Practical Aspects of Animal Models of Psychiatric Disorders

Mental illness and dementia are associated with tremendous personal and societal cost. Twenty percent of the population in Canada experiences a mental illness or addiction every year, and 50% of the population experiences mental illness by age 40. ¹ Mental health care and reduced productivity are estimated to cost the Canadian economy $51 billion every year. ² Similarly, more than 500,000 Canadians are currently living with dementia, with an estimated $10.4 billion per year in health care costs. ³ However, despite the prevalence and burden of these brain disorders, effective treatments remain elusive for many individuals. Psychiatric disorders are particularly difficult entities to diagnose and treat because of the heterogeneity of their causes and symptoms. Many of the drugs used to treat these disorders have long delays before efficacy, remain ineffective for many patients, and present with adverse side effects that decrease patient compliance. Thus, there is an urgent need for new and more effective pharmacotherapies.

An improved understanding of the etiology of mental illness including dementia facilitates the development of new drug targets. Animal models of psychiatric disorders and of dementia offer an opportunity to investigate the neurobiological bases of brain pathophysiology with rigor and control not possible in clinical settings. Such experimental control is essential when investigating the relative contributions to varying underlying etiologies to the symptoms of these disorders. Moreover, animal models allow for more expedient monitoring of disease progression and treatment responses; the ability to use more invasive tools to define molecular, structural, and functional changes in the brain associated with different etiologies or therapies; and highly controlled experiments to test potential therapeutic agents. However, a major challenge with animal models of psychiatric disorders is the inability to model and assess symptoms that are uniquely human (e.g., paranoid delusions or auditory hallucinations in schizophrenia). Nonetheless, a greater understanding of the environmental and genetic risk factors that lead to brain disorders has improved our ability to model their etiology, and sophisticated behavioral assays with translational relevance to clinical manifestations have been developed. These tools provide hope for improved preclinical research and clinical translation of novel therapies for dementia and mental illness.

To be meaningful, animal models must have face validity, construct validity, and/or predictive validity. A model with face validity possesses similarity in symptoms, such that the observed characteristics of the model have clinical correlates in patient populations. Construct validity means that the model should replicate the neurobiological and pathological bases of the disorder in humans, preferably using a known etiological factor. Finally, an animal model with predictive validity will show the expected pharmacological response to efficacious drugs currently used to treat the disorder. In this issue of the journal, animal models of schizophrenia and Alzheimer’s disease (AD) are reviewed. In both cases, animal models have been developed to replicate the known etiology of these complex disorders. These models are used to provide insight into the neuropathophysiology and behavioural characteristics of the disorders and to identify new targets for drug development to improve treatment.

In modeling psychosis and schizophrenia, different approaches are used to assess risk factors and pathophysiology in the disorder. The onset of schizophrenia typically occurs in adolescence and early adulthood, but there is extensive research linking the origin of schizophrenia to prenatal events. Developmental models use environmental or pharmacological manipulations during gestation or the early postnatal period to recapitulate known risk factors for schizophrenia and then examine offspring for pathological and behavioural outcomes. Similarly, genetic models focus on mutations known to increase the risk of developing schizophrenia in humans. By manipulating candidate genes in murine models, the consequences on brain structure and function as well as the behavioural manifestations are ascertained and compared against clinical data. While many developmental and genetic models have excellent face, construct, and predictive validity, they are limited by the variability in environmental risk factors and the variable penetrance of mutations in the clinical population. For more controlled models of acute psychosis, pharmacological models that replicate known neurophysiological dysfunction in schizophrenia can be used in mice and rats. These models thereby provide mechanistic insight into the pathophysiology and behavioural abnormalities in schizophrenia.

In the first review article here, Winship and colleagues provide an overview of the most prevalent developmental, pharmacologic, and genetic models of schizophrenia, highlighting common cognitive and behavioural abnormalities and shared pathology as well as identifying controversies in the field. Given the heterogeneity of the disorder, multiple models should be used to identify core pathophysiology associated with particular behavioural phenotypes and to validate potential therapies. In the second review article, Nyarko and colleagues discuss the potential relationship between major depressive disorder and AD, which has been the subject of considerable debate. Recent analyses suggest that major depressive disorder is a risk factor for subsequent AD. ⁴ In their review, Nyarko and colleagues focus on the potential role of the amyloid protein precursor (APP) in depression and compare and contrast the clinical findings with preclinical data, particularly data generated using genetic mouse models of AD-related amyloidosis. Cleavage of APP by secretases produces β-amyloid peptides (Aβ), which ultimately accumulate in the brains of patients with AD in the form of amyloid plaques. However, whether mice with mutations of APP known to be involved in AD show alterations in a variety of behavioural tests and monoamine levels related to major depressive disorder has not been systematically reviewed. There is contention in the relevant literature, and Nyarko and colleagues highlight that a part of the ambiguity might reflect a lack of consideration of sex-dependent mechanisms and phenotypes in the clinical data and in any supporting mouse models being studied. The authors report the depressive-like phenotypes in these mouse models, including the 3xTg-AD, J20, and TgCRND8 mouse models of AD and in the Ts65DN mouse model of Down syndrome.