Interview with Naheed Kurji, Cofounder and CEO of Cyclica

We cannot achieve the Sustainable Development Goals unless we end the ‘war on drugs’. Today we meet Naheed Kurji, Co-Founder, President, and CEO of Cyclica, a Toronto-based biotechnology company that leverages artificial intelligence and computational biophysics to reshape the drug discovery process.

Indeed, I have no formal academic background in the area, but I’ve always been interested and passionate about biology, specifically neurobiology. I entered my undergraduate studies many years ago studying biology and biomedical science. While I was studying for entrance exams to medical school, the entrepreneurial world called on me and I knew I had to pursue that, drawing me away from medicine. So I went to business school at the University of Toronto where I met with my co-founder Jason. Jason is a bioinformatics scientist, and he was already thinking, back in 2011-2012, about how to use data to prevent counter reactivity in medicines, well before there was a company around any of that idea. He put together a pitch for a case competition – I heard the compelling story and was hooked. I graduated from the University of Toronto Rotman School of Management and went to cut my teeth in banking at the Bank of Montreal in corporate finance. Around 2013, Jason and I started speaking when he was starting to put Cyclica together and invited me to join as a co-founder and CFO, I finally joined the company in early 2014.

Part of the DNA of Cyclica is also deeply linked to the early part of my career in the humanitarian sector. I worked in strategy and finance at Focus Humanitarian Assistance (FOCUS), the disaster response and risk management agency of the Aga Khan Development Network (AKDN). I spent a lot of time strategizing and working with people in South and Central Asia, East Africa, West Africa, and other remote communities in the world, thinking about education and providing them the tools and resources that the people in these communities needed to manage in the face of natural or human-made disasters. Fast forward to Cyclica, while we are a for-profit entity, a lot of the work we do now has no financial benefit; in certain situations, we offer our platform in kind, as part of our corporate social responsibility. We do not take revenue from these contracts, we do not have any economic upsides, we provide it for use and wish to make an impact.

Yes, of course. We kicked off the Covid-19 stimulus plan well before the world thought it was an issue back in February 2020. Since then, we have engaged with over 20 institutions globally, academic and biotech companies who are progressing science on covid. We have no contract, no IP, no economic benefit

We have also worked in many infectious disease areas, we are partnering with leading malaria researchers globally. My heritage is South Asian and East African and we have scientists at Cyclica that come from other developing or emerging countries, so for us, it’s our moral obligation to have that level of impact. We are trying to impact the discovery of medicines for all diseases.

Caroline Lair: Your computational approach to polypharmacology is pretty unique in the industry. You use a proteome-wide lens to evaluate multiple target interactions simultaneously, can you tell us a bit more about your approach and present one of your methods?

Naheed Kurji: At Cyclica, we think polypharmacology is critically important to understand early in drug discovery. Indeed, over the past 40 to 50 years, innovation in our space has been focused almost exclusively on the paradigm that when you understand a disease, you find the biological target that goes with that disease, and then you design a drug to interact with that target. It is really like a reductionist, let’s say narrow centric approach, to screen millions of molecules against one protein target, physically.

Now, fast forward the past 10 to 15 years, a number of companies have emerged in our marketplace using knowledge-based or AI-driven techniques to parse through the wealth of data that now exist, and make predictions on a specific molecule potentially interacting with a specific protein. What they are not doing is physically determining where a molecule and a protein interact, they’re not bringing physics together, they’re bringing AI to make predictions based on available data.

So, there is the classical technology relying on protein structure physics -molecules, proteins at atomic-level, and there is the knowledge-based AI ML -bunch of data, non-obvious association, let’s make a prediction.

There are two underlying themes that persist through our innovation effort:

So, while a vast majority of the AI companies are incrementally innovating on target-centered virtual screening, we focus on polypharmacology against the entire proteome, which means basically that we study how one drug will interact with all proteins in our body.  To do so, we have adopted a different approach, instead of going narrow, we go wide. We’ve built a proprietary database on all protein structures, there are tens of thousands of them.  We built our proprietary engine called MatchMaker™ by training it on features of molecules and features of proteins, so it can reveal how small molecules interact with all protein structures. Basically, we combine computational biophysics with AI and approach a fundamentally different problem than has classically been addressed.  As a result, with MatchMaker™, we’ve flipped the problem on its head and come at it with a different strategy; i.e. instead of focusing a drug design campaign narrowly on one target, we take a panoramic view of the proteome and bring that into drug design.

In terms of application, we are using MatchMaker™ in our platform Ligand Express® which is used to investigate the mechanism of action of a given molecule that comes from a biological assay like a phenotypic screen. A phenotypic screen is an assay that represents a disease. Molecules are screened through the phenotypic assays and those molecules that show activity are further evaluated. The challenge is that even for those molecules where activity is shown in the phenotypic screen, the underlying mechanism is unknown – that is, they do not know why a molecule is showing activity. They do not know the target that the molecule is interacting with to cause that phenotypic expression. Ligand Express® powered by MatchMaker™ is able to take that molecule and screen it against the entire proteome, to find potential proteins that it is interacting with, in seconds. Ligand Express® brings a wealth of downstream information of systems biology data, genomics data, converting them into interactive maps so we can start contextualizing the disease areas or the genes – as such it is effective in elucidating the mechanism of action from a phenotypic screen.

Really good question. So, when there is a disease, the first goal is to find the target to get the drug, and that target is generally the proteins. You and I have the same protein structures but there may be mutations within our proteins that are different due to our genetics. Now, once you have designed the drug for a protein structure, it may be only conducive to a certain population because within the binding site of the protein there may be mutations within a certain population group compared to another. We’re talking now of pharmacogenomics. As I mentioned previously, at Cyclica we are interested in designing drugs that harness this genomic data and contextualize it across the proteome – we call this structural pharmacogenomics. With our technology, we’re able to provide insight into how a particular gene within the pocket of a protein may have deleterious effects on the way in which the molecule binds there for a specific population.

It can also have a more immediate impact at the preclinical level. When you go from a disease to the protein, to a drug, the next thing you do is putting it in an animal; that is the classical paradigm. We have multiple species, rats, mice, etc, Each one has its own genetic profile, so within the world of structural pharmacogenomics, you can consider interspecies differences which is useful to know when considering what animal model to proceed with. So yes it has an impact on population but it also has a more immediate impact in terms of preclinical animal studies species to get to the right animal model much faster.

That’s really the difference between how we’re thinking about drug discovery than others as we’re taking a much more holistic view, we are considering the steps to bring a molecule, turn it into a medicine beyond just identifying a molecule for a target.

Because of our proteome-wide approach, we become disease agnostic in a sense. We have examples in neurodegenerative, neuropsychiatric, oncology, rare and infectious disease In terms of speed for a drug design campaign, we can carry out an in silico drug design campaign in days. When chemical synthesis or acquisition of the chemical compound is factored in alongside the experimental biophysical and functional tests, we can generate advanced lead-like molecules in a couple of months and at a fraction of the cost. We’ve also created a company with ATAI Life Sciences in neuropsychiatric diseases. We turned on our engine, and in a matter of a few months, we went from the identification of targets to very exciting data in a very short period of time.

Yes, we’ve created dozens of programs and a number of companies via spinouts or joint ventures. This model helps us move faster, do more, go wild and hopefully have a bigger impact.

There are 2 types of business models in our industry. Those that license their technology to pharma, or get into a collaboration with upfront fees and potentially massive downstream milestone payments, and possibly royalties on the back end. Then, there are those companies that apply their platform to a specific use case and become a disease-focused biotech company. In the first scenario, you don’t capture that much value because you don’t own the IP. In the second scenario, you own the IP but it is very risky because the chance of success is 5-7% and it is very costly to go down that path.

So we brought together these 2 approaches as we do not want to invest in only one program because that is not doing justice to the platform. We’ve created companies, each one is a biotech company, we contribute our technology, capital, and people in the creation and in the advance of these companies. We’ve created a diversified portfolio of assets that are owned by separate companies, all of which we own a substantial stake in, that have their own management teams that operate. And, by virtue of that, we’re able to scale across many diseases, we’re not just confined to one thing.

We’ve raised a CA$23M Series B in June 2020, grown our team by 2 times over the past 20 months and we’re now close to 60 people.

In the next 5 years, our vision is to scale out the biotech pipeline of the future. We will continue working with multinational pharma companies, building relationships and brands, cultivating a deep affinity towards scientific integrity, such that when we have a portfolio company that is at an advanced stage, we can position with our pharma partners and present a strategic opportunity. Basically, by working at both ends of the market we’ve created a marketplace and we’re the link between the two. So, we create, we position and then we partner off.

Big pharma is really good at taking molecules and turning them into medicine, but it’s with early-stage hyper innovative biotech companies where innovation in drug discovery will be found. So, we’re going to take that on. We will create hundreds of programs that will be held by a few dozens of companies that we will own. A portfolio approach like this allows us to mitigate downside risk, and given that we will be at the forefront of the innovation strategy on the back of our platform we’ll be able to capture substantial value.

As we progress the biotech pipeline of the future, we’ll be able to contribute our platform to advance therapeutic response for diseases that go unnoticed due to a fallacious economic model that puts profits over people. While Cyclica is a for-profit organization, we are going to do both: we will create value for our shareholders and we will also ensure that we have an impact in other areas where the economic impact isn’t our profit. People and profits can and should go hand in hand – it doesn’t have to be a tradeoff.  This is really part of our ethos, it is how we are operating.

So, right now, our talent strategy consists of finding the best drug hunters in the world. People who are disenfranchised by the traditional way of doing business and see an opportunity of partnering with a company that is as enabling, flexible, and visionary as Cyclica, to jump in and use our technologies, to go find new opportunities, and to progress it fast. What is required is a deep level of empathy and complexity of the sciences and also experience in bringing molecules down the screen.

So the next wave of folks, our Chief Strategy Officer, Melissa Landon joining us from Schrödinger, our Chief Partnership Officer Vern De Biasi, joining us from GSK, are people that have that experience from other institutions but are looking to have an impact in a different context. We’re also always looking for more senior leadership, senior experts in oncology, in infectious disease to help run companies that we are creating there.

I’m pretty optimistic. There has been substantial progress in the last years. Thanks to public policy, education, and the impact of drug discovery and drug development technologies. I do believe that with a concerted effort and continuous support from institutions, like the Bill and Melinda Gates Foundation, ongoing education, and investment to support AI-powered biotech companies such as Cyclica, we absolutely can achieve those goals.

Latest news about Cyclica:  Cyclica Teams Up with Top-Tier Academic Institutions to Identify a Repurposed Drug for COVID