The DyNaMics of Excitation and Inhibition Govern Epileptic Encephalopathies and Their Comorbidities

Commentary

Independent Neuronal Origin of Seizures and Behavioral Comorbidities in an Animal Model of a Severe Childhood Genetic Epileptic Encephalopathy.

Asinof SK, Sukoff Rizzo SJ, Buckley AR, Beyer BJ, Letts VA, Frankel WN, Boumil RM. PLoS Genet 2015;11(6):e1005347..

The childhood epileptic encephalopathies (EE's) are seizure disorders that broadly impact development including cognitive, sensory and motor progress with severe consequences and comorbidities. Recently, mutations in DNM1 (dynamin 1) have been implicated in two EE syndromes, Lennox-Gastaut Syndrome and Infantile Spasms. Dnm1 encodes dynamin 1, a large multimeric GTPase necessary for activity-dependent membrane recycling in neurons, including synaptic vesicle endocytosis. Dnm1^Ftfl or “fitful” mice carry a spontaneous mutation in the mouse ortholog of DNM1 and recapitulate many of the disease features associated with human DNM1 patients, providing a relevant disease model of human EE's. In order to examine the cellular etiology of seizures and behavioral and neurological comorbidities, we engineered a conditional Dnm1^Ftfl mouse model of DNM1 EE. Observations of Dnm1^Ftfl/flox mice in combination with various neuronal subpopulation specific cre strains demonstrate unique seizure phenotypes and clear separation of major neurobehavioral comorbidities from severe seizures associated with the germline model. This demonstration of pleiotropy suggests that treating seizures per se may not prevent severe comorbidity observed in EE associated with dynamin-1 mutations, and is likely to have implications for other genetic forms of EE.

Recent advances in human genetics have identified new epilepsy-causing genes at a rapid pace in the past several years; however, in the case of the dynamin-1 (DNM1) mutation, the animal model led the way. The first link between Dnm1 and neurologic disease was reported in 2010 when a mutagenesis screen found that mice homozygous for a missense mutation in Dnm1, referred to as the “fitful mutation,” had lethal seizures, severe ataxia, and neurosensory deficits (1). Then, in 2014, a genetic study confirmed that de novo mutations in DNM1 cause epileptic encephalopathy in humans (2). The primary function of DNM1 is thought to be to enable the formation and recycling of synaptic vesicles (3); thus, DNM1 is a member of the growing list of genes that are thought to cause severe epilepsies via impairment of synaptic transmission (4).

The concept of an imbalance in excitation over inhibition has long been invoked to characterize the epileptic brain (5), and this paradigm has more recently gained traction to explain other neurologic diseases, such as autism (6). In the case of voltage-gated ion channel genes such as SCN1A and KCNQ2, an alteration of the excitatory to inhibitory balance is likely caused by changes in neuronal membrane excitability, but in the case of GABA and NMDA receptor subunit mutations, and now DNM1 mutations, the cause of the imbalance is likely alterations in synaptic signaling. GABA and NMDA receptor mutations are specifically linked to inhibitory or excitatory synaptic transmission, respectively, whereas DNM1 is ubiquitously present in synapses. Because Asinof et al. wanted to understand the cellular-level mechanisms through which the fitful mutation causes epilepsy and other cognitive deficits, they began by asking whether introducing the fitful mutation in glutamatergic or GABAergic neurons was sufficient to cause the observed phenotypes.

To achieve this cell-type specificity, the authors drove cre recombinase expression in either GABAergic or glutamatergic neurons to delete floxed Dnm1 from one allele, leaving these populations with only the fitful mutation on the other allele. They first used the Gad2 promoter to drive cre expression in most inhibitory neurons in the brain, presumably leaving almost all GABAergic neurons with only the mutant form of DNM1. This resulted in early-onset lethal seizures, with age at death ranging from postnatal day 15 to day 27, showing that DNM1 dysfunction in inhibitory neurons is sufficient to cause the severe epilepsy. Interestingly, however, these animals did not show the hunched posture and ataxic gait of animals with homozygous expression of the fitful mutation in all cells. Next, the authors targeted excitatory neurons by using Emx1-cre to delete the floxed Dnm1 allele, which should leave glutamater-gic neurons in the cortex and the hippocampus with only the mutant DNM1. In contrast to the Gad2-cre mice, the Emx1-cre mice did not have early-onset lethal seizures and had a normal life span and EEG, suggesting that the mechanisms through which the fitful mutation causes epilepsy is not dependent on glutamatergic neurons in the forebrain.

GABAergic neurons are a diverse group of cells that can be further subdivided based on localization, function, and gene expression. Therefore, the authors next used promoters active in subpopulations of inhibitory neurons to determine whether exclusive expression of the fitful allele in parvalbumin (Pvalb)-, somatostatin (Sst)-, or cortistatin (Cort)-expressing GABAergic neurons would recapitulate the phenotypes seen in the Gad2-cre mice. They found that fitful expression in Pvalb neurons, which are thought to comprise about 40% of the total GAB-Aergic neurons, phenocopied the more extensive Gad2-cre deletion. The Sst-cre and Cort-cre mice also had seizures and premature lethality, but the seizures had a later onset and the mice lived longer than the Gad2-cre or Pvalb-cre mice. Furthermore, EEG recordings showed differences depending on the cre driver, suggesting that the same genetic mutation can cause different electrical signatures, depending on the neural circuits that it impacts.

Although dysfunction of DNM1 in forebrain excitatory neurons was not sufficient to cause epilepsy, several other phenotypes of the fitful mutation were evident in the Emx1-cre model. These mice were visibly smaller, had postural abnormalities, and were hyperactive in several contexts. Although the results of several cognitive tests were inconclusive due to their extreme hyperactivity, the Emx1-cre mice spent significantly more time at the perimeter of their cages in the open field test and had higher immobility time in the tail suspension test, suggesting an anxiety and depressive-like phenotype.

In addition to intractable seizures, epileptic encephalopathies, such as those caused by DNM1 mutations, are often accompanied by severe cognitive and behavioral deficits in humans. The severity and onset of seizures correlate with the extent of the comorbidities, leading to the assumption that the seizures cause the neurologic deficits, and thus, controlling the seizures should improve the comorbidities. Although substantial support for this position exists, the results of Asinof et al., demonstrating behavioral deficits in the absence of seizures, contribute to a growing body of literature that suggests that this is not always the case. In the majority of epilepsies in which genetic mutations cause or contribute to seizures, it is likely that the genetic insult is also causing the accompanying neurologic deficits through its effects on discrete neural circuits. This implies that focusing on controlling the seizures may not improve comorbidities, and may even exacerbate them, depending on complex interactions between the molecular effects of the mutation, developmental changes as well as effects on distributed neurotransmitter systems and circuits throughout the brain.

How this new information can eventually be used to design therapies for patients with epileptic encephalopathy, particularly those with DNM1 mutations, is still unclear. A straightforward conclusion may be to assume that if loss of DNM1 at inhibitory synapses causes epilepsy because the function of these synapses is compromised, then pharmacologically boosting inhibitory transmission would lessen the seizures. However, many antiepileptic drugs attempt to achieve this by enhancing GABA receptor function, increasing GABA uptake, or preventing its degradation. These are molecular strategies that are doomed to fail if synapses run out of synaptic vesicles to fill with GABA in the first place, which is likely the case in DNM1 mutations. One of the hopes for genetic discovery in epilepsy is that we can pinpoint the molecular starting points and drivers of the disease; then, we can use experimental models to understand the pathway from these molecular changes to the development of epilepsy, which would then lead to the discovery of critical mechanisms that are amenable to intervention. Whether synaptic vesicle dynamics is one of these mechanisms remains to be seen.