Can a Mouse Help Us Unravel the Mysteries of CDKL5-Related Epilepsy?

Abstract

CDKL5 Deficiency in Forebrain Glutamatergic Neurons Results in Recurrent Spontaneous Seizures

Wang HT, Zhu ZA, Li YY, Lou SS, Yang G, Feng X, Xu W, Huang ZL, Cheng X, Xiong ZQ. Epilepsia. 2021;62(2):517-528. PMID: 33400301. doi:10.1111/epi.16805

Objective:

Mutations of the cyclin-dependent kinase-like 5 (CDKL5) gene cause severe neurodevelopmental disorders characterized by intractable epilepsy, intellectual disability, and autism. Multiple mouse models generated for mechanistic studies have exhibited phenotypes similar to some human pathological features, but none of the models has developed one of the major symptoms affecting CDKL5 deficiency disorder (CDD) patients: intractable recurrent seizures. As disrupted neuronal excitation/inhibition balance is closely associated with the activity of glutamatergic and γ-aminobutyric acidergic (GABAergic) neurons, our aim was to study the effect of the loss of CDKL5 in different types of neurons on epilepsy.

Methods:

Using the Cre-LoxP system, we generated conditional knockout (cKO) mouse lines allowing CDKL5 deficiency in glutamatergic or GABAergic neurons. We employed noninvasive video recording and in vivo electrophysiological approaches to study seizure activity in these Cdkl5 cKO mice. Furthermore, we conducted Timm staining to confirm a morphological alteration, mossy fiber sprouting, which occurs with limbic epilepsy in both human and mouse brains. Finally, we performed whole-cell patch clamp in dentate granule cells to investigate cell-intrinsic properties and synaptic excitatory activity.

Results:

We demonstrate that Emx1- or CamK2α-derived Cdkl5 cKO mice manifest high-frequency spontaneous seizure activities recapitulating the epilepsy of CDD patients, which ultimately led to sudden death in mice. However, Cdkl5 deficiency in GABAergic neurons does not generate such seizures. The seizures were accompanied by typical epileptic features including higher amplitude spikes for epileptiform discharges and abnormal hippocampal mossy fiber sprouting. We also found an increase in spontaneous and miniature excitatory postsynaptic current frequencies but no change in amplitudes in the dentate granule cells of Emx1-cKO mice, indicating enhanced excitatory synaptic activity.

Commentary

The X chromosome housed cyclin-dependent kinase-like 5 (CDKL5) gene encodes a serine-threonine kinase. CDKL5 transcripts are especially abundant in the forebrain and the protein is important for axonal outgrowth, dendritic spine development, synapse formation, and maintenance of synaptic function. ¹ Pathogenic variations in the CDKL5 gene underlie the CDKL5 deficiency disorder (CDD), a rare and typically severe neurodevelopmental disorder with an estimated incidence of 2.35 per 100 000 livebirths with female to male ratio estimated at 4:1. ^1,2 The clinical hallmark of CDD is a medically intractable epilepsy that typically occurs shortly after birth and manifests in 90% of children by 12 months of age. Other common comorbid clinical features include hypotonia, global developmental delay, autism, and cerebral visual impairment. ¹ Several mouse models have been developed and they reproduced neurocognitive pathologies, such as autistic-like behavior, poor learning, and impaired motor control. However, modeling of human CDD-related seizures has been more challenging. ³ While hyperexcitability and an increased sensitivity to chemoconvulsants ⁴ has been reported, the first overt stimulus-induced seizure-like phenotype akin to sensory-reflex seizures in humans ⁵ was observed only relatively recently in aged heterozygous female mice either carrying the human CDKL5 R56X mutation or an exon 6 deletion. Median age at event onset was 28 and 32 weeks, respectively. ³ Interestingly, the behavioral reflex myoclonic and tonic–clonic seizure-like events were specific to heterozygous females but absent in animals expected to be far more susceptible to the expression of model epilepsy, that is, hemizygous males or homozygous females. ³ Prolonged electroencephalographic (EEG) monitoring with subdural electrodes aided in the capture of epileptic spasms in both the exon 6 knockout and the CDKL5 R56X knock-in genetic models. Curiously, the electroclinical epileptic spasms manifested again in aged adult mice, while in humans they typically occur in CDKL5-deficient neonates. The interspecies phenotypic difference seems to be modulated by the X-linked mosaicism at play in CDD. ⁶

Given the challenges in the generation and capture of seizures in these models of CDD, the work by Wang et al delivers important new insights and opens additional avenues for mechanistic investigations. ⁷ Cdkl5 is abundantly expressed in neurons and the authors affirmed the ample Cdkl5 expression in the forebrain and cerebellum in both the glutamatergic (V-glut1-positive) and GABAergic (Gad1-positive) cells. In pursuit of identifying the driving mechanism of CDD-related epilepsy, the authors generated 3 conditional model lines; two lines with glutamatergic forebrain-specific Cdkl5 deletion occurring either embryonically (Emx1-Cre) or postnatally (CamK2α-iCre) and one line with a postnatal deletion of Cdkl5 in GABAergic neurons (VGAT)-Cre. They observed spontaneous electroclinical behavioral seizures or status epilepticus in both lines of males deficient in Cdkl5 in glutamatergic but not GABAergic neurons between 2 and 7 months (8-28 weeks) of age. To confirm that the seizure phenotype was due to a Cdkl5 deletion in forebrain glutamatergic neurons and not a result of homologous recombination in the glia or due to a Cdkl5 deficiency elsewhere in a nervous system, Wang et al confirmed the absence of a consistent seizure phenotype in GFAP-driven Cre-mediated cKO of Cdkl5 mice and in mice Cdkl5 deficient in central and peripheral nervous system, the Nestin-cKO model. The presence of seizures in a model with glutamatergic forebrain-specific Cdkl5 deletion but not in mice with a more broader Cdkl5 absence in the central nervous system (Nestin-cKO model) is puzzling and warrants further research. Nevertheless, the discovery that Cdkl5 deficiency in forebrain glutamatergic neurons was sufficient to bring on a seizure phenotype is an important finding. Excessive glutamatergic transmission, hyperexcitability, increased levels of postsynaptic NMDA receptors, and an autistic phenotype were observed previously in animals with selective loss of Cdkl5 in GABA-ergic neurons. ⁸ The full scope of molecular and electrophysiological changes due to Cdkl5 deficiency in glutamatergic neurons remains to be elucidated. An additional important observation was the impact of model epilepsy on survival given the previous report of a sudden unexpected death in epilepsy (SUDEP) in a patient affected by CDD. ¹ While the onset of seizures in male mice deficient in Cdkl5 in glutamatergic neurons was delayed, once seizures occurred they significantly affected survival, with two-third of the mice dying by 7 months of age. This is an important phenomenon to note as Cdkl5 knockout mice were previously found prone to sleep apnea ⁹ which depending on its severity is considered a potential risk factor for sudden unexpected death. ¹⁰ Lastly, the authors observed mossy fiber sprouting in the dentate gyrus of the murine hippocampus and altered synaptic activity of the granule cells in mice with Cdkl5 deficiency in glutamatergic neurons. Mossy fiber sprouting has been observed in human patients, experimental limbic epilepsy models, and in many other animal seizure models. These structural hippocampal changes are known to contribute to the emergence of recurrent excitation. ¹¹ Interestingly, Yennawar et al showed an increase in GluA2-lacking AMPA receptors (AMPAR) in the hippocampus of adult Cdkl5R59X knock-in mouse, an increased rectification ratio of AMPAR-mediated excitatory postsynaptic currents (EPSCs) and an elevated early-phase long-term potentiation. ⁴ Future studies will hopefully clarify whether similar changes occur in animals with the selective Cdkl5 deficiency in glutamatergic neurons.

There are other unanswered questions about the mechanisms of CDD. Nevertheless, the work by Wang adds a valuable model system to the existing array of Cdkl5 mouse models and opens avenues for further investigations of molecular and electrophysiological changes involved in CDD. It also delivers important evidence that CDKL5 deficiency in glutamatergic neurons is sufficient for the emergence of model CDD-related epilepsy and the animals will likely be useful for investigations of underlying mechanisms and testing of candidate therapies. In the current study, the authors intentionally focused on hemizygous male animals in order to avoid the expected confounding effect of a random X-inactivation in females and to maximize the chance for the development of an overt electroclinical epilepsy as a more severe phenotype has been typically observed in CDKL5-deficient human male patients. ² Future studies evaluating and comparing the clinical phenotype of similarly genetically modified female mice will be of interest. Last but not least, the work of Wang et al also suggests that CDKL5 may be an important gene for an epilepsy-related premature mortality and future studies will be necessary to better understand these functions.