Impaired mitochondrial function abolishes gamma oscillations in the hippocampus through an effect on fast-spiking interneurons

Sir, Synchronous neuronal oscillations in the gamma frequency band (30–80 Hz) are implicated in a wide range of cognitive processes, including memory formation (Varela et al., 2001) and sensory processing (Singer, 1993; Gray, 1994). This rapid, temporally co-ordinated activity in spatially distributed areas of cortex is highly energy dependent, producing large changes in bloodoxygen level dependent signals in vivo and showing exquisite sensitivity to hypoxia in vitro (Huchzermeyer, 2008). Gamma oscillations are generated by reciprocal excitation and inhibition in networks of electrically coupled pyramidal cells and GABAergic interneurons (Fisahn et al., 1998; Cunningham et al., 2003; Hájos et al., 2004). Synchronous firing of networks of basket-cell interneurons at gamma frequency produces phasic inhibition of pyramidal cells, the periods when this inhibition is minimal, providing temporal windows for pyramidal cell firing. Pyramidal cell firing in turn provides reciprocal tonic excitation onto interneurons, allowing the cycle to continue (Traub et al., 1996). The research article by Kann et al. (2011) demonstrates the critical dependence of gamma oscillations upon mitochondrial function. However, little is known about the relative energetic requirements of the different cellular components of this network. We have addressed this issue with intracellular recordings of pyramidal cells and interneurons undergoing gamma oscillations under conditions of metabolic stress in vitro. Rat hippocampal slices (450 mm horizontal sections) were maintained in a recording chamber at the interface between 95% oxygen/5% CO2 and artificial CSF (in mM: 10 glucose, 126 NaCl, 3 KCl, 24 NaHCO3, 1.25 NaH2PO4, 2 MgSO4 and 2 CaCl2). Stable gamma oscillation was induced by adding low dose (100 nM) kainic acid to the perfusate. Local field potential recordings were made from the stratum pyramidale layer in CA3 region of the hippocampus using glass microelectrodes (10–300 Hz band-pass filtered, digitized at 10 kHz). We also performed simultaneous intracellular recordings from CA3 pyramidal cells using sharp microelectrodes (70–130 M ) containing 2 M potassium acetate. Cells were recorded in current clamp mode, and identified electrophysiologically on the basis of their low resting firing rate, accommodating response to sustained depolarization and gamma frequency phasic inhibitory post-synaptic potentials. Excitatory post-synaptic potentials (EPSPs) and inhibitory post-synaptic potentials (IPSPs) were recorded at 70 and 20 mV, respectively. We then measured the effect on gamma oscillation power and corresponding intracellular activity using a variety of mitochondrial respiratory chain inhibitors. Potassium cyanide (KCN; 100 mM), an inhibitor of cytochrome oxidase in complex IV of the mitochondrial respiratory chain, caused a 72.1% reduction in gamma power (1221.3 207.8 mV control versus 340.8 70.5 mV KCN, n = 9, P50.05; Fig. 1A(i)) within 15 min. There was no significant change in the resting membrane potential ( 57.9 1.29 mV versus 60.6 2.05 mV KCN, P4 0.05; Fig. 1A(ii)) of pyramidal cells. Despite this, there was a marked reduction in pyramidal cell firing rate during the concurrently recorded gamma frequency oscillation (3.02 0.70 Hz versus 1.43 0.64 Hz KCN, n = 9, P50.05). We therefore examined the synaptic inputs to pyramidal cells. There was no significant change in excitatory input (mean EPSP amplitude 3.18 0.70 mV control versus 2.43 0.72 mV KCN, P40.05; mean EPSP frequency 12.4 2.6 Hz control versus 11.9 1.9 Hz KCN, P40.05). The mean IPSP amplitude was reduced, (6.80 1.66 mV control versus 5.21 0.74 mV KCN, P50.05); however, the most doi:10.1093/brain/awr018 Brain 2011: 134; 1–3 | e180