Cannabidiol exerts antiepileptic effects by restoring hippocampal interneuron functions in a temporal lobe epilepsy model

Abstract

Background and purpose: A non-psychoactive phytocannabinoid, cannabidiol (CBD), shows promising results as an effective potential antiepileptic drug in some forms of refractory epilepsy. To elucidate the mechanisms by which CBD exerts its anti-seizure effects, we investigated its effects at synaptic connections and on the intrinsic membrane properties of hippocampal CA1 pyramidal cells and two major inhibitory interneurons: fast spiking, parvalbumin (PV)-expressing and adapting, cholecystokinin (CCK)-expressing interneurons. We also investigated whether in vivo treatment with CBD altered the fate of CCK and PV interneurons using immunohistochemistry.

Experimental approach: Electrophysiological intracellular whole-cell recordings combined with neuroanatomy were performed in acute brain slices of rat temporal lobe epilepsy in in vivo (induced by kainic acid) and in vitro (induced by Mg²⁺ -free solution) epileptic seizure models. For immunohistochemistry experiments, CBD was administered in vivo (100 mg·kg^-1 ) at zero time and 90 min post status epilepticus, induced with kainic acid.

Key results: Bath application of CBD (10 μM) dampened excitability at unitary synapses between pyramidal cells but enhanced inhibitory synaptic potentials elicited by fast spiking and adapting interneurons at postsynaptic pyramidal cells. Furthermore, CBD restored impaired membrane excitability of PV, CCK and pyramidal cells in a cell type-specific manner. These neuroprotective effects of CBD were corroborated by immunohistochemistry experiments that revealed a significant reduction in atrophy and death of PV- and CCK-expressing interneurons after CBD treatment.

Conclusions and implications: Our data suggest that CBD restores excitability and morphological impairments in epileptic models to pre-epilepsy control levels through multiple mechanisms to reinstate normal network function.

Figure 1

CBD dampens excitation. (A) Spontaneous EPSPs, recorded at a membrane potential of −60 mV in healthy control tissue and KA model of epileptic tissue. Concentrations of 5, 8 and 10 μM CBD were bath‐applied. sEPSP frequency and peak amplitudes were only significantly changed with 10 μM of CBD. (B) Unitary recording between two pyramidal cells synaptically connected. Triple presynaptic action potentials elicited by the pyramidal cell resulted in EPSPs that displayed synaptic depression recorded at three different membrane potentials (−70, −60 and −55 mV). These EPSPs were reduced in amplitude and duration following bath application of 10 μΜ CBD (red traces). (C, D) Plots of change in peak average amplitude of EPSPs per synaptic connection (n = 5) with bath application of CBD recorded at −55 and −70 mV. CBD had a more pronounced action at more positive membrane potentials. (E) Plot of PPR (second‐amplitude EPSP/first‐amplitude EPSP) at −55 and −70 mV; no significant change in PPR was measured in CBD‐treated cells.

Figure 2

CBD enhances inhibition. (A, panels i, ii) sIPSPs, recorded at a membrane potential of −55 mV. The frequency and peak amplitudes of sIPSPs increased with CBD (10 μM) and remained unchanged with subsequent addition of a selective CB₁ receptor antagonist AM4113 (1 μM). (B, panels i, ii) Unitary connection between a presynaptic FS PV containing interneuron (FS‐PV) and a postsynaptic excitatory cell is shown. Averaged IPSPs are shown as black traces (in control) and after bath application 10 μM CBD (20–30 min, red traces). (B, panels i, ii.) IPSPs elicited by CCK, SCA interneuron onto postsynaptic pyramidal cells. The top traces in (A) and (B) left panel show the averaged IPSPs elicited by one presynaptic action potential, whereas the bottom panel shows the inhibition elicited by a train of presynaptic interneuron action potentials; as a result, there is summation of the inhibitory event. (C, panels i,– ii) Plots show the average change in IPSP amplitudes, IPSP RTs and width at half amplitude elicited by FS‐PV and CCK, SCA presynaptic interneurons onto pyramidal cells with bath application of CBD. The line plots show change in peak average IPSP amplitude with CBD per individual synaptic connection.

Figure 3

Pyramidal cell's hyperactive membrane properties are restored by CBD. The intrinsic membrane properties of pyramidal cells recorded in healthy rats in (A, panel i) control and (A, panel ii) after 10 μΜ CBD bath application. Similarly, pyramidal cells recorded brain slices in (B, panel i), (B, panel ii) Mg²⁺‐free and from (C, panel i), (C, panel ii) KA epileptic models. The input–output curves displayed a pseudo‐linear relationship between the number of spikes generated by pyramidal cells of (A, panel iii) healthy rodents (n = 13), (B.iii) Mg²⁺‐free (n = 5) and (C, panel iii) KA (n = 12) epileptic models with increasing current injections. The firing of pyramidal cells was decreased after application of CBD. (D) Bar plots representing intrinsic membrane properties of pyramidal cells in Mg²⁺‐free (n = 5) and KA epileptic models (n = 12) in comparison to healthy rodents (n = 13). (D, panel i) CBD produced no apparent changes of action potential (AP) threshold of pyramidal cells. (D, panel ii) The input resistance and (D, panel iii) time constant were decreased by CBD. Results are expressed as mean ± SEM. *P ≤ 0.05.

Figure 4

CBD reduces intrinsic excitability of CCK adapting interneurons in CA1. The membrane properties of SCA interneurons, which were CCK positive, were recorded in healthy rats (A, panel i) in control and (A, panel ii) after 10 μΜ CBD bath application. Similar recordings are shown in Mg²⁺‐free condition (B, panel i), (B, panel ii) and in the KA epileptic model (C, panel i), (C, panel ii). The input–output curves displayed a pseudo‐linear relationship between the number of spikes generated by adapting cells of (A, panel iii) healthy rodents (n = 11), (B, panel iii) Mg²⁺‐free (n = 5) and (C, panel iii) KA (n = 12) epileptic models with increasing current injections. The number of action potentials of adapting interneurons was decreased after application of CBD. (D) Bar plots representing intrinsic membrane properties of adapting cells in Mg²⁺‐free (n = 5) and KA epileptic models (n = 12) in comparison to healthy rodents (n = 11). (D, panel i) CBD produced no apparent changes of action potential (AP) threshold of adapting cells. (D, panel ii) The input resistance and (D, panel iii) time constant were decreased by CBD in the disease models. Results are expressed as mean ± SEM. *P ≤ 0.05.

Figure 5

CBD enhances excitability of FS‐PV interneurons in the CA1 region of the hippocampus. The membrane properties of FS‐PV interneurons recorded in healthy rats (A, panel i) in control and (A, panel ii) after a 10 μΜ CBD bath application. Similarly, FS‐Pv cells were also recorded in Mg²⁺‐free (B, panel i), (B, panel ii) and in KA epileptic models (C, panel i), (C, panel ii). The input–output curves displayed a pseudo‐linear relationship between the number of spikes generated by FS PV cells of (A, panel iii) healthy rodents (n = 9), (B, panel iii) Mg²⁺‐free (n = 4) and (C, panel iii) KA (n = 7) epileptic models with increasing current injections. The firing of FS‐PV interneurons was dramatically increased after application of CBD. (D) Bar plots representing intrinsic membrane properties of FS PV cells in Mg²⁺‐free (n = 4) and KA epileptic models (n = 7) in comparison to healthy rodents (n = 9). (D, panel i) CBD produced a reduction of action potential (AP) threshold of FS PV cells. (D, panel ii) The input resistance and (D, panel iii) time constant were increased by CBD in healthy rodents and the seizure models. Results are expressed as mean ± SEM. *P ≤ 0.05.

Figure 6

CBD halts PV cell death in the hippocampi of KA‐induced epileptic rats. (A) Reconstructions and photomicrographs of all the PV neurons present in the CA1 hippocampal subfield of a randomly picked rat in (A, panel i) control, (A, panel ii) epileptic, (A, panel iii) CBD‐treated at zero time (CBD_time0) and (A, panel iv) at 90 min (CBD_time90) post‐SE conditions using a 40× magnification in a light microscope with an attached drawing tube, from three consecutive 100‐μm‐thick coronal sections of rat hippocampi. The CA1 hippocampal subfield including, SO, SP, and stratum radiatum (SR) and stratum lacunosum moleculare (SLM). All photomicrographs taken of PV neurons were from brains sections processed using immunoperoxidase labelling. Scale bars set at 250 μm for all the reconstructions and 50 μm for all the photomicrographs. (B, panel i) Sholl plot exhibiting the pattern of dendritic arborisation of PV neurons in the CA1 hippocampal subfield of rats in all experimental conditions, control, epileptic and CBD‐treated at zero time (CBD_time0) and at time 90 min (CBD_time90) post‐SE (n = 20, for all). Morphometric analyses of PV neurons (n = 20) were performed, which are represented as bar graphs comparing the (C, panel i) length of primary dendrites (μm), (C, panel ii) secondary dendrites (μm) and (C, panel iii) area of soma (μm²) between all the experimental groups. Results are expressed as mean ± SEM (*P ≤ 0.05; one‐way ANOVA with post hoc Tukey's test).

Figure 7

CBD prevents CCK cell death in the hippocampi of KA‐induced epileptic rats. (A) Reconstructions and photomicrographs of all the CCK neurons present in the CA1 hippocampal subfield of a randomly picked rat in (A, panel i) control, (A, panel ii) epileptic, (A, panel iii) CBD‐treated at zero time (CBD_time0) and (A.iv) at 90 min (CBD_time90) post‐SE conditions using a 40× magnification in a light microscope with an attached drawing tube, from three consecutive 100‐μm‐thick coronal sections of rat hippocampi. All photomicrographs taken of CCK neurons were from brains sections processed using immunoperoxidase labelling. Scale bars set at 250 μm for all the reconstructions and 50 μm for all the photomicrographs. (B, panel i) Sholl plot exhibiting the pattern of dendritic arborization of CCK neurons in the CA1 hippocampal subfield of rats in all experimental conditions, control, epileptic and CBD‐treated at zero time (CBD_time0) and at time 90 min (CBD_time90) post‐SE (n = 20, for all). Morphometric analysis of CCK neurons (n = 20) were performed which are represented as bar graphs comparing the (C, panel i) length of primary dendrites (μm), (C, panel ii) secondary dendrites (μm) and (C, panel iii) area of soma (μm²) between all the experimental groups. Results are expressed as mean ± SEM (*P ≤ 0.05; one‐way ANOVA with post hoc Tukey's test).