Instead of relying on traditional, often biased, human-led discussions to select drug targets, Scape Bio leverages Large Language Models. This AI-driven approach systematically scores thousands of potential target-indication combinations to identify the most promising opportunities with the highest unmet need and biophysical advantage for mini-proteins.
A key reason GPCRs are difficult targets for antibodies is the antigen generation process. GPCRs sit within a lipid membrane; removing them to create an injectable antigen often alters their conformational shape. Antibodies developed against these distorted versions are then frequently non-functional when encountering the receptor in its natural cellular environment.
Instead of pursuing novel biological targets, Scape Bio's initial strategy focuses on de-risking its platform. They target GPCRs already clinically validated by existing small-molecule drugs but that suffer from off-target toxicity or selectivity problems. This allows them to prove their mini-protein modality by creating a superior version of a known therapeutic mechanism.
Mini-proteins are framed as a superior drug modality that merges the key strengths of traditional therapies. They possess the high selectivity characteristic of biologics like antibodies, while also having the stability and formulation advantages of small-molecule drugs. This combination allows them to precisely target difficult receptors while avoiding common off-target effects or instability issues.
As AI tools have largely solved the initial design of functional mini-proteins, the critical technical challenge has shifted. The primary bottleneck is now engineering a suitable half-life. In their raw form, these molecules are cleared from the body in 10-20 minutes, necessitating strategies like FC fusions or lipidations to make them therapeutically viable.