Immunology drug development poses several challenges to scientists, who are tasked with managing a delicate balance. They must find ways to control inflammation without also tamping down the immune system to the point of compromising a patient’s immunity, explained Jo Viney, CEO of Seismic Therapeutic. Adding to that challenge is the chronic nature of these diseases, which means any drug that treats them likely must be a lifelong therapeutic.
Viney has steered several drugs into clinical trials in a career that includes stops at Amgen and Biogen. She was most recently chief scientific officer of Pandion Therapeutics, an autoimmune disease biotech acquired by Merck last year in a $1.85 billion deal. At Watertown, Massachusetts-based Seismic, Viney has something she didn’t have at any of those companies: the capability to apply machine learning to drug discovery. Her new venture is now unleashing that technology on the challenges of immunology drug research, an effort backed by a $101 million Series A round of financing.
“It’s a magic moment of time for machine learning,” Viney said. “We really now have the opportunity as an industry to think about machine learning to accelerate the drug discovery and development process.”
Drug development has historically been a trial and error endeavor. As an example, Viney pointed to making amino acids substitutions in a protein, changes that can alter a protein’s function. After making a substitution, an experiment would be run to test it. Each subsequent change would be followed by another test.
Machine learning enables Seismic to look at an entire protein and evaluate what happens when multiple changes are done. Viney said this “parallelization,” borrowing a term from the technology sector, improves the speed and efficiency of the entire process. Data from this iterative design and testing process is fed back into the company’s technology, enabling it to learn from the amino acid substitutions. Seismic has dubbed its proprietary platform “Impact.”
A Deep-dive Into Specialty Pharma
A specialty drug is a class of prescription medications used to treat complex, chronic or rare medical conditions. Although this classification was originally intended to define the treatment of rare, also termed “orphan” diseases, affecting fewer than 200,000 people in the US, more recently, specialty drugs have emerged as the cornerstone of treatment for chronic and complex diseases such as cancer, autoimmune conditions, diabetes, hepatitis C, and HIV/AIDS.
Seismic is one of a growing number of life science companies either adding new machine learning capability or building an entire business model around it. In the realm of proteins, Generate Biomedicines uses machine learning to find patterns and make predictions about the targets to which a protein will bind, informing its protein drug designs. Last November, Generate closed a $370 million Series B financing to fuel its plans to reach the clinic in two years. Last month, Sanofi paid AI-biotech Exscientia $100 million up front to begin a drug discovery alliance focused on cancer and immunology. And this week, startup Congruence Therapeutics launched with $50 million and technology that uses machine learning to design and develop molecules that stabilize misfolded proteins.
Viney contends Seismic is offering something distinct. With Impact, machine learning isn’t replacing other drug research techniques. Instead, Seismic integrates machine learning with structural biology, protein engineering, and translational immunology. Viney said that machine learning also enables the startup to alter the properties of its drugs to fine-tune their activity.
“We believe it’s differentiated,” Viney said. “The central thesis is not machine learning for machine learning’s sake, but [putting it] right in the middle of the drug discovery and development process.”
Viney declined to say what types of drugs Seismic is developing, other than to describe them generally as biologics. She did say that those drugs will target dysregulation of adaptive immunity, which is the immune response from white blood cells called lymphocytes. These responses are carried out by two types of cells, B cells and T cells. Seismic will be developing drugs addressing both types.
Specific disease indications also remain undisclosed, but Viney said that beyond autoimmunity, the Impact technology has potential applications in immuno-oncology, as well as gene therapies or enzyme replacement therapies that are limited by the immune responses that they trigger. Viney added that the financing enables Seismic to work on multiple programs concurrently.
Seismic traces its roots to Timothy Springer, a professor of biological chemistry and molecular pharmacology at Harvard Medical School. Springer’s research has led to the formation of several biotech companies; he is a founding investor of Moderna and Editas Medicine. Viney said Springer saw the opportunity for machine learning to accelerate and enhance drug discovery, particularly in immunology. He took his ideas to venture capital firm Polaris Partners. Alan Crane, a partner at the firm, took interest. The pair founded Seismic.
After Merck’s acquisition of Pandion closed last year, Viney stayed on to integrate the biotech’s programs into their new home. When that work was done, she was considering options for her next move when she reconnected with Crane, who was a Pandion founder. Crane told Viney about what Springer’s ideas could bring to immunology drug research. Viney formally joined Seismic as its CEO last October. The biotech employed 14 people at the end of 2021. Viney said she projects headcount will more than double, perhaps even triple by the end of this year.
The Series A round of funding was led by Lightspeed Venture Partners. Springer and Polaris Partners, the lead founding investors, also participated. New investors that joined the financing include GV, Boxer Capital, and Samsara BioCapital.
