Accelerating drug discovery for Leigh syndrome

Two recent studies shine a light on a rare neurodegenerative disease and showcase how the integration of artificial intelligence and experimental validation can accelerate the discovery of new therapeutic options. An international research team, combining clinical expertise from the University Hospital Düsseldorf and several partners, with the computational know-how of teams from the University of Luxembourg and Spain, identified promising drugs for the treatment of Leigh Syndrome.

A devastating rare disease and a lack of models for research

Leigh Syndrome is a progressive disease that affects the brain. It usually manifests in childhood, with severe symptoms leading to motor impairment, intellectual disability and early death. Belonging to a group of rare genetic conditions called mitochondrial diseases, Leigh Syndrome concerns 1 in 36,000 people.

It is a devastating disease for which therapeutic options are currently extremely limited. As often with rare diseases, the small number of patients makes research more complex. In the case of Leigh Syndrome, the issue is exacerbated by a lack of cellular or animal models that can faithfully mimic the disease in the lab. To address this challenge, a large group of international collaborators has worked on developing innovative strategies to accelerate the drug discovery process.

Combined approaches to improve drug discovery: in vitro, in vivo & in silico

The two studies, recently published in Cell and Nature Communications, demonstrate the potential of combining approaches for therapeutic discovery. “There is a growing interest in developing computational tools to fast-track the drug screening process, which is costly and time-consuming,” underlines Prof. Antonio Del Sol, head of Computational Biology groups at the Luxembourg Centre for Systems Biomedicine (LCSB) and at CIC bioGUNE in Bilbao. “With our artificial intelligence expertise, we can build innovative pipelines that help reduce the number of potential drugs to be assessed in the lab and identify novel compounds of interest.”

Collaborations with experimental teams and clinicians are then key to validate computational predictions in relevant disease models. There, cell cultures and brain organoids – complex 3D cellular systems – derived from skin cells donated by patients are instrumental to uncover pathological mechanisms and to test the effects and mode of action of promising compounds.

The results obtained for Leigh Syndrome by the interdisciplinary consortium showcase how the integration of computational drug discovery, what is called in silico screens, and experimental validation in vitro and in vivo can accelerate translational research. It led to the identification of several existing drugs that could be repositioned for treating patients affected.

Identifying promising repurposable compounds

In the first study, the researchers tested a library of over 5500 drugs already approved for other conditions or for which extensive safety and efficacy data are available. Among these, Sildenafil, currently approved for the treatment of erectile dysfunction in adults and for the treatment of pulmonary hypertension in infants, was identified as a potential therapeutic candidate for Leigh Syndrome. By combining multi-omics analyses, an approach that integrates data about entire sets of molecules, from RNA to lipids, with computational modelling, the team uncovered the mechanism of action of Sildenafil in Leigh syndrome organoids. “Following the positive effects on metabolism, cell function and lifespan observed in cellular and animal models, Sildenafil was used as a compassionate treatment in six patients,” reports Prof. Alessandro Prigione, head of the Stem Cell Metabolism group at University Hospital Düsseldorf. “They showed preliminary improvements in clinical condition and motor function. We are now planning larger clinical trials to confirm its safety and efficacy.”

Using a deep learning algorithm to accelerate the process, the researchers performed another drug repurposing screening in a second study. They identified an additional compound – talarozole, initially developed for acne and psoriasis – as other promising candidates for Leigh Syndrome and filed a patent application for its use in mitochondrial diseases.

In addition to representing significant progress toward the treatment of a rare disease, these findings highlight the powerful contribution of computational methods when it comes to drug discovery. “From facilitating the prioritisation of the most promising compounds and pinpointing novel molecules of interest, to analysing the mode of action of drugs and guiding the design of more effective therapies, in silico approaches are becoming a key ingredient to accelerate translation to patients,” concludes Prof. Del Sol.

The methodological pipeline developed by this international & interdisciplinary group of researchers could represent a blueprint for uncovering new therapeutic opportunities in other rare neurological disorders.

—

References:

• Accelerating Leigh syndrome drug discovery through deep learning screening in brain organoids, Carmen Menacho, Satoshi Okawa et al., Nature Communications, April 2026.

• Pluripotent stem-cell-based screening uncovers sildenafil as a mitochondrial disease therapy, Annika Zink, Dao-Fu Dai, Annika Wittich, Marie-Thérèse Henke, Giulia Pedrotti, Sonja Heiduschka et al., Cell, March 2026

Funding and partners:

The publication in Cell is the result of a multinational collaboration within the framework of the CureMILS consortium funded by the European Joint Programme on Rare Diseases (EJP RD).

These two studies were conducted in collaboration with several partners. In addition to the LCSB, the CIC bioGUNE, Heinrich Heine University Düsseldorf and University Hospital Düsseldorf, Charité and the Fraunhofer Institute for Translational Medicine and Pharmacology were also involved, together with other research groups in Germany, Spain, Austria, Finland, the Netherlands, Poland, Italy, Greece and the USA.