Abstracts

Rationale: Status epilepticus (SE) describes a state of persistent seizures that are unrelenting, and is considered a medical emergency.  These prolonged seizure states more commonly occur in children and if not managed effectively can lead to severe neurological sequela.  The preferred first-line treatment of SE includes benzodiazepines (BZP), a class of GABAA receptor (GABAAR) agonists. In a subset of patients however, BZP prove to be ineffective in terminating SE.  Previous data from in vitro models has demonstrated that during single seizures, instead of being inhibitory, GABA can in fact become excitatory (J Neurosci; 34(46);15208-15222).  This is due to a disruption in the transmembrane chloride (Cl-) gradient that follows periods of neuronal hyperexcitability. To date, it is unknown whether this shift in GABAergic function contributes to prolonged seizure states, nor how it may modulate the anticonvulsant function of BZPs.The aim of this study is investigate the role of this excitatory GABAergic signalling during an in vitro replica of SE, termed the late recurrent discharge (LRD) phase.  Furthermore, we aim to investigate how this transient shift in GABAergic function may affect the efficacy of the commonly used BZP, diazepam (DZP).  Methods: Organotypic hippocampal brain slices were prepared from a GAD2-Cre transgenic mouse-line and transfected with channelrhodopsin (ChR2).  Whole-cell patch-clamp and local field potential recordings were made from the CA1 region from both interneurons and pyramidal cells.  Results: During baseline network activity, DZP significantly increased GABAAR-mediated inhibitory signalling (n = 9; p = 0.02) with this effect being reversed by the competitive antagonist, flumazenil (n = 9; p = 0.004).Next we show that when using the 0 Mg2+ proconvulsant model, if DZP is washed in before the onset of seizure-like events (SLEs), DZP significantly reduced the duration of SLEs (n = 10, p = 0.01) while causing a decline in SLE frequency (n = 10; p = 0.0028).  However, when applied during the LRD phase, DZP appears to significantly increase burst duration (n = 9; p = 0.02) without changing the interburst interval (n = 9; p = 0.25).We then demonstrate that during LRD, interneuron and pyramidal cell activity become strongly correlated compared to baseline activity (r = 0.25 vs r = 0.82; n = 9; p = 0.004). In addition, ChR2-mediated activation of interneurons during LRD drives depolarising GABAAR currents (n = 7, p = 0.03) whilst there being a significant decrease in GABAAR conductance before and after LRD (n = 14, p = 0.0001). Lastly, we show that ChR2-mediated activation of interneurons during LRD can entrain the network (n = 7, p = 0.002).  Furthermore, this entrainment can be blocked by washing in the GABAAR blocker, picrotoxin (n = 6, p = 0.0017). Conclusions: Our preliminary results show that during baseline activity, DZP enhances GABAAR inhibitory signalling. However, DZP appears to have a differential and time-dependent anticonvulsant effect during prolonged seizure-like activity. Furthermore, it appears that interneurons drive the propagation of the LRD through a strongly depolarising GABAAR activation.  We postulate that this shift in GABAergic function and resultant change in DZP efficacy may be caused by activity-driven changes in the transmembrane Cl- gradient. Funding: Mandela Rhodes Foundation; South African Medical Research Council; Newton Research Grant, National Research Foundation, Human Brain Project