IUPHAR THEMED ISSUE: New Strategies for Medications to Treat Substance Use Disorders

Abstract

Substance use disorders (SUDs) and drug overdose are a public health emergency and safe and effective treatments are urgently needed. Developing new medications to treat them is expensive, time-consuming, and the probability of a compound progressing to clinical trials and obtaining FDA-approval is low. The small number of FDA-approved medications for SUDs reflects the low interest of pharmaceutical companies to invest in this area due to market forces, characteristics of the population (e.g., stigma, and socio-economic and legal disadvantages), and the high bar regulatory agencies set for new medication approval. In consequence, most research on medications is funded by government agencies, such as the National Institute on Drug Abuse (NIDA). Multiple scientific opportunities are emerging that can accelerate the discovery and development of new medications for SUDs. These include fast and efficient tools to screen new molecules, discover new medication targets, use of big data to explore large clinical data sets and artificial intelligence (AI) applications to make predictions, and precision medicine tools to individualize and optimize treatments. This review provides a general description of these new research strategies for the development of medications to treat SUDs with emphasis on the gaps and scientific opportunities. It includes a brief overview of the rising public health toll of SUDs; the justification, challenges, and opportunities to develop new medications; and a discussion of medications and treatment endpoints that are being evaluated with support from NIDA.

Graphical abstract

1. Introduction

The development of medications to treat substance use disorders (SUDs) has been a public health priority for the past two decades but given the recent dramatic increases in mortality due to drug overdoses, the need to develop medications to treat SUDs has become particularly urgent. However, the development of medications that succeed and make it into the clinic is limited and costly. The probabilities for discovering a new medication and progressing it to clinical trials is in the range of 35%, for successfully taking it from Phase 1 trials to regulatory approval is < 15%, and the process takes 12 to 15 years [1]. Medications development is driven by market forces and therapeutics for SUDs struggle to secure adequate private funding. In consequence, most research of medications for SUDs is supported by government agencies, such as the National Institute on Drug Abuse (NIDA).

The purpose of this review is to provide a general description of new research strategies for the development of medications to treat SUDs with emphasis on the gaps and scientific opportunities. It includes a brief overview of the public health extent of SUD, the justification, challenges, and opportunities to develop medications for SUDs, and a discussion of treatment endpoints that are being evaluated with support from NIDA.

2. Public health extent of SUD

Results from National Survey of Drug Use and Health (NSDUH) of 2021 [2] showed that among people 12 and older and based on the American Psychiatric Association (APA) Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) criteria [3], 16.3 million people met criteria for cannabis use disorder (CUD), 5 million for prescription opioid use disorder (which include products containing hydrocodone, oxycodone, tramadol, codeine, morphine, prescription fentanyl [not illicitly produced fentanyl], buprenorphine, oxymorphone, and hydromorphone, as well as Demerol^®, methadone, or any other prescription pain reliever), 4.1 million for stimulant use disorder (StUD) which includes cocaine, methamphetamine, and any other prescription stimulant (i.e., amphetamine products [except methamphetamine], methylphenidate products, anorectic stimulants, modafinil, or any other prescription stimulant), and 1 million people with heroin use disorder. In total, it is estimated that approximately 24 million Americans have a drug use disorder [2].

Furthermore, the current drug overdose mortality rates are at the highest level that they have ever been. Provisional data from the Centers of Disease Control (CDC) National Center for Health Statistics (NCHS) indicate that during the 12-month period ending in March of 2023, there were approximately 110,000 deaths due to drug overdoses. Of them, about 76,000 were associated with synthetic opioids, mainly fentanyl, 35,000 to methamphetamine and other psychostimulants, and 28,000 to cocaine, most of them in combination with fentanyl [4]. The authors recognize that nicotine and alcohol are also significant public health problems; however, due to space constraints this review focuses mostly on OUD, StUD and CUD.

Although there are medications approved by the Food and Drug Administration (FDA) to treat opioid use disorder (OUD) (i.e., methadone, buprenorphine, and naltrexone^*1), to treat acute opioid withdrawal (i.e., lofexidine), and to reverse opioid overdose (OOD) (i.e., naloxone and nalmefene), there are no FDA approved medications for StUD (cocaine or methamphetamine) nor for CUD or polysubstance use disorder (PSUD).

Medications for OUD (MOUD) are effective in decreasing craving, withdrawal, and opioid use and in preventing overdoses. However less than 20% of those who need them are prescribed MOUD and of those who get them 40–50% discontinue them. There are multiple barriers to MOUD initiation including lack of knowledge and stigma about these medications, insufficient providers trained to prescribe them, inadequate reimbursement by insurance providers and, in the case of methadone, difficulties accessing treatment programs [5, 6].

Given the enormous morbidity and mortality associated with SUD, the small number of FDA approved medications, and the barriers to prescribe them, it is urgent to develop interventions to treat these disorders.

3. Challenges to develop medications for SUDs

The limited number of FDA approved medications for SUDs reflects in part the lack of interest of pharmaceutical companies to invest in this area. Multiple reasons account for their disinterest, including the fact that clinical trials for SUDs are particularly challenging since the patient population tends to have a greater burden of social, economic, and legal challenges and tend to be distrustful of the medical system due to prior experiences of discrimination as well as persistent stigma of addiction as a disorder. Additionally, requirements for evidence of abstinence for approval of a medication is a very high bar to achieve that discourages sponsors and industry to pursue medications development for SUD.

Another hurdle has been the lack of standards regarding clinical trial designs and outcomes. A recent FDA draft guidance for industry to develop medications to treat moderate or severe StUD emphasizes that the study population should be carefully selected according to preferred route of administration (oral, smoked, intravenous, or intranasal), motivation to use the stimulant, and motivation to seek treatment (drug abstinence initiation versus relapse prevention) [7]. The guidance also recommends that clinical trials should be “of sufficient duration to achieve a meaningful change in stimulant use” and suggest that clinical trials should last 3 months or longer and involve a controlled observation period of 6 months. Although the FDA guidance is an important attempt to set trial standards, it may increase the costs of medication development that might further discourage investments in SUD treatment. Hopefully, this guidance and the one for OUD [8] will be revised as new scientific evidence emerges.

4. Scientific opportunities to develop new medications

Among the multiple scientific opportunities relevant to SUD medication development are the availability of fast and efficient tools to screen new molecules, discovery of new targets and medications, use of big data to explore large clinical data sets, use of machine learning and artificial intelligence (AI) applications, and use of precision medicine approaches to uncover individual factors associated with treatment response.

4.1. Molecular screening

Molecular screening is a high-throughput method to identify the biological activity of small molecules and identify new drug candidates for treatment of specific health conditions. Multiple compounds can be tested against a specific biological target, such as a protein or enzyme, in a short time. It allows scientists to rapidly and efficiently discover compounds that activate or inhibit the target, identify new targets, and study the mechanism of action of molecules.

The most common molecular screening assays include: 1) binding assays that measure the ability of a compound to bind to a specific biological target, 2) functional assays that evaluate the activity of a compound on a specific biological target, and 3) cell-based assays to determine the ability of a compound to affect the growth or survival of cells. All assays have the risk of producing false positives when compounds bind to a target but are devoid of biological activity and yield false negatives when active compounds are not identified by the assay.

There are multiple publicly available small-molecule screening databases. For example, PubChem (https://pubchem.ncbi.nlm.nih.gov) is the world’s largest collection of freely accessible chemical information searchable by molecule name, formula, structure, chemical and physical properties, biological activities, safety and toxicity, patents, literature citations and more [9]. Similarly, the Chemical Entities of Biological Interest (ChEBI, https://www.ebi.ac.uk/chebi/) is freely available and provides a dictionary of molecular entities, with emphasis on ‘small’ chemical compounds, which is part of the MBL’s European Bioinformatics Institute (EMBL-EBI) and includes analysis tools, including AI [10]. Molecular screening can help in the identification of new drug targets and molecules for SUD pharmacotherapies and biomarkers to monitor disease severity, evaluate treatment response, and predict prognosis [11]. The increasing expansion of AI will undoubtedly accelerate the role of molecular screening in the discovery and development of treatments for SUDs [12].

4.2. New targets

Advances in the understanding of the neurobiology of SUDs and the use of molecular screening tools are providing opportunities to advance development of SUD medications. Particularly exciting are the findings from epigenetic studies [11]. Repeated use of addictive drugs can alter the expression of genes in reward-related brain regions via epigenetic modifications that are associated with molecular, cellular, and neurocircuit function and drug using behaviors. Preclinical studies have shown that compounds targeting epigenetic modifiers such as non-selective histone deacetylase (HDAC) inhibitors reduced cocaine self-administration [40]. For example, the HDAC3-selective inhibitor, RGFP-966, enhanced extinction of cocaine conditioned place preference (CPP) [41], and sirtinol, a class III histone deacetylase inhibitor reduced cocaine conditioned place preference [42].

A potential limitation of epigenetic-based pharmacological interventions is the risk of inducing unpredictable genotypical and phenotypical changes and associated adverse effects. Further preclinical and in vitro research is needed to ensure the safety of epigenetic modifiers.

In 2019 Rasmussen et al. provided a list of what was considered then the ten most relevant SUD targets. Continuous research findings have expanded the list. Table 1 and 2 summarize the pharmacological targets for StUD and OUD, respectively, that are currently being evaluated by NIDA supported studies.

Table 1.

Pharmacological Targets for StUD

11-beta-hydroxysteroid dehydrogenase1 inhibitor

Cholecystokinin receptors (CCK) – A and -B antagonists

Dopamine transporter (DAT) inhibitor

Ghrelin antagonist

Kappa opioid receptor (KOR) antagonist

Kv7.2–7.5 voltage-gated potassium channel opener

Metabotropic glutamate receptor 5 (mGluR5) Negative Allosteric Modulator

Muscarinic acetylcholine receptor (mAChR) M4 subtype agonist and partial agonist

N-methyl-D-aspartate (NMDA) receptor antagonist

Nociceptin opioid peptide (NOP) partial agonist

Phosphodiesterase-4B inhibitor

Sigma receptor antagonist/ DAT inhibitor

Table 2.

Pharmacological Targets for OUD

5-hydroxytriptamine (5HT) 2A agonist

Acetate-dep acetyl-coenzyme A (CoA) synthetase-2 inhibitor

Arylepoxamide agonist

Biased mu opioid receptor (MOR) agonist Cannabinoid 1 antagonist

Biased mu opioid receptor (MOR) agonist

CCK-A and -B antagonists

Dopamine D2/5-hydroxytriptamine 1A receptor (5HT1A) partial agonist

Ghrelin antagonist

Kappa opioid receptor (KOR) antagonist

mGluR 2 Positive Allosteric Modulator (PAM)

mGlu2/3 PAM

MOR Partial Agonist and 5HT2A/Dopamine D1 Antagonist

Mu-Opioid Receptor (MOR) biased agonist

NaV1.7 inhibitor

N-methyl-D-aspartame (NMDA) receptor antagonist

NOP receptor/MOR agonist

NOP/MOR partial agonist

Orexin (OX)-1 receptor antagonist

OX-1/2 and KOR antagonist

Phosphodiesterase-4B inhibitor

Soluble epoxide hydrolase inhibitor

4.3. Medication Repurposing

Medications repurposing is the process of identifying new therapeutic uses of medications already investigated or approved for other indications. It can include medications that failed in clinical trials for an indication, but the pharmacological mechanism may be appropriate for a different indication. This approach has been widely used for SUDs, in particular medications tested for other psychiatric disorders such as depression or obsessive-compulsive disorder. Depending on the mechanism of action and known safety toxicology of the compound, the evaluation may require additional testing to evaluate its safety in participants with SUDs, even if it had already been administered to humans. This is because of potential drug-drug interactions between the medication and the addictive drug. FDA guidance includes language in this respect [7].

Medication repurposing can be guided by overlap in the neurobiology of SUDs and the action of the repurposed medication, or by in vitro assays, animal models, clinical trials results, analyses on databases of health records, or serendipitous findings from patients in treatment for other conditions. AI may be a formidable tool to identify and test repurposed medications. One of the advantages of repurposing is that sponsors already collected pre-clinical or clinical data that can be used to support the medication approval and, thus, the risk of failure in late stages of development may be lower. Medication repurposing can also help accelerate the development of SUD treatments, for which there is limited industry investment and urgent needs.

Some repurposed FDA approved medications currently being tested for StUD include zolmitriptan (migraine), mirtazapine (depression), clavulanic acid (antibiotic), cariprazine (atypical antipsychotic and antidepressant), pioglitazone (type 2 diabetes), rotigotine (Parkinson’s and restless legs syndrome). For OUD, some of the repurposed medications being investigated include brexpiprazole (antipsychotic), cannabidiol (antiepileptic), dexmedetomidine (anesthetic), guanfacine (antihypertensive), ketamine (anesthetic and antidepressant), Lemborexant (insomnia), liraglutide and semaglutide (diabetes), lofexidine (antihypertensive), olanzapine (antipsychotic), pramipexole (Parkinson’s and restless legs syndrome), pregabalin (analgesic, antiepileptic), and suvorexant (insomnia). Repurposing medications can be a cost-effective approach to expedite development of SUD treatments.

4.4. New formulations

Evaluation of new formulations of medications already approved by the FDA for SUDs is another cost-effective method to improve SUD treatment outcomes. Specifically, a major clinical challenge is the difficulty of SUD patients to adhere to their medication. Development of long-acting formulations of medications is a good option to enhance retention and improve outcomes. NIDA is currently supporting development of long-acting formulations of naltrexone (3-month, 6-month, and 12-month), buprenorphine (weekly oral), methadone (weekly oral), and nalmefene (6-months).

4.5. Artificial Intelligence

AI allows unprecedented exploration of information in pre-clinical and clinical pharmacology. Pharmaceutical industry, biotechnology companies, investors, and biomedical research funding agencies have growing interest in using AI to accelerate and improve medication development [1]. AI is likely to accelerate the discovery, development, and dissemination of new treatments for all kinds of diseases. This includes but not limited to identification of new targets and molecules, biomarker development, identification of disease pathways, new diagnostics, pre-clinical and clinical toxicology assessments, clinical trial design and implementation, risk stratification, evaluation of treatment endpoints, adverse event predictions, personalized medicine, pharmacogenomics, pharmacovigilance, mitigate health disparities, treatment delivery services, etc.[13–15]. The SUD field is not an exception.

Among the multiple types of AI currently available these might be particularly relevant for SUD: 1) machine learning (ML), which involves development of computer algorithms using historical data as input to predict new output values, 2) deep learning, a subcategory of ML that uses artificial neural networks to learn from data and is inspired by the structure and function of the human brain using multiple layers of interconnected processing nodes, 3) natural language processing (NLP), which deals with the interaction between human language and computers and includes answering questions and summarizing text, 4) computer vision (CV) to program computers to capture and interpret visual information, and 5) robotics that combine and integrate information from sensors with actuators using AI applications [13, 15, 16].

A recent review provides an overview of practical foundations of ML for addiction research [17, 18]. The first part focuses on methods and description of AI, including open-source programming tools, and examples of use in addiction research. The second part focuses on workflow and use cases and describes the application of ML in study design, data collection, data pre-processing, modeling, and results communication in addiction research. The review highlights the limited literature on ML approaches and invites investigators to advance the use AI in the SUD field.

Among the AI tools for drug discovery, AlphaFold is freely available and used to predict the 3D structure of proteins from their amino acid sequence. It has over 200 million entries and allows downloads for the proteome of humans and of 47 other organisms [19]. Other systems include, DeepAffinity [20], which predicts drug-protein interactions and DeepTox [21], which predicts toxicity of candidate molecules.

For clinical trials, AI can revolutionize their design, implementation, data analysis and interpretation of results. For example, for several years, clinical trials reporting guidelines such as the Consolidated Standards of Reporting Trials (CONSORT) and the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) have been instrumental in ensuring transparency in the evaluation of new interventions. However, there has recently been recognition that interventions involving AI also need to follow the guidelines to ensure the impact on health outcomes. For that purpose, both systems have been expanded to include reporting guidelines for clinical trials of interventions with an AI component (CONSORT-AI and SPIRIT-AI) [22, 23]

AI methods are increasingly applied in neuroimaging research, including in SUD. A recent review, published by Yang L. et al [24], summarizes the results of research that applied ML in neuroimaging to evaluate electrophysiological, morphological and functional brain measures in the diagnosis, risk prediction, and treatment of SUDs. It reports the use of supervised learning and deep learning in the diagnosis and prediction of SUD for alcohol, nicotine, methamphetamine, heroin, cocaine, cannabis, and codeine use disorders. Most studies focused on the diagnosis of SUD (n=23), some on prediction of relapse (n=6) and treatment responses (n=3). The authors propose the integration of multimodal large data sets from brain imaging and clinical assessments to explore hidden, nonlinear, and high-order features that cannot be revealed by traditional statistical models and that are clinically relevant.

AI can also contribute to advancing the understanding of the social determinants of health (SOHD) in SUD, which has implications to the delivery of care, prevention, and public health. For example, patients in rural areas could benefit from predictive models that assign medications for OUD based on resources and individual patient needs. AI can be used for patients in remote areas, with limited access to care, to send reminders to take medication and improve adherence to the treatment plan [14]. AI can also be used to match patients to sociocultural preferences and potentially improve outcomes.

Challenges of AI in medications development include 1) limited access to large data sets that are of good quality with low missing values, and free of bias or discrimination (crucial since ML algorithms learn from data). 2) poor definition of variables and heterogeneity in data collection, 3) risks of breach of confidentiality, 4) legal liability from adverse outcomes because treatment providers are accountable for decisions regarding patient care, yet there is no liability framework for AI in patient care. The lack of demand of AI services can become a barrier to train and fund professionals that gain the knowledge and expertise to apply AI in their field [1]. Therefore, there is a need to increase the use of AI in the SUD field and simultaneously develop a cadre of professionals in this area that can apply AI and make a significant contribution to advance its use.

To protect against some of the risks of AI, an Executive Order was issued on October 30th, 2023, to ensure that America leads the way in seizing the promise and managing the risks of AI. It establishes new standards for AI safety and security, protects privacy, advances equity and civil rights, stands up for consumers and workers, and promotes innovation and competition. This order builds up on previous voluntary commitments from 15 leading companies to drive safe, secure, and trustworthy development of AI [25]. In sum, AI has the potential to revolutionize the way research is conducted, data are analyzed and interpreted, and care is provided to patients with SUDs, within an acceptable legal and ethical framework.

4.6. Precision Medicine

Precision medicine or personalized medicine is based on the notion that not all patients respond the same way to a given treatment. It uses genetic, environmental, and other individual factors to determine the best treatment for patients with specific characteristics. Precision medicine is greatly benefiting from multiple scientific and technological advances including: 1) gene sequencing to conduct genetic analysis more rapidly and at lower costs, 2) data and computational sciences to allow the collection, management, and analysis of enormous data sets at unprecedented speed, and 3) mathematical modeling to support AI tools that are increasingly popular and more affordable. Not surprisingly, there has been an exponential increase in publications focused on precision medicine.

Precision medicine has the potential to determine the most cost-efficacious intervention to prevent or treat SUD. Research is ongoing trying to identify associations between specific gene variants with predictions of clinical manifestations and treatment outcomes [26]. For example, the single-nucleotide polymorphism(SNP) rs73568641) in the gene that codes for the mu-opioid receptor OPRM1 was associated with both a need for a higher dose of methadone for OUD treatment and higher doses of morphine to treat pain [27]. Precision medicine will require that it considers social determinants of health to reduce health disparities among SUD populations [14]. Advances in statistical estimation approaches incorporating machine learning while retaining valid inference can allow more precise estimates of the effectiveness of certain treatment regimens among individuals who share some specific characteristics [28, 29].

5. Treatment approaches and clinical endpoints

A major challenge in medication development for SUDs is the required clinical endpoint of abstinence for New Drug Application (NDA) approval by the FDA. Though the StUD FDA guidance recognizes the need for a certain amount of reliance on self-report it highlights that it is not persuasive by itself. With regards to urine drug tests, the guidance states that there is no evidence to support a frequency of biological testing and suggests a balance between burden to the subject and some degree of biological confirmation of drug use self-report. It also states that “sustained period of negative urine toxicology findings, indicating abstinence, could be a valid surrogate for clinical benefit” [7].

An alternative endpoint that has long been discussed in the field is “reduction in drug use”, which is an endpoint that acknowledges the difficulty of most patients with SUDs to achieve abstinence. There are several reports showing that reduction in drug use is associated with clinical improvements. In fact, a recent publication by Kiluk [30] and others showed that patients with cocaine use disorder (CocUD) with at least 75% of their urine drug screens negative for cocaine showed short-term and long-term clinical benefits. However, the FDA guidance does not recommend the use of the phrase “reduction in stimulant use” and prefers “change in pattern of stimulant use” and suggest that subjects achieving a target pattern of use days per period could be an “acceptable endpoint with a prespecified target pattern of use that defines a relevant within-subject response” [7].

The DSM-5 criteria have been widely used to determine eligibility of subjects to enter clinical trials and, for some disorders, as endpoints. However, the FDA guidance does not recommend using changes in the number of DSM-5 diagnostic criteria as an endpoint. It accepts early or sustained remission, which are defined as meeting no DSM-5 criteria for StUD for between 3 and 12 months or at least 12 months, respectively. Clinicians would agree that this endpoint is unattainable in most patients. Therefore, sponsors are very unlikely to select this endpoint.

Within the backdrop of the FDA guidance for OUD [8] and for StUD [7], we next provide insights about some endpoints that are used in SUD treatment research. These include drug craving, relapse prevention, sleep disturbances, polysubstance use disorders, drug overdose, and medication combinations. They were selected because patients report them as important factors in drug recovery and research supports their relevance.

5.1. Drug Craving

A treatment endpoint of interest to patients and providers is drug craving, which is commonly defined as the strong and sometimes overwhelming desire to use a drug [31]. The DSM-5 includes craving as one of the diagnostic criteria for SUD. Drug craving can be triggered by physiological, psychological, and environmental cues and it has been associated with drug use relapse and poor treatment outcomes [32, 33]. A metanalysis of 237 studies representing 51,788 human participants found a significant association of all cue and drug craving measures with future drug use and relapse [34]. The findings cement the importance of drug craving in addiction and relapse and strongly support its use as an outcome in treatment development.

Craving has been associated with dopamine release in striatum, amygdala, and prefrontal cortex (anterior cingulate and medial orbitofrontal cortex) and opioid peptides release in the anterior cingulate and frontal cortex [35, 36]. Imaging studies have documented reactivation of dopamine signaling during acute craving episodes and impaired activity in frontal cortex, which may be associated with reductions in inhibitory control, working memory, and self-regulation [37]. Craving has also been associated with activation of the insula, which possibly reflects interoceptive cues, and may involve CRF activation [36, 38].

Based on these and other findings, multiple pharmacological compounds have been evaluated to reduce craving. A recently published metanalysis evaluated the endpoints used in 130 clinical medications trials for CocUD. Most of the studies considered abstinence and retention as the main outcome in cocaine treatment and craving a secondary outcome. They found that in short-term treatment, acute phenylalanine-tyrosine depletion, clonidine, fenfluramine, meta-chlorophenylpiperazine (m-CPP) and mecamylamine had promising effects. In long-term treatment, amphetamine, biperiden, carbamazepine, lisdexamfetamine, lorcaserin, methamphetamine, mirtazapine, pioglitazone, progesterone, guanfacine, levodopa, nefazodone showed promising effects against craving. The authors note the high heterogeneity of craving assessments across the studies and report that they found a direct association between craving and cocaine use, which highlights the importance of craving as a treatment target [39].

From the regulatory perspective, the FDA guidance for StUD states “Craving has not been consistently defined or understood, but it is viewed as a significant source of distress for patients and could be a suitable target for treatment. FDA encourages the development of a suitably developed, fit-for-purpose measure of craving and envisions incorporating claims about effects on craving as secondary endpoints for drugs that are effective treatments for stimulant use disorder. We are also open to data demonstrating the ability of craving modification to predict clinical benefit to consider craving as a potential primary endpoint.” For OUD, the FDA guidance is less specific and states “Sponsors interested in using a reduction in craving endpoint should contact the division about developing a fit-for-purpose instrument to measure craving or the urge to use opioids to complement other endpoints and to determine how the endpoint correlates with sustained clinical benefit.” There is no FDA guidance for the development of medication to treat CUD. In consequence, there is an urgent need to produce the data and results necessary to support the use of drug craving as a treatment endpoint for FDA approval of a medication.

Despite the uncertainties about the regulatory approval of medications to reduce drug craving, several compounds are being tested for that indication (Table 3).

Table 3.

Medications under evaluation to reduce cocaine craving.

Medication	Mechanism of Action
OMS182399	Selective inhibitor phosphodiesterase 7 (PDE7
PPL-138	Non-selective opioid receptor partial agonist at nociception opioid receptor (NOP) and mu receptors, and antagonist activity at kappa and delta
Cariprazine	High dopamine D3 affinity, serotonin, and histamine
Clavulanic Acid	Activation of GLT-1
Pioglitazone	Peroxisome proliferator-activated receptor gamma (PPAR-γ) agonist

5.2. Relapse prevention

Definitions of relapse vary but in general it refers to the use of an addictive drug after a period of voluntary abstinence. Similarly, relapse prevention is not clearly defined and there are significant discrepancies between what regulatory agencies consider relapse (and thus, relapse prevention) for medication approval purposes and the perspective of relapse for clinicians and investigators. Moreover, there are questions about the validity of animal models of relapse and ethical challenges for conducting human laboratory studies of relapse. Thus, the development of medications to prevent drug use relapse is challenging.

The FDA guidance on endpoints for OUD medications intended to reduce the risk of relapse mentions that “patients already stable on other treatments for OUD should be studied, and the comparator should be an approved therapy”. The guidance further defines that stable patients as those who “have ceased problematic drug use or use only very sporadically while receiving treatment” while the individuals may or may not meet DSM-5 criteria [8]. The guidance also suggests that clinical trials could use the proportion of patients meeting DSM-5 criteria for remission of OUD at the end of the trial as a primary or secondary efficacy endpoint. Interestingly, a long-acting intramuscular injection of the mu-opioid antagonist naltrexone was approved by the FDA for relapse prevention of patients with OUD who have been abstinent from opioids for a minimum of 7 to 10 days before receiving the injection. The label indicates that some patients may require a naloxone challenge before the injection to confirm that opioid withdrawal is not precipitated.

The FDA guidance to develop medications to treat StUD [7] suggests that some medications may be more suitable for preventing relapse in subjects who are abstinent at baseline and that a suitable primary endpoint could also be the proportion of subjects meeting DSM-5 criteria for remission at the end of the clinical trial. Therefore, sponsors interested in developing medications for relapse prevention should consult with the FDA about the requirements to obtain approval for that indication.

Despite the lack of clarity about the definition of relapse, relapse prevention, and the endpoints required for approval of a relapse prevention medication, research is ongoing, both at the pre-clinical and clinical level. At the pre-clinical level, the reinstatement model is commonly used to investigate relapse and three models have been proposed: drug-, cue- and stress-induced reinstatement of drug seeking. Animal models have helped identify the neurobiological mechanisms underlying relapse to drug seeking.

Drug-induced reinstatement seems to involve glutamatergic projections from the prelimbic prefrontal cortex to the nucleus accumbens and these projections seem to be modulated by dopamine D1 and D2 receptors in the frontal cortex. Cue-induced reinstatement also appears to affect glutamatergic projections but, not only from the prelimbic prefrontal cortex but also basolateral amygdala and ventral subiculum to the nucleus accumbens, where there is dopamine modulation. Stress-induced reinstatement seems to be associated with activation of the corticotropin release factor (CRF) and norepinephrine, in the central nucleus of the amygdala and bed nucleus of the stria terminalis. Thus, it is possible to hypothesize that based on animal models, craving may be associated with strong glutamatergic responses in the pre-frontal cortex and medications that modify the glutamatergic systems may have an anti-relapse effect [40].

Several clinical trials testing the efficacy of compounds that showed effects in the reinstatement models failed to show benefit. Extensive reviews of these and other animal models and the neurobiology of relapse and the effects of different compounds have been published and a remaining big question is about the predictive validity of animal models for drug relapse in humans and, more importantly, the predictive validity for FDA-approval [41–43].

NIDA is supporting clinical studies with several new formulations of naltrexone and nalmefene to evaluate their efficacy for relapse prevention. In addition, methocinnamox an opioid antagonist that is long-lasting, potent, and selective is being evaluated for the treatment of OUD and OOD and that, based on its pharmacology, has the potential of being efficacious for relapse prevention [44]. More research is needed to increase the crosstalk between pre-clinical and clinical investigators as well as regulatory agencies to direct the science of drug relapse more clearly toward obtaining FDA approval of medications for this indication, which will in turn serve to determine the predictive validity of preclinical models.

5.3. Sleep Disturbances

Bidirectional relationships between sleep problems and substance use problems have long been described [45]; such that sleep disturbances can be a risk factor for SUD [46] or a consequence of SUD. Sleep disturbances affect all phases of the addiction cycle, including initiation, maintenance, and relapse [47]. In addition, sleep disturbances may negatively influence the outcome of SUD treatment.

The most frequent reported sleep problems in SUD are insomnia, circadian rhythm sleep disorder–delayed sleep phase type (CRSD-DSP), and sleep-related breathing disorder (SRBD). Insomnia is characterized by problems with sleep continuity, such as falling asleep and/or staying asleep, CRSD-DSP is characterized by going to bed later in the night and awakening later in the morning, and SRBD is a disruption of sleep by respiratory events [48].

SUDs affect the physiology of multiple neurotransmitters systems, including those that regulate the sleep-wake cycles such as acetylcholine, GABA, dopamine, glutamate, norepinephrine, and the orexin systems [49–51]. For example, chronic opioid use can produce multiple sleep disturbances, including SRDB because the opioid-induced respiratory depression interferes with breathing, decreases arousal, reduces amplitude of reflex responses, and can result in upper airway disfunction [52]. Estimates of up to three-quarters of individuals receiving methadone or buprenorphine treatment for OUD have a sleep problem [53] and SRBDs occur in most chronic opioid users. A metanalysis showed that the prevalence of central sleep abnormalities in people taking opioids chronically was 24%, while the prevalence in the general population is 1% [52].

Acute cocaine use increases arousal, and binge use is associated with periods when the individual does not sleep or sleeps briefly to recover and continue using. The effect of cocaine withdrawal on sleep has been widely investigated but the results are not conclusive, in part due to variability in experimental design, their outcomes, and the interpretation of the results. A recent comprehensive review summarizes findings on cocaine use and sleep problems and concludes that, while cocaine withdrawal is associated with sleep dysfunction the specific nature of the dysfunction cannot be determined [54].

With regards to CUD, it has been reported that sleep problems are a frequent reason for people to use cannabis and sustaining its use. Cannabis use has been associated with reduced latency to sleep onset, a decrease in rapid eye movement sleep, and an increase in Stage 3 sleep; however, there is evidence of developing tolerance to these effects. Interestingly, a clinically relevant feature of cannabis withdrawal is difficulty to fall sleep and an increase in vivid/strange dreams [55, 56]. Cannabis users who seek treatment frequently have significant sleep dysfunction, and poor sleep at treatment entry has been associated with relapse [57]. Therefore, cannabis as a hypnotic agent is not recommended due to the development of tolerance to its hypnotic properties, risk of long-term sleep disturbance, and subsequent withdrawal symptoms on abrupt cessation with compounding sleep problems [55–57]

Improving sleep during abstinence can reduce relapse. However, most conventional sedatives have high abuse liability and, thus, are contraindicated for persons with SUDs. Accordingly, there is significant interest in novel non-abusable sleep medications that can be used as standalone treatments for SUD or as adjuncts to medications such as buprenorphine or methadone for the management of sleep disturbance in SUD populations. The hypothalamic orexin (hypocretin) neuropeptide system is a promising new target to treat SUD or improve sleep disturbances in SUD. Orexin neurons widely project to reward and motivation‐related brain regions, including the dopaminergic nuclei involved in drug rewards and SUD. Studies conducted with the dual orexin receptor 1 and 2 antagonist (DORA), suvorexant, have shown that it may have two therapeutic effects: 1) a “top down” effect of normalizing sleep and improving executive control over drug seeking, and 2) a “bottom up” effect of reducing craving by suppressing the activity of brain networks implicated in craving. A recently conducted clinical trial showed that suvorexant (0, 20, and 40 mg) during a buprenorphine/ naloxone taper was associated with improvements in total sleep time and reduction of opioid withdrawal symptom. DORAs have been designated schedule IV medications though reports suggest limited abuse liability. Currently, NIDA is funding several studies with DORAs to manage sleep disturbances and to evaluate its effects in withdrawal and craving in SUD [58–60].

Adequate treatment of sleep problems may prevent the development of SUD or help improve outcomes of those with SUDs. More research is needed to elucidate the relationship between SUD and drug-seeking behaviors, the long-term consequences of SUDs on sleep quality, whether improvements in sleep quality are associated with improvements in SUD outcomes, and the effect of DORAs and single orexin receptor antagonists (SORA) on sleep function and treatment outcomes in SUDs.

5.4. Polysubstance Use Disorders

Poly-substance use is a major challenge in the treatment of SUD. Reports indicate that individuals who seek SUD treatment use an average of 3.5 substances and only 9% use only one. The rest use 2 (18%), 3 (22%), 4 (13%) and 5 (38%) substances [61]. Patients with PSUD have poorer treatment outcomes, including substance use and retention in treatment, higher rates of relapse, more medical and psychological comorbidities, and higher mortality rates compared to those having a single SUD [62]. Research on treatments for PSUD is an urgent priority.

Most clinical trials of medications for SUD focus on a single drug and multiple comorbid substance use disorder is often an exclusion criterion. This is because most animal models of addiction focus on individual drugs, and it is more likely to find a treatment effect in more homogeneous populations. In addition, regulatory agencies prefer a single diagnosis to claim treatment efficacy and obtain approval. However, limiting the inclusion criteria for clinical trials to a single substance use risks missing drug-drug interactions, decreases the generalizability of the results, and will not reflect the substance use situation in the real world [63].

Neurobiological research has uncovered overlap between various SUDs and investigators are now evaluating medications that may be effective for more than one disorder. Such strategies can be focused on common targets that influence addictive behaviors in general, for example, D3R antagonists and GLP-1 receptor agonists, or common clinical symptoms such as medications to reduce craving, anxiety, or insomnia among many others.

Currently, the only available FDA approved medication with potential to treat PSUD is naltrexone because it is approved separately for alcohol use disorder and for OUD. Naltrexone blocks the effects of alcohol on the endogenous opioid release and attenuates the euphoria produced by ethanol [64]. In clinical trials, naltrexone reduced cravings, relapse rates, and frequency of drinking [65, 66]. One study found that depot naltrexone was safe, acceptable, and feasible in patients with OUD and AUD in the context of HIV treatment [67]. There is need for more studies of naltrexone for PSUD.

Nalmefene has an oral formulation approved in Europe for AUD and an intranasal formulation approved for OUD. Nalmefene appears to selectively decrease cocaine self-administration in male rhesus monkey and may be effective for OUD relapse prevention. Thus, nalmefene may have the potential to be effective for comorbid AUD, OUD, and StUD. Also, buprenorphine has been demonstrated to be effective for comorbid OUD and cocaine use disorder in preclinical and clinical studies [68, 69]. Confirmation of these results is still pending.

Based on the mechanism of action, some compounds might be beneficial for PSUD. For example, dopamine type 3 receptor (D3R) antagonists appear to reduce the reinforcing effects of drugs such as cocaine and alcohol, and opioids [70–72]. Orexin receptor antagonists appear to block both cocaine- and morphine-induced neuroplasticity. Glucagon-like-peptide 1 (GLP-1) agonists appear to prevent the ability of multiple addictive drugs to activate the mesolimbic dopamine system in rodents and to attenuate cocaine-mediated behaviors [73], oxycodone seeking without compromising its anti-nociceptive effects [74], and the reinforcing effects of alcohol [75, 76]. Research to evaluate the safety and efficacy of these compounds in preclinical models and clinical trials of PSUD is needed.

5.5. Medication Combinations

Medication combinations (MC) can be a useful strategy to treat SUDs. Sometimes more than one medication might be needed because of medical and psychiatric comorbidities, to improve the efficacy of SUD medications, or to treat specific clinical manifestations of SUD. For example, craving, withdrawal, or be combined to obtain a synergistic effect or to prevent undesirable effects of a medication. In fact, to address the latter, the combination of buprenorphine plus naloxone was developed and approved by the FDA to prevent the potential diversion of buprenorphine [77].

Several MC are being evaluated to treat SUD. A combination of oral bupropion (450 mg/day) and injectable extended-release (XR) naltrexone (380 mg every 3 weeks) was evaluated in a 12-week, multisite, double-blind, placebo-control clinical trial for MUD and the results showed that it was significantly better than placebo [78]. Moreover, a secondary analysis of this clinical trial showed that this MC was associated with significant reductions in cigarette smoking, suggesting that bupropion plus depot naltrexone may be effective for comorbid MUD and nicotine use disorder [79].

Based on some promising preliminary results [80], a combination of buprenorphine plus XR-naltrexone is being evaluated for different OUD. The rationale is that naltrexone will block the opioid agonist effect while maintaining the kappa antagonist effects of buprenorphine. For OUD the aim is to evaluate the effect of the MC on relieving protracted opioid withdrawal, reducing craving, improving mood, and preventing opioid relapse (NCT05011266). Anther MC under evaluation is the combination of metyrapone and oxazepam. The hypothesis is that the combination may increase neuroactive steroids, most notably tetrahydrodeoxycorticosterone, in the medial prefrontal cortex and amygdala and thus, reduce cocaine seeking and taking by decreasing activity within the HPA axis [81]. A small pilot clinical trial showed promising results [82] and recently a double-blind phase 2 clinical trial was completed but the results are not publicly available (NCT02856854).

Combinations of biologics and medications for SUD also hold promise. Either vaccines or monoclonal antibodies could be combined with, for example, medications to reduce craving or withdrawal symptoms. There is no clinical research yet since there are currently no biologics approved by the FDA. Currently, a multivalent anti-opioid vaccine (NCT04458545) and an anti-fentanyl monoclonal antibody (NCT06005402) are being evaluated in clinical trials. If successful, they will permit evaluation of their combinations with MOUD. This type of combination has a distinct regulatory pathway with the FDA that will require further exploration [83].

MC has some challenges, including higher risk of drug-drug interactions, side effects, and potentially a more challenging regulatory approval pathway; however, they may be a promising approach as has been the case for other complex medical conditions (e.g., HIV, cancer).

5.6. Drug overdose

Drug overdose is the leading cause of death for Americans under the age of 50. Currently, with support from NIDA, naloxone and nalmefene are FDA approved antidotes for opioid overdose. However, there are no approved antidotes for overdose with stimulants and for overdoses that result from drug combinations such as fentanyl and xylazine or fentanyl and alcohol. Some of the clinical manifestations and cause of death of drug overdose include cardiovascular, renal, and respiratory failures. NIDA is supporting research to develop safe and effective antidotes to antagonize the effects of drugs or revert the respiratory depression, which is the commonest cause of death.

Antidotes are being evaluated using pharmacological approaches such as mu-opioid receptor antagonists, pharmacokinetic antagonists, drug sequestrants, and respiratory stimulants. With regards to opioid antagonists, several formulations of naloxone are being investigated and new MOR antagonists are being developed. For example, methocinnamox appears to reverse and prevent the respiratory depression of MOR agonists and a single injection provides protection for up about two weeks [84].

Pharmacokinetic antagonism is being investigated using vaccines and monoclonal antibodies. Anti-drug antibodies form an antigen-antibody complex that cannot cross the blood brain barrier and prevents the drug from entering the brain. Currently, vaccines are being developed against opioids (fentanyl, heroin, and oxycodone), cocaine, and methamphetamine. Monoclonal antibodies are being studied for opioids and methamphetamines that may serve as effective antidotes, particularly for methamphetamine overdoses.

Another approach is the use of sequestrants, which are molecules that immediately inactivates drugs by binding and encapsulating the drug. It rapidly reverses its toxic effects and accelerates the clearance from plasma into urine. Currently, CS-1103 is a sequestrant that is being evaluated for fentanyl and methamphetamine intoxication. A phase 1 clinical trial of CS-1103 for methamphetamine intoxication is expected to start soon [85–87].

Respiratory stimulants increase or improve breathing and they work via central of peripheral mechanisms to increase the rate and depth of respiration [88]. Pharmacologic strategies proposed include AMPA receptor agonists, 5-hydroxytryptamine receptor agonists, phosphodiesterase inhibitors, dopamine D1 receptor agonists, leptin receptors, D-cysteine ethyl ester (D-CYSee), and almitrine.

5.7. Additional medication strategies

Due to space limitations, we cannot describe all the strategies currently being pursued to develop interventions for SUD and drug overdoses. Instead, we refer to the Supplemental Tables that summarize strategies currently being funded by NIDA to expand the armamentarium of safe and effective medications to treat OUD and StUD. These tables can be found in the website of the NIDA’s Division of Therapeutics and Medical Consequences at https://nida.nih.gov/about-nida/organization/divisions/division-therapeutics-medical-consequences-dtmc. These tables are “living documents” that are regularly updated. So, we encourage the readers to consult them.

6. Conclusions

SUD and drug overdose deaths are significant public health concerns. Although there are challenges in the development of medications for SUD and drug overdose, there are also extraordinary new opportunities. Major advances in pharmacology and neuroscience alongside with new tools for molecular screening, the discovery of new pharmacological targets, and the expansion of databases and AI models offer an unprecedented prospect to accelerate medication development in SUD. Furthermore, advances in precision medicine and statistical techniques are opening new ways to match treatments to the individualized characteristics of patients. In the process, care should be taken to ensure that SDOH are considered in AI algorithms and in reimbursement models to ensure that advances benefit all individuals suffering from SUD and to avoid current systems that lead to the large health disparities in SUD outcomes.

Significance Statement.

This publication is highly relevant because it provides a general overview of the obstacles and opportunities to develop medications to treat substance use disorders (SUDs), a discussion of the clinical endpoints to demonstrate treatment efficacy, and new tools available to study new medications. It also provides information about medications currently investigated with support from National Institute of Drug Abuse (NIDA) and access to “living documents” where this information is regularly updated.

Highlights.

• Development of medications to treat substance use disorders (SUD) a public health priority.

• Multiple scientific opportunities to accelerate the development of medications.

• Clinical trial endpoints required by regulatory approval of new medications are challenging

• The National Institute on Drug Abuse (NIDA) leads the discovery and development of medications for SUDs.

• Medications pipelines available in NIDA website https://nida.nih.gov/about-nida/organization/divisions/division-therapeutics-medical-consequences-dtmc