Taking 'trial and error' out of epilepsy treatment

Neural architecture of an organoid. Neuronal cell bodies are red,
dendrites are green, axons are white, blue labels the nucleus of
astrocytes and senescent cells.
Image: Ben Rollo, PERSIST Associate Investigator

Enabling innovative research that cuts across multiple scientific and clinical disciplines in order to address urgent unmet health care needs is what the Medical Research Future Fund (MRFF) scheme was designed to do, and is doing exactly that with its recently announced funding for a project to help patients with epilepsy.

By combining their expertise in  human stem cell biology, clinical epilepsy and artificial intelligence scientists from the University of Queensland and Monash University will identify effective treatments for people with epilepsy more efficiently.

This project, funded through the MRFF's Stem Cell Therapies Mission, is called PERSIST, which stands for "Personalising Epilepsy Regimes with Stem cells and artificial Intelligence models for Superior Treatment outcomes."

Professor Ernst Wolvetang from the Australian Institute of Bioengineering and Nanotechnology at the University of Queensland is the team leader. He said, “Nearly a million Australians will develop epilepsy and finding effective drug treatment can be a long journey of 'trial and error'. More than 30% of epilepsy patients will not experience improved seizure control.”

“At present we don't know which particular anti-seizure drugs would be most effective for individual patients at a personal level, and this makes evidence-based drug selection difficult,” said Professor Patrick Kwan, the Monash University lead on the mission.

The research will use stem cells from people with drug-resistant epilepsy to grow lentil-sized mini-tissues called brain organoids. They are miniature models that mimic human brains, but in vitro, i.e. in a dish. See Prof Wolvetang's video explainer of organoids.

The organoids will be used to screen a library of approved medications to identify those able to reduce seizure-like activity for each person.

An artificial intelligence model will be developed, using laboratory, clinical and genetic data to assist in predicting the most promising anti-seizure medications.

"So the questions we're asking are, can in vitro cultured human stem cell derived brain organoids from drug resistant epilepsy patients be used to predict anti-seizure drug regimens that are effective in real patients? And will artificial intelligence neural network approaches and genetic analysis provide insight into the processes that cause drug resistance and enable these predictions to be more accurate?" said Prof Wolvetang.

“Our approach, because it is personalised to the patient (with the organoids grown from their stem cells) and highly efficient (with parallel testing of drugs and AI to speed up the analysis and decision-making), has the potential to fundamentally transform epilepsy care. This will facilitate evidence-based drug selection, reduce health care costs and most important, improve patient health outcomes,” said Prof Kwan.

See University of Queensland media release: aibn.uq.edu.au/article/2021/07/overcoming-trial-and-error-epilepsy-treatment