Hippocampal cytotoxic lesion effects on species-typical behaviours in mice

Abstract

The behavioural effects of hippocampal lesions have been extensively documented in rats. However, paradigms developed for rats cannot be assumed to transfer straightforwardly to mice; the behaviour of the two species differs in many respects. Mice are currently the species of choice for targeted genetic manipulations. A number of these programs aim to modulate hippocampal function. The present studies were therefore designed to provide a behavioural profile of selective, cytotoxic hippocampal lesions in tasks appropriate for mice. The lesions abolished food hoarding from a source outside the home base, and reduced the tendency to displace food pellets from a tube inside the home cage (burrowing). Lesioned mice showed reductions of directed exploration (rearing and head dipping), but not locomotor activity, in a holeboard and open field, and explored the edges of their home cages less when the lids were removed. Nest construction was also impaired. These effects were not due to gross motor impairments, as formal tests revealed no deficiencies in co-ordination or strength. There were suggestions of changes in emotionality, although a more consistent finding was that lesioned mice were often slower to initiate behaviour in novel surroundings, which may be congruent with the other deficits we observed. These results may aid interpretation of the many genetic manipulations that target the hippocampus, and of neurodegenerative conditions that induce hippocampal pathology.

Introduction

The role of the hippocampus in behaviour has been extensively studied in the last half century. This was prompted by the discovery of the dramatic impairments in episodic and declarative memory produced by hippocampal lesions in man [33]. While selective hippocampal lesions in animals appear not to produce such global amnesia, in that some forms of learning are spared, spatial memory is generally impaired. Most of the evidence for this comes from rat studies [30], as rats are the animal species most often used in experimental psychology. At the molecular level of neuroscience, however, the mouse genome is much better characterised than that of the rat. Thus many current genetic studies on learning, memory and the hippocampus, are performed in the mouse [7], [8], [9], [28], but their design and interpretation is influenced by rat lesion work, much of which was performed before the development of the more selective cytotoxic stereotaxic lesion techniques [22].

The present study was therefore designed to characterise the behavioural profile of cytotoxic hippocampal lesions in mice, using a battery of tests. The importance of carefully assessing the behavioural effects of selective hippocampal lesions in different species and strains has also been noted in other reports [2]. It was expected, on the basis of the rat lesion studies, that these tests would demonstrate impairments in spatial cognition, while basic sensorimotor abilities would be unaffected. The results were as expected: lesioned mice were found to be impaired on spatial working memory (spontaneous and rewarded alternation) and spatial reference memory (Morris water maze, a shallow water maze and an elevated Y-maze). Learning of a Lashley maze, which appeared not to be dependent on spatial cues, was unaffected [10].

The present report describes a number of other tests performed on these mice. Firstly, it was considered important to ensure that the behavioural profile of the lesion included assessments of motivational, motor and emotionality effects, which could themselves have affected performance on our other tasks. Secondly, we have reported that prion disease initiated by focal injection of scrapie agent into the hippocampus of the mouse produces a reduced tendency to displace (burrow) material from a tube in the home cage, a reduction in polydipsia induced by a palatable solution (glucose), reduced spontaneous alternation in a T-maze, increased open field activity and impaired nest building ability [4], [12], [20]. We therefore wished to assess directly the possibility that these effects resulted from the hippocampal component of the disease. Thirdly, a comprehensive behavioural profile of the effects of hippocampal lesions would also facilitate interpretation of genetic manipulations that affect hippocampal function. This is both because they identify a range of behaviours that should be affected, and, perhaps equally importantly, because they identify a range of behaviours that should not, if the effects of the manipulation are truly selective. Interpretations built on single behavioural assays may not adequately characterise the effects of novel manipulations.

During the course of these tests it was noted that the hippocampal lesioned mice seemed to explore the edges of their home cages less than controls when the cage lids were removed. As this was an exploratory activity, possibly linked to the reduced rearing seen in the open field, it was formally assessed. For comparison, the mice were also tested in a holeboard, as this is a more standard and established test of exploration.

Although several of the tests, for example burrowing and hoarding, investigated spontaneous behaviour in semi-natural circumstances, our aim was not to perform detailed ethological investigations. Rather, since our own and other work has suggested the utility of an ethologically oriented approach [17], and it is known that hippocampal lesions in rats may affect such behaviours (reviewed in [30]), we incorporated ethologically based tests into an otherwise standard behavioural neuroscience program. Thus we tested the mice for hoarding behaviour, simply weighing the amount of food transported to the home base, as this technique has been used in rats to assess dysfunction of the hippocampus, and the prefrontal cortex, to which it is connected [5], [6], [26], [36], [37], [38].

The term ‘species-typical’ is used in the present context to describe behaviours spontaneously emitted by, but not necessarily specific to, mice. They are typical behaviours seen in semi-naturalistic situations. Species-typical tests may be distinguished from more formal tests commonly found in the behavioural neuroscience literature, such as the open field and plus-maze.

C57BL/6J mice were used in these studies, as they are probably the strain most widely used in behavioural work [13], [28] and frequently provide the blastocysts into which embryonic stem cells are introduced in the creation of transgenic mice. Moreover, the resulting offspring are frequently back-crossed into other C57BL/6J mice to create a congenic strain [9].

Cytotoxic hippocampal lesions require a considerable investment of experimental time, as does production of genetically modified mice, so multiple behavioural tests on the same animals maximise their potential. Positive convergent results from a test battery may be as compelling as single results from naı̈ve animals, although inevitably there are increased risks of Type 1 statistical errors. Given the desirability of running lesioned mice through a test battery comparable to those used in genetic studies, if similar results are seen in two or more tests designed to measure the same factor(s), stronger conclusions on the effects of the treatment may be possible than from a single result. Negative results, however, should be treated with more caution, particularly when considering tests of emotional behaviour, which, almost by definition, will be affected by the prior history of the animal. Lesioned animals and their sham-operated controls need to be weighed and assessed several times post-surgery to monitor their recovery. The resulting habituation to handling means that for certain tests, notably those involving anxiety, the reaction of operated animals will depend, far more than naı̈ve mice, on specific aspects of apparatus and procedure. Naı̈ve animals will be influenced to a greater extent by more general aspects of the experimental procedure, especially the relative novelty of handling, resulting in potentiation of intrinsic properties of the apparatus and procedure. For example, it has been suggested that open field behaviour may represent attempts to reinstate contact with conspecifics, and avoidance of the experimenter as a potential predator [35]. Assessing the affective profile of lesioned mice therefore presents different problems to, for example, drug or strain difference testing in relatively naı̈ve mice. Accordingly, some modifications to standard tests were made to extend their potential range and effectiveness. For example, extensions were added to the open arms of the plus-maze, with the aim of differentiating between animals with very low baseline levels of anxiety. Similarly, only mice with very low levels of anxiety would be expected to spend much time on the most exposed of the successive alleys. These tests have proven sensitive to AMPA receptor deletion which impairs hippocampal LTP ([39], unpublished results).

Therefore, in the present studies, hippocampal lesions were performed on a cohort of mice, which were then assessed on tests of species-typical, affective, motor and cognitive behaviour. Behaviour is dependent on the integrity of motor systems, so tests of co-ordination and strength were performed to clarify the specificity of any lesion effects [9]. Motor performance on these tasks is severely impaired in the later stages of prion disease (scrapie) in mice [4], [12], [20].