Many newly developed anti-cancer treatments rely on identifying cancer unique signatures that can be used to harness the immune system against the tumor. The internal state of all cells is constantly represented on their surfaces by unique molecules called Human Leukocyte Antigen (HLA) bound peptides. Consequently, many efforts focus on identifying cancer‑unique HLA-bound peptides. However, currently available methods only allow considering the peptides’ primary sequence, overlooking other potential chemical modifications known as post-translational modifications (PTMs). PTMs greatly affect the binding of these peptides to HLA molecules and thus the subsequent response by interacting immune cells. There is a need for a method to identify cancer-unique HLA-bound peptides that takes into account the peptides’ PTMs. The group of Prof. Yifat Merbl successfully developed a computational platform; where an improved pipeline allows for simultaneous in‑parallel detection of multiple PTMs, expanding the landscape of HLA bound-peptides by up to 20% by identifying chemically modified peptides.
The internal state of cells is represented on their surfaces by Human Leukocyte Antigen (HLA), molecules that bind and display samples from the cells’ waste disposal system on the cell surface. These waste samples are short sections of different proteins from inside the cell, peptides, which are unique and specific to that cell. This is used by the immune system to check for any abnormalities in the cells. Consequently, HLA-bound peptides serve as the immune system’s window into the cellular proteome and, as such, are key components of immune system defenses. HLA-bound peptides that are unique to a patient’s tumor and are not presented in healthy tissue, can be used to design various types of cancer therapies (e.g. cancer vaccines or CAR-T therapy). Thus, identifying peptides that are antigenic, common for several cancer types, and bound by several HLA haplotypes holds tremendous potential for the scientific community and pharmaceutical industries. Current efforts for antigen discovery focus on identifying peptides with tumor-specific genomic mutations, through profiling tumor immunopeptidome or with transcriptomic data, both combined with HLA binding predictions using computational tools. As these methods rely only on the encoded amino acid sequence of proteins, they overlook the realm of protein modifications (PTMs) on HLA-bound peptides. Post-translational processing and modifications such as phosphorylations, citrullinations, or glycosylations, have been shown to play a large part in modulating both peptide presentation and subsequent T-cell recognition. However, identifying modified peptides remains a challenge because the combinatorial analysis of PTMs increases the theoretical number of peptide possibilities tremendously. Therefore, there is a need for a method that predicts cancer-specific modified HLA-bound peptides, which takes into account the unique PTM on these peptides.
The Merbl lab developed a computational pipeline to detect endogenous peptides and their PTMs. The improved pipeline allows for the combinatorial detection of multiple PTMs at the same search, a task that is not feasible by currently available tools.
The team developed PROMISE (PROtein Modification Integrated Search Engine), which enables efficient search against combinatorial reference data with multiple modifications. Compared to an existing search tool (MaxQuant), PROMISE decreased search time by 100-fold. Looking for insight into PTM-driven antigenicity, the team used PROMISE to analyze 29 different PTM combinations simultaneously in immunopeptidomics data obtained by patient tumors tissues, healthy adjacent tissues, cancer cell lines, and tumor-infiltrating lymphocytes (TILs) (Figure 1). The search expanded peptide identification by 22% and enriched the modified peptide fraction threefold, yielding hundreds of novel modified antigens derived from cancer-associated proteins and cancer-specific immunopeptide signatures. Importantly, the team further confirmed their results by in vitro and ex vivo assays, showing the identified cancer-associated modified peptides can bind to MHC, are immunogenic, and can stimulate CD8 T cell-mediated killing of tumor cells.
- Identifying novel cancer neoantigens for developing new anti-cancer treatments.
- Identifies endogenous modified peptides at biological levels, which make it usable for clinical settings.
- Additional human pathologies such as infections and autoimmunity
The team fully developed PROMISE, a computational platform for identification of endogenous modified peptides. By applying their platform tested on public HLA–profiling available data, new targets were identified, confirmed experimentally, and proved to induce anti-tumor immunological response in a murine model system.