Working With Patients To Use AI Against Rare Diseases
An Interview With Tim Guilliams, CEO of Healx
Executive Summary
Tim Guilliams, CEO of Healx, talks about the company’s mission of creating from existing drugs new options for rare disease patients. The company is using artificial intelligence to find combination treatments against rare conditions, while working closely with patient groups as key partners in the development process. The goal: faster development timelines and cheaper therapies for patients.
Tim Guilliams, a 2020 In Vivo Rising Leader, co-founded UK-based Healx Ltd. in 2014 with a mission of developing affordable treatments for rare diseases. In October 2019, the artificial intelligence firm raised $56m in a Series B financing round, led by Atomico, one of Europe’s largest tech-focused venture capital firms. As CEO, Guilliams has grand plans for the future of Healx, as a tech-driven drug developer, not a service provider.
“We started the company in 2014 with a few people around Cambridge University,” Guilliams told In Vivo in a recent interview. He completed a PhD in the university’s chemistry department. Alongside David Brown, now chair of the company’s board, Guilliams formed Healx around the idea of using machine learning as an engine for drug discovery.
Tim Guilliams
“We me a chap called Nick Sireau, who has two children with rare diseases. He lives in Cambridge. When we met him, he was trying to repurpose a component in a weed killer to save his children,” Guilliams recalled. “We were just amazed by how ingenious he was … and so that's how the different pieces came together and how our mission came about.”
Guilliams and Brown are both passionate about the idea of repurposing drugs in combination approaches, using the speed and capabilities of artificial intelligence. Sireau presented them with an unmet therapeutic need they believed could be targeted with the technology at Healx’s core.
Healx has around 55 employees but it functions more as a tech company than a typical biotech. The business is completely virtual, and many of its staff come from technology backgrounds. While not having any of its own labs, Healx works with contract research organizations and academic lab partners. “It is a really lean biotech company,” Guilliams noted.
The Technology
There are two main parts to the Healx platform. “On the one hand, we analyze a lot of information captured in text,” Guilliams explained. “Then we build what's called a biomedical knowledge graph.” A knowledge graph is a database relationship, like the technology Google uses for a search. “You search something on Google and you almost always get the best answer in the first three suggestions. We are training a knowledge graph for rare disease.” The technique behind this approach is natural language processing.
Healx’s knowledge graph technology, Healnet, integrates and analyzes biomedical data from multiple sources. Healnet has proven success; it has been able to identify eight repurposed drug candidates with potential benefits for fragile X syndrome, which were not considered for the condition before.
The second aspect of Healx’s platform is focused around data and computational biology. “Looking at your genome, transcriptome, metabolome, the list continues. You use these data to match the drug to the right disease,” Guilliams said.
The company is focused on finding treatments for rare diseases as cheaply and as quickly as possible. “The aim is to take a drug and have it ready for the clinic within 24 months of partnering a project. We look at existing drugs, but more importantly we look at drug combinations,” the CEO said. “If you look at combinations of two or three drugs, you have more than 12 billion possibilities by diseases, which is a huge drug base. That’s where we focus our computational capabilities. We believe you are much more likely to have an effect on a disease with the ability to combine two or three modes of action.”
There are a couple of other companies developing this kind of AI-driven discovery technology. One example is Oxford, UK-based Exscientia. Though Guilliams says Healx’s approach is different. Exscientia, for instance, is more focused on novel targets and using AI to find new drug candidates. “With our focus on rare diseases, and our machine learning approach,” Healx is more unique, said Guilliams. “I don't think we have a direct competitor.”
Another AI drug developer, Utah, US-headquartered Recursion, is also working in the area of rare diseases, but it has a different machine learning approach from Healx. Recursion has built a proprietary drug discovery platform that combines high-throughput biology and automation with AI technology.
“There are 7,000 rare diseases and 95% don’t have a treatment,” Guillaims said, noting the huge unmet need. “The only way to try to solve that problem in a more scalable, cost-effective way is to apply machine learning, but also be able to connect drugs and combinations.” Healx’s chief executive believes business models for drug discovery are ready for change. “The current cost versus the probability of success is not sustainable, and the current drug prices seen for rare diseases will not be sustainable. The industry needs to figure out a way to lower that cost and shorten development timelines significantly.” He said Healx was able to build a sustainable model because it had focused on existing drugs alongside data and technology to drive the business forward in a cost-effective way. “There is a huge opportunity here for patient impact but also for return on investment,” Guilliams said.
Guilliams does not believe AI should, or needs to, replace all the other discovery tools. “There are amazing tools being developed.” But whatever tools and approaches are used, the critical point is to be able to capture the data in a feedback loop, to enable drug discovery as a function to become smarter. “You will get to a point where, hopefully, you'll have a much higher likelihood of success in your clinical trials, in your preclinical development, but also a much lower cost basis and a much more scalable approach … In drug discovery, it is about integrating these tools and really making the most of the information that is available.”
Working With Patients
For every project it starts, Healx partners or works with a patient group. Healx’s most advanced program is for fragile X syndrome, a genetic disorder that causes a range of developmental problems including learning disabilities and cognitive impairment.
Within this program a patient group helped the company better understand the unmet need, find key opinion leaders and get in touch with the right clinicians to discuss clinical design and to be able to recruit patients. “It is a real partnership” with the patient group, Guilliams said. “We need to work with patient groups, they are invaluable.”
The company had been preparing to start a clinical trial in fragile X patients in April 2020; however, that study has been delayed due to the COVID-19 pandemic. Healx’s pipeline covers eight other rare diseases (see Exhibit 1).
|
Indication |
Development Stage |
Therapeutic Area |
|
Fragile X |
Phase I Ready |
Rare Neurology |
|
Pitt Hopkins |
Preclinical Trials Completed |
Rare Neurology |
|
CDKL5 Syndrome |
Preclinical Trials Completed |
Rare Neurology |
|
Neuroblastoma |
Preclinical Trials Completed |
Rare Oncology |
|
Barth Syndrome |
Preclinical |
Rare Metabolic |
|
4 Undisclosed Programs |
Preclinical/In Silico Discovery Completed |
Rare Oncology |
Working With Pharma
Healx will seek pharma partners for its programs to move through larger clinical trials. Guilliams hinted at a co-development deal in the works with a pharma partner. The company also has earlier-stage partnerships. In December 2019, it announced a deal with Boehringer Ingelheim to identify indications related to rare neurological diseases.
“We were receiving too much interest from pharma companies,” Guilliams said. “What is important for our business model is not selling our technology platform. So, we are only interested in co-development partnerships. We are being approached on a regular basis about access to our platform, but that is not our model,” he explained. “In a way we are a tech company, but we are focused on using that tech to invest in the development of rare disease therapies. For companies who want to then partner on that, and do co-development, we are very open,” he added.
Action Against COVID-19
Healx is using its technology platform to try to predict a safe drug combination for use in patients with COVID-19. “We believe the most likely approach, given the time and given the severity for patients, will be repurposing existing drugs and combinations.” If one drug is able to stop the virus entering the cells, it is not likely it can also help the patient to recover, according to Guilliams. “We believe it has to be a combination approach” to combat the Sars-COV-2 virus.
Healx has around 27 staff focused on a COVID-19 project. “We are in touch with labs that will be able to test these combinations. Once we have been able to validate a combination in a preclinical setting, we will be able to move quickly to proof of concept.”
Guilliams said the pace of activity during the pandemic had been amazing. “We have never seen such a coordinated global effort behind one disease and one purpose. It is great. Also, a lot of people who were not focused on vaccines or virology have been looking into COVID-19 from different perspectives.”
The company is also using its machine learning technology to analyze all the suggested approaches that are being tried against COVID-19 around the world. “We will have more to say on this topic at a later date,” Guilliams stated. “But every day there is a new data set, a new approach. Machine learning will help to keep track of this and make sense of information that is moving very quickly. Our technology will help us see the gaps and where we could have the biggest impact.”
Healx is not set up to be a “fee for service” provider. “Many AI companies are doing that,” Guilliams noted. He went on to say that the company has sought investors who believe on the one hand that technology is disrupting discovery, but on the other hand that the best way to create value is by partnering to co-develop drug candidates.
Tech Talent And Biology Expertise
Guilliams said a key bit of advice given to him was to “focus on hiring better, smarter people than you.” He said Healx has managed to be attractive as an employer. “We have a hybrid staff between tech and drug discovery. It is difficult to build that kind of team and find the right people. We need experts in their domain but those who are curious and are early tech adopters.”
For example, the company wants to employ those with drug discovery expertise. But they must also have the mind-set that machine learning and technology can change the traditional development model. However, Healx also seeks technology experts who can understand the biology. “Usually the best tech people do not come from pharma or drug discovery – that is not where the best technology has been developed so far,” Guilliams noted, adding that the company has a balance of discovery and technology talent.
A second bit of advice: “Keep the mission and make no trade-offs or concessions.” This has been particularly tough when talking to investors, Guilliams said. “We have sought the right investors for the company who believe in our mission. They believe in a business model focused on getting a drug to market more cheaply and quickly, and therefore being able to charge a price that is not a million dollars per patient. This is one of the best decisions we made early on, to say no to investors who were not aligned with the mission.”
Healx has secured more than $70m in financing so far, “with more to come,” according to its leader.
Leaders today need to have a deep expertise in one domain, but be curious about other areas, according to Guilliams. Leaders today “need to be comfortable with trying and failing” and they also “need to be able to switch between getting stuff done immediately and the tactical strategic, longer-term stuff.” Guilliams said it was harder to find people who can do both things.
Healx has targeted people who have worked in large companies but then moved over to start-ups, where traditionally there is a need to be more hands-on and strategies have to evolve quickly. “It’s that mix [we look for], people with profiles that show they have done really well in a large organization but also have shown that they can adapt in a scale-up, start-up environment.”
As one example, Kate Hilyard, Healx’s chief operating officer, was previously corporate vice president at Charles River Discovery, leading a team of more than 400 drug discovery scientists across three countries, and earlier she was head of inflammatory diseases at Roche. However, her extensive career also includes senior leadership positions at the biotech, Cambridge Antibody Technology.
Looking ahead at the road map for Healx, Guilliams said, “It's really about achieving the mission.” He wants to be able to look back in 10 years’ time and be able to count the number of rare diseases, and patients, that have benefited from the company’s discovery and development work. “We want to add scale to drug discovery. We want to create options for rare disease patients that they don’t have today.”