All levels of the translational spectrum must be targeted to advance psychopharmacology and improve patient outcomes

Correll et al1 correctly state that many psychiatric disorders remain insufficiently treated despite advances in psychopharmacology, and attribute this to the limited knowledge of pathophysiology of these disorders, the lack of biological markers precluding tailored treatment selection, the few mechanistic targets for treatment development, and the challenges with clinical trial design and conduct. Here I address the chasms at the various levels of the translational spectrum that should be targeted through innovations in order to advance psychopharmacology and improve outcomes for patients. Drug discovery in psychiatry has been mostly driven by the pharmaceutical industry. The discovery of selective serotonin reuptake inhibitors and second-generation antipsychotics ushered in a “new era” of psychopharmacology in 1980s and 1990s. However, these drugs and their modifications, while claiming to provide better safety and tolerability, primarily targeted monoaminergic systems, similar to tricyclic antidepressants and first-generation antipsychotics. Any attempts to develop new drugs with novel targets, such as metabotropic glutamate receptors, CRF1 receptors, and tackykinin NK1 receptors, were met with failures. As the pipeline for drug development in psychiatry was drying out, many major pharmaceutical companies announced ceasing further investments in this area, citing “very low probability and disproportionately high cost for attaining success”2. Indeed, it takes nearly nine years to bring a psychotropic drug to the market, and the likelihood of drug approval in psychiatry – which includes success in all phases of development leading to regulatory approval – is only 6.2%, which is the lowest amongst non-oncology diseases3. Thus, novel strategies to enhance success of drug discovery in psychiatry are urgently needed. Pre-clinical assays – such as forced swim test and chronic mild stress, as well as stimulant induced locomotor activity and reduced prepulse inhibition – have been used to screen drugs for prediction of antidepressant and antipsychotic activity, along with positron emission tomography (PET) studies in humans to estimate receptor occupancy in order to determine appropriate dosing for therapeutic efficacy. These strategies have worked well in general for drugs that targeted the monoaminergic systems. However, drugs with actions on novel targets (such as NK1 receptors, CRF1 receptors and glutamatergic system), while demonstrating activity in some pre-clinical assays, did not succeed in phase 3 clinical trials. The general consensus is that newer pre-clinical tests that have better construct and predictive validity are urgently needed. Attempts to improve construct validity by developing mouse models with knockout of genes implicated in schizophrenia have not proven to be helpful in consistently detecting drugs with antipsychotic activity4 . Whether CRISPR-based gene editing to create knockout animal models might be more useful remains to be seen. Similarly, human induced pluripotent stem cells and brain organoids are being used to screen drugs for their effects in disease relevant cells, but their full potential is yet to be documented. Phenotypic screening has been more suc­cessful than target-based approach­es for drug development in central nerv­­ous system disorders. To this end, PsychoGen­ics has developed a phenotypic drug discovery platform called SmartCube, which uses a target-agnostic approach to screen com­pounds. This automated testing platform, through its customized hardware, pre­sents a sequence of challenges to a mouse, collects massive amounts of data points, and uses proprietary machine learning algorithms to detect the potential for efficacy of compounds. SEP-363856 (ulotaront) was developed using this platform; it has trace amine-associated receptor 1 (TAAR-1) and serotonin 5-HT1A receptor agonistic properties, and has shown efficacy in a phase 2 clinical trial for schizophrenia5 . The results of the phase 3 trials for this drug, and the efficacy of other compounds identified using this platform for other indications, will indicate whether it represents a significant advance over the previous models. The success rate in phase 2 trials for drugs tested for psychiatric disorders is only 24%, which is the lowest among 14 disease areas3. Further, many psychotropic drugs that succeed in phase 2 fail in phase 3 trials. Correll et al1 outline various reasons for such outcomes and suggest use of adaptive trial designs and strategies for minimizing placebo response to reduce the risk of failure. Given that a high placebo response is a major contributor to failed trials, setting a priori a threshold for excluding all patients from centers with an improbable placebo response might be worth considering. In addition, academia must work in close collaboration with the industry to develop innovations in trial designs, and conduct in-depth analyses to take lessons from failed trials which will inform further drug development. For instance, the first trial of cariprazine for bipolar depression6 failed due to a high placebo response rate of 60%. Knowledge from this and other trials was used to design subsequent phase 2/3 studies, all of which were positive, leading to cariprazine's approval by the US Food and Drug Administration (FDA)7. Despite a signal for efficacy in post-hoc analyses, a similar strategy was not pursued for agomelatine, which also had a 60% placebo response rate in a bipolar depression trial8. This illustrates the impact of business decisions by the industry on drug development in psychiatry. While development of new drugs with novel mechanisms of action would be a wel­come addition to the therapeutic arma­mentarium, there are limitations to the gen­eralizability of data from randomized pla­cebo-controlled trials. Real-world data com­ing from a variety of sources must be gathered in order to understand the effectiveness of treatments and tailor them to the needs of each individual. Most currently approved treatments for various psychiatric indications work for about 50% of patients, but there is little information to guide clinicians with regards to what treatment is most likely to work for which patient, and, if the first treatment is ineffective, what is the next most appropriate intervention. Thus, there is an urgent need to incorporate approved treatments into real-world clinical practice protocols/algorithms, similar to cancer treatment protocols, to generate evidence and move the field towards precision psychiatry. Such efforts could be further bolstered by using learning health care systems in clinical practice settings and collecting data that could be analyzed for discovery of biomarkers that predict response to each treatment. Moving along the translational spectrum, patients need to access care, and evi­dence-based treatments need to be used ap­propriately by clinicians. Although several evidence-based treatment options exist for some psychiatric disorders, such as major depressive disorder, unfortunately only 8% to 33% of patients with this disorder use mental health services, and only 3% to 23% receive minimally adequate treatment9. Fur­ther, even in developed countries such as the UK, adherence to evidence-based care pathways for treatment of depression is poor, with many patients not receiving guideline-concordant care. In order to address this translational chasm, governments must invest funds to bolster mental health services and support education aimed at addressing stigma. Moreover, health care organizations must make every effort to establish an infrastructure that promotes and supports evidence-based practices to optimize outcomes. In conclusion, innovations need to occur at all levels of the translational spectrum to advance psychopharmacology and improve patient outcomes.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.