Stress responsiveness and cognition: revealing the impact of a genetic predisposition to increased HPA axis reactivity on learning and memory
Cognitive deficits in depressed patients have recently received increasing attention. These symptoms are mainly in the realm of executive functioning deficits and include hippocampus-dependent and prefrontal cortex (PFC)-dependent tasks. The stress reactivity (SR) mouse model is of potential interest in modeling these deficits as glucocorticoid exposure has detrimental effects on both the hippocampus and PFC as these brain regions express abundant amounts of glucocorticoid receptors. Furthermore, the expression of brain-derived neurotrophic factor (BDNF), a key neurotrophic factor in cognitive functioning, is regulated by glucocorticoids, and depressed patients have repeatedly been reported to have decreased BDNF levels. Moreover, an increase in BDNF in response to antidepressant treatment is believed to be a key factor in the functional mechanism of the treatment. It was therefore our aim in the first instance to examine cognitive abilities in the SR mouse model and furthermore measure their hippocampal BDNF levels with the hypothesis that highly stress reactive mice would have cognitive deficits and reduced BDNF levels. To this end we subjected HR, IR and LR mice to a cognitive test battery including a test of recognition memory (novel object recognition) and a spatial memory task (Y-maze). In addition, BDNF levels were measured in the hippocampus of these mice. Indeed, the results revealed that, compared to the other two lines, male and female HR mice have recognition memory deficits (see Figure a) as well as spatial memory deficits (see Figure b) and exhibit a corresponding decrease in hippocampal BDNF levels (see Figure c).
With this established link between stress reactivity and hippocampus-dependent memory tasks we further probed the functional integrity of the hippocampus by measuring hippocampal N-acetylaspartate (NAA) levels using magnetic resonance spectroscopy. This revealed that the HR mice have a significant decrease in hippocampal NAA levels, a marker of neuronal integrity (see Figure d).
In order to also measure the capacity of HR, IR and LR mice in a PFC-dependent task, a novel reversal learning paradigm was developed. The classical reversal learning task, the water T-maze, would not have been suitable to be use in mice differing in stress reactivity as swimming would be acutely stressful, and an acute effect of stress on memory could have confounded our results. In this novel task that we developed, the mice were trained in a dry land T-maze with the reward being to find an escape tunnel back to their home cage at the end of one of the arms. This task demonstrated that, while HR mice where able to learn the task and recall the correct arm 24h later, they required significantly more trials to complete the reversal learning task compared to IR and LR mice.
Taken together, these results demonstrate that HR mice exhibit hippocampus-dependent memory deficits along with decreased hippocampal BDNF levels and decreased levels of NAA in the hippocampus. In addition, these animals show deficits in PFC-dependent reversal learning. HPA axis hyperactivity could thus be linked to cognitive deficits as well as specific neurobiological correlates of learning and memory. This corroborates that HR mice are a putative new model for cognitive deficits in major depression. Modeling these endophenotypes in mice provides a valuable tool in the search for new targets for the treatment as well as for testing new potential drugs to further characterize and, finally, cure major depression.
Current and future experiments include hippocampal volume measurements, NAA measurements in the PFC as well as attempts to reverse the cognitive deficits and emotional phenotype of HR mice by administration of antidepressants or intrahippocampal BDNF infusion. Furthermore, the molecular mechanisms involved in the described phenotype are under investigation.
Performance of high (HR), intermediate (IR) and low (LR) stress reactive mice in a spatial and a non-spatial hippocampus-dependent memory task and two neurobiological correlates of hippocampal neuronal integrity. HR mice demonstrate an inability to discriminate between a novel and a familiar object in the non-spatial novel object recognition task with a 2-hour inter-trial interval (a). HR mice are also unable to discriminate between novel and familiar arms of the Y-maze spatial learning task (b). In comparison to LR mice, HR mice had reduced amounts of brain-derived neurotrophic factor (BDNF) in the hippocampus (c) and reduced hippocampal N-acetylaspartate (NAA) levels (d).