CytoReason, an artificial intelligence (AI) disease modeling company, has made a deal with Sanofi for several years to help them find new targets in the field of inflammatory bowel disease (IBD). Under the terms of the expanded agreement, Sanofi will pay the Israeli company an undisclosed multimillion-dollar amount to use CytoReason’s AI platform, which enables in silico clinical trials.
“We built a platform that allows biopharmaceutical companies to manage their portfolios and run synthetic trials for making portfolio decisions in their precision medicine group,” CytoReason co-founder and CEO David Harel told GEN Edge.
“We take a lot of data—multiple data sets and types and all sorts of information that can be aggregated on the disease—and reconstruct the disease biology on the mechanistic level, giving the researcher a tool to be able to understand the disease better, either for understanding the functionality of targets or the mechanism of action of drugs or comparing across diseases.”
Sanofi has previously applied Cytoreason’s cell-centered model to understand asthma patient subtypes. In 2021, Cytoreason and Sanofi announced that they would collaborate to make cell-centered models that would help explain how asthma endotypes work.
This is the second notable deal for CytoReason in the past few months, following a deal announcement with Pfizer that could be worth up to $110 million over the next five years. That deal also follows an initial deal struck in 2019, in which Pfizer and CytoReason agreed to create a cell-based model for immune responses.
The (baskets &) umbrella academy
Founded in 2016, CytoReason’s technology is accessible in three ways—a graphical user interface (GUI), API, or straight into the database—depending on the familiarity of customers with data science and programming. Once accessed, CytoReason’s technology can be used for two immediate in silico applications: synthetic basket trials and synthetic umbrella trials.
In a basket trial, a compound’s mechanism of action is compared across multiple indications. Because it is synthetic, it can be done in preclinical stages. “When we talk about basket trials, one might think about an actual compound that is given to multiple patients in multiple indications, but when it’s in silico, you can do that synthetically in preclinical or discovery phase assets,” said Harel.
For example, CytoReason’s platform helped Pfizer figure out how to apply an experimental compound that interferes with CCR6, a protein that is elevated in certain autoimmune diseases. As a result, Pfizer was able to develop a drug specifically geared toward treating IBD, which entered Phase I clinical trials in July 2020.
The other application—the synthetic umbrella trial—can compare the efficacy of two different treatments for an indication in a specific patient population or subpopulation. “When you think about prioritizing your targets, every large company has several dozen targets in each indication, and they’re evaluating whether to license or develop in-house,” said Harel.
“What they can do with our models in each disease is to prioritize the drugs that are more likely to be relevant for the disease and effective [on] the clinical side. In cancer clinical trials, you look at how your drug fares against the standard-of-care in different subpopulations. The results of a larger trial may not be clear in all of the arms, but the question is whether you can find subgroups of patients who benefit from the specific combination. That would be a typical use of our technology.”
Brain on fire
The two diseases that the CytoReason-Sanofi deals are centered around—asthma and IBD—are notoriously heterogeneous in their makeup and immune-mediated, which is one of CytoReason’s three sweet spots. The other two are immuno-oncology (non-small cell lung cancer, renal cell carcinoma, and melanoma) and fibrosis (non-alcoholic steatohepatitis, systemic sclerosis, and idiopathic pulmonary fibrosis).
The Israeli company’s deals with Sanofi and Pfizer both start with a small number of diseases that are a focus for CytoReason. As the comfort level with the technology has grown, Pfizer has expanded the number of diseases studied within each therapeutic area and is broadly applying CytoReason’s technology across the company’s portfolio.
As for what to tackle next, Harel said it highly depends on the availability of data. “The challenge for data science in life sciences is the availability the relevance of the data to the disease mechanism,” said Harel.
For example, the challenge with neuropsychiatric diseases and many neurological conditions is that there is just not a lot of data that is associated with the disease mechanism. “In many of these cases, the use of imaging and peripheral blood doesn’t provide a close enough understanding of disease biology.”
But CytoReason isn’t avoiding all indications related to the nervous system. “We’re probably going to shy away from neuropsychiatric disease and neurological conditions, but we are starting to work on neurodegenerative diseases—the immediate next step to immune-mediated diseases,” said Harel.
Pharma teams are starting to rely more on AI to make key R&D decisions, including selecting the right indication or patient population for a drug candidate. The key to success in the AI industry, according to Harel, comes down to matching the tool to the problem. He says that, in a way, AI is a lot like transportation—there are many ways to get from point A to point B. The decision to take a plane, train, automobile, or boat all have very different purposes. Likewise, there are many tools being developed in AI, but they’re only useful if applied to traverse the right landscape.
