a population of individuals with a specific genetic background, including specific peptides most effective in ethnic populations.

We created a knowledgebase by building a comprehensive taxonomy of immunological information linking HLA genetics to tumor-specific antigens. Our Peptide Target Knowledgebase currently contains 10^8 true HLA-epitope pairs derived from 1,300 tumor antigens and HLA class I and class II molecules covering the HLA genotype of 26,000 subjects.

Our peptide immunotherapy development algorithm selects from the knowledgebase peptides matching with the patient’s HLAs and tumor type. Another algorithm predicts peptide-induced immune responses and excludes autoimmunity. We continually use the latest published scientific knowledge and proprietary learnings to expand the knowledgebase and further refine the algorithms. Ongoing quality control ensures data integrity and reliability.

IN SILICO TRIALS

Our  in silico trials are conducted with peptide cancer therapeutics in a Model Population of HLA genotyped individuals. The in silico trials predict the Immune Response Rate and Objective Response Rate of clinical trials. At TREOS we validated the in silico trial model on 2,338 subjects participating in 94 vaccine trials. We developed this model to minimize the risk and expedite the clinical development of our peptide cancer immunotherapies.