The Institute of Medicine’s series of three workshops was aimed at understanding why treatments for central nervous system (CNS) disorders that were developed based on animal models often fail in human clinical trials. Animal models can help elucidate the basic biology of CNS disorders. However, they are unable to fully represent these complex conditions, often lead to failed human trials, and should be replaced or supplemented with new technologies.

The current workshop focused on the urgent need to change how CNS drugs proceed to human testing. Currently, few drugs in the development process make it to market rates as low as 6.2% – in part because psychiatric disorders cannot be adequately modeled in animals. How do you measure diagnostic criteria such as apathy, delusions, or re-experiencing a traumatic event? The challenge of this workshop was to understand what level of evidence is necessary to test new treatments in humans when not first using animal models and what new models might replace traditional animal models. Perspectives were shared by representatives from industry, investors, regulatory agencies, ethics committees, academic researchers, and other public and private agencies. And some of the technologies and models proposed included computational modelling, using genetic data to find drug targets, or building cellular models of disease based on cells grown directly from a patient with the disease.

Industry representatives noted that non-animal models could help facilitate bringing new compounds for testing in human clinical trials by validating and confirming viable drug targets. Such models could help determine a drug’s therapeutic index, reducing potential false negatives during development. Building these models with clinical input could ensure that they are translatable to humans or that the same non-invasive assessments used in humans could be developed for use in these preclinical studies.

Representatives from FDA and EMA further confirmed that animal models are not necessary to proceed to testing a drug’s effect in humans. For instance, an IND can be opened for testing in humans given sufficient information on safety and dosing. This point was clarified by Dr. Robert Temple, Deputy Center Director for Clinical Science Center at CDER who said it is up to the Sponsor to determine their own risk level to invest in a clinical development program without predictive preclinical efficacy data.

Beyond the Sponsor’s level of risk and regulatory agency requirements, it will be critical to determine from an ethical perspective the level of evidence needed to test new compounds in humans. Because most first-in-human studies do not evaluate efficacy, one suggestion was that studies should first enroll subjects capable of consent, with potential risks and benefits clearly described in the informed consent form.

Despite excitement generated by novel techniques, the group agreed animal models may still play a critical role in developing CNS therapeutics in terms of understanding disease biology and drug target selection. However, human studies are necessary to complement these findings. For instance, human data on the natural history of CNS disorders are needed to advance our understanding of disease biology and progression.

Finally, as modeling technologies rapidly develop and require increasingly in-depth technical knowledge and expertise, academic institutions must be enabled to provide the necessary education, tools, and incentives to build future scientific professionals. One overarching question: How does all of this get done? Public-private partnerships may be one important method, including our own Cohens Veterans Bioscience, Innovative Medicines Initiative, Foundation for NIH Biomarkers Consortium, and the FasterCures Consortia-pedia with 49 active consortia.