Objective: To test the efficacy of 3 drugs predicted by IBM Watson for Drug Discovery (WDD) as having the potential to treat L-DOPA-induced dyskinesia (LID).

Background: Using the machine learning capabilities of WDD we analyzed ~1.3 million Medline abstracts using natural language processing and created a model to rank 3539 existing drugs based on predicted ability to reduce LID. 3 drugs from the top 5% of the 3539 candidates, were prioritized for in vivo validation studies based on i) having a novel mechanism of action, ii) having not been previously validated for the treatment of LID, iii) being blood-brain-barrier penetrant, orally bioavailable and iv) clinical trial ready.

Method: Drugs were initially screened in the 6-OHDA-lesioned rat model of LID for efficacy to reduce L-DOPA-evoked rotational asymmetry and Abnormal Involuntary Movements (AIMs), established by 15 days pre-treatment with L-DOPA (10mg/kg). For each compound, 3 doses or vehicle were administered in combination with L-DOPA using a randomized, partial Latin-square design. Drugs with anti-dyskinetic efficacy in 6-OHDA-lesioned rats were further assessed in the MPTP-lesioned non-human primate model of LID. MPTP-lesioned cynomolgus macaques had been previously rendered dyskinetic by daily treatment with L-DOPA (30 mg/kg) for at least 90 days. 3 doses of test drug or vehicle were administered in combination with L-DOPA using a randomized, partial Latin-square design. Parkinsonism and LID were quantified using validated rating scales via post hoc analysis of high definition video recordings by a rater blinded to treatment.

Results: In the 6-OHDA-lesioned rat, UHNWDD115 (3mg/kg) resulted in a 58% reduction in rotational asymmetry (P<0.05, 20-120 min) compared to vehicle, with no significant effect on AIMs. In the MPTP-lesioned macaque, UHNWDD115 (10mg/kg) resulted in an 82% reduction in LID (P<0.05, 0-3 hours post administration) accompanied by a 306% increase in parkinsonian disability. The two remaining drugs failed to demonstrate efficacy in reducing rotational asymmetry or AIMs.

Conclusion: We provide proof of concept that IBM WDD can predict novel treatments to reduce LID based on natural language processing and machine learning. However, the suitability of drugs tested to date is limited by lack of significant efficacy or confounding effects on the anti-parkinsonian actions of L-DOPA. Analysis of further highly ranked drugs is ongoing.

References: Acknowledgements The authors would like to acknowledge: the Ontario Brain Institute and the Government of Ontario for providing access and training to the IBM Watson Drug Discovery platform.

« Back to 2019 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/preclinical-efficacy-of-drugs-identified-by-ibm-watson-for-repurposing-to-treat-l-dopa-induced-dyskinesia/