How This Harvard Professor Is Using AI to Heal Our Wounds

Jess Whited, a regenerative scientist at Harvard University and the co-founder of Matice Biosciences, is using AI to cure scarring. She sat down with Jessica Abo to talk about her company.

Jessica Abo: Jess, you're the co-founder of Matice Biosciences. Tell us a little bit about what you do.

Matice Biosciences is a regenerative medicine company founded to harness the power of super regenerators. These are species in nature that are amazing healers and possess some important similarities to humans. We're mining data from these species to identify pathways and peptides that will help humans heal from injuries without scarring and reimagine wound and skin care.

What's your background and what inspired you to start your company?

For my professional life, I lead an academic research lab at Harvard where we're devoted to understanding how salamanders regenerate. And while I'd always known that they do this, I wasn't focused on the scarring aspect of it until my son had a terrible bicycle accident. He was eight years old and suffered serious cuts to the middle of his face, which required plastic surgery. In talking with the surgeons, I learned there wasn't really much that I could do to shift the outcome of his scars. So as a mom, I felt compelled to rethink the information that we were generating in our limb generation studies and to consider a fortuitous byproduct of these studies that we might actually be able to repurpose the data to develop human anti-scarring treatments. It is also important to note that many of the regenerative abilities that these super-regenerators possess are also innately present in humans, though they are only active in-utero, and for a short period during infancy. If you are a parent, you may have noticed how quickly and often scarlessly a baby will heal from a wound. Part of our job is to figure out how to reactive this capability through what we learn from the super regenerators.

What made you land on salamanders?

I started searching around for what animals are really great at replacing entire parts of their body. And to me, the axolotl salamander was the most compelling because it regenerates its limbs throughout its whole life. And these are very complex appendages and they're extremely similar mechanically to human limbs. I felt like this was the most extreme example of limb regeneration in nature, and I really wanted to go after that.

And what can you make cells do that will make the scar outcome better?

From the moment that the wound happens, we want to focus on shifting the outcome of the scar. There are three main differences between how a super regenerator heals a wound, and how a human heals a wound. Firstly, the super regenerators do it really fast. They form a skin barrier over the wound very quickly. Then, instead of just remaining in an inflamed state, they shift into an anti-inflammatory mode much faster than we do. Lastly, humans tend to make these kinds of deleterious cells called myofibroblasts, and they're basically responsible for secreting the tissue that creates the appearance of a scar, but super regenerators are really good at staving that off and not making very many myofibroblasts.

How are you using AI to solve these issues?

Our proprietary platform, RegenProAI, uses computational biology and artificial intelligence to sort through billions of data points from many super regenerators that are out in the world to find the molecules, or peptides that would be the best at having effects on human skin. We've developed an entire pipeline of computational analyses where we feed in the raw data and then we refine it according to our predictions about which peptides might have the best effects on human cells. Then we synthesize them in vitro, in a totally vegan way, and obtain these molecules. We test them on human cells and then we use the results of these tests to feedback into the AI platform and make it a little smarter every time. In the end, we find these needles in a haystack, which we test again on humans, and the ultimate compounds go into formulations.

What have you been able to prove so far?

One thing that I'm super excited about is that we have discovered a significant number of peptides that inhibit the formation of the bad types of cells, the myofibroblasts, which are largely responsible for scarring. We have some molecules that can actually reduce the formation of myofibroblasts by almost 70% compared to controls.

Some of these same peptides also speed up the growth of skin cells, also called proliferation, which means more cells grow faster, which in turn speeds up healing, and some of these even accelerate the speed at which the cells migrate to, and cover a wound site by more than 150%, meaning that they can form a protective barrier much faster following an injury.

All of this cell-based work has allowed us to select some commercial targets that have been tested on humans and demonstrated some fantastic clinical results. From intentionally induced wounds, we have demonstrated that we can speed recovery time by 76% and reduce inflammation by 142%

What are you doing differently from what's out there?

Matice is the first company to really use AI and super regenerators in this way to create products that have actual proven outcomes for humans. There's really nothing else available like this right now.

We are marrying deep expertise in the realm of regenerative biology with massive data mining, bioinformatics and artificial intelligence to generate our own ingredients. The peptides we've discovered are all some combination of nature, often from multiple organisms, and machine learning. As such, they are all de-novo, meaning they don't exist in nature and are all unique to us and are patentable. They're inspired by the salamander, and possibly a zebra fish or spiny mouse, but they're optimized to affect elements encoded in the human genome.

Jess, how do you think this will positively impact people, both in the short term and long term?

In the short term, Matice Biosciences is developing peptides that will be used in formulations that people will be able to buy at a dermatologist or drugstore for wound care. In the long term, many of these peptides and others may go into FDA-approved medical devices or drugs that can be used to treat organ fibrosis, which mechanistically shares similarities to human skin scarring, or used for internal and external post-surgical wound healing.

What's next for you?

So far, Matice has screened billions of unique protein fragments and discovered and designed more than 200 novel and patentable peptides. But what I'm really excited about is what the future of Matice holds as we create and test more of these peptides, which are inspired by nature but perfected by science.