Immune checkpoints are inhibitory proteins of the immune system that are crucial for controlling the duration and magnitude of the immune responses. These proteins are the key to minimizing collateral tissue damage during an immune response. Novel members of this class may be utilized as targets for cancer immunotherapy and as a basis for immunomodulatory therapeutics for autoimmune diseases.
Recently, the blockade of the immune checkpoints CTLA4 and PD-1 emerged as "game changers" in cancer therapy, leading to durable clinical responses in patients with advanced melanoma, lung and kidney cancers, thus forming the foundation for a new era of cancer immunotherapy. However, many patients and many types of cancers do not respond to these treatments, indicating that there are additional immune evasion mechanisms, mediated by unknown members of this protein family.
As this protein family shares very low sequence similarity, traditional sequence similarity analysis approaches are not effective. Therefore, our discovery platform was designed to specifically address this challenge by developing specialized algorithms designed to identify a set of the unique characteristics of this protein family at genomic and protein levels as well as unique gene expression profiles.
The platform predicted multiple novel proteins as possible additional members of the B7/CD28 family. In addition, the output of the platform also included a non-novel group of candidates such as TIGIT and VISTA, before they were published in the scientific literature, providing a strong validation of the predictive power of the model. Two of these novel proteins, CGEN-15001T and CGEN-15022, are the focus of our collaboration with Bayer HealthCare.
Immunomodulatory proteins are proteins capable of modifying or regulating one or more immune functions. Immune checkpoints, including inhibitory receptors and ligands, are one type of immunomodulators.
The immunomodulatory platform models two distinct biological phenomena related to the role of the immune system that are conceptually different from those employed in the discovery of Compugen’s B7/CD28-like candidates.
The first biological phenomenon that was modeled exploits the interplay between the immune system and intruding pathogens. As a result of such interplay, some immune proteins tend to evolve differently. Compugen devised an evolutionary model to detect such immune proteins, and the predictive algorithm was incorporated into Compugen’s discovery infrastructure and integrated with its existing tools for the discovery of target candidates for cancer immunotherapy.
The modeling of the second biological phenomenon relies heavily on the Company’s MED Platform, which was employed to predict proteins that play a role in the biology of tumor-associated macrophages (TAMs).
TAMs are an important component of the tumor microenvironment and play a major role in creating the immunosuppressive environment that enables tumor development. Proteins having the potential to modulate the tumor microenvironment may serve as potential targets for cancer immunotherapy.
Antibody-drug conjugates (ADCs) are a new class of drugs designed as a targeted therapy for the treatment of cancer. ADCs are composed of an antibody linked to a highly potent cytotoxic drug. The antibody binds to a cell surface protein on the target cell and is thereby internalized, carrying its toxic cargo into the cell. ADC is therefore a guided molecular missile aimed at specifically destroying cancer cells.
An ideal ADC target is a surface protein that is highly abundant on tumor cells but has low to no expression on normal healthy cells. Another key attribute of the target protein is internalization into the cell upon binding of the antibody.
Our ADC target discovery platform predicts membrane proteins having the potential to internalize. In addition, an algorithm was devised to employ the MED platform for the detection of membrane proteins which are highly expressed on cancer cells and have low expression on healthy cells. The fidelity of the ADC platform was demonstrated by predicting correctly several ADC targets currently in clinical trials. Five novel target candidates were selected as potential new ADC targets which are currently under experimental evaluation.