AI-enabled drug discovery approach identified potentially game-changing treatment which has been advanced from the lab bench to an FDA Orphan Drug Designation in record time
By Benjamin Boettner
(BOSTON) — Rett syndrome is a devastating rare genetic childhood disorder primarily affecting girls. Merely 1 out of 10,000 girls are born with it and much fewer boys. It is caused by mutations in the MeCP2 gene on the X chromosome, leading to a spectrum of cognitive and physical impairments, including repetitive hand motions, speech difficulties, and seizures.
However, besides severe impairment of neurological functions, which has been the primary focus of researchers, Rett syndrome also upsets the functions of many non-neurological organs, including the digestive, musculoskeletal, and immune systems. This complexity has made the development of an effective cure able to treat the disease across the multiple tissues an extreme challenge.
Now, a highly multi-disciplinary research team at the Wyss Institute at Harvard University has made a significant breakthrough by leveraging an AI-driven drug discovery process in combination with innovative disease modeling. Their study identified a drug known as vorinostat as a promising treatment for Rett syndrome, demonstrating disease-modifying abilities across multiple neuronal and non-neuronal tissues in preclinical models of Rett syndrome that were superior to trofinetide, the only approved treatment for Rett syndrome. The findings are published in Communications Medicine .
Because vorinostat has already been approved by the Food and Drug Administration (FDA) to treat a blood disease, the Wyss-enabled startup Unravel Biosciences has been able to rapidly repurpose this drug as a therapy for Rett syndrome. Unravel Biosciences' lead pipeline asset, RVL-001, is a proprietary formulation of vorinostat, which recently received an Orphan Drug Designation from the FDA. Unravel will be initiating a proof-of-concept clinical trial to assess the drug's efficacy and safety in 15 female patients with Rett syndrome in Colombia later this year, and will test an "n-of-1 trial design" to evaluate different vorinostat treatments within individual patients, which is more appropriate to the complexity of the disease, and to rare disease communities in general.
"The identification and further development of vorinostat as the potentially first curative treatment for Rett syndrome would not have been possible without our unique AI-enabled computational approach to drug discovery, and its combination with an innovative disease model that broadly mimics the features of Rett syndrome," said senior author and Wyss Founding Director Donald Ingber , M.D., Ph.D. "This new target-agnostic approach for drug discovery proved to be extremely fast and effective and, together with our unique technology translation capabilities, creates a model for us to tackle other diseases with unmet need that pose similarly enormous challenges." Ingber is also the Judah Folkman Professor of Vascular Biology at Harvard Medical School and Boston Children's Hospital, and the Hansjörg Wyss Professor of Biologically Inspired Engineering at Harvard John A. Paulson School of Engineering and Applied Sciences.
A new paradigm for drug discovery
Key to the discovery of vorinostat as a potential Rett syndrome therapy was the Wyss Institute's computational nemoCAD pipeline that enabled the team to predict drug candidates not based on a specific target molecule of the disease – as most traditional drug discovery approaches do – but on changes that occur in the entire gene network across multiple organ systems in Rett syndrome. Richard Novak , Ph.D., then a Staff Scientist on Ingber's team at the Wyss Institute and now CEO of Unravel and other members of the team originally developed nemoCAD as part of the Wyss-led DARPA THoR Project to be able to identify why some patients are more tolerant to infection with pathogens than others. This DARPA-funded project led the path to other successful demonstrations of drug discovery by the Wyss Institute across diverse medical challenges, from neuropsychiatry to artificial hibernation .
As a starting point to develop a treatment for the full clinical spectrum of symptoms experienced by Rett disease patients, the team created a small animal model consisting of tadpoles from the frog Xenopus laevis, in which they used CRISPR genome-engineering technology to generate various mutations inactivating the MeCP2 gene to reflect the diverse patient population. The engineered tadpoles recapitulated a range of critical features of Rett syndrome, including developmental and behavioral delay, seizures, as well as intestinal, muscle and brain anomalies. Importantly, the researchers could analyze this new Rett model for changes in gene expression, across multiple organs that are associated with Rett syndrome-specific neurological and non-neurological changes in behavior and tissue function.
The researchers then used nemoCAD to compare all gene expression changes that occurred in MeCP2-defective tadpoles vs. healthy tadpoles to predict drug compounds from a public data base curated by the NIH that could reverse the pathological changes in the same gene expression networks. LINCS, as the data base is called, contains gene expression signatures induced by more than 19,800 drug compounds in a large variety of human cell lines, including drugs that already have been approved by the FDA for the treatment of other diseases. This type of analysis far exceeds traditional gene expression analysis, which determines the expression changes of individual genes or smaller groups of genes in isolation from all other changes.
"Critically ill patients demand accelerated discovery of new treatment options for their understudied disorders. Computing how entire gene expression networks are changing in a concerted fashion allowed us to predict which drugs are the most likely to push the Rett-specific gene expression network back to its normal state across multiple organs," said co-first author Novak, who led the project together with co-author and Unravel co-founder Frederic Vigneault , Ph.D. "We then went from a list of candidates predicted in silicoto directly validating the top candidates in vivo in our tadpole model within a few weeks, demonstrating an efficient way of identifying previously unknown therapeutic mechanisms," Novak added.
Vorinostat scored the highest on the list and produced the strongest therapeutic effects in the genetically engineered tadpoles, which showed an impressive reversal of their disease features on a whole organism level. Symptoms such as seizures, unusual swimming motions that resemble repetitive behaviors in patients with Rett syndrome, as well as gastrointestinal and muscular symptoms were all potently suppressed by the drug; and vorinostat was much more effective at suppressing these symptoms than trofinetide.
"Important for its translation toward patients, vorinostat also consistently reversed multiple symptoms of Rett syndrome in a preclinical mouse model, even when administered after symptoms were already in full progression, which trofinetide was unable to do. The FDA has considered success in this model as a critical milestone before deciding moving trofinetide to human clinical trials in the past," said Tiffany Lin, a co-first author on the new study and Wyss Staff Scientist who worked closely with Novak and Vigneault in Ingber's team. "Excitingly, with some additional formulation work, vorinostat produced significant therapeutic outcomes even as an oral treatment."
New drug, new insights
Through their gene network predictions and in-depth analysis of the molecular and cellular processes affected in MeCP2-defective tadpoles and following treatment with vorinostat, the researchers discovered what could be an unexpected driving force of the disease. MeCP2 encodes a protein that regulates the expression of hundreds of genes. It does so by binding to DNA regions that carry so-called methyl groups, and forming complexes with other proteins. One such class of proteins is known as histone deacetylases (HDAC), which modify other proteins by removing a different small chemical group known as an acetyl group from them. The acetylation status of proteins such as histones and other proteins is misregulated in Rett syndrome due to the inactivation of MeCP2.
However, the team's study in whole organisms utterly redefined this view. They found that while histones in the Rett syndrome models indeed were under-acetylated in brain cells, they, surprisingly, were over-acetylated in other tissues affected by the disease, like the gastrointestinal (GI) tract. Importantly, the team's network analysis predicted vorinostat to also affect the acetylation of a-tubulin, a protein that plays important roles in neurodegenerative disorders and other diseases. a-tubulin assembles diverse cytoskeletal structures in cells that critically contribute to their functions, such as surface cilia that help move mucus in the lung and GI tract.
"In our models, a-tubulin in cilia was also hypo-acetylated in brain tissue but hyper-acetylated in the other tissues such as the GI tract, which correlated with signs of functional abnormalities, including inflammation," said Lin. "Vorinostat was able to reverse this dysregulated a-tubulin acetylation pattern in both directions, which tells us that it must have targets beyond the HDAC family for which it is known, and that Rett syndrome, across multiple organs, is caused by mechanisms that should be explored further."
Unravel Biosciences, which was founded by Novak and Vigneault, together with Ingber and Wyss Institute Associate Faculty member Michael Levin , Ph.D., another author of the new study, is building on vorinostat and the team's discovery of a new therapeutic mechanism to advance the possibly first curative therapy for Rett syndrome in a first patient cohort. "We are extremely excited to be reaching clinical stages following this rapid discovery and development journey and hope to be able to impact the lives of Rett syndrome patients in an unprecedented way," said Novak.
The study was also authored by Shruti Kaushal, Megan Sperry, Erica Gardner, Sahil Loomba, Kostyantyn Shcherbina, Vishal Keshari, Alexandre Dinis, Anish Vasan, Vasanth Chandrasekhar, Takako Takeda, Rahul Nihalani, Sevgi Umur, and Jerrold Turner.
