Healx has raised $47 million in a Series C round co-led by Silicon Valley-based R42 Group and Atomico with participation from new and existing investors including Balderton, Jonathan Milner, Global Brain, btov, Ayana Capital, o2h, and VU Venture Partners. In conjunction with the financing, Stanford Medicine adjunct professor Ronjon Nag, PhD, founder of R42 Group and 2024 Silicon Valley Hall of Fame AI inductee will join Healx’s board.
The funds will help the company accomplish its goal of using AI to identify new treatments for rare diseases. Healx said that the proceeds of the financing will be used to advance its pipeline of medicines in rare oncology, renal, and neurodevelopmental disorders. This includes advancing its lead program HLX-1502 through a Phase II clinical trial for the treatment of neurofibromatosis type 1 (NF1). Tim Guilliams, PhD, Healx’s co-founder and CEO, said that the company intends to begin testing HLX-1502 in patients by December 2024.
Healx has been cleared by the FDA to move forward with a Phase II trial that will test its drug in adults with NF1 and inoperable plexiform neurofibroma.
HLX-1502 is a tablet taken orally that Healx claims works differently from other NFI treatments. The drug has received orphan drug and rare pediatric disease designations from the FDA for treating NF1. The company has also signed an investment agreement with its long-term research partner, Children’s Tumor Foundation (CTF). Under the terms of the agreement, Healx will receive milestone-driven payments from CTF that it will use to move its drug programs forward including HLX-1502.
Healx Co-founder and CEO
Several of those drugs will target a few of the approximately 10,000 rare diseases, about 95% of which currently don’t have an approved treatment. Guilliams said that Healx opted to focus on rare diseases because of this unmet need. Aside from the commercial opportunity, “there’s the ability to make a breakthrough and find something new that’s really going to be impactful for the patient community and the families,” he told GEN.
Healx is confident that the approach used by its machine learning solution, HealNet, can help with finding effective candidates for rare diseases. “Broadly it’s really understanding disease biology in a different way, in a more complex way,” Guilliams explained. Specifically, it defines “complex patterns and signatures around disease” as well as “complex drug patterns and signatures that basically reverse the disease signature into a healthy one.” These definitions are based on a combination of data types including information from transcriptomic, metabolomic, and phenomic experiments.
The system then predicts drug candidates that are most likely to target and reverse the disease signature, and Healx scientists test them in the lab. “The first wave AI technology that we built was around a biomedical rare disease knowledge graph. That was basically based on the Google search model,” Guilliams said. Now with the advent of generative AI and large language models, Healx is looking to leverage the technology to enhance its platform further using some of the funds raised in the Series C. Their plans include generating more data to better understand rare disease biology as well as using techniques such as retrieval augmented generation to try to fill in gaps in their knowledge about diseases.
At the moment, “drug discovery is largely focused on a target-based paradigm [and] in certain cases it works,” he said. However, that approach may not work for most rare diseases whose biology “is still poorly understood. And then if you look at more complex diseases, maybe your single target will not work to … modulate the disease biology.” This is why Healx approaches drug discovery for rare diseases in “a target agnostic manner” from the very beginning.
Testing HLX-1502 in patients
Affecting approximately 1 in 2,500 individuals, NF1 is a rare genetic disorder that is associated with a predisposition to develop multiple benign and malignant tumors. Two notable types of tumors associated with NF1 are plexiform neurofibromas and cutaneous neurofibromas.
Plexiform neurofibromas grow aggressively along nerves which often leads to significant morbidity and greater risk of malignant transformation. These tumors can cause functional impairments, disfigurement, and pain, and require multidisciplinary management. Currently, a single treatment option is available for some children with plexiform neurofibromas. But these treatments have tolerability and safety concerns including GI, heart, eye, and skin toxicity. Cutaneous neurofibromas, for their part, are benign tumors but they can cause significant symptoms. There are currently no approved treatments for this subtype.
Currently, chemotherapies seem to be the treatment approach of choice for the pharma companies that are working on therapies for NF1. So far, AstraZeneca’s Koselugo (selumetinib), a MEK inhibitor, has been approved for the treatment of NF1 patients in 32 countries including the United States and China, and is in clinical trials for other tumor types that occur in NF2-related schwannomatosis and other types of schwannomatosis.
But chemotherapies have significant side effects. If Healx’s drug is successful, the company claims it will have a much better safety profile than existing inhibitors, which is good since the drug is intended for chronic use.
And early data about the molecule is promising. There is documented safety data on the compound from previous studies in healthy volunteers. Results from those studies also describe properties about the compound that indicate a high affinity for nerve cells.
In NF1 cases, HLX-1502 works by modulating mitochondrial function, through various pathways, to control tumor growth. It is also drawn to signals from the environment around nerve cells with tumors, including higher levels of oxidative stress. “It’s a small molecule with very few side effects and shown preclinically to be exceptionally effective at reducing tumor growth and taking out the tumor cells,” Guilliams said.
Healx launched its HLX-1502 program in 2019 and plans to dose its first of 20 patients participating in the trial later this year. They’ll be evaluating the participants for tumor reduction along the nerve in response to the drug as the clinical trial endpoint.
For Guillams, AI was critical for helping his team “analyze things quickly, match drugs, and patterns, and then select the best one with the best properties” but so was its partnership with CTF. “They really helped us accelerate things further [and] they also invested in the program themselves.” Healx hopes to wrap up the trial in the first half of 2026.
