Access
This article is part of Nature's premium content.
Published online 31 October 2007 | Nature | doi:10.1038/450006a
Model behaviour
The brain is no longer the black box it used to be, and neuroscientists are starting to put new knowledge to good use, developing better animal models for psychiatric disorders. Alison Abbott reports.
To read this story in full you will need to login or make a payment (see right).
Comments
Reader comments are usually moderated after posting. If you find something offensive or inappropriate, you can speed this process by clicking 'Report this comment' (or, if that doesn't work for you, email webadmin@nature.com). For more controversial topics, we reserve the right to moderate before comments are published.
This is a decent review of animal models in psychiatry, but it probably does not mention critical issues that have hampered psychiatric drug development and indeed hampered psychiatric thinking about fundamental core syndromes. The biggest issue frankly is that psychiatry has trying to develop an understanding of psychiatric syndromes without any real consideration as to the fundamental nature of brain emotional systems, particularly in terms of prototype emotional systems that we share with all mammals. Although the addition of lots of complex cognition clearly alters those emotional systems, it does not change their fundamental quality. In other words humor, a highly desirable quality, is a cognized extension of the mammalian prototype of rough and tumble play, only now we tickle each other with words instead of with somatosensory roughhousing. Behavioral neuroscience has been hamstrung in its efforts to avoid anthropomorphism by the complementary danger of speciesism-we think we are more different from mammals than I believe we really are. There are extensive homologies with respect to the prototype emotional systems, and extensive homologies with respect to virtually all the basic subcortical systems, but the evolution of the cortex frankly has not changed those subcortical emotional systems fundamentally, other than providing additional top-down modulation/inhibition/activation. I would strongly disagree with the comments of Laurence Tecott. We have not had fundamentally new drugs in psychiatry not because of the intrinsic inadequacy of all animal models but because of the cookie-cutter nature of design strategies by big Pharma. Nowhere is this more in evidence than in relationship to depression, where animal models implicate large contributions from brain opioids systems (both kappa and mu opioids) in depressive phenomena but where market-driven cookie-cutter drug design philosophy has dictated one serotonergic and noradrenergic agent after another. We have plenty of evidence also that stress cascades and that the modulation of cytokines play a large role in depression as well again from robust animal models. CRF antagonists still currently in the development pipeline are one important offshoot of this research. many other novel non-aminergic drugs are in the pipeline also from work in animal models. Although we probably never have good animal models for religious experience, meditation, language disorders, or suicidality, animal models have still been an absolutely critical component, in some ways perhaps the most critical component, in developing a better sense of underlying etiologies for many DSM-IV disorders. I think that effort could be further assisted by as much vigilance about the dangers of species-ism as we traditionally have paid to the dangers of anthropomorphism. Behavioral neuroscience has also been hamstrung I believe by continuing behaviorist assumptions that animals have no version of conscious experience, a position that already has an enormous amount of data against it. Although animals certainly do not think about their experiences, or generate other highly cognitive forms of consciousness, it seems extremely unlikely that mammals do not possess some basic form of sentience, or that they do not feel a basic version of rage, fear, separation distress, etc.. Douglas F. Watt, Ph.D. Boston University School of Medicine Harvard Medical School
Dr. Watt's recognition that animals are able to feel "a basic version of rage, fear, separation, distress, etc." offers some relief to those of us who cringe over photos of mice struggling to keep from drowning. But the article on which we are commenting shows how far away most scientists are from that same recognition. Dr. Hyman's observation about the usefulness of fMRI is reason to be hopeful that there is at least some elegant research being done to help humans who suffer from mental illnesses. In an August 2007 letter to me about the use of animals for research into schizophrenia and bipolar disorder, E. Fuller Torrey, M.D., of The Stanley Medical Research Institute wrote: "For these disorders, animals, even primates, are of limited value because I believe that the part of the brain affected evolved relatively recently and thus there is no comparable brain area in primates." It is disheartening to know that there are so many researchers willing to use millions of tax dollars and decades of their careers fiddling about with knocking out this gene and that to get a wholly unsatisfactory model of a particular human mental illness. From a layperson's perspective, it seems as if the research community is besotted and blinded by a method that has not lived up to its expectations but is now a runaway train. When the FDA tells us that 92% of all drugs tested safe and effective in animals fail when they get to human clinical trials we shouldn't just go knock out another gene to see if we can make it better. Mice, monkeys, and everyone in between may feel basic versions of emotions we humans feel, but they still have their own operating systems that we don't understand any better than we understand our own. Far better to use our big brains and advanced technology to see what we can find out about ourselves. Mary Beth Sweetland, Director of Research & Investigations, In Defense of Animals, San Rafael, Calif.
I would like to agree with Dr. Watt that the standard animal models in common use have demonstrated utility in facilitating studies of neurobehavioral phenomena and in early stage drug development for neuropsychiatric disorders. It is understandable that this piece would lead him to conclude that I view all animal models to be intrinsically inadequate. Unfortunately, the comment attributed to me, that “woefully inadequate� animal models have been the major factor stifling advances in psychiatric therapeutics, is not an accurate portrayal of my views. I had indicated to the author that although limitations exist in the implementation of classical tests, these tests nevertheless have utility. I had better believe this, since we use such assays all the time in my laboratory. I had also indicated my view that hindrances to psychiatric drug development are not solely attributable to limitations in behavioral assays. Dr. Watt provides interesting food for thought about what some of these other factors may be. Finally, we appreciate that the author makes mention of our home cage behavioral monitoring initiative. We would like to make clear that we view this approach as a powerful addition to, and not a global replacement for, standard behavioral assessment methods in drug discovery. The combination of improved behavioral assay methods, with a willingness to explore nontraditional drug targets (as described in this piece and advocated by Dr. Watt), represents a promising formula for success. -- Laurence Tecott M.D., Ph.D., University of California, San Francisco Department of Psychiatry
I would like to thank Dr. Tecott for his clarifications on this, and apologize if remarks taken out of context were criticized unfairly. Certainly the use of animal models raises lots of concerns, including concern about the degree of suffering imposed on the animals themselves. Although both animal rights advocates, and more cortically and cognitively oriented researchers would perhaps both want to emphasize the great differences between humans and mice, the differences are primarily confined to the higher forebrain, presumably through genes that regulate cortical neuroplasticity and development. While I would agree that each species has a unique overall "operating system", all mammals share a handful of affective systems (what Jaak Panksepp has called basic emotional/behavioral operating systems) that determine a sense of internal value(s), pleasure and pain. We will neither make progress in unraveling core emotional issues in psychiatry, nor treat animals well, if we are seduced by the old siren's song of human grandiosity - that we are just so different and so unique that whatever we share with other mammals is dwarfed by our cognitive complexity and superiority. This is starkly revealed in still prevalent views in the minds of some neuroscientists that animals don't feel pain. The genome project has instead underlined that our original view (that we had 100,000 genes or more - way more than any other mammal) was just way off. We are special and uniquely gifted, but not that different from other mammals in terms of primary emotional systems. Personally, I think this is a larger and in the end more important similarity than our obvious differences. Douglas F. Watt, Ph.D. Harvard Medical School/BUSM
All these comments illustrate the need to add new variables in setting animal models of e.g. depression and get a sharper view on the results published so far. First, a great amount of data indicates that genotype x environmental interactions dictate how an individual, animal or human, will deal with potentially depressogenic/anxiogenic events. Environment-related issues are numerous and complex of course, especially when translated in laboratory rodents. However,it seems to me that key developmental stages such as the prenatal and early postnatal life conditions are not systematically addressed in preclinical research. As far as genotype is concerned, there has been a boom since the last two decades but we are still facing the likely possibility that it is the intrication of multiple genes, none of each bearing a clearcut influence per se, that favours depressive outcomes. The discovery of these genes, and thus the recognition of the systems they belong to or they control is hard due to intra- and inter-assay variabilities. Of course, animal models have shown that key genes may play a role but in most cases these genes were seeked on the basis of our present knowledge of the systems involved in stress reactions and/or antidepressant therapy. The finding that the genetic variability in the size of the serotonin transporter promoter has great influence on neuroticism indices illustrates that statement. Second, do we use the most relevant paradigms to assess stress reactions in animal models? Clearly, numerous models are not ethologically relevant and the outcomes of the results brought about through these models should be carefully analyzed with regard to that issue. Third, one other key issue is the baseline level of the behaviour we address in animal studies. Investigations on how environmental conditions affect adult neurogenesis have shown that environmental enrichment or exercise favour neurogenesis. This result is primordial but it should be reminded that this holds true for laboratory animals that do not have naturally access to these two positive stimulators, as opposed to the wild rodent. This could illustrate the need to reset the conditions under which laboratory animals are kept so that discrete but enduring consequences of negative stimuli could be observed. Lastly, we should now convince the whole community that the tail suspension test or the forced swimming test are by no means animal models of depression but tests which respond to antidepressants. This difference, which may appear subtle for the non specialist, is huge in terms of consequences for the construct of adequate animal models of depression.
Abbott correctly highlighted the value of recent mouse models, particularly of DISC1, for understanding the mechanisms underlying schizophrenia. What the piece failed to note or comment on however was our earlier success (Calpcote et al 2007 ‘Behavioural phenotypes of Disc1 missense mutations in mice Neuron’, 54, 387-402) using a powerful alternative strategy to the transgenic route taken by Sawa, Pletnikov and Silva. We believe that our original observations following the ENU mutagenesis approach deserves to be highlighted. Our study took advantage of a very large library of ENU induced mutations in the mouse assembled by the RIKEN Institute, Japan (http://www.gsc.riken.jp/PQG/RGDMSavailability.htm [over 10,000 mutagenised G1 male mice archived]). Other such ENU mouse libraries are available for screening via the MRC Mammalian Genetics Unit Harwell (www.mgu.har.mrc.ac.uk/facilities/dnaarchive [over 4,000 DNA samples from individual F1 animals]) or Ingenium Pharmaceuticals (http://www.ingenium-ag.com/content_tech_technology.html [16,000 mutagenised mice]). By screening exon 2 of Disc1 in a fraction of the RIKEN ENU library, two independent missense mutations, Q31L and L100P, were identified, mice rescued and studied. We estimate that the RIKEN library alone most probably contains 10-20 further missense Disc1 mutants. Strikingly, both mutants showed reduced volumes by magnetic resonance imaging in relevant sub-regions of the brain (L100P by 13%, Q31L by 6%). The behavioural effects of the two Disc1 mutations were assessed using a battery of relevant standardised tests (including pre-pulse inhibition, latent inhibition, open field activity, forced swim test, conditioned choice, social interaction and sucrose consumption). Mice with an L100P mutation exhibited schizophrenia-like behaviour, with profound deficits in prepulse inhibition and latent inhibition that were reversed by antipsychotic treatment. The L100P mutants also showed hyperlocomotion and impaired attention and learning. In contrast, Q31L mutants showed depressive-like behaviour in social interaction measures and a profound deficit in the forced swim test (a measure of ‘despair’) that was reversed by the antidepressant buproprion, but not by the prototypical antidepressant rolipram, a PDE4 inhibitor. This last observation is of considerable significance given our earlier demonstration that DISC1 interacts with PDE4B dynamically to regulate cAMP (Millar et al 2005 ‘DISC1 and PDE4B are interacting genetic factors in schizophrenia that regulate cellular cAMP signalling.’ Science, 310:1187-91). Both Q31L and L100P mutant proteins showed reduced binding to PDE4. Moreover, PDE4B activity was significantly reduced in the brains of Q31L mutants, suggesting this might be a contributory factor in depression. Collaborating again with Houslay (Galsgow), we have now shown that the Q31L and L100P missense mutations map to binding sites for PDE4 on full length DISC1 (Murdoch et al 2007 ‘Isoform-selective susceptibility of DISC1/phosphodiesterase-4 complexes to dissociation by elevated intracellular cAMP levels.’ J Neurosci. 27:9513-24.). Thus these ENU mutants are not only providing valuable behavioural and developmental models for schizophrenia and mood disorder, but can address directly mechanistic issues that do not follow naturally from conventional transgenic or ‘knock out’ strategies. That missense mutations in Disc1 can give rise to either schizophrenia-like or depression-like phenotypes in mice, further validates the role of DISC1 in major mental illness and points a way forward in human studies aimed at explaining how different DISC1 variants can lead alternatively to a diagnosis of schizophrenia, schizoaffective disorder, bipolar disorder, major depression or Asperger’s syndrome. In a complementary way, we intend to identify and study additional ENU mouse mutations in Disc1 and for DISC1 interactors and recommend this strategy to others as a powerful complement to more traditional approaches to mouse modelling. David Porteous, Medical Genetics Section, University of Edinburgh Centre for Molecular Medicine, Western General Hospital, Edinburgh EH4 2XU, UK John Roder, Mount Sinai Hospital Research Institute, Toronto, Ontario, M5G 1X5, Canada Yoichi Gondo, RIKEN Genomic Sciences Center, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan Steven Clapcote, Medical Genetics Section, University of Edinburgh Centre for Molecular Medicine, Western General Hospital, Edinburgh EH4 2XU, UK