The present study generated several novel findings. Firstly, we did not observe an increase in anxiety in our murine model of CRF; on the contrary, CRF mice were less anxious in the dark/light test. Secondly, recognition is impaired after 10 weeks of CRF. Thirdly, CRF enhances the severity of ischemic stroke in mice, as assessed by the infarct size after 34 weeks of CRF. Fourthly, the neurological test results in non-CRF mice tended to improve from day 1 to day 3 after ischemic stroke, whereas the results in CRF mice tended to worsen.
3.2. Anxiety and Exploratory Behavior
One of our study’s main findings is that anxiety was no higher in CRF mice than in non-CRF mice; on the contrary, CRF mice had lower anxiety levels when assessed in the dark/light box. This finding contradicts with several clinical studies in which anxiety levels are high in both predialysis [17
] and hemodialysis CRF patients [17
]. This apparent discrepancy may be linked to the fact that the results in patients with CRF are related to factors such as low self-esteem, feelings of uselessness and disturbance of the body image. These psychosocial factors may not be relevant in mice. In future work, it would be interesting to examine depression in our murine model of CRF. In fact, depression is linked to the disturbance of neurotransmitter systems. Furthermore, an association between levels of uremic toxins and depression in hemodialysis patients has already been reported [19
In the present study, CRF mice did not show abnormally low levels of locomotor activity or exploratory behavior. This finding agrees with a previous study (in the same murine model) in which locomotor activity in the dark was normal [20
]; this was considered to indicate that uremic toxins do not impair the central circadian pacemaker [20
3.4. The Severity of Ischemic Stroke
One of our study’s most important findings is that infarct cerebral volume was higher in CRF mice than in non-CRF mice. Our results after 34 weeks of CRF are in agreement with clinical reports of a stroke-induced increase in morbimortality in CRF patients [22
]. Factors that may contribute to poor post-stroke outcomes in CRF patients might include uremic toxins [23
], oxidative stress [24
], inflammation [25
] and endothelial dysfunction [26
]. In a previous study in our murine model, we found that endothelial dysfunction in the cerebral arterioles was associated with an increase in the eNOS inhibitor ADMA [14
]. Further investigation of the mechanisms responsible for the increased brain infarct volume in CRF mice is thus required.
Another interesting finding is that neurological test results tended to improve in non-CRF mice and to worsen in CRF mice. This is also in agreement with clinical data from CRF patients [6
]. Impaired performance in neurological tests in CRF mice may be due to a lack of recovery of the ischemic penumbra—the part of the brain in which neurons suffer during ischemia but may recover upon reperfusion. Factors possibly associated with post-stroke worsening of the ischemic penumbra again include oxidative stress and inflammation [27
]. However, one must interpret these data with caution because we observed post-stroke mortality in the CRF mice; this may have artificially improved the neurological test results in the CRF mice by eliminating the most severely affected individuals. Furthermore, factors other than post-stroke worsening of the ischemic penumbra may have contributed to the decrease in neurological test performance. Firstly, the increase in infarct volume in CRF mice may have altered the animals’ motor and sensory abilities and could have resulted in undernutrition and dehydration. Secondly, one of the characteristics of our ischemic stroke model is the ligature of the external carotid. Ligature of the external carotid makes it more difficult for the animal to feed [29
]; we did not measure food and fluid intakes in the present experiments and thus where unable to test this hypothesis.
3.5. Study Limitations
Our study had several limitations. Firstly, we did not perform recognition tests before the induction of CRF (i.e., before the electrocoagulation at 8 weeks of age). However, the food restriction implemented before Y maze experiments would probably have increased the perioperative mortality rate.
Secondly, the tests used to examine anxiety and recognition in the present study are used by many other laboratories worldwide. However, other tests of anxiety and recognition might yield different results.
Thirdly, we did not perform neurological tests before stroke induction. In most studies on ischemic stroke, neurological tests are performed both before and after stroke induction so that the mice with the worst pre-stroke scores are not included in the experiments.
Fourthly, the animals assigned to the CRF group differed significantly from the animals assigned to the non-CRF group in terms of the time spent respectively in the closed arm and in the center of the elevated maze prior to induction of CRF; this interferes with interpretation of the results. Great care was taken to ensure that all mice were exposed to the same experimental conditions. However, a number of extrinsic factors (such as small differences in noise levels) and/or intrinsic factors (such as differences between individual mice) may have influenced the results. For practical reasons, we did not have time to analyze the pre-CRF results before the induction of CRF. This analysis would have enabled us to randomize the mice and ensure that the various groups had similar pre-CRF results.
Fifthly, the mortality rate in the CRF mice was high after induction of ischemic stroke. As mentioned above, this may have biased the results by decreasing the number of mice evaluated in the neurological tests.
Sixthly, we did not measure parameters such as blood pressure, blood glucose levels and body temperature in the present experiments. We have reported previously that our murine model of CRF is not associated with elevated blood pressure [14
]. During ischemic stroke induction, mice were kept on a heating pad and then in a warm environment until they had fully recovered from the anesthesia. However, we cannot rule out the possibility that parameters other than uremic toxins were involved in the greater infarct size in CRF mice. Lastly, we did not ascertain whether uremic toxins were specifically involved in the observed alterations. As mentioned above, levels of several uremic toxins are known to be elevated in our murine model of CRF [16
]. Further experiments are necessary to fully explore the role of uremic toxins and the underlying disease mechanisms.