UPDATE, 5/19/2021: twoXAR, a company focused on discovering and bringing first-in-class molecules to market, has rebranded as Aria Pharmaceuticals. The name reflects the artistry and complexity that defines how the company and its scientific teams approach medicine from discovery to FDA approval in a way that is scientifically sound. Since its founding in 2015, the company has built its own proprietary pipeline as well as focused its partnership strategy on working with biotechnology companies on drug discovery research. With successful preclinical efficacy and safety data emerging from several of the company’s proprietary candidates, Aria Pharmaceuticals will transition from discovery collaborations with pharmaceutical companies to advancing its own pipeline into clinical research.
Since being named to the inaugural STEX25 class in 2017, the aptly named twoXAR—serendipitously, both co-founders are named Andrew Radin (hence, two times AR)—has moved from proof of concept to commercialization, signing a number of important deals, closing their series A funding with SoftBank, and developing a substantial portfolio of discovery-stage disease programs across a variety of therapeutic areas. Typical drug discovery relies on a linear and stepwise approach that often iterates on several steps, which compounds the time, cost, and failure rate to develop a new drug. This process involves first understanding pathogenesis (how the disease works), selecting a target protein to modulate, and then creating a bioassay to test molecules against the target protein. With that in place, researchers test thousands of drug candidates in High-throughput screening to find hits that they then use medicinal chemistry to turn into drug candidates. It’s often represented that this process takes about 5 years.
The pharmaceutical industry has faced declining R&D productivity for many years. Recent estimates indicate that the total cost to develop a new drug is ~$2 billion and can take 10-15 years with <5% likelihood of approval. In an effort to keep the return on investment of R&D above $0, executives at biopharmaceutical companies have increasingly stated that there needs to be greater efficiencies in R&D. These efficiencies can be incremental (optimizing a step in the traditional process) or transformational (redefining the process). TwoXAR is seeking to bring transformational efficiency changes to the industry. The twoXAR team is hyper focused on two things: making drug discovery more efficient and maximizing the probability of a twoXAR-discovered drug benefiting patients. The twoXAR team developed software that leverages existing disease and drug candidate data to enable rapid, more accurate decision-making around launching a program; this allows them to define a different drug discovery process. “Our platform takes an understanding of pathogenesis and generates a computational equivalent of disease theories we can run to examine whether or not a new medication will work, creates those tests digitally, and ultimately runs the equivalent of High-throughput screening.” Andrew A. is referring to an essential aspect of the traditional drug discovery process whereby experimental samples are subjected to simultaneous testing under given conditions. The software identifies data-driven novel biological hypotheses and in vivo-testable drug candidates and the team validates efficacy and safety in standard animal model – all in a matter of months.
Our platform allows us to take multiple complex, very expensive steps that cost millions of dollars over the course of many years, and produce that same output in a matter of minutes.
“Our platform allows us to take multiple complex, very expensive steps that cost millions of dollars over the course of many years, and produce that same output in a matter of minutes; our success rate—the likelihood that we’ll show efficacy in an animal study—is approximately 30 times higher than the old process,” explains Andrew A., referring to data from their drug candidate validation studies where approximately three candidates in ten demonstrated efficacy as opposed to the oft-cited number of one or two in one hundred at this stage of discovery. Even with these efficiencies, drug discovery is still a numbers game. To increase the probability of discovering and developing drugs that address unmet medical needs, the twoXAR team is taking a portfolio approach. “In the last year, we’ve launched 12 disease programs,” says Andrew M., who holds an MBA from MIT Sloan. “Some of those are internal and some are co-development partnerships with biopharmaceutical companies. The National Cancer Institute recently granted twoXAR a $225,030 Phase I Small Business Innovation Research (SBIR) award. The grant will fund a program for twoXAR to identify and validate new, first-in-class drug candidates for pancreatic ductal adenocarcinoma in an effort to improve patient survival outcomes. “With all of our collaborations, we focus on complimentary capabilities,” says Andrew M. “TwoXAR brings rapid and accurate drug discovery, and our partners bring development capabilities. The combination dramatically improves efficiency and speeds along the discovery timeline.” In January 2019, twoXAR announced a partnership with 1ST Biotherapeutics, Inc., a preclinical-stage biotechnology company focused on neurodegenerative diseases, immuno-oncology, and orphan diseases. With twoXAR’s proprietary predictive algorithms driving the discovery phase, the South Korean company hopes to develop novel treatments to address unmet medical needs in an aggressive brain and spinal cord cancer known as glioblastoma. Adynxx is another biotech company set to benefit from partnering with twoXAR, as they seek to develop oral, non-hormonal drug treatments to address endometriosis, an inflammatory disease that affects 176 million women worldwide. According to twoXAR CEO Andrew A., who studied biomedical informatics at Stanford University, while the disease programs are diverse, there are common threads they pursue. First, is the presence of quality data in that disease, a prerequisite for accurate predictions. Second is a definitive pre-clinical study that helps to characterize how the drugs being tested will perform in human subjects. Finally, a key signifier is opportunity: are there clear unmet medical needs and are biopharmaceutical companies committing resources to developing drugs for that disease - that gives us a good indication that we’re working in an area that we can both create value and capture value,” says Andrew A.
Historically, drug discovery has been guided by biologists, which has led to incremental improvements in the process. The Andrews and the team at twoXAR, on the other hand, are seeking meaningful impact in society by rethinking the process itself. Andrew A. draws a parallel to the advent of autonomous vehicles. Viewing passenger safety from the perspective of the automotive industry is what led to the invention of seat belts, airbags, and antilock brakes. Meanwhile, a computer scientist considering passenger safety decides to rethink the whole paradigm and replace the driver entirely. Andrew M. likes to make the point that the technology, while not irrelevant, is the conduit not the end goal: “The goal is efficient drug discovery; AI-driven technology is what allows us to be the most efficient in discovering new drug candidates,” he says. But clearly, the technology is key. Which is where Andrew A. comes in. “If you bring a computer scientist to the table to solve drug discovery problems, a computer scientist can reimagine whether some of the steps are even necessary, and possibly replace them,” he says.
If you bring a computer scientist to the table to solve drug discovery problems, a computer scientist can reimagine whether some of the steps are even necessary, and possibly replace them.
While twoXAR is not the first company to apply computer science to drug discovery, the team is certainly approaching it from a different perspective. Andrew A. notes that while researching how computer science was being used in drug discovery, he noticed that the software approaches he came across fell into two categories: either someone had created data that no one else had access to or someone had created a shrewd algorithm capable of extracting “hidden” information from available data sources. Regardless, both approaches suffered from the same problem: an inordinate number of false positives. “My thought was, rather than going after single data sources, why not combine multiple data sources from many different disciplines, many different angles, different lenses, and combine them to get clearer picture of what is real and what is a false positive,” says Andrew A. The result has been an influx of industry suitors interested in discussing the possibilities of efficiency change for their own drug discovery programs. “Since being named to STEX25, we’ve benefitted tremendously from the support offered by the Industrial Liaison Program,” says Andrew M. “In conjunction with MIT STEX and ILP we’ve had the opportunity to meet biopharma partners across the world, including Korea, Japan, the UK, and of course, here in Boston.” “Our ideal partner is a company looking to bring innovation to their drug development systems, which means a company culture that is open to external partnerships and overall flexibility with regard to the drug discovery process as a whole,” says Andrew A. “Our technology allows us to rapidly identify potential new therapies and validate them quickly, which requires our partners to think differently about the process and be able to go from selecting a disease of interest to in vivo efficacy in a just few months.”