Human Tissue Testing Beats Animal Trials | Andrei Georgescu

MTS (0:00)

This is like the reinforcement learning for biology.

Andrei Georgescu (0:03)

We have this perfect kind of RL problem where you have an extremely complex state, you have an extremely deep view into that state, right? So we do single cell sequencing and 3D imaging of these tissues, a thousand plex proteomics. We can extremely deeply characterize tissues at all these different states that they can take when we perturb them.

And we end up with an environment where we can take a healthy tissue and a diseased tissue, and we can say, take this diseased tissue and make it healthy. And then the throughput that we have, these tens of thousands, hundreds of thousands of different tissues we can grow at once, gives us huge rollouts that we can do. And then the ability to cycle that quickly in human tissue, over a matter of a couple of days of experimentation, means that we can move extremely quickly through this. We can take very many steps at once in this complex space and try to get there.

MTS (0:45)

Hello everyone and welcome back to MTS. Today I am joined by Andrei Georgescu, who is the co-founder and CEO of Vivodyne. Vivodyne grows living human tissues and uses robotics labs to run drug and genetic experiments on them at scale, generating human biological data before a drug ever reaches a patient. Wow, what a mouthful. That is a pretty cool company, but Andrei, welcome to MTS.

Andrei Georgescu (1:08)

Thanks so much for having me.

MTS (1:09)

All right, if you had to actually say what your company does, maybe a little bit better than my little one-liner, what are you guys actually building?

Andrei Georgescu (1:16)

We're enabling the testing on humans without testing in people.

And so if we want to understand how all this complexity that comprises biology really works, and then how to restore diseased function to patients with disease, how to enhance human function, increase health span, lifespan, there's a lot of complexity we have to unpack. And you can only do that by experimenting on the actual thing. But we can't test in people. And so what do we do? And so our approach is to grow human tissue that's functional, that's lifelike, it's patient-derived, responds just like real tissue at massive, massive scale. So like data center scale for what experimentation.

MTS (1:53)

No, I'm so excited about that, because what that actually enables is a whole slew of things. Like for example, it's getting a lot easier now with AI to actually do drug discovery, but the development process, if we can't fix that, a lot of these cures to cancers, cures to Alzheimer's, cures to all these different diseases never actually get implemented in society. So tell me a little bit about what the actual impact of these human tissues can be for all of us.

Andrei Georgescu (2:18)

So the major problem now in making new drugs is the only time you figure out how they perform in actual people is when you go to a clinical trial, and you've locked in the drug candidate, you've locked in the specific patients you're going to dose, and it's a hell of a wait to figure out if it works or not. And if it doesn't work, not only is it bad for the company, but all those patients you've enrolled, I've taken a drug that doesn't work for some pretty serious diseases. So how can we move that to the beginning of the drug development process and like start in human tissue? So the impact is when you enter clinical trials in people, you already have human data. And I think a lot of people think like, you know, I think to most people a mouse is pretty different from a person, right? But you start to think, you know, primates and monkeys look a lot like us, right? Maybe they're pretty similar in the drug responses, but they are marginally better than mice. So, and even something as simple as liver toxicity, for example, the sensitivity of mice to catching positives that are going to impact people, right? So, like severe liver toxicity events is about 43 percent. So, like, they miss a pretty good chunk. Primates only bring that up to about 63 percent on the high end, right? It could be as low as 50, 63 percent.

By growing human tissue, we have 92 to 94 percent accuracy on those metrics because it's just human tissue that we're testing in, right? And so, the ability to identify toxicities and to look at whether a drug actually works in tissue is a huge, huge increase, like much larger than even the increase from mice to primates.