Rationale:
The dentate gyrus is critically involved in temporal lobe epilepsy. Within the dentate gyrus, granule cells (GCs) and mossy cells (MCs) establish a recurrent excitatory circuit. Given the extensive MC projections contralaterally and along the longitudinal axis of the hippocampus and the large proportion of MCs active during exploratory behaviors, MCs have a great potential to destabilize this circuit. In addition, MCs promotes epileptic activity during early stages of epilepsy. MC repetitive activity induces BDNF/TrkB-dependent long-term potentiation at the MC-GC synapse (MC-GC LTP), which increases GC output. This plasticity promotes behavioral seizures
in vivo and, left unchecked, may interfere with pattern separation, a DG-dependent form of learning that relies on sparse GC firing. Strikingly, MC axon terminals express high levels of type-1 cannabinoid receptor (CB
1R). We recently reported that CB
1R activity dampens MC-GC LTP, via multiple synaptic mechanisms, which may be a critical mechanism to suppress seizure-induced runaway excitation. We hypothesized that CB
1Rs, acting via endogenous or exogenous cannabinoids, can protect against runaway excitation and seizure generation by controlling GC output.
Methods:
Here, we used multiple complementary approaches, such as a conditional knockout strategy,
in vitro electrophysiology,
in vivo optogenetics,
in vivo calcium imaging.
Results:
We found that selective deletion of CB
1Rs from MCs (MC CB
1R cKO) increased the severity and susceptibility to acute seizures in adult mice. Using miniature scopes to assess neuronal activity from large cell populations in the dentate gyrus, we analyzed population-level activity in MC CB
1R cKO mice. We found that adult mice injected with intraperitoneal kainic acid (KA) injection, a well-established model that leads to experimental epilepsy, displayed robust calcium waves which were exacerbated in MC CB
1R cKO mice, suggesting a strong, synchronized neural activity. Moreover, selective deletion of CB
1Rs from MCs impaired DG-dependent learning processes including spatial memory and contextual memory. Finally, we found that endocannabinoid release from GCs is sufficient to depotentiate previously strengthened synapses in a model of experimental seizures in mice. To confirm this observation
in vivo, we have developed and tested a photoactivable CB
1R (opto- CB
1R) to mimic CB
1R in MCs.
Conclusions:
Our findings reveal that CB
1Rs are critical to control seizures and may be used as a novel therapeutic strategy to alleviate or prevent the development of seizures.
Funding:
National Institute of Health (NIH) R01-NS113600, R01-MH125772, R01-NS115543, R01-MH116673 to PEC. American Epilepsy Society Postdoctoral Award and Einstein Junior Investigator Neuroscience Research Award (JINRA) to C.B.