Hippocampal slice preparation in rats acutely suppresses immunoreactivity of microtubule-associated protein (Map2) and glycogen levels without affecting numbers of glia or levels of the glutamate transporter VGlut1

Misprogramming of glucose metabolism impairs recovery of hippocampal slices from neuronal GLT-1 knockout mice and contributes to excitotoxic injury through mitochondrial superoxide production

We have previously reported a failure of recovery of synaptic function in the CA1 region of acute hippocampal slices from mice with a conditional neuronal knockout (KO) of GLT-1 (EAAT2, Slc1A2) driven by synapsin-Cre (synGLT-1 KO). The failure of recovery of synaptic function is due to excitotoxic injury. We hypothesized that changes in mitochondrial metabolism contribute to the heightened vulnerability to excitotoxicity in the synGLT-1 KO mice. We found impaired flux of carbon from 13C-glucose into the tricarboxylic acid cycle in synGLT-1 KO cortical and hippocampal slices compared with wild-type (WT) slices. In addition, we found downregulation of the neuronal glucose transporter GLUT3 in both genotypes. Flux of carbon from [1,2-13C]acetate, thought to be astrocyte-specific, was increased in the synGLT-KO hippocampal slices but not cortical slices. Glycogen stores, predominantly localized to astrocytes, are rapidly depleted in slices after cutting, and are replenished during ex vivo incubation. In the synGLT-1 KO, replenishment of glycogen stores during ex vivo incubation was compromised. These results suggest both neuronal and astrocytic metabolic perturbations in the synGLT-1 KO slices. Supplementing incubation medium during recovery with 20 mM D-glucose normalized glycogen replenishment but had no effect on recovery of synaptic function. In contrast, 20 mM non-metabolizable L-glucose substantially improved recovery of synaptic function, suggesting that D-glucose metabolism contributes to the excitotoxic injury in the synGLT-1 KO slices. L-lactate substitution for D-glucose did not promote recovery of synaptic function, implicating mitochondrial metabolism. Consistent with this hypothesis, phosphorylation of pyruvate dehydrogenase, which decreases enzyme activity, was increased in WT slices during the recovery period, but not in synGLT-1 KO slices. Since metabolism of glucose by the mitochondrial electron transport chain is associated with superoxide production, we tested the effect of drugs that scavenge and prevent superoxide production. The superoxide dismutase/catalase mimic EUK-134 conferred complete protection and full recovery of synaptic function. A site-specific inhibitor of complex III superoxide production, S3QEL-2, was also protective, but inhibitors of NADPH oxidase were not. In summary, we find that the failure of recovery of synaptic function in hippocampal slices from the synGLT-1 KO mouse, previously shown to be due to excitotoxic injury, is caused by production of superoxide by mitochondrial metabolism.

Dissection method affects electrophysiological properties of hippocampal slices

The rodent hippocampal slice preparation has long been a critical tool for studying the electrophysiological effects of pharmacological and genetic manipulations. Slices can be prepared with several different slicing methods including the tissue chopper, vibratome, and rotary slicer. To examine how slicing methods affect slice integrity, we generated hippocampal slices by these three methods and compared their histology and electrophysiological responses. Although all three methods generate histological alterations, the time course is slowest in slices generated with a rotary slicer. Furthermore, although paired-pulse facilitation in dendritic field EPSPs was observed in slices generated by all three methods, paired-pulse potentiation in population spikes, which is common in chopper- and vibratome-generated slices was seldom observed in rotary-generated slices, suggesting less disinhibiton. For preservation of hippocampal slice integrity, the rotary slicer may offer advantages over the other two devices.

Hippocampal slice preparation in rats acutely suppresses immunoreactivity of microtubule-associated protein (Map2) and glycogen levels without affecting numbers of glia or levels of the glutamate transporter VGlut1

INTRODUCTION

With its preservation of cytoarchitecture and synaptic circuitry, the hippocampal slice preparation has been a critical tool for studying the electrophysiological effects of pharmacological and genetic manipulations. To analyze the maximum number of slices or readouts per dissection, long incubation times postslice preparation are commonly used. We were interested in how slice integrity is affected by incubation postslice preparation.

METHODS

Hippocampal slices were prepared by three different methods: a chopper, a vibratome, and a rotary slicer. To test slice integrity, we compared glycogen levels and immunohistochemistry of selected proteins in rat hippocampal slices immediately after dissection and following 2 and 4 hr of incubation.

RESULTS

We found that immunoreactivity of the dendritic marker microtubule-associated protein 2 (Map2) drastically decreased during this incubation period, whereas immunoreactivity of the glutamate transporter VGlut1 did not significantly change with incubation time. Astrocytic and microglial cell numbers also did not significantly change with incubation time whereas glycogen levels markedly increased during incubation.

CONCLUSION

Immunoreactivity of the dendritic marker Map2 quickly decreased after dissection with all the slicing methods. This work highlights a need for caution when using long incubation periods following slice preparation.