Variety of approved drugs target PINK1 protein, may aid Parkinson’s
Study into repurposing therapies showing potential to protect mitochondria
An artificial intelligence (AI)-based algorithm identified more than a dozen U.S.-approved medications that might be repurposed to treat Parkinson’s disease, all based on their ability to increase levels of the PINK1 protein.
PINK1 is involved in the health and workings of mitochondria, the cells’ energy production centers, and has been closely linked to Parkinson’s.
The identified molecules included some with no previous ties to Parkinson’s and others that showed benefit in preclinical disease research. Among the listed compounds was Comtan (entacapone), an approved Parkinson’s treatment.
“In addition to revealing the new potential drug candidates against [Parkinson’s disease], the obtained results additionally confirm the utility of already known antiparkinson drugs, in the new context of PINK1 expression,” the scientists wrote.
PINK1 protein a target for its protective effects on mitochondria
The study, “AI molecular property prediction for Parkinson’s Disease reveals potential repurposing drug candidates based on the increase of the expression of PINK1,” was published in Computer Methods and Programs in Biomedicine.
PINK1, formally known as the PTEN induced putative kinase 1, plays a pivotal role in mitochondria via mitophagy, the process cells use to recycle damaged or dysfunctional mitochondria, replacing them with new ones.
Mutations in the gene that encodes PINK1 lead to an insufficient production of cellular energy and are a cause of familial Parkinson’s disease. Mitochondrial dysfunction, however, is implicated in all forms of the neurodegenerative disease.
“Mitophagy is a key pathway for maintenance of mitochondrial and neuronal health. Impairment of the process of mitophagy causes accumulation of damaged mitochondria, leading to neuronal death,” the researchers wrote.
Accordingly, medicines that aim to boost the activity of the PINK1 gene, thereby increasing PINK1 protein production, are of interest in potentially treating all with Parkinson’s.
As AI and machine learning techniques advance, they increasingly are being used in the design or discovery of therapeutic targets for disease. Machine learning is a form of AI in which an algorithm is “taught” to predict certain outcomes based on available information.
Its uses include screening molecules approved for one disease to determine if they also could treat another condition.
20 molecules show a potential for treating Parkinson’s disease
Scientists in Austria and Poland took this approach to search for potential PINK1-targeted therapies, using a machine learning algorithm to screen U.S. Food and Drug Administration-approved medications that could enhance PINK1 gene activity in mitochondria.
“The starting point for our study was the choice of PINK1 as a target for the discovery of the potential neuroprotective drugs. Existing published research concerning drug design related to this target is very limited,” the researchers noted.
Data to train the algorithm came from a high-throughput cell culture screen of about 400,000 molecules, 825 of which increased, to some degree, PINK1 protein levels on or inside mitochondria.
With this acquired knowledge, the algorithm then screened approved drugs and ranked them.
Among the 20 molecules identified as having the highest potential as a PINK1-targeted Parkinson’s therapy, 11 previously were shown to have neuroprotective properties, the scientists reported.
Top-ranked compound, nitazoxanide, among those spotted in earlier studies
A number of the candidates also were reported to show benefits in early research in Parkinson’s models.
Among them was nitazoxanide, which consistently ranked highest as a potential repurposed drug. Research in a Parkinson’s mouse model found it enhanced mitochondrial function and showed neuroprotective properties.
Comtan, which ranked 18th, is the only identified molecule approved specifically for Parkinson’s motor symptoms. But as a COMT inhibitor, its established mechanism of action relates to levodopa’s breakdown and not PINK1.
A number of other listed molecules have shown neuroprotective, anti-inflammatory, or mitochondrial-protective properties in preclinical studies. Some also had an ability to inhibit monoamine oxidases, a family of enzymes that are a known target for Parkinson’s therapeutics.
Given that “diseases like [Parkinson’s] usually derive from multifactorial disorders … combining multiple critical mechanisms of action in one drug, may be of high significance,” the researchers wrote.
Beyond the specific medications identified in the study, they noted that its findings offer broad insights into those families of molecules that could be of benefit.
Studies in disease cell and animal models — and, potentially, clinical trials in patients — are needed to confirm these molecules’ potential in treating Parkinson’s disease, the team concluded.