Human natural killer cell. (National Institute of Allergy and Infectious Diseases Image)

Modulus Therapeutics, a Seattle-based cellular therapy company spun out of the Allen Institute of Artificial Intelligence (AI2), announced $3.5 million in seed funding.

The company combines laboratory research with machine learning to engineer immune cells with enhanced anti-tumor powers. Modulus focuses on generating performance-enhancing genetic changes in natural killer (NK) cells, an up-and-coming immune cell type in the cell therapy field.

Modulus is building a platform that could potentially provide early cell product candidates to more mature cellular therapy companies for further testing.

Seattle is known for its strength in cell therapy, notes Modulus chief scientist Max Darnell, with growing biotech companies such as Sana Biotechnology and Lyell Immunopharma, as well as an arm of Bristol Myers Squib. “We want to be this discovery player that can plug into these larger incumbents,” he said.

Last year, Darnell, who has a bioengineering background, teamed up with bioinformatician and chief data scientist Bryce Daines to found the company. The pair met at AI2, where they were entrepreneurs-in-residence.

Modulus Chief Scientist Max Darnell. (Modulus Photo)

Natural killer cells are shorter-lived than T cells, which are the basis for several approved cell therapies for blood cancers, as well as numerous clinical trials. Approved T cell therapies sell for hundreds of thousands of dollars per patient, in part because the cells are harvested from each individual before being engineered in petri dishes and re-infused.

But though they last only about two weeks in the body, NK cells have some unique advantages. They are more amenable to a less expensive, “off-the-shelf” approach that does not require harvesting from individual patients. Numerous companies are developing NK cell products.

Modulus aims to generate perturbations, or mutations, in NK cells that give them extra cancer-fighting ability — including the power to kill solid tumors such as breast cancer. Solid tumors, with their tightly stuck-together and hard-to-reach cells, present an extra challenge to cellular therapies, which have shown more success for blood cancers.

To power up NK cells, Modulus starts in the lab. The researchers first create perturbations in the NK cells to alter their function. The cells are then injected as a pool into mice implanted with human tumors and later collected from the mice. The researchers assess which perturbations confer positive functions, such as ramped up immune capabilities, by profiling each cell’s RNA, a readout of active genes.

The AI comes in next. The information on the perturbations and the RNA profile of the NK cells is fed into Modulus’ machine learning algorithms, which predict which combination of perturbations makes the best NK cell. The researchers then repeat the process, making the recommended perturbations in NK cells and injecting the cells again into mice. The process is an “iterative feedback loop to converge on better and better designs,” said Darnell.

The end result, if it all works, will be a super-powered NK cell. “We would love to get our hands on a therapeutically interesting candidate that is compelling either for further internal development or compelling to a partner,” said Darnell.

He notes that other cellular therapy companies are honing laboratory practices and tests to ease the regulatory landscape for gene-edited cell products. Regulators will be asking, for instance, if such products have unexpected effects on the body or are likely to become cancerous themselves.

“I’m somewhat glad we’re not one of the first pioneers going to the FDA with that challenge,” said Darnell. “Almost all major cell therapy players are working on gene edited therapies now.”

Modulus chief data scientist Bryce Daines. (Modulus Photo)

In the future, Darnell and Daines anticipate adapting their approach to different cell types, including T cells. They are currently focused on breast cancer, but they also plan to engineer cells against other tumor types in the future.

Collaborator and scientific advisor Alana Welm, senior director of basic science at the Huntsman Comprehensive Cancer Center in Salt Lake City, partners with Modulus for the animal studies and investigates a range of tumor types. The CoMotion innovation program at the University of Washington also provides lab space.

Darnell and Daines will use the new funding to scale up their experiments on NK cells and hire new employees on both the laboratory and computational sides of the company.

Other companies generating platforms to optimize therapeutic cell design include San Francisco-based Arsenal Bio, though it focuses mainly on T cells. Ideal future partners would be cellular therapy companies that could plug Modulus’ cell candidates into their preclinical testing pipeline, said Darnell and Daines.

“As we think about potential competitors, it’s really a group of potential customers or collaborators as well, because we’re in that early stage of discovery,” said Daines.

The funding round was led by Seattle-based Madrona Venture Group, which has historically invested in tech companies, but more recently has backed several life sciences companies, including Seattle protein analysis company Nautilus Biotechnology, which recently went public. Madrona also recently backed another AI2 spinout, Ozette, which profiles immune responses through analyses of protein datasets.

Raphael Gottardo, who directs a translational data science center at Fred Hutchinson Cancer Research Center, is a co-founder of Ozette, and is the other scientific advisor for Modulus.

KdT Ventures and the AI2 also participated in the funding round. Other AI2 spinouts include Kitt.ai (acquired by Baidu in 2017), Xnor.ai (acquired by Apple in 2020), Why LabsBlue CanoeWellSaid Labs, and Panda AI.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline

Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.