Cancer is one of the first areas where precision medicine was applied in a clinical setting. Today, particularly in developed countries such as the U.S. and U.K., it is becoming increasingly common to undergo precision diagnostic testing and subsequent targeted therapy depending on the tumor mutations and type of cancer.
It is widely acknowledged that targeted therapy guided by suitable diagnostics has a beneficial impact on patient outcomes and can more than double survival times, depending on the cancer and treatment in question.
While there have been large steps forward in cancer therapy, there is much left to be achieved and a pressing need for new and innovative startups to address unmet needs in the field. For example, Zahra Jawad, PhD, founder and CEO of Cambridge-based biotech Creasallis, noticed an unmet need to improve the efficacy of antibody therapies used for treating different cancers and to reduce side effects from these drugs.
In contrast, London-based techbio company Concr, led by CEO Irina Babina, PhD, is taking a computer-led approach and developing analytical tools to predict the effectiveness of potential cancer therapies both at the drug-development stage and in patients in the clinic.
Both CEOs spoke to Inside Precision Medicine senior editor Helen Albert about their journeys and what they are trying to achieve in their respective companies.
In conversation with Zahra Jawad, CEO and founder of Creasallis
Q: Did you always want to be a scientist?
CEO and founder, Creasallis
Definitely! I have always wanted to be a scientist ever since I was 11 years old and started doing science at school. It was my science teacher who really inspired me. He thought I was good at science, and he would always encourage me to do more projects on the side through the summer holidays. I only ever wanted to do science, so that means I essentially ignored every other subject!
I got to A levels, and my studies were still too broad. I still wanted to be more specialist. It was only during my final year at university that I felt like I was finally doing stuff I really want to. And then my PhD at the University of Cambridge was even better, because I was finally doing what I really wanted to do.
I was interested in proteins and how you can engineer them from an early age, for example, by changing their structures by just one or two amino acids. It was a totally fascinating subject for me. The more I could engineer them myself, the happier I became in my education. I just love the magic of the lab. I mean, it still fills my heart with joy every day that I see a gel!
Q: How did you decide to move to industry from academia?
My first job in industry was at a biotech called Domantis, which was acquired by GSK in 2006. I always thought I was going to be an academic, a researcher or a lecturer, that’s what I was aiming for. I finished two postdocs, and then I applied for loads of postdocs and fellowships, and then I saw this one job in industry, and I just thought, ‘Okay, I’ll apply for that as well, because I’m at a crossroads. I’m going to apply for everything.’
The only job I got was the one in industry. I thought it was going to be a boring, monotonous job. I remember when I applied, the hiring manager said to me, “You won’t regret this.” And he was right, because I loved it.
I like the industry way of working together to achieve your aim. When you’re a postdoc, you’re doing your own cloning, your own expression, your own testing of your project. It takes you years to get some data, whereas in industry, it’s set up in such a streamlined way that we’re getting data back in weeks. I like the sort of camaraderie that you get in industry, which I think you don’t get in academia, and developing things that are actually really useful to humanity.
Q: Why did you decide to launch Creasallis?
I was head of an innovation group at one of the other companies that I’d worked for, and I had a young woman in my group who developed a stage four lymphoma. One day, I think she was on her fourth cycle of treatment, I said to her, “Look, can I come to the hospital with you?” I took the day off work, and I went with her. She was taking an antibody treatment as well as chemotherapy.
I spent the whole day in the hospital seeing how bad it was and how the drugs we were designing were actually getting to the patients. She told me, “It’s the antibody that’s really bad, not the chemotherapy. It’s what gives me the side effects.” When I looked around there were other people taking these therapies as well and it was the same thing, when the antibody infusion came in the patients got really sleepy, really nauseous, and that’s what was most toxic, which surprised me.
I was always told “We’re close to curing cancer, we’re almost there” and that was when I realized we’re not even close. When I looked into it more, I realized that the actual antibody doesn’t get into the tumor very well. It always just accumulates around the blood vessels.
We knew quite early on from Domantis that making antibodies smaller does make them penetrate organs really well. But because they’re so small, they also filter from the kidney really quickly. So, we can’t just make things smaller, we’ve got to keep them in the serum and get them to bypass the kidney.
Proteases cause havoc in cancer. They start overexpressing in the tumor, and the cells are able to escape into the blood vessels and they metastasize. We were designing antibodies to block them, and I thought, “What if we use the proteases to break antibodies into smaller parts?” Then you would get a size improvement, but only inside the tumor. It wouldn’t happen anywhere else because these are only expressed in the tumor.
Q: What was founding the company like?
I spent three years going through a cycle of thinking about this idea and not telling anybody about it. Then I was headhunted for a company called bit.bio, which has nothing to do with antibody engineering, as head of research project management. I got a project management framework set up but missed the science after a year.
I was getting more and more anxious, and I felt like I was in the wrong role. A friend said to me, “Well, Babraham’s got this startup competition, and you should apply. If you get a place, they’ll give you £10,000, and it might just be useful to get some feedback on your idea. See what they think, and maybe they’ll help you figure out what investors you need to talk to.”
I filled out the application form. It said at the bottom I needed a Companies House number, so I paid £15 and set up a company. Then I won a place, I was so surprised! They taught me how to start a company. I had absolutely no idea, I didn’t have a bank account, I didn’t know how to pitch, they taught me everything from scratch.
I was very open with bit.bio, and I told them, “Look, I’ve applied for this, and I need to satisfy my scientific curiosity.” They were super supportive. We left on really good terms in the end. I got offered some seed investment and I quit my job the next day to start Creasallis properly in 2022.
Q: How have things gone since then?
The founder’s journey is a tough one. It’s a huge amount of responsibility, especially when you start hiring people. It can be quite a lonely journey as well, because nobody quite understands the passion that you have, and nobody’s on that same level as you in the organization. It does help to connect with other founders.
I think that the biggest learning for me is that it can be so easy to just throw yourself into the company that you forget about yourself. Forget to eat well, forget to exercise. You kind of neglect yourself and it’s best not to do that. We’re now five people, still based at the Babraham Research Campus near Cambridge. I’ve tried to purposefully keep the team as lean as possible. I’m a big believer in getting high-quality people that can just get things done. I’ve kept the team purposefully very small. I’m still in the lab, even though I’m the CEO. It’s a really important part of my happiness and my productivity. I don’t do the major stuff anymore, but I do maxi preps and other non-critical stuff. I want to be in there every day.
Since we started, we managed to prove that we can re-engineer antibodies with protease, and they cut inside tumors and improve how much they penetrate into tumors, all in a mouse model, which has been fabulous.
We incorporate these tumor-specific protease sites into the antibody, so that they’re fine in the serum when they inject it in, but as soon as the antibody goes into the tumor, there are these proteases inside it that chop it up into smaller parts. It’s those smaller parts that are now able to diffuse inside the tumor. It really is a very simple modification of the antibody itself, rather than making a complex sort of drug or a formulation, or a delivery mechanism.
We’re now gearing the team up to apply this to our first indication, our first asset. We’re at this point scrambling to plan what that’s going to look like and what that data package is going to look like for the next two years.
Q: Do you think you will pick a pre-existing antibody treatment to modify or design a brand-new therapy?
I’m more inclined to the former rather than the latter. I do think there’s been a lot of great antibodies that have shown such promise in the pre-clinical stages and then failed in clinical trials because they were too toxic for the patient.
I think there’s an opportunity there to repurpose drug candidates, rather than coming up with our own novel drugs. To try and improve those ones that were doing so well preclinically. We are interested in going into the antibody drug conjugate (ADC) space as well, because they are also very toxic. We’re not planning to come up with brand new antibodies with brand new targets. That’s not what the focus of the company is. We’re more looking at improving things that have failed in the clinic, to try to get them through the clinical stages again.
There’s so many people looking for that magic cancer target that’s going to be as good as the anti PD-1 drugs like Keytruda [pembrolizumab] and it’s been a struggle. Every company I’ve worked at has a target discovery group where they’re also looking for better targets, and it’s been a struggle for everyone to find those targets.
I think we do need to start getting creative about what we already have and improving them. Keytruda only benefits 20–30% of patients. That’s the overall response rate. There’s still 80% of people that cannot tolerate Keytruda.
Q: What are your future plans?
The part of the antibody we are engineering typically hasn’t been engineered before, so at the moment we are swimming in our own lane. I don’t have plans to exit the company anytime soon. My plan is to build as high a value company as I possibly can and see where the environment and the people on our journey want to take us.
Moving forward, there’s going to be so many more people that come on board, who all will help us build something. I’m not here for short-term gain, I am really here to bring creative solutions to solve these antibody discovery bottlenecks that we have. For example, the blood brain barrier is still a big problem. Getting antibodies into the brain is a challenge. In so many disease areas we see limitations with the current drugs that we have. There’s so much scope for improvement and that’s the philosophy of what I’m trying to build.
In conversation with Irina Babina, CEO of Concr
Q: How did you end up where you are today?
I trained as a geneticist and spent about 12 years in academia developing targeted therapeutics against oncogenic mutations in breast and gastric cancer. I trained at the College of Surgeons in Ireland, and then did postdocs at the Institute of Cancer Research and the Royal Marsden Hospital, and then I went into funding management.
I decided to move from doing science to funding science, and that’s why I went into government funding. I helped to deliver about £120 million in government funding via the U.K.’s National Institute for Health and Care Research into research programs at the Imperial NHS Trust and Imperial College.
That was fascinating, because you got to see really cool ideas come to life, and being able to actually support it was great. During that time, I did an executive MBA and ended up in venture capital consulting.
I consulted for Concr at the start and then I got more and more involved. The company was founded in 2018, but it didn’t really get going until 2021. I realized it was a really great company that I wanted to be a part of. Then I went full time, and I eventually succeeded the CEO in January this year.
Q: What made you want to move from academia
into investment?
I saw how much good science died. We’re always told as academics that industry is the dark side, but ultimately, we all want to make drugs or help target drugs to the right patients and nobody has the scale and the expertise and the cash to be able to do that in academia. You have to collaborate with industry or other organizations. So I thought, okay, what makes good science go around is not the science itself, it’s the funding behind it. In order to influence that, I needed to go and work in funding. I started with government funding, because that was the opportunity at the time. Then I went into private funding.
Q: What has it been like being a CEO for the first time?
I must say, the team has been really supportive. All the founders are still with the company, and we have been working really closely together. The transition was also very smooth and was happening slowly for quite a while. But it’s certainly one thing when you invest and another when you are actually in the trenches, building the venture from the ground up! They are definitely different things, and I’m humbled by the experience.
Q: Do you think your experience on the investment side has helped you to be a better CEO?
Not necessarily a better CEO, but it really helped me assess how to deliver a message. I’m continuously working on this, because we’re talking about astrophysics, medicine, and biology, but being able to talk about science in business terms really helps. So I think I definitely benefited from the experience.
Q: What attracted you to working at Concr?
What really attracted me to Concr, which is headquartered in London, is that I don’t like waste. That’s one of the reasons that I moved out of academia into funding. What Concr has done is adopted a technology that already existed in a different field but applied it to biology.
In biology, the data is really fragmented. It’s all over the place and each one of these data parcels and datasets carries information about responses to a therapeutic. What Concr was founded to do is predict an individualized response to cancer treatment. To be able to predict this, you need to model a person’s biology, because we’re all fundamentally different at a molecular level and you need to be able to model interactions between so many different things and take into account learnings from so many datasets. So, it’s not just about real-world data or big data. You need to be able to connect it all and apply it in a particular setting.
The company founders realized that we need to be able to deal with a lot of uncertainty, because we don’t understand enough about biology. We need to be able to interpret it, but we don’t understand enough about it. This was the very problem cosmologists faced when they were studying the distribution of dark matter to discover black holes. They can’t observe them directly, but they can observe the effect of that invisible thing on the surrounding environment.
For cancer, it’s exactly the same. You can’t observe it directly, but you can measure different things about it and infer properties about it. So that’s why Concr adopted methodologies from astrophysics and adapted them to cancer biology. To me, that was really appealing, because it’s not just convergence of science, but is also saving so much time. Rather than developing conventional machine learning algorithms that have a lot of limitations, you leapfrog that development process.
Q: What is Concr trying to achieve?
We are a predictive analytics provider. We create tools to predict the efficacy of cancer drugs, now in a preclinical setting, using preclinical models in patients in the clinical trial space, but we’re aiming to eventually penetrate the clinic as well.
We are positioned as a platform. We have a predictive engine based on machine learning, but using a Bayesian approach, as opposed to conventional foundation models. Hence the uncertainty distribution element, so that predictive engine is packaged in a software piece called FarrSight named after William Farr, a statistician. This is to enable a biologist to interact with their data and the results of our predictions. Because biologists don’t just take things at face value. They like to drill into the data, understand the data.
It’s a little bit like a comparison between a paper map and GPS. If you want to get from A to B, through a dense forest, and you have a paper map you follow the map and that’s great. You will get from A to B because it’s been done before. But, if you encounter, let’s say, a fallen tree, you might have to go back and reassess and see where you need to go next. With a Bayesian approach, it happens almost in real time. You start off, you will assess the data that surrounds you right now and you will follow the most likely scenario at every step of the way. That’s where the unknown bit comes in. You’re taking into account the data that’s available to you at a given moment. That’s why it makes our approach quite efficient in that you don’t have to retrain the model every time and we don’t require large data sets to work with.
The uniqueness is that as a biologist, you can interact with a really sophisticated system, ask it questions, and then infer kind of results yourself. At this point in time, there is an element of collaboration between us and the partner or the client, because they know everything about their drugs. There is a lot of interaction and tailoring going on there, but it is a hybrid model in that it’s not a purely tech read out. It is a prediction that’s made, delivered via a software platform, but it’s also put in a context by our clinical and scientific experts. So you get a service, and a SaaS, and AI as a service all in one.
Q: How can your technology help cancer drug developers?
We use the minimal amount of research data that the client provides us with to make predictions. To give you an example, they have an emerging small molecule that they’ve tested in ten cell lines. They will give us that data, and we will make a prediction based on the structure of the drug and the little data that they give us about future efficacy in all available cell lines, or the cell models that we have access to, in order to identify what the biology behind response across all cancers is. We would then make a prediction of what would be the best indications for this particular therapeutic.
In a way, at every step of the development pathway, we’re predicting what the most likely scenario will be for the next step. So instead of the company carrying out all of the experiments, they’re kind of guided through that process.
As soon as there is a therapeutic, or a structure of the therapeutic, we can already start making predictions. The minimum data required is the structure of the drug. At the moment, it is small molecules and ADCs only, we don’t do immunotherapy just yet.
The biggest gap at the moment is moving from pre-clinical testing into patients. This is, of course, what everybody struggles with. But at least we can make a prediction with broad confidence intervals. As soon as there is any clinical data, those intervals shrink, and then we’ll be able to make predictions about how to identify responder populations and the biology of response, or the biology of resistance.
Q: How do you hope to reach patients in the clinic in the future?
What we want to do is be able to predict an individualized response to all available treatments in order for patients to be put on the best first-line treatment for their disease.
What happens at the moment is all based on average response rates. Everybody gets the same. Only about 30% of cancer patients respond to first-line therapy and what we want to do is really challenge that. This is no mean feat, especially for a small company.
As an individual patient, do you want to know what’s good for you, or what’s good for an average population? To be able to not just match patients to treatments based on clinical characteristics, but to be able to model their individual biology to make that decision more accurately—that will transform cancer care.
We currently have a prospective clinical trial running at the Royal Marsden Hospital, funded by Innovate UK, but we will need a lot more of that evidence before we can get into the clinic, and that is okay. The path to the ultimate goal is long, heavily regulated, and expensive, but at least we can already see that regulators are jumping on board. One example would be the FDA-led Project FrontRunner. That’s a project where they want pharmaceutical companies to test their cancer drugs in earlier lines of therapy.
Q: Where do you see this field going in the next five to ten years?
I do think that these solutions, including our own, apply in other fields. I think we will become a lot smarter in the way we develop therapeutics, or in how we identify patients. Our technology can also be used to review large datasets of failed assets to repurpose them for other uses, for example.
The rare disease space is an interesting one. More data doesn’t necessarily mean more informative outputs or more insights. With linear machine learning models, you will hit a plateau of predictability. So actually, you need diverse data. Rare diseases, especially rare cancers, are a wealth of biological information that we don’t use. The focus of pharma and biotech is the “big four” cancer types [breast, lung, prostate, and colorectal cancers]. But that other group is so informative in terms of biology of response that we have to use it.
With Concr’s technology, we can make predictions on a smaller dataset. This is where I think using this kind of technology will really add value. It’s not just going to be analyzing loads of data quicker. It’s going to really give useful insights.
Ultimately, it’s about patients and about improving outcomes. There are some good drugs out there—patients just don’t get them, or don’t get them at the right time, and we want to help change that.
Helen Albert is senior editor at Inside Precision Medicine and a freelance science journalist. Prior to going freelance, she was editor-in-chief at Labiotech, an English-language, digital publication based in Berlin focusing on the European biotech industry. Before moving to Germany, she worked at a range of different science and health-focused publications in London. She was editor of The Biochemist magazine and blog, but also worked as a senior reporter at Springer Nature’s medwireNews for a number of years, as well as freelancing for various international publications. She has written for New Scientist, Chemistry World, Biodesigned, The BMJ, Forbes, Science Business, Cosmos magazine, and GEN. Helen has academic degrees in genetics and anthropology, and also spent some time early in her career working at the Sanger Institute in Cambridge before deciding to move into journalism.
