Stem cells meet AI

Micha Breakstone  credit: Jonathan Bloom
Micha Breakstone credit: Jonathan Bloom

Together with Harvard and MIT scientists, Micha Breakstone founded, which combines stem cell and AI techniques to develop new therapies.

Two years ago, Micha Breakstone, founder of Chorus.Ai, a company that created a sales call analysis system, presented a new company, a startup to treat ALS. The company, NeuraLight, was launched with great fanfare, but Breakstone has since reduced his involvement with it. This was because of disagreements with the management of the company, but also because he is busy with a new company,, that he is now ready to introduce to the public. He believes it can revolutionize stem cell therapy.

Somite was founded on the basis of years of work by one of the most prominent teams of scientists in the stem cell field, some of them Israeli. Breakstone's first connection to this group was through a personal friend, Prof. Allon Klein, head of the Klein lab at Harvard Medical School.

"I met Allon when we were both in the army. He was then, and still is, one of the smartest people I know", says Breakstone. "I approached him and asked whether his research was ready for the market. Initially, he said the same thing he's been saying to me for 25 years - let's stay friends, why complicate things? But eventually I convinced him and he connected me with two other scientists he works with at Harvard, Prof. Olivier Pourquié and Prof. Cliff Tabin. Olivier had already founded a few companies, among them Trophos, which had been acquired by Roche for $500 million."

After three months of conversations, he says, the four of them decided to found the company together. Despite all the discussions, Breakstone's final decision to invest was spontaneous: "And all the way home I wondered what it would be like to sleep on the sofa, since I had forgotten to consult with my wife, and the amount I invested was not a sum we usually spend without each other's blessing".

Breakstone is originally from the AI sector, and so is the fifth founder who joined the company, Dr. Jonathan (Johnny) Rosenfeld, originally from Jerusalem. Another Israeli involved in the company but not as a founder is Ilan Ganot, formerly an executive at JP Morgan. About a decade ago, Ganot founded a company in hopes of saving his son from the genetic disease Duchenne Muscular Dystrophy. Breakstone: "That company had some successes but has not yet managed to treat the disease, so Ganot directed some scientists to us, since our technology may be relevant to DMD."

The company was officially founded a few months ago, but the science has been in development for years. Klein: "My lab is focused on the question of how cells differentiate. We all begin as one fertilized egg, and develop to contain many different types of cells based on the same DNA, but with very different functions - skin cells, bone cells, blood cell and so on."

Klein was always in awe of this wonder. On the day of our interview with Klein, Breakstone's second daughter was born. Klein, excited for his friend and colleague, said: "At such moments, it is difficult to be a developmental biologist. How did the universe create such a miracle? We have learned so much about the process, and still it is amazing to see it every time, each and every molecule in place."

Klein began his education as a theoretical physicist, and then went on to research how cells supposedly 'choose' to develop into a certain type. "We realized over the years that these processes are much less arbitrary than we thought. Signals from the cell's surroundings drive them to make the 'decision'". These signals might be the presence of a certain substance, a change in acidity, a hormonal stimulus, or hundreds of other factors that can effect the fate of a cell.

Porquié was also interested in the same questions, but from an engineering perspective. Breakstone: "He spent fifteen years trying to urge cells to differentiate in a specific direction. After twelve years, he had gotten to 25% purity, meaning that one of every four cells turns into the cell we are aiming for. And then he decided to spend some time at Allon's lab. They used AI, and after a few months, they had doubled the purity, after understanding many of the factors that affect the differentiation of the cell".

Klein: "The surroundings of the cell cause certain genes to acquire a 'tag' that dictates if and to what degree that specific gene will be expressed. The scientific community has developed over the years methods for mapping 20-30 genes in a cell to determine which one of them is switched on and off. This way, we can see the cell’s 'story' unfold before us. We know how each cell is programmed to change until it reaches its final role".

One important breakthrough was the ability to use these methods at the single cell level, since each cell differentiates differently. If we can predict the cell's potential, and then understand how signals from the surrounding direct it to realize some of it's potential futures rather than others, we can influence the differentiation. "We can build not only the street map, but also understand the traffic lights".

Clinical trials in two years

When the gene map is understood, it can be possible to influence either directly by gene editing, or by manipulating the surroundings of the cell. Somite.AI focuses on the latter, in order to reap cells that are as similar as possible to those that appear spontaneously in the human body.

The first goal is muscle stem cells, but the company believes it can create many types of cells needed for different types of treatment.

The first clinical indication will be Duchenne Muscular Dystrophy. "When we inject the muscle stem cells into an existing muscle, even if it is a weak one, the muscles can rebuild and regain function", says Breakstone. So far this has been demonstrated in animals, and the company believes it can reach clinical trials in two years.

Has there ever been an attempt to treat muscular dystrophy with stem cells?

Breakstone: "Attempts have been made to use cells called Myoblasts, that are muscle building cells, but not stem cells. The effect lasted only about two months. Why? Because these cells are limited in their differentiation and they cannot support the muscle over time".

The next indication is creating brown fat cells for the treatment of diabetes. This has also been shown in animals, in experiments performed by Pourquié but not yet published.

The company has raised a few millions of dollars from Breakstone and investors such as Lerer Hippeau, Mark VC and DarkMode.

What can you tell us about NeuraLight, your previous company, and your reasons for leaving them?

"NeuraLight preferred to focus on one disease, while I wanted to address a broader spectrum of conditions. Some of the staff continued with me to Somite. I am still a stockholder in NeuraLight, and am rooting for their success. When I left they had $20 million in the bank".

What kind of companies do you invest in?

"I believe in investment in people, and can sometimes invest after a single meeting. I try to avoid two shortcomings of the venture capital world. One, investors who shortchange the founders, and the second, investing only in people that are similar to you, especially alpha males. I was involved in the early days of companies like ImmunAI and Verbit. Lately I invested in Freed, a company that helps doctors write up reports after consultations".

You mentioned that your path has not always been easy.

"A classic entrepreneurship story. I founded a company in 2009. I was a PhD student at MIT at the time, and lost all my money. The whole story caused a breakup between me and my then fiancée. But when you realize you know how to build companies, you can't turn your back to that. I found out it's the best way for me to impact the world."

Published by Globes, Israel business news - - on February 29, 2024.

© Copyright of Globes Publisher Itonut (1983) Ltd., 2024.

Micha Breakstone  credit: Jonathan Bloom
Micha Breakstone credit: Jonathan Bloom
Twitter Facebook Linkedin RSS Newsletters גלובס Israel Business Conference 2018