You are here

Harnessing Artificial Intelligence and Polypharmacology to Discover Pharmacotherapeutics for Substance Use Disorders (R43/R44 Clinical Trials Not Allowed)


Background: Drug discovery and development is a high-cost, high-risk, and time-consuming endeavor where the failure rate of therapeutic candidates that enter clinical trials is 90%. Because of this, in recent years, artificial intelligence (AI), including machine learning ( ML) technologies, has been embraced to reduce clinical failure rates and to speed up drug discovery. AI/ML technologies are used to identify novel targets, design new molecules, conduct virtual preclinical studies, and analyze preclinical and clinical data. Multiple AI/ML-discovered compounds have advanced into Phase II clinical trials. While AI/ML is transforming drug discovery and promises to deliver new classes of medications, AI/ML-driven drug discovery for substance use disorders (SUDs) has lagged. A contributing reason is that the traditional drug discovery paradigm involves identifying highly potent and selective molecules that act on specific targets (i.e., enzymes and receptors) in a "one disease, one target, one molecule" model. Although single-target-based drug discovery has been successful in developing new medications for some indications, SUDs can be caused by complex mechanisms and, therefore, can be outside the traditional single-target-drug paradigm. Another complicating factor for SUDs is that polysubstance use is increasingly more common. Polysubstance use involves the intentional or unintentional co-use of two or more drugs, such as alcohol, tobacco, benzodiazepine, cannabis, cocaine, fentanyl, xylazine, or nitazene. Due to polysubstance use, individuals face a greater risk of toxicity from the drug combinations, leading to increased morbidity and mortality. For instance, drug poisoning deaths involving fentanyl and stimulants climbed from 0.6% in 2010 to 32.3% in 2021. Polypharmacology is emerging as a new paradigm that can advance drug development for multifactorial diseases such as SUDs. Polypharmacology is the study of how drug molecules interact with multiple targets. In the context of this notice of funding opportunity (NOFO), polypharmacology is defined as a single drug acting on multiple targets of a unique disease pathway or multiple targets of multiple disease pathways. These molecules are multi-target directed ligands (MTDL). The concepts for polypharmacology and MTDL design are supported by several approved drugs that elicit their therapeutic effect through complex polypharmacology. Some key challenges in polypharmacology are identifying the target combination, predictions of the off-target toxicities, and the rational design of MTDLs , especially when the targets of interest are not phylogenetically or structurally related. Implementing a polypharmacology strategy in drug discovery relies on advances in computational approaches such as data mining, ligand-based analysis and virtual screening. Since drug discovery is a multidimensional search and optimization process, artificial intelligence and machine learning (AI/ML) technologies trained in polypharmacology could significantly improve SUD discovery and development efforts. By unlocking insights from the network of drug-drug, protein-drug, and protein-protein interactions that drive substance use, AI/ML tools can efficiently evaluate and predict the effects of binding to multiple biological targets, thereby enhancing potential clinical efficacy ("beneficial polypharmacology"). AI/ML can also identify pathways and mechanisms leading to side effects caused by drug binding to unintended off-targets ("adverse polypharmacology"), thereby reducing unmanageable toxicity. Such AI/ML tools can identify the best targets, design effective MTDLs, and inform the in vitro and in vivo assays to characterize the effects of these ligands on targets and functions. Research Objectives The goal is to leverage AI/ML tools to identify pharmacotherapeutic development candidates with lower toxicity and higher efficacy to prevent or treat SUDs. Molecules may include new chemical entities, investigational compounds, and repurposed marketed medications. AI/ML tools can pinpoint the most promising targets, design effective ligands based on predicted drug-likeness, and guide in vitro and in vivo assays to assess the effects of these ligands on biological targets and functions. Applicants should propose and conduct activities that use AI/ML tools to streamline, enhance decision-making, and accelerate the identification of SUD pharmacotherapeutics. Applications may aim to conduct the following process: Identify and validate disease targets. Screen potential compounds to develop preliminary hits. Develop assays to test the activities of candidate compounds in vitro. Synthesize novel series of compounds; test efficacy and toxicities in vitro. Test pharmacokinetics and toxicity of selected compounds in relevant in vivo models on a non-GLP level. Conduct non-GLP in vivo toxicity and efficacy of lead compound; pharmacokinetic studies. Application Not Responsive to this NOFO The following types of projects are not responsive to this NOFO and will not be reviewed : Applications that pursue a single target. Applications solely focused on alcohol use disorders. Applications pursuing pain as a sole focus without addressing substance dependence and SUD. Applications that do not propose using computer-based approaches to augment drug discovery efforts. The Small Business Innovation Research (SBIR) program is a phased program. An overall objective of the SBIR program is to increase private sector commercialization of innovations derived from federally supported research and development. The main objective in SBIR Phase I is to establish the technical merit and feasibility of the proposed research and development efforts. In contrast, the SBIR Phase II objective is to continue the R&D efforts to advance the technology toward ultimate commercialization. Beyond the scope of this NOFO, it is anticipated and encouraged that the outcomes of successful SBIR projects will help attract strategic partners or investors to support the ultimate commercialization of the technology as a publicly available product or service. The following types of applications are accepted in response to this NOFO: Phase I. The objective of Phase I is to establish the technical merit, feasibility, and commercial potential of the proposed R/R&D efforts and determine the quality of performance of the small business recipient organization before proceeding to Phase II. Fast Track. The NIH Fast Track process allows Phase 1 and Phase II grant applications to be submitted and reviewed together. It expedites award decisions and funding of SBIR and STTR Phase II applications for scientifically meritorious projects that have high potential for commercialization. Importantly, before Fast Track Phase II can start, the National Institute on Drug Abuse (NIDA) conducts an administrative review and evaluates the achievement of the stated milestones. In addition to Approach and Investigator(s), Fast-Track milestones are assessed as Additional Review Criteria: Does the Phase I application specify milestones that should be achieved prior to initiating Phase II? Applicant’s failure to provide milestones and specific, measurable, achievable, relevant, and time-bound milestone deliverables may be sufficient reason for the peer review to exclude the application from the Fast Track review. Fast Track applicants must propose two separate sets of milestones and associated with the specific, measurable, achievable, relevant, and time-bound milestone deliverables, one set for Phase I and another set for Phase II. It is important to clearly state the go/no-go milestone decisions that will determine transition to Phase II. Failure to adequately address these criteria may negatively affect the application’s impact score. Based on peer review recommendations, NIDA Program Officer may negotiate the Phase I milestones with the Fast Track potential awardees before they are included in the terms of the award. Direct to Phase II. NIH can issue a Direct to Phase II award for a small business that has already demonstrated scientific and technical merit and feasibility but has not received a Phase I award for that project. The NIH SBIR Direct to Phase II is appropriate for Phase II submissions regardless of the funding source for the proof of principle work on which the proposed Phase II research is based. Small businesses eligible to submit Phase II applications for projects supported with a Phase I SBIR award are expected to submit the regular Phase II application as a "Renewal" application based on the awarded Phase I SBIR or Small Business Technology Transfer (STTR) project. Only one Phase II application may be awarded for a specific project supported by a Phase I award.
US Flag An Official Website of the United States Government