Harnessing the power of AI technology for drug discovery treatments

The buzz about the power of artificial intelligence (AI) to save or damn humanity is pretty deafening. One way to explain what people mean by AI is that it encompasses the technologies we employ for computers to analyze large data sets using programs (“algorithms”) to find and extract patterns, help humans understand those patterns, suggest decisions regarding future choices and predict future events or suggest results not previously conceived.

Although some expect AI to mimic human intelligence, some AI systems show superior performance to humans (e.g. AlphaGo1). On the other hand, unlike humans, most AI systems need a lot of data and training to achieve acceptable performance (for example, we only need to test strawberries once to know if we like them or not).

One area of ​​research that has attracted a lot of interest and funding is AI-based design of new drugs, particularly small molecules.two A simple way to classify the work that is carried out in this area is shown in the Figure.

Focusing on a goal

Goal-based approaches are based on a first selection of a receptor or enzyme to activate, inhibit or modulate.3 Targets for psychiatry, for example, could be monoamine transporters (eg, SSRIs) or newer targets such as TAAR1 (eg, Ulotaront4). A receptor model involves knowing or hypothesizing its structure based on known crystal structures.5 A ligand-based model requires a data set containing the structure of targeting molecules (“ligands” and “chemical probes”) and the result of their interaction with a target protein (eg, a neurotransmitter receptor). in an appropriate assay (eg, serotonin receptor 1A6).

Importantly, both receptor-based and ligand-based approaches to again drug design implicitly assumes that action on this particular target is the most important action driving the therapeutic value of a drug.

The advantage of these modeling platforms is that you can take advantage of existing data in silico to predict the action of a new molecule on a target, as well as its physical properties, before triggering expensive preclinical work. Since establishing AI expertise also comes at a considerable cost, several partnerships between AI and pharmaceutical companies have been initiated to drive the development of AI-based models for again drug design, avoiding the need to develop in-house AI expertise.

Associations such as Genentech/GNS Healthcare, GSK/Insilico Medicine, Takeda/Numerate AI, Atomwise/Abbvie, CrystalGenomics/Standigm, and Cloud Pharmaceuticals/TheraMetrics are using novel machine learning techniques to design or identify molecules that act on biological targets of particular interest.7 However, most of these associations do not focus on psychiatric indications or CNS disorders in general.

Phenotypic Drug Discovery: A Holistic Approach

Mental health drug discovery, whether traditional or AI-based, has lagged, while mental health needs have continued to rise and remain a huge societal burden. Although single-targeted AI drug design may be adequate for some indications, evidence from psychiatry research points to the need for polypharmacology (Figure). In fact, action on different objectives may be necessary, with the aim of achieving an appropriate balance.8

However, to quantify the effects of compounds on multiple targets in a way that encompasses subsequent actions and interactions, it is necessary to explore drug activity using a Live phenotypic drug discovery (iPDD) approach.9 An iPDD platform can be used for polypharmacology against multiple known targets or in a target-independent manner (Figure), as the organism used for drug screening acts as an amplifier for the actions of all compounds, providing a complete profile of the drug. drug. iPDD platforms include those high-throughput displays based on drosophila, zebrafish, and mice.10-12

Phenotypic screening using an iPDD platform allows characterization of the full range of behavioral effects of reference drugs and data-driven target-agnostic comparison with novel compounds. The potential of machine learning-based analysis of behavioral phenotyping data can be harnessed in many drug discovery applications, such as mining libraries of iPDD-selected compounds for new hits, or screening new candidate analogs. to drugs, to speed up lengthy drug discovery. process. Furthermore, the use of animal models of disease in iPDD platforms opens up opportunities to explore phenotypic drug screening for psychiatric, neurodegenerative, and rare disorders.13

iPDD drug discovery projects can progress agnostic or become single or multi-target programs as needed. Although not knowing the mechanism of action makes the path to the clinic more difficult, target agnostic programs can be eliminated by “anti-target” panels, such as avoiding D2 antagonism in the development of new antipsychotics. Biomarkers can also be used to assess target compromise in the clinic when the mechanism of action is unknown.

Perhaps the most convincing evidence will be available very soon. Ulotaront, an antipsychotic with a novel mechanism of action now undergoing Phase III testing, was discovered and developed in partnership between Sunovion and PsychoGenics,4 using an iPDD platform (SmartCube®).

concluding thoughts

In short, AI based on targets and ligands again Drug design approaches hold promise for indications with validated hypotheses regarding the required therapeutic mechanisms of action. For complex CNS disorders, on the other hand, phenotypic examinations and associated AI methods (such as iPDD platforms) are needed. The phenotypic approach in the next era of AI-based drug design promises to provide new insights and accelerate drug discovery for CNS disorders.

Dr Bruner is director of innovation at PsychoGenics Inc. and an adjunct associate professor at the Mt. Sinai School of Medicine. He is a member of the CTF Business Advisory Council and the CureVCP Scientific Advisory Council.

References

1. AlphaGo. deep mind. Accessed October 12, 2022. https://www.deepmind.com/research/highlighted-research/alphago

2. Vedantam K. AI is making its way into drug discovery. What does it mean for biotechnology? crunchbase. October 4, 2022. Accessed October 13, 2022. https://news.crunchbase.com/health-wellness-biotech/artificial-intelligence-venture-drug-discovery/

3. Mouchlis VD, Afantitis A, Serra A, et al. Advances in de novo drug design: from conventional to machine learning methods. Int J Mol Sci. 2021;22(4):1676.

4. Correll CU, Koblan KS, Hopkins SC, et al. Safety and efficacy of ulotaront (SEP-363856) in schizophrenia: results of a 6-month open-label extension study. NPC Schizophre. 2021;7(1):63.

5. Schwartz TW, Hubbell WL. Structural biology: a moving history of receptors. Nature. 2008;455(7212):473-474.

6. Czub N, Paclawski A, Szlek J, Mendyk A. Do AutoML-based QSAR models meet the OECD principles for regulatory assessment? A case of 5-HT1A receptor. Pharmacy. 2022;14(7):1415.

7. Buvailo A. How Big Pharma Embraces AI to Drive Drug Discovery. BioPharmaTrend.com. October 8, 2018. Accessed October 12, 2022. https://www.biopharmatrend.com/post/34-biopharmas-hunt-for-artificial-intelligence-who-does-what/

8. Kondej M, Stepnicki P, Kaczor AA. Multi-target approach to drug discovery against schizophrenia. Int J Mol Sci. 2018;19(10):3105.

9. Leahy E, Brunner D. We need a new Prozac: the demand for brain drug innovation. The pharmaceutical letter. August 15, 2022. Accessed October 12, 2022. https://www.thepharmaletter.com/article/we-need-a-new-prozac-the-demand-for-brain-drug-innovation

10. His TT. Drug detection in drosophila; Why, when and when not? Wiley Interdiscip Rev Dev Biology. 2019;8(6):e346.

[ PubMed ]11. McCarroll MN, Gendelev L, Keizer MJ, Kokel D. Harnessing large-scale behavioral profiling in zebrafish to explore neuroactive polypharmacology. ACS Chem Biol. 2016;11(4):842-849.

12. Roberds SL, Filippov I, Alexandrov V, et al. A machine vision-enabled rapid mouse screening system identifies the neuropharmacological potential of two new mechanisms. frontal neuroscience. 2011;5:103.

13. Kabitzke P, Morales D, He D, et al. Autism Spectrum Disorder Mouse Model Systems: Replicability and Informatics Signature. Genes Brain Behavior. 2020;19(7):e12676.

Leave a Reply

Your email address will not be published. Required fields are marked *