“Seismonauts” Stretch Frontiers of Immunology with ML and $121 Million

Astronauts are real-life space travelers unless they are Russian, in which case they are cosmonauts. Argonauts are figures in Greek mythology who pursued all manner of daring quests, such as the crew that accompanied Jason in his search for the Golden Fleece.

Then there are “Seismonauts”—researchers and other staffers at Seismic Therapeutic, a developer of machine learning (ML)-based immunology therapies derived from its IMPACT platform, which integrates ML with structural biology, protein engineering, and translational immunology.

“We go into new frontiers, exploring the galaxy of previously undiscovered and unknown areas of science,” Seismic’s founder, president and CEO Jo Viney, PhD, explained to GEN Edge.

Privately-held Seismic recently fueled up for its ongoing mission by completing a $121-million Series B financing. Combined with its Series A of $101 million, Seismic has raised $222 million since Viney and six co-founders established the company in 2021.

Since then, Seismic has built a pipeline anchored on two lead programs, each designed to modulate one of the two arms of the adaptive immune system, humoral immunity and cell-mediated immunity. Both are in IND-enabling studies.

One is a pan‑immunoglobulin (Ig) G protease sculpting (Sc) enzyme candidate, S-1117. For that lead IgSc enzyme program, Seismic engineered novel pan-IgG proteases with reduced B cell and T cell immunogenicity and chemical liabilities while maintaining enzyme activity and stability. The result is a protease therapeutic created to address multiple pathogenic mechanisms in a single drug for the treatment of acute and chronic autoantibody mediated diseases.

The other lead program is a PD-1 agonist:Fc gamma receptor IIb selective Dual-cell Bidirectional (DcB) antibody candidate, S-4321. Seismic’s DcB approach is designed to restore homeostasis by targeting dysregulated cell-mediated immunity through optimally engaging T cells through PD-1 agonism and antigen presenting cells such as B cells via the inhibitory Fc gamma receptor.

Seismic reasons that activating these pathways may control multiple autoimmune diseases, such as multiple sclerosis, lupus and rheumatoid arthritis.

No favorites

“Our two programs target different mechanisms in different patient cohorts. Both of them are very exciting for us,” Viney said. “We haven’t got a favorite at the moment.”

Seismic plans to use proceeds from the Series B financing in part to advance both lead programs through Phase I proof-of-mechanism trials. “We’re targeting in the next couple of years getting into the clinic,” Viney said, without offering any specific timelines.

“We’ve got an emerging pipeline behind in both of those different drug product areas. We’ve already got discovery programs going on with other Ig-sculpting enzymes and with other dual-cell bidirectional antibodies as well,” Viney added. “Our goal is to keep up the pipeline of development and platform development and make sure that we end up with a pipeline that’s very attractive for many development candidates for many years to come.”

Seismic also plans to use the Series B proceeds to expand its pipelines of both IgSc and DcB candidates beyond the lead programs—as well as further expand the IMPACT platform by developing additional methods to enhance biologics drug discovery in immunology. Among IgSc candidates, Seismic has disclosed a “next” program and three exploratory programs, all with undisclosed targets and all in discovery phases.

In antibody-mediated disease drugs, Viney acknowledged as a leader argenx, the Amsterdam-based developer of Vyvgart^® a neonatal fragment crystallizable receptor (FcRn) blocker indicated to treat generalized myasthenia gravis (gMG) in adults who are anti-acetylcholine receptor (AChR) antibody positive.

“We’re looking at those diseases that are being investigated, with the FcRns as a place where we would like to take our molecules and see if we can add some additional benefit to the patients who are responding inadequately,” Viney said.

Commercial success

Vyvgart has proven to be commercially successful, marketed as a monotherapy (efgartigimod alfa-fcab) and in combination with hyaluronidase-qvfc under the name Vyvgart Hytrulo. They are on track to achieve combined “blockbuster” status of $1 billion or more in product net sales this year, having generated $816.43 million in the first three quarters of 2023. That’s nearly quadruple the $227.32 million racked up in Q1–Q3 2022 (Vyvgart finished 2022 with $400.7 million in product net sales).

Among other developers of FcRn treatments are Immunovant and Johnson & Johnson (J&J). Last month J&J announced that its selective FcRn inhibitor monoclonal antibody (mAb) nipocalimab had established proof of mechanism in the Phase IIa IRIS-RA trial (NCT04991753) in adults with some forms of moderate to severe active rheumatoid arthritis (RA).

Immunovant has two anti-FcRn candidates in its pipeline. One is lead candidate batoclimab, for which Phase III topline data in gMG is expected in the second half of 2024. Immunovant has additional batoclimab programs under development in thyroid eye disease (Phase III topline data in first half of 2025), chronic inflammatory demyelinating polyneuropathy (Phase IIb initial data in the first half of 2024), and Graves’ disease (Phase II proof-of-concept data expected by year’s end).

Immunovant’s other anti-FcRn is IMVT-1402, for which the company last month announced positive Phase I data which it said showed potential  as a best-in-class FcRn inhibitor. Four subcutaneous doses of 600 mg produced a mean IgG reduction similar to high dose batoclimab, but with minimal changes in albumin and LDL-C similar to those in placebo.

Seismic’s DcB pipeline includes four additional programs with undisclosed targets: A “next” program in lead optimization phase and three exploratory programs in discovery phases. “For the dual-cell bi-directional antibodies, we’re really targeting inhibitory receptors on T-cells. We’re trying to turn off the activation of those immune cell types,” Viney said.

DcB drug development builds upon successes seen in monodirectional antibodies. One recent example: Eli Lilly researchers in May published data in The New England Journal of Medicine  from a Phase IIa trial showing its RA candidate peresolimab met the study’s primary endpoint for efficacy and had similar rates of adverse events in both treatment and placebo arms. Peresolimab is a humanized immunoglobulin G1 monoclonal antibody designed to stimulate human programmed cell death protein 1 (PD-1).

“Little bit differentiated”

“Our molecules a little bit differentiated because we’re not trying to just agonize inhibitory receptors on T cells, but we’re also binding to inhibitory receptors on other cell types like antigen presenting cells and B cells.”

Among diseases of interest, she said, are those that involve both T cells and B cells, including inflammatory bowel disease, arthritis, and lupus—chronic autoimmune disease indications with large patient populations.

IMPACT is designed to generate drug molecules on an unprecedented scale that are engineered to be improved therapeutics for treating autoimmune diseases, by simultaneously optimizing multiple interdisciplinary drug discovery properties in parallel:

• Opening new protein sequence space by discovering previously unexplored sequences and designing proteins with optimal drug-like properties.
• Removing B and T cell epitopes from novel biologic drugs to make them invisible to the immune system, while retaining the fitness and function of these drugs.
• Reducing the number and duration of design/test cycles by shortening the time of each cycle; and rapidly identifying top candidates by removing several cycles of iteration and mutation combinations.

One example of how ML and IMPACT have helped Seismic, Viney said, is its Ig sculpting enzyme program. The company identified proteases that can degrade immunoglobulins, but needed to turn them into drugs.

“There’s about 3,000 or so proteases that we trained our ML models on. We needed to make the protease more stable to last longer, so it would have a longer pharmacokinetic or pharmacodynamic effect, and just turn it more into something that was like a drug that could be administered easily to humans,” Viney recalled.

A protein engineer changing a single amino acid risks collapsing the protein—but with a series of changes resulting in a three-dimensional structure, “I’ve still got my protein, but I’ve got something that has got a better drug-like property. Instead of having tens and hundreds of protein engineers working for years, we can now use machines to in silico predict what happens with this, based on what we’re teaching the algorithm about with what exists in nature,” Viney said. “It just opens up new frontiers for us.”

Career focus

Seismic’s focus on immunology reflects Viney’s career focus in discovering and developing drugs in the specialty. During her first internship at St Bartholomew’s Hospital in London, Viney worked in an inflammatory bowel disease research lab.

“My first morning I went to see an endoscopy, and that was the day that I was compelled to spend my career thinking about bringing new medicines to patients with autoimmunity. So it’s been a lifelong passion of mine,” Viney recalled.

After holding positions at Immunex, Amgen, and Biogen (where she was senior VP of the drug discovery organization), Viney in 2017 co-founded Pandion Therapeutics with Alan Crane, an entrepreneur partner with Polaris Partners. Four years later, Merck & Co. shelled out $1.85 billion to acquire Pandion, whose pipeline focused on Treg-targeting treatments for autoimmune disorders.

Viney co-founded Seismic with Crane, who is also chairman of Seismic’s board; Timothy A. Springer, PhD, of Harvard Medical School and Boston Children’s Hospital; and four other academic co-founders: Debora Marks, PhD, and Andrew Kruse, PhD, both of Harvard Medical School; and Eric J. Sundberg, PhD, of Emory University School of Medicine.

Springer is an immunologist and professor of biological chemistry and molecular pharmacology at Harvard Med who has founded and invested in numerous biotechs. Most notably in 2010, he invested $5 million as a founding investor in Moderna when the company was formed in 2010; he now owns 3.5% of Moderna’s shares. Springer ranked No. 1,434 on Forbes’ World’s Billionaires List published earlier this year, with an estimated net worth of $1.3 billion.

“We launched the company with just ideas. We didn’t license any assets or technology. It’s all homegrown,” Viney said.

How did the company come to be called Seismic? “We want to make a seismic shift in how we think about developing new proteins. That was the rationale behind the name, and we certainly feel like we’re living up to that name,” Viney said.

Leading the Series B was a new investor, Bessemer Venture Partners with participation from five other new investors: Amgen Ventures, Codon Capital, Alexandria Venture Investments, Gaingels and GC&H. Andrew Hedin, MBA, biotech and healthcare partner with Bessemer, will join Seismic’s board of directors.

“Seismic has made impressive achievements in a short time integrating the power of machine learning with its deep drug development expertise to create two differentiated lead programs, each offering a compelling opportunity to address unmet medical needs in autoimmune diseases in a new way,” Hedin stated.

Also joining in the Series B were Springer, Seismic’s other founders and management, and several other existing investors that included Lightspeed Venture Partners, Polaris Partners, Boxer Capital, GV, and Samsara BioCapital.

Based in Cambridge, MA, Seismic has more than tripled its headcount since the end of 2021, growing from 14 to 47 people. “We’re anticipating adding about another dozen or so over the next year or two as we start to build out our development organization as we get closer and into the clinic,” Viney said.

The post “Seismonauts” Stretch Frontiers of Immunology with ML and $121 Million appeared first on GEN – Genetic Engineering and Biotechnology News.

therapeutics
antibodies
biologics
pharmacodynamic
medical
biotech
healthcare
machine learning
medicine
hospital
shares