A Fireside Chat with Alex Zhavoronkov, founder of Insilico Medicine, on life, death and ageing

October 12, 2022

4

mins read

Let's start with Insilico Medicine's mission and your story - what led you to found the company? 

I was historically very interested in ageing and longevity because nature is very unfair and does not allow for continuous improvement. You're born, you grow up, you reproduce, you reach your peak, and then you continuously decline and die. I'm actually fine with dying; I'm just not fine with declining! When I was growing up in Eastern Europe during Perestroika, there weren't many opportunities to do science to change this, so I did my grad work at Johns Hopkins and Moscow State University in parallel. I worked for multiple companies, supported multiple projects, and participated in conferences. Then, around 2014 I was in charge of my first big lab, and I decided that I wanted to go full-scale into AI-powered drug discovery and founded Insilico Medicine. We started offering AI services to a large number of pharmaceutical companies and developed a platform that allows you to identify targets for a variety of diseases, and also to generate small molecules with the desired properties, and predict clinical trial outcomes. 

"Im actually fine with dying; I'm just not fine with declining!"

We realised quickly that pharma wants validated, clinical-stage assets. That's when we realised that if your molecules are good, they are also very valuable, and if your targets are novel, those molecules are even more valuable. So we decided to focus on our own pipeline rather than help pharma develop theirs! That's when I met Dr Ren, who joined me as Chief Science Officer, and we started prioritising our pipeline above everything else and started setting, to my knowledge, world records in the speed and cost of drug discovery. 

Most of our pipeline targets disease and ageing at the same time. We call these dual-purpose therapeutics. Our first programme, in fibrotic disease, was discovered using ageing research and AI and is now in phase 1. I think that's a really cool model because targets are often very heterogeneous, and the patient might respond weakly. With a dual-purpose therapeutic, even if you miss the target for primary indication, you will still hit the ageing target and provide substantial benefit by addressing the ageing pathways. So that's kind of my modus operandi. I want to extend human life, which you can do both by treating disease and the underlying causes of ageing. 

"I want to extend human life, which you can do both by treating disease and the underlying causes of ageing. "

Do you think traditional pharma has been slow to research and develop longevity assets? Why?

 

Pharma historically has been focused on diseases because we usually do not really care about our health until disaster strikes! But once you get a disease, that's when you start fighting it, and the pharmaceutical companies are addressing a problem that is top of mind. 

There have been several attempts in the past few years by pharma to go after ageing. For example, GSK acquired Citrus Therapeutics, a metabolic diseases company focusing on ageing. However, that exercise failed because the molecules and targets were of poor quality. So, that $650 million plus acquisition was not exactly successful.

Novartis tried something similar with ResTORbio, which failed predominantly because they changed the drug substantially between phase 2 and phase 3. Here and there, you will see pharmaceutical companies trying to prioritise ageing in one way or another, but there has never been a centralised push or common theme in the pharma industry to go after ageing. 

One of the reasons for that is that there is no quick money in ageing; it's not reimbursable at this point in time; even if you develop a compound that prevents disease, it's very difficult to get reimbursed and also very difficult to conduct clinical trials on. In general, there is a pretty standard approach for developing therapeutics, and they are pursuing the standard approach. You do not see pharma companies often thinking out of the box; this is the job of biotechs!

"You do not see pharma companies often thinking out of the box; this is the job of biotechs!"

Have longevity or regenerative medicine trials historically been difficult? 

It can be very difficult to design a pure longevity trial. If you want to test a therapeutic for just ageing, you have to define the endpoints. You also need to work with healthy volunteers basically, in phase two and phase three, where you're going to be testing efficacy, but what are you going to test? 

One of the most pioneering ageing trials is the TAME (Targeting Ageing with MEtformin) trial by Nir Barzilai and his collaborators. They looked at the onset of different diseases, and they measured ageing biomarkers. Nowadays, there are many pretty established biomarkers of ageing, and you can use an ageing clock as a biomarker. Other longevity companies may start by targeting a specific disease and then collect data that can be used to make an ageing claim secondarily. For example, collecting proteomics data or epigenetic data, or even just regular blood tests, and then running those ageing clocks or other ageing biomarkers as an auxiliary endpoint.

What longevity targets are you most excited about at the moment?

Well, many of the targets were identified by my team! We recently published a really fun paper on 145 dual-purpose targets that are implicated in ageing and specific diseases. We also evaluated the novelty of each target and its druggability. So we created a target wheel and even created a dartboard where you can throw darts at some of those targets and create a portfolio of possible longevity therapeutics. Pretty much every target on that dartboard is exciting! Ok, if I had to pick one, it is PEG 2 HIF-1 Alpha, currently in IND enabling for anaemia and for IBD for Inflammatory Bowel Disease. I know that BioAge has in-licensed compounds for the same target (PEG 2). The fact that two companies identified one target and prioritised that for ageing gives you a little bit more confidence that this target is important. I think that PEG 2 HIF-1 Alpha is going to be a very famous longevity target.

"I think that PEG 2 HIF-1 Alpha is going to be a very famous longevity target."

You've said before that Asia is leading the way in longevity research and clinical development. Why do you think that is? 

Biotech in Asia is very similar to software in the early 2000s - a lot of the hard engineering and science is being carried out here. Many of the leading longevity CROs developing assays and biomarkers to support the longevity ecosystem are here in Asia. And, of course, we at Insilico Medicine are partly an Asian company; we are entrenched in Taiwan, Hong Kong and Shanghai. It's a huge community, a huge platform, and everything is nearby. 

Are you excited about the potential impact of decentralised drug development and DAOs (e.g. Vibe Bio, Molecule)?

I think the concept is really promising, but we haven't had enough time to see results yet. Unlike many general crypto projects, which don't tend to attract people with credible scientific profiles, VitaDAO has some very credible scientists on board. I think the concept could still be taken to the next level. Patients themselves are very interested in being involved in clinical studies and incentivised with cryptocurrencies that would allow them to share in the experimental results and value created by those trials. We actually published a paper in 2018 on converging blockchain and AI to decentralise biomedical research and clinical trials. We also came up with the concept of live data economics with George Church and published a paper on this. However, this idea has not really found a good home because of the crypto hype and crypto bust cycle, which affected the credibility of the entire field.

The impact of machine learning on drug discovery is well known. Do you think machine learning can have a similar impact on accelerating clinical trials? 

Well, I would actually rephrase this a little bit. AI is being very actively used in clinical research and was really the first low-hanging fruit application of AI. Many companies are using large datasets of patient data to stratify patients, identify patients and even build actuarial models. We have a tool called inClinico, which predicts the outcomes of Phase 2 to Phase 3 transitions. But it's not easy to develop as, of course, it requires a long time to validate predictions. You have to wait not only for clinical studies to be complete but for the results to be published. 

However, in early-stage discovery, I think that only very few companies managed to demonstrate for real that AI can substantially accelerate and derisk the process. Currently, there are very, very few cases where AI has been proven to accelerate and derisk asset development at the early stage. There are only maybe five AI-designed assets past pre-clinical. One of these companies is, of course, Insilico Medicine, and others like BioAge and Recursion. Others that focus on chemistry, like Exscientia and Relay Therapeutics, take a very old established target and make better small molecules for hitting that target. 

"Currently, there are very, very few cases where AI has been proven to accelerate and derisk asset development at the early stage."

At Insilico, how do you design clinical trials to ensure there is a data feedback loop? For example, how do you ensure data from a clinical trial will be useful for drug discovery? 

We mostly do this at the pre-clinical stage. At pre-clinical, data has so much more granularity; you get gene expression analysis, proteomics, enzymatic assays, and of course, you can do so much more with mice than you can with humans. At the clinical stage, the potential to refine the molecule is limited, but you can still adjust it, so it better fits the target. We try to study the pre-clinical data as much as possible to prepare for phases 1 - 3. We work with huge amounts of patient data and companion biomarkers to ensure we're stratifying patients by their ability to respond.  

You are an Angel investor in Lindus Health. Why did you invest?

Well, I invested because Lindus Health is cool! (Your branded socks are in the mail, Alex! - LH). I think the industry can handle a lot more disruption. There are not that many players innovating in the clinical trial stage, but there's a lot of excitement around companies like yours in Asia, where next-generation CROs are attracting large multiples. People in the US and Europe may not understand the potential of the field yet, so you are ahead of the curve. Being able to execute really solid, fast clinical trials is a no-brainer that any drug development or biotech company would like to have. And I got to invest alongside my friends at FirstMinute capital and Brent Hoberman. I see many, many companies every week, and we very often have to make decisions, even at Insilico, about which biotech to work with, so you train the AI over the years to find people with and companies with a higher probability of success. So that's why I thought that Lindus Health was a pretty good bet.

"I invested because Lindus Health is cool!"

What advice would you give to someone early in their career who is considering entering the longevity field? 

Really carefully choose a mentor, try to meet as many people as possible at conferences like ARDD to understand the work ongoing in the field, but make sure to specialise in something concrete, for example, extracellular matrix stiffness, DNA damage repair or epigenetic drift. And it goes without saying to stay in research early in your career and consider a PhD! We have a programme called Inspire Longevity where high schoolers can interview Key Opinion Leaders in the field and apply for internships in leading longevity labs. Recently several participants presented at the ARDD conference, using our PandaOmics tool to complete a target discovery exercise! The students presented at the professor level; it was seriously impressive. It goes to show there is huge enthusiasm for ageing research in younger generations, which makes me hopeful for the future of the field! 

Thank you, Alex!

View more