Abstract / Summary
The cerebellum, long studied as the principal center for motor coordination, is now recognized as a participant in epileptogenic network dynamics through the pathway running from the cerebellar cortex to the deep cerebellar nuclei and on to thalamic and cortical targets. Purkinje cells are the sole inhibitory output of the cerebellar cortex, and Bergmann glia, a radial astrocyte population apposed to Purkinje cell dendrites and parallel-fiber synapses, sustain the perisynaptic environment on which Purkinje cell firing depends through glutamate clearance by GLAST/EAAT1, potassium buffering by Kir4.1, and calcium-dependent signaling. In cortex and hippocampus, loss of the equivalent functions in protoplasmic astrocytes lowers seizure threshold and produces spontaneous seizures, but no comparable experiment has been performed in Bergmann glia. Molecular genetics has identified numerous epilepsy-associated genes, of which only SLC1A3, KCNJ10 and AQP4 have cell-type-resolved expression in Bergmann glia; the remainder act on cerebellar circuits through neurons. In human epilepsy, cerebellar atrophy with Purkinje cell loss and Bergmann gliosis tracks cumulative seizure burden and antiseizure medication exposure rather than a primary cerebellar lesion and is therefore best read as reactive. Reactive astrogliosis is nonetheless accompanied by reduced glutamate transport and potassium buffering in sclerotic hippocampus, and whether cerebellar reactive gliosis carries the same homeostatic cost has not been measured. This review sets out the circuit-level, genetic and neuropathological evidence bearing on that question and specifies the experiments that would answer it.