Animal models of movement disorders: species versus investigator–dependent limitations

Sirs

Bezard et al.¹ provide a comprehensive review of the pathophysiology of levodopa-induced dyskinesia, one of the most debilitating complications of chronic levodopa pharmacotherapy in Parkinson´s disease. Although the system-level changes addressed in this paper are derived from primate studies, most of the reviewed molecular findings have been obtained in the rat. Yet, the paper raises doubts about the possibility that dyskinesia can be adequately modelled in rats. It is suggested that only primates might be physically capable of showing the various types of levodopa-induced dyskinesia that are seen in patients. This point highlights a problem of paramount importance to all studies that attempt to model neurological symptoms in rodents. It is certainly warranted to ask how accurately rodent behaviours reflect complex human disorders. But the answer to this riddle cannot simply rely on the physical identity between the neurological symptom in humans and its corresponding behavioural manifestation in rodents. Because of obvious differences in osteoarticular, muscular and locomotor organization between rodents and humans, there might never be complete physical identity between motor manifestations in these two species. Unattended expectations of physical identity will cause uncertainty and arbitrariness in the choice of behavioural tests, and will not favour efforts towards a clear definition of neurological deficits in rodents. This problem is well exemplified by the widespread use of rotational tests in the analysis of rats lesioned unilaterally with 6-hydroxydopamine. Although very convenient, objective and useful, rotational tests are an unspecific measure of behavioural outcome, which has been used variably in the literature to model either parkinsonian disability, dyskinesia or anti-akinetic effects of antiparkinsonian drugs²,³. Bezard and colleagues are right in pointing out that there are different types of dyskinesia, and that when examining animal models — rat or primate — investigators have to be clear on what aspect of the human condition they are modelling. The authors are also correct in pointing out that a full distinction between choreiform hyperkinesias and dystonia has not been provided in rats so far⁴. However, this and other limitations might be more related to the behavioural tests and lesion models that are used, rather than to the species per se. It is not at all clear that, in response to intermittent dopamine receptor agonists, dopamine-depleted rats will fail to develop detectable patterns of sustained, abnormal muscle contractions (dystonia) that can be distinguished from hyperkinetic, dance-like movements (chorea). On the other hand, if the rat model ultimately shows only hyperkinetic movements, but not dystonia-like muscular reactions, investigators might well take advantage of the molecular and cellular insights gained from exploring rat and primate differences. In the meantime, the rat model of Parkinson´s disease has proved to parallel the human condition more closely than expected. For example, if the severity of the nigrostriatal dopamine cell loss is great enough, dopamine-depleted rats show akinesia, and in response to intermittent levodopa, they also develop chorea-like limb and trunk movements that seem strikingly reminiscent of dyskinesias found in parkinsonian patients (akinesia and dyskinesias are displayed unilaterally if the degeneration is unilateral³,⁴). Rather than assuming that rodents cannot be as useful as primates in the study of parkinsonian dyskinesias, it might be more prudent to examine in more detail the pathophysiology of this movement disorder in both rat and primate models. Rat models would not only be convenient, but also amenable to the use of less expensive methods to investigate the neural mechanisms of dyskinesia development —in vivo fast-cycle voltammetry, microelectrophysiology, and molecular and microanatomical approaches.

Like any other field of neurobiology, research on movement disorders is critically dependent on the availability of manageable and well-standardized animal models to address prime molecular and cellular questions. In addition to practical and logistical reasons, the high degree of genetic homology between rodents and humans justifies the use of rats and mice for the modelling of human disorders⁵. Valid animal models must be able to mimic key functional features of human conditions, and must be sensitive to drugs or treatments that alleviate or exacerbate the corresponding symptom in humans. Eventually, the most crucial factor to ensure predictive validity of animal models might not necessarily be the choice of species so much as it is the selection of behavioural measures.