Using artificial intelligence to identify drugs for repurposing to treat l-DOPA-induced dyskinesia.

Repurposing regulatory agency-approved molecules, with proven safety in humans, is an attractive option for developing new treatments for disease. We identified and assessed the efficacy of 3 drugs predicted by an in silico screen as having the potential to treat l-DOPA-induced dyskinesia (LID) in Parkinson's disease. We analysed ∼1.3 million Medline abstracts using natural language processing and ranked 3539 existing drugs based on predicted ability to reduce LID. 3 drugs from the top 5% of the 3539 candidates; lorcaserin, acamprosate and ganaxolone, were prioritized for preclinical testing 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 and orally bioavailable and iv) being clinical trial ready. We assessed the efficacy of acamprosate, ganaxolone and lorcaserin in a rodent model of l-DOPA-induced hyperactivity, with lorcaserin affording a 58% reduction in rotational asymmetry (P < 0.05) compared to vehicle. Acamprosate and ganaxolone failed to demonstrate efficacy. Lorcaserin, a 5HT2C agonist, was then further tested in MPTP lesioned dyskinetic macaques where it afforded an 82% reduction in LID (P < 0.05), unfortunately accompanied by a significant increase in parkinsonian disability. In conclusion, although our data do not support the repurposing of lorcaserin, acamprosate or ganaxolone per se for LID, we demonstrate value of an in silico approach to identify candidate molecules which, in combination with an in vivo screen, can facilitate clinical development decisions. The present study adds to a growing literature in support of this paradigm shifting approach in the repurposing pipeline.

Expert-Augmented Computational Drug Repurposing Identified Baricitinib as a Treatment for COVID-19.

The onset of the 2019 Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic necessitated the identification of approved drugs to treat the disease, before the development, approval and widespread administration of suitable vaccines. To identify such a drug, we used a visual analytics workflow where computational tools applied over an AI-enhanced biomedical knowledge graph were combined with human expertise. The workflow comprised rapid augmentation of knowledge graph information from recent literature using machine learning (ML) based extraction, with human-guided iterative queries of the graph. Using this workflow, we identified the rheumatoid arthritis drug baricitinib as both an antiviral and anti-inflammatory therapy. The effectiveness of baricitinib was substantiated by the recent publication of the data from the ACTT-2 randomised Phase 3 trial, followed by emergency approval for use by the FDA, and a report from the CoV-BARRIER trial confirming significant reductions in mortality with baricitinib compared to standard of care. Such methods that iteratively combine computational tools with human expertise hold promise for the identification of treatments for rare and neglected diseases and, beyond drug repurposing, in areas of biological research where relevant data may be lacking or hidden in the mass of available biomedical literature.

Using artificial intelligence to identify anti-hypertensives as possible disease modifying agents in Parkinson's disease.

PURPOSE: Drug repurposing is an effective means of increasing treatment options for diseases, however identifying candidate molecules for the indication of interest from the thousands of approved drugs is challenging. We have performed a computational analysis of published literature to rank existing drugs according to predicted ability to reduce alpha synuclein (aSyn) oligomerization and analyzed real-world data to investigate the association between exposure to highly ranked drugs and PD. METHODS: Using IBM Watson for Drug Discoveryâ (WDD) we identified several antihypertensive drugs that may reduce aSyn oligomerization. Using IBM MarketScanâ Research Databases we constructed a cohort of individuals with incident hypertension. We conducted univariate and multivariate Cox proportional hazard analyses (HR) with exposure as a time-dependent covariate. Diuretics were used as the referent group. Age at hypertension diagnosis, sex, and several comorbidities were included in multivariate analyses. RESULTS: Multivariate results revealed inverse associations for time to PD diagnosis with exposure to the combination of the combination of angiotensin receptor II blockers (ARBs) and dihydropyridine calcium channel blockers (DHP-CCB) (HR = 0.55, p < 0.01) and angiotensin converting enzyme inhibitors (ACEi) and diuretics (HR = 0.60, p-value <0.01). Increased risk was observed with exposure to alpha-blockers alone (HR = 1.81, p < 0.001) and the combination of alpha-blockers and CCB (HR = 3.17, p < 0.05). CONCLUSIONS: We present evidence that a computational approach can efficiently identify leads for disease-modifying drugs. We have identified the combination of ARBs and DHP-CCBs as of particular interest in PD.

Applications of machine learning to diagnosis and treatment of neurodegenerative diseases.

Globally, there is a huge unmet need for effective treatments for neurodegenerative diseases. The complexity of the molecular mechanisms underlying neuronal degeneration and the heterogeneity of the patient population present massive challenges to the development of early diagnostic tools and effective treatments for these diseases. Machine learning, a subfield of artificial intelligence, is enabling scientists, clinicians and patients to address some of these challenges. In this Review, we discuss how machine learning can aid early diagnosis and interpretation of medical images as well as the discovery and development of new therapies. A unifying theme of the different applications of machine learning is the integration of multiple high-dimensional sources of data, which all provide a different view on disease, and the automated derivation of actionable insights.

Identifying drugs with disease-modifying potential in Parkinson's disease using artificial intelligence and pharmacoepidemiology.

PURPOSE: The aim of the study was to assess the feasibility of an approach combining computational methods and pharmacoepidemiology to identify potentially disease-modifying drugs in Parkinson's disease (PD). METHODS: We used a two-step approach; (a) computational method using artificial intelligence to rank 620 drugs in the Ontario Drug Benefit formulary based on their predicted ability to inhibit alpha-synucleinaggregation, a pathogenic hallmark of PD; and (b) case-control study using administrative databases in Ontario, Canada. Persons aged 70-110 years with incident PD from April 2002-March 2013. Controls were randomly selected from persons with no previous diagnosis of PD. RESULTS: A total of 15 of the top 50 drugs were deemed feasible for pharmacoepidemiologic analysis, of which seven were significantly associated with incident PD after adjustment, with five of these seven associated with a decreased odds of PD. Methylxanthine drugs pentoxifylline (OR, 0.72; 95% CI, 0.59-0.89) and theophylline (OR, 0.77; 95% CI, 0.66-0.91), and the corticosteroid dexamethasone (OR, 0.72; 95% CI, 0.61-0.85) were associated with decreased odds of PD. CONCLUSIONS: Our findings demonstrate the feasibility of this approach to focus the search for disease-modifying drugs. Corticosteroids and methylxanthines should be further investigated as potential disease-modifyingdrugs in PD.

Zebrafish oxytocin neurons drive nocifensive behavior via brainstem premotor targets.

Animals have evolved specialized neural circuits to defend themselves from pain- and injury-causing stimuli. Using a combination of optical, behavioral and genetic approaches in the larval zebrafish, we describe a novel role for hypothalamic oxytocin (OXT) neurons in the processing of noxious stimuli. In vivo imaging revealed that a large and distributed fraction of zebrafish OXT neurons respond strongly to noxious inputs, including the activation of damage-sensing TRPA1 receptors. OXT population activity reflects the sensorimotor transformation of the noxious stimulus, with some neurons encoding sensory information and others correlating more strongly with large-angle swims. Notably, OXT neuron activation is sufficient to generate this defensive behavior via the recruitment of brainstem premotor targets, whereas ablation of OXT neurons or loss of the peptide attenuates behavioral responses to TRPA1 activation. These data highlight a crucial role for OXT neurons in the generation of appropriate defensive responses to noxious input.

Identification of pharmacodynamic biomarker hypotheses through literature analysis with IBM Watson.

BACKGROUND: Pharmacodynamic biomarkers are becoming increasingly valuable for assessing drug activity and target modulation in clinical trials. However, identifying quality biomarkers is challenging due to the increasing volume and heterogeneity of relevant data describing the biological networks that underlie disease mechanisms. A biological pathway network typically includes entities (e.g. genes, proteins and chemicals/drugs) as well as the relationships between these and is typically curated or mined from structured databases and textual co-occurrence data. We propose a hybrid Natural Language Processing and directed relationships-based network analysis approach using IBM Watson for Drug Discovery to rank all human genes and identify potential candidate biomarkers, requiring only an initial determination of a specific target-disease relationship. METHODS: Through natural language processing of scientific literature, Watson for Drug Discovery creates a network of semantic relationships between biological concepts such as genes, drugs, and diseases. Using Bruton's tyrosine kinase as a case study, Watson for Drug Discovery's automatically extracted relationship network was compared with a prominent manually curated physical interaction network. Additionally, potential biomarkers for Bruton's tyrosine kinase inhibition were predicted using a matrix factorization approach and subsequently compared with expert-generated biomarkers. RESULTS: Watson's natural language processing generated a relationship network matching 55 (86%) genes upstream of BTK and 98 (95%) genes downstream of Bruton's tyrosine kinase in a prominent manually curated physical interaction network. Matrix factorization analysis predicted 11 of 13 genes identified by Merck subject matter experts in the top 20% of Watson for Drug Discovery's 13,595 ranked genes, with 7 in the top 5%. CONCLUSION: Taken together, these results suggest that Watson for Drug Discovery's automatic relationship network identifies the majority of upstream and downstream genes in biological pathway networks and can be used to help with the identification and prioritization of pharmacodynamic biomarker evaluation, accelerating the early phases of disease hypothesis generation.

Repurposing drugs to treat l-DOPA-induced dyskinesia in Parkinson's disease.

In this review, we discuss the opportunity for repurposing drugs for use in l-DOPA-induced dyskinesia (LID) in Parkinson's disease. LID is a particularly suitable indication for drug repurposing given its pharmacological diversity, translatability of animal-models, availability of Phase II proof-of-concept (PoC) methodologies and the indication-specific regulatory environment. A compound fit for repurposing is defined as one with appropriate human safety-data as well as animal safety, toxicology and pharmacokinetic data as found in an Investigational New Drug (IND) package for another indication. We first focus on how such repurposing candidates can be identified and then discuss development strategies that might progress such a candidate towards a Phase II clinical PoC. We discuss traditional means for identifying repurposing candidates and contrast these with newer approaches, especially focussing on the use of computational and artificial intelligence (AI) platforms. We discuss strategies that can be categorised broadly as: in vivo phenotypic screening in a hypothesis-free manner; in vivo phenotypic screening based on analogy to a related disorder; hypothesis-driven evaluation of candidates in vivo and in silico screening with a hypothesis-agnostic component to the selection. To highlight the power of AI approaches, we describe a case study using IBM Watson where a training set of compounds, with demonstrated ability to reduce LID, were employed to identify novel repurposing candidates. Using the approaches discussed, many diverse candidates for repurposing in LID, originally envisaged for other indications, will be described that have already been evaluated for efficacy in non-human primate models of LID and/or clinically. This article is part of the Special Issue entitled 'Drug Repurposing: old molecules, new ways to fast track drug discovery and development for CNS disorders'.

A convergent and essential interneuron pathway for Mauthner-cell-mediated escapes.

The Mauthner cell (M-cell) is a command-like neuron in teleost fish whose firing in response to aversive stimuli is correlated with short-latency escapes [1-3]. M-cells have been proposed as evolutionary ancestors of startle response neurons of the mammalian reticular formation [4], and studies of this circuit have uncovered important principles in neurobiology that generalize to more complex vertebrate models [3]. The main excitatory input was thought to originate from multisensory afferents synapsing directly onto the M-cell dendrites [3]. Here, we describe an additional, convergent pathway that is essential for the M-cell-mediated startle behavior in larval zebrafish. It is composed of excitatory interneurons called spiral fiber neurons, which project to the M-cell axon hillock. By in vivo calcium imaging, we found that spiral fiber neurons are active in response to aversive stimuli capable of eliciting escapes. Like M-cell ablations, bilateral ablations of spiral fiber neurons largely eliminate short-latency escapes. Unilateral spiral fiber neuron ablations shift the directionality of escapes and indicate that spiral fiber neurons excite the M-cell in a lateralized manner. Their optogenetic activation increases the probability of short-latency escapes, supporting the notion that spiral fiber neurons help activate M-cell-mediated startle behavior. These results reveal that spiral fiber neurons are essential for the function of the M-cell in response to sensory cues and suggest that convergent excitatory inputs that differ in their input location and timing ensure reliable activation of the M-cell, a feedforward excitatory motif that may extend to other neural circuits.

Whole-brain activity mapping onto a zebrafish brain atlas.

In order to localize the neural circuits involved in generating behaviors, it is necessary to assign activity onto anatomical maps of the nervous system. Using brain registration across hundreds of larval zebrafish, we have built an expandable open-source atlas containing molecular labels and definitions of anatomical regions, the Z-Brain. Using this platform and immunohistochemical detection of phosphorylated extracellular signal­regulated kinase (ERK) as a readout of neural activity, we have developed a system to create and contextualize whole-brain maps of stimulus- and behavior-dependent neural activity. This mitogen-activated protein kinase (MAP)-mapping assay is technically simple, and data analysis is completely automated. Because MAP-mapping is performed on freely swimming fish, it is applicable to studies of nearly any stimulus or behavior. Here we demonstrate our high-throughput approach using pharmacological, visual and noxious stimuli, as well as hunting and feeding. The resultant maps outline hundreds of areas associated with behaviors.