Pharma startup Quris aims to use a ‘patient on a chip’ to target drug delivery

Nobel Laureate, Aaron Ciechanover, is one of several notable names behind pharma startup Quris. The company aims to bring together artificial intelligence, the industry’s vast knowledge of the human genome, and the concept of the “patient on a chip” to improve the effectiveness of drug delivery.

Last month, the startup announced the launch of its AI platform and a $9 million seed round, led by Moshe Yanai (the mind behind EMC Symmetrix) and Dr. Judith Richter, and Dr. Jacob Richter (founders of Medinol, which has sold more than 2 million cardiovascular stents).

Ciechanover, as well as Moderna cofounder, Robert Langer, are among Quris’ noteworthy advisers.

For decades, medical research has successfully cured cancer and treated rare diseases in innumerable quantities of mice – but has not done so as frequently in humans. As Nobel Laureate Aaron Ciechanover points out, mice are different from humans in nearly every way, from genetics to diet. “It’s no wonder that 92% of the drugs that are successful in mice are failing in clinical trials in humans,” Ciechanover says.

“At the base of it, pharma is amazingly ineffective,” says Quris founder and CEO Dr. Isaac Bentwich, whose previous companies have focused on computerized medical records, genomic data analysis, and analyzing soil conditions to grow more crops using fewer resources. “We need to be able to better predict which drug candidates will work safely in humans.”

Testing drugs can be similar to predicting elections

The concept of the patient on a chip drives the Quris approach. It’s similar to the practice of testing different chemotherapy options on a tumor that has been removed from a patient, Bentwich says – only the patient on a chip uses stem cells to create miniaturized versions of internal organs and arteries.

Hundreds of patients can fit on a chip.  By applying the Quris AI platform to enough patients, it’s possible to test more than 1,000 drugs on the same patient cohort at the same time, Bentwich explains. Quris has developed its AI tool to predict things like which patients are the best candidates for certain drugs, as well as which drugs will be most effective on those particular patients based on various genome subtypes. Patients with a similar subtype of cancer or rare genetic disease are likely to respond to similar therapies.

“You’re not just selecting an individual. You’re selecting a representative group, so you can get the right balance of age and gender, and occupation. It’s like predicting elections,” Bentwich says.

At launch, Quris has two projects in the works. One is a partnership with a large pharma firm to test the platform, with an option to purchase it to help develop a single drug over a five-year period. Quris anticipates charging a rate of $60 million to $100 million per drug or indication — a fraction of the potential losses due to safety and efficacy failures during drug development. The approach also aligns with the recent call from the European Parliament to move away from animal testing for scientific purposes.

The other is applying the platform to develop its own drug, which would treat a cause of hereditary autism known as Fragile-X syndrome. While admittedly not a large market, Bentwich describes it as an “archetypal example” of the type of conditions Quris hopes to address, as there’s no equivalent model in mice. The company hopes to begin clinical trials for the drug in 2022. “We see it as a proof of concept, as a validation of our engine.”

Standing out among phrama companies using AI

There is no shortage of companies both large and small applying AI to pharma R&D. Biopharma Dealmakers, a Nature Research publication, points to more than a dozen that have received funding or completed an IPO since the start of 2020, along with a similar number of partnerships inked since the spring of 2019.

This pace doesn’t surprise Natalie Schibell, a senior healthcare analyst at Forrester. She notes that pharma is leaning on AI tech for a host of tasks, including recruiting patients, cleansing real-world data, monitoring biomarkers, and automating administrative tasks.

“Historically, it takes an average of ten to 12 years to bring a new drug to market, including five to seven years for clinical trials. The benefits of AI can reduce risk and save an abundance of effort, cost, and overall time to market,” she says. “The capacity for AI to automate data capture, digitize clinical assessment, and share data across multiple systems far outweighs the speed and volume that can be performed manually.”

What differentiates Quris from other AI companies targeting pharma, Bentwich says, is taking the step beyond analytics and working to bring an AI-developed drug into the clinic. With the work on the Fragile-X syndrome treatment, “the Quris engine has yielded a drug that allows us to validate our approach.”

A ‘Mini-Me’ for personal testing

The current cost of generating a set of miniature organs on a chip is roughly $15,000, Bentwich says. It’s expensive, sure, but a far cry from $2 million about a decade ago. And it’s poised to drop to $100 within the next decade – following a trajectory similar to the cost of sequencing the human genome.

At that price point, consumers could acquire a “Mini-Me” that lets them personally test not just pharmaceutical therapies, but over-the-counter medications like vitamins and antibiotics.

“If you can identify the genomic markers that make the differences between us, think of the power that can have,” Bentwich says. “Predicting which drugs will work best in humans is a trillion-dollar problem. It’s the most lucrative AI challenge of our time.”