Effects of neonatal stress on gamma oscillations in hippocampus.

Abstract

Repeated separation of rat pups from the dam during a critical period of hippocampal development alters the physiological profile of the stress response in a lifelong manner. This model has been used to study the etiology of mental illness, for which there is increased risk in adults who were exposed to chronic trauma, neglect or abuse in early life. The results of early life adversity seen in adulthood for both humans and rats include chronic and excessive exposure to corticosteroids. Studies in adult rats have shown that chronic stress — i.e., bodily restraint, exposure to cold, or social challenge by a dominant male, recurring over time — reduces the number of parvalbumin-immunoreactive interneurons in hippocampus. Approximately 60% of that population of interneurons is comprised of parvalbumin-expressing basket cells, whose rapid, synchronous firing induces the rhythmic inhibitory postsynaptic potentials observed as gamma oscillations. Gamma oscillations are fundamental to the processes of cognition and memory, and are aberrant in bipolar disorder and schizophrenia, as are the number of parvalbumin-expressing cells in cortex and hippocampus. We asked whether gamma oscillations in neonates would be affected by lengthy separations from the dam; if so, whether that effect would be mediated by compromised development of the parvalbumin-expressing basket cell network, which is currently thought to underlie the onset of schizophrenia and bipolar disorder later in life. Using in vitro slice electrophysiology, we recorded local field potentials of cholinergically-induced gamma oscillations in hippocampus. In control pups, separated from their dams for 15 minutes during postnatal days 2 through 14 (HMS15), gamma power increased with — and was very significantly correlated with — both age (p = 0.0022) and weight (p = 0.0024). Conversely, in pups separated from the dam for 180 minutes during the same period (HMS180), there was no correlation between gamma power and either age or weight. Analysis of covariance determined that the correlations between gamma power and both age (p = 0.0342) and weight (p = 0.0541) were, respectively, significantly and near-significantly different between the two groups. Differences between the two groups increased with age. Immunohistochemistry revealed, in both groups, strong and significant or near-significant correlations between weight and the number of parvalbumin-immunoreactive cells in each subfield of hippocampus, and also that gamma power was not correlated to the number of parvalbumin-immunoreactive cells. Our findings are consistent with critical-period disruption of electrophysiological development of the parvalbumin-expressing basket cell network as a result of chronic early life stress.