Drug development is currently expensive and inefficient, taking 12-15 years and $2.6 billion dollars with a high failure rate of at least 90%, often due to unforeseen toxicity issues. The reliance on expensive live cell-based screens for toxicity testing, data reproducibility problems, and error-prone screens highlight the urgent need for a more cost-effective and accurate approach to early-stage toxicity prediction in drug development. Providentia Technologies links cell membrane damage to drug toxicity and introduces a synthetic cell membrane screening platform for early toxicity prediction. They validated this technology with a high success rate and are working on a second product that uses machine learning to pinpoint and modify the drug properties causing membrane damage and toxicity.
What is the problem?
Currently drug development is failure fraught. Developing one new medication takes 12-15 years, $2.6 billion dollars, and yet, the failure rate is still at least 90%. A lot of failure is due to unforeseen toxicity issues. Live cell-based screens are the go-to standard for early-stage in vitro toxicology; however, they are often so expensive that cytotoxicity testing can be deprioritized. This delay in identifying toxicity wastes years of effort and hundreds of millions of dollars. Additionally, issues with data reproducibility can interfere with data interpretation. Cell-based screens are also error-prone: today’s screens are, on average, incorrect at least 1 out of 5 times. Addressing these inefficiencies and providing a mitigating solution is critical because it has far-reaching consequences for drug expense, drug safety, and perhaps even the ability to work towards novel therapies for more diseases. A better approach is needed for early-stage toxicity prediction and correction.
What is their solution?
Providentia's research demonstrates that the extent of cell membrane damage caused by a small-molecule drug is strongly correlated with and predictive of that drug’s toxicity. So, they created the cell membrane component of the cell in the lab yielding their initial product offering: a synthetic cell membrane screening platform that enables early-stage toxicity prediction. They validated their technology in a blinded study against high-content cytotoxicity screens of ~500 drugs and found that precision of their approach exceeds 90%. Their second product offering in development combines their unique data type (measured membrane perturbation: that is aggregate changes of membrane properties (elasticity, curvature, thickness)) with machine learning approaches to identify the specific part of the drug driving the membrane damage/toxicity and provide corrective modifications to avoid this toxicity.
