How AI and Genomics Can Treat Epilepsy

Source: geralt/pixabay

Epilepsy is among the most common neurological disorders that affects 65 million people of all ages globally. In the United States, 3.4 million Americans have epilepsy according to the CDC. Epilepsy can interfere with a person’s ability to drive a car, play sports, swim, or exercise. It is a non-contagious brain disorder where recurrent, unprovoked seizures occur.

Epilepsy may be caused by many factors, including traumatic brain injuries, stroke, loss of oxygen to the brain, brain tumor, parasitic brain infections (malaria, neurocysticercosis from tapeworms), viral infections (Zika, dengue, influenza), bacterial brain infections, neurological diseases, genetic predisposition, and other causes.

Artificial intelligence (AI) and genomics are being deployed to help those suffering from epilepsy. Solutions for epilepsy that use AI include personalized medicine, seizure management, and drug discovery.

Predicting Seizures

Last month, the FDA approved Embrace, the first-ever non-EEG based physiology signal seizure monitoring system for children who have tonic-clonic seizures due to epilepsy.

Tonic-clonic seizures, also known as grand mal seizures, are characterized by violent muscle contractions and may include loss of consciousness, issues with breathing, and loss of bladder control, among other symptoms.

Embrace is made by Empatica, a Massachusetts Institute of Technology (MIT) Media Lab spin-off. Last year the FDA approved Embrace for adults. Embrace is an epilepsy smartwatch that not only detects possible seizures through movement electrical impulse in the skin, but also alerts caregivers and provides location GPS data using artificial intelligence technology. Embrace stores and analyzes patterns during sleep and wakefulness to provide important insights.

Drug Repositioning: Finding New Medications with Genomics and Bioinformatics

Researchers at the University of Iowa have identified FDA-approved medications that are not currently being used to treat seizures that can be repurposed for anti-seizure treatment.

In a study led by Alexander Bassuk, M.D., Ph.D., a professor of pediatrics and neurology at UI, a team of researchers used gene expression profiling and bioinformatics to identify potential new anti-seizure medications.

The scientists identified expression signatures by studying the gene expression of 25,000 genes in human epileptic brain tissue. This analysis was then compared to a big data repository that contained the patterns of gene expression resulting from medication on cells. This produced 184 candidate therapeutic compounds, of which 91 are already approved by the FDA for purposes other than to control seizures. The team then tested the anti-seizure capabilities of four candidate compounds on zebrafish.

The result was that the team identified three drugs with promising anti-seizure capabilities—a blood pressure medication called nifedipine, a diabetes drug called metformin, and an antiparasitic drug called pyrantel tartrate. These medications are already FDA approved for other conditions not related to epilepsy.

As a next step, the team plans to test the other therapeutic candidate compounds in zebrafish, than in mammals—first using mice models. Eventually, the team hopes to conduct clinical trials on human epilepsy patients for the medications that perform well on in the zebrafish and mice studies.

Personalized Medicine for Epilepsy

Epilepsy is a heterogeneous medical condition—it has many causes or etiologies. According to figures from the University of Chicago Medicine, approximately 30 to 40 percent of epilepsy is due to genetic predisposition, and doctors are unable to determine the exact cause of epilepsy in half of all cases.

Artificial Intelligence Essential Reads

Is "Digital Thought" the Next Cognitive Domain?

What if Artificial Intelligence Replaces Human Therapists?

It is a medical disorder that can benefit from the applied artificial intelligence with genomics, for more targeted health care.

Studies have shown that for certain types of epilepsy, precision medicine has produced favorable results. This is the case for tuberous sclerosis complex (rapamycin), GRIN2a mutations (memantine), and Glut 1 deficiency (chetogenic diet) types of epilepsy syndromes according to a study published in June 2018 in the Journal of Translational Genetics and Genomics.

Pharmacogenetics and Pharmacoresistance

Antiepileptic drugs (AED) typically treat the seizures. As with any medication, there is the chance that some patients may experience adverse reactions. For certain individuals, AED medications may exacerbate seizure control, or even be life-threatening.

Furthermore, genetic mutations may impact the efficacy of antiepileptic medication. The enzymes from the cytochrome P450 genes metabolize medications, among other functions. Studies have shown that individuals with polymorphisms (gene variation) of the gene encoding CYP enzymes can negatively impact serum antiepileptic concentrations and result in drug toxicity according to Italian researchers at the University Hospital of Pisa in Italy.

An estimated one-third of those with epilepsy have seizures that are not controllable with the existing available medications. Although there have been a number of studies, the exact reason for pharmacoresistance to antiepileptic medications remains to be understood.

This remains an area that requires additional research. Applying the pattern-recognition capabilities of AI deep learning with big data from pharmacogenetics and antiepileptic pharmacoresistence, may lead to more precise treatment for epilepsy patients in the future.

Copyright © 2019 Cami Rosso All rights reserved.