Two new HHS programs to help with complex health challenges

Tomek Temin Well, let’s start with the first one, artificial intelligence. I don’t think there’s any agency grant research project today that doesn’t involve artificial intelligence, drug testing, and getting drugs approved in a 10,000-page, 10-year effort. What are you trying to do here?

Andy Kiliński Yes. Thank you again for the opportunity to be here. And you’re right, a lot of the efforts we’re making across the government have to do with the world of artificial intelligence. So what we’re trying to do with Catalyst, AI-enabled drug development tools, really increases our chances of success when drug candidates reach clinical trials. It is usually preceded by several years of preclinical research and is the case with computational models, but also with animal models and other types of wet lab experiments. This process leads to licensing for less than 10% of all drugs entering clinical trials. So we have had a lot of failures both in terms of the effectiveness and the safety of these drugs. So what we really want to do with Catalyst is create an environment where we can use computational tools to better predict outcomes in humans based on preclinical experiments.

Tomek Temin So the idea is to avoid that 90%, or at least a significant portion of the 90%, from entering the pipeline in the first place.

Andy Kiliński Exactly. And what we really don’t do well in clinical trials is representing the diversity of populations across the country. So what we’re going to be doing in the cannabis program is really going to be emphasizing bringing in baseline human data that’s representative of this country, genetically and environmentally, to be able to again model in silico with computers how these candidate drugs can affect a wider range of people.

Tomek Temin Is it just something like DEI making you feel good, or do we know that a given drug has different effects on different segments of the population?

Andy Kiliński NO. The effects of drugs vary greatly depending on genetic makeup, environment, and many different factors. Typically, in preclinical models we use animals, inbred animal models and the like, which do not have any population genetic diversity. We use these tools that have been shown to not actually reflect human physiology. Next, as we get into something like a phase one clinical trial. Usually there are dozens or even hundreds of patients. And based on the safety rating, these are healthy people. Young adults, middle-aged adults, 18 to 35 types, and typically do not accurately reflect the broader diversity of the population, both because of their genetic background and the diseases they suffer from. And that they carry comorbidities as well as environmental factors.

Tomek Temin Well, the chimpanzees will thank you for this, but it looks like there are datasets that, in effect, could be used to digitize what was once done to humans and animals.

Andy Kiliński Exactly. So we’re trying to really innovate in how we collect data from human models and human beings, but also from the wide range of 3D printed organs that we’re also investing in. Ways to actually use data from these tools to better simulate our physiology as humans.

Tomek Temin So is it assumed that what we can know about what has already happened, given all the data on effects, can be immediately incorporated into the design of new drugs?

Andy Kiliński Yes, 100%. The aim of the program is to introduce real innovations in obtaining consent for human-first research. So we do all of these experiments in advance, and then the drug developer goes to the FDA and submits a package called an investigational new drug package, which includes detailed information about the safety evaluation of its drug candidate. They are also asking for consent to participate in the first phase of the clinical trial. Typically, these experiments involve animals, various types of animal models. Little has changed in this area of ​​drug development over the past 50 years. We still use many of the same tools that we used to rapidly innovate in the early stages of drug discovery, select a target, take a billion chemicals and narrow them down to 10 potential candidates for a given target. We would really like to bring this innovation to the safety assessment as part of the drug development process.

Tomek Temin We are talking to Dr. Andy Kilianski. He is a program manager at the Agency for Advanced Research Projects for Health. It is part of the Health and Social Care Service. The second program we talked about is the computational systems needed for vaccine design, which I think also require a lot of computational power. What’s going on with this program?

Andy Kiliński Yes. Tomek, you see a certain theme here in the programs we run. My team and our portfolio are really focused on making biomedical research in development and health more predictable. So we do a lot of trial and error on these issues, especially when designing vaccines and therapies. Therefore, based on our collective experience with Covid-19 and other emerging infectious diseases, we would like to get everything needed on the shelves ahead of any future public health threat. We need solutions that address multiple infectious diseases simultaneously and do so in a way that makes these products commercially viable so that we can maintain our ability to respond to problems in the future. So the Apex program is focused on using these computational tools and generally designing immunogenic proteins in computers and then evaluating them in animal models and in Congress.

Tomek Temin And I heard during the last pandemic that someone said – I don’t remember who it was – that the ability to quickly develop the Covid-19 vaccine that we received was because this virus belonged to a family of viruses that we had already developed vaccines for. That’s why it only took a year or something. But there are 15 other types of viruses that have appeared on the scene and have been cooked as humanity.

Tomek Temin Is it true? Could the work you’re doing branch out from the viruses we’re used to to ones that haven’t yet made it into the wild and started killing people?

Andy Kiliński Yes. I’m actually a coronavirus researcher by training, so I did some of my work on the original virus in the run-up to the Covid-19 pandemic. You’re absolutely right anyway. We knew a lot about the coronavirus and the structure and function of the proteins, so we could design these vaccines relatively quickly. There are other families of viruses that we know much less about, but we generally know what this evolutionary diversity looks like. So our program is trying to climb up that evolutionary tree to find those really common denominators among virus families and target them from a vaccine design standpoint. This way we can prepare for things we already know exist and develop drugs and vaccines that will work against the various threats that exist today. But if nature throws something else at us that we don’t see today, we will have a reasonable expectation that the tools we have developed will also work against these threats.

Tomek Temin Yeah, so it’s kind of like it’s not quite artificial intelligence, but replacing something wet with something digital.

Andy Kiliński Exactly. Starting with the use of AI-enabled tools and starting with our best possible multi-target design, as opposed to the way we currently do it, which is really one pathogen for one drug or vaccine. We would like to change this paradigm in the next program.

Tomek Temin Now we have discussed two different programs. What forms do they take? These will be research grants. How to do it? Competitive with financing opportunities, that sort of thing?

Andy Kiliński Yes. So the Apex program was a competitive candidate acquisition process. We have just reached the end of contract negotiations and the program has started. And so we have five teams spread across the country, made up of academic labs, non-governmental organizations and industry partners, that are already out there, generating data and really trying to take advantage of this moonshot by building an effective vaccine or drug against everything. In the case of our Catalyst program, we just announced that it launched a few weeks ago. Therefore, this is an open invitation to participate in the program and we are actively seeking the participation of industry representatives, as well as academic partners and non-governmental organizations. These are therefore competitive processes. We look at what is happening in the biomedical ecosystem and select the best proposals we receive.

Tomek Temin And how do you know that in this biomedical ecosystem you always get the best minds? Because there is a really wide level of players. There are large and well-known pharmaceutical companies. And they have some pretty good chops. There are large and well-known medical schools that do this too. But there are a lot of startups and a lot of mid-market companies that can do something, they will rent the research and buy it because they can scale it to something they can produce, on many levels. What mechanisms do you have in place to ensure that all potential brains are reached?

Andy Kiliński I really appreciate you inviting me because this is one of those mechanisms. Yes, I think we have a challenge as a brand new agency in the U.S. government in disseminating information in marketing. What we do, what type of research we fund, and why it’s cool to work with people our age and why it’s so important to pursue targets that go to the moon. So we’re getting proposals from some of the larger companies, larger academic labs that you’ve probably heard of. But we do receive proposals from the startup world. And in fact, as part of the Apex program, we finance start-ups of biotechnology companies that lead one of such teams. So we really want to reach, as you said, every layer of research that’s going on across the country and especially in the area of ​​AI-based machine learning. Many of these companies may be developing a tool that will work for an engineering problem or a cybersecurity problem. Through our site, we would like to help companies with non-dilutive capital truly pursue biomedical and health outcomes.

Tomek Temin And for someone who is a lifelong researcher, just like you. When you come to ARPA-H as a program manager, you need to feel like you’re in the cockpit for a while.

Andy Kiliński It’s amazing work. That’s why I worked in the government for many years. I was out of government for several years. It’s very exciting to be back as an aid program manager. I have the best job in the government. I can go around and find brilliant people to work with to solve these really difficult biomedical challenges. And yes, I am very happy. I was in government and realized what a great job I had here.