The Role of PRRT2 in Synaptic Transmission May Not Be So Benign

Commentary

PRRT2 Is a Key Component of the Ca²⁺ Dependent Neurotransmitter Release Machinery.

Valente P, Castroflorio E, Rossi P, Fadda M, Sterlini B, Cervigni RI, Prestigio C, Giovedi S, Onofri F, Mura E, Guarnieri FC, Marte A, Orlando M, Zara F, Fassio A, Valtorta F, Baldelli P, Corradi A, Benfenati F. Cell Rep 2016;15:117–131.

Heterozygous mutations in proline rich transmembrane protein 2 (PRRT2) underlie a group of paroxysmal disorders, including epilepsy, kinesigenic dyskinesia, and migraine. Most of the mutations lead to impaired PRRT2 expression, suggesting that loss of PRRT2 function may contribute to pathogenesis. We show that PRRT2 is enriched in presynaptic terminals and that its silencing decreases the number of synapses and increases the number of docked synaptic vesicles at rest. PRRT2 silenced neurons exhibit a severe impairment of synchronous release, attributable to a sharp decrease in release probability and Ca(²⁺) sensitivity and associated with a marked increase of the asynchronous/synchronous release ratio. PRRT2 interacts with the synaptic proteins SNAP 25 and synaptotagmin 1/2. The results indicate that PRRT2 is intimately connected with the Ca(²⁺) sensing machinery and that it plays an important role in the final steps of neurotransmitter release.

PRRT2 occupies a unique niche in the world of epilepsy genes. Gene variants in PRRT2 are causative for approximately 80% of cases of benign familial infantile epilepsy (BFIE), an autosomal dominant seizure disorder in which seizure onset occurs at a mean age of 6 months but usually remits by 2 years (1). These same variants, however, can also cause paroxysmal kinesigenic dyskinesia (PKD), a disease characterized by short attacks of irregular involuntary movements triggered by sudden voluntary motion. These two disorders can occur separately or together, a presentation that is termed infantile convulsions with choreoathetosis (ICCA), where patients present with BFIE during infancy and then PKD later in life (2).

The majority of PRRT2 variants that cause these diseases are thought to lead to a premature stop codon and loss of PRRT2 protein function. But despite the strong genetic evidence that these PRRT2 gene variants cause neurological disease, the function of the protein has remained mysterious. Like a handful of other epilepsy genes (STXBP1, DNM1, LGI1), PRRT2 is enriched in the presynaptic terminal and interacts with SNAP 25, a key member of the SNARE protein complex that enables calcium evoked synaptic vesicle endocytosis (3). This led Valente and colleagues to hypothesize that the physiological function of PRRT2 was to regulate the calcium dependent fusion of synaptic vesicles.

After confirming that PRRT2 was primarily expressed in the presynaptic compartment, the authors set out to test their hypothesis by using shRNAs to reduce the expression of PRRT2 in mouse hippocampal neuron primary cultures. Although the extent of the reduction in PRRT2 was not quantified, a representative Western blot showed a reduction in PRRT2 protein. The bulk of their dataset employed whole cell voltage clamp analysis of autaptic neuronal cultures, a preparation in which isolated neurons synapse with themselves, making the normally difficult task of measuring presynaptic function more manageable. When these neurons were stimulated and the synaptic currents measured, both glutamatergic and GABAergic neurons showed a reduction of about 80% in the peak current amplitude. An 80% reduction suggests a severe deficit in synaptic transmission, especially for a knockdown, and is in the range usually seen with impairment of the most essential synaptic proteins.

According to the quantal theory of neurotransmission (4), reductions in evoked synaptic currents can be caused by: 1) a change in the number of available synaptic vesicles (n), 2) a change in the current in response to single vesicle fusion (q), or 3) a change in the probability of vesicle release (p). The authors tested these possibilities with several electrophysiological assays and concluded that the 80% reduction in evoked release was accounted for by a small reduction in q and a larger reduction in p. This reduction in the probability of release could be rescued by overexpressing a shRNA resistant form of Prrt2, but not by altering the calcium concentrations in the bath solution, suggesting that reduction of PRRT2 impairs synaptic vesicle release by interfering with the calcium sensing machinery. Further biochemical experiments showed that, in addition to an interaction with SNAP 25, PRRT2 may also interact with two forms of synaptotagmin, the protein that is the major calcium sensor for evoked vesicle release, providing a plausible molecular mechanism for the effect of Prrt2 knockdown on synaptic transmission.

The data also show, however, that the effects of Prrt2 knockdown did not seem to be limited to an effect on calcium triggered exocytosis. For example, there was the aforementioned decrease in q, a reduction of over 50% in synapse density, and even effects on synaptic vesicle recycling that were investigated by using a fluorescent reporter of vesicle reacidification. Although these effects could all possibly be secondary to a primary deficit in calcium sensing, it is more likely that the normal role of PRRT2 extends beyond maintaining efficient calcium triggered neurotransmitter release. These roles could include maintenance of calcium homeostasis at the synapse, or even a more general role in neuronal viability. Neurons cultured from Snap–25 knockout mice, for example, appear normal after 7 days, but after that, the dendrites begin to degenerate and the neurons die (5). Thus, although the authors provide strong evidence that Prrt2 knockdown impairs the calcium sensitivity of evoked neurotransmitter release, the specificity of this phenotype and its relationship to epilepsy and dyskinesia are open questions.

The authors have taken steps toward addressing these questions in a recently published follow up paper, in which they generated a Prrt2 knockout mouse and looked at regional Prrt2 expression, synaptic physiology, and dyskinesia and seizures (6). Using a lacZ reporter, they found that these mice showed the highest levels of Prrt2 expression in the cerebellum and hindbrain. Hippocampal expression was evident, but mostly localized to the hilar region of the dentate gyrus. Although the mice had normal growth and general health, they showed motor impairments such as loss of balance events starting at 8 days old that persisted through adulthood. The mice did not display spontaneous seizures, but they exhibited audiogenic seizures and the duration of seizures induced by PTZ was longer in the knockout mice than in controls. Surprisingly, whole cell patch clamp analysis of Prrt2 knockout dentate granule neurons and cerebellar Purkinje neurons in acute slices did not show the same large reductions in evoked synaptic responses that were seen with knockdown of Prrt2 in the primary hippocampal neurons, although there were more subtle differences, including an increase in facilitation in Purkinje neurons that mirrored what was found in the cultured hippocampal neurons.

Taken together, these two studies provide important new information about the potential roles of PRRT2 in neuronal physiology and in disease but leave many unanswered questions about the relationship between the two—questions that, to be fair, are outstanding for most of the genes in which we now know that variants can cause epilepsy. One of the most pressing questions in the case of PRRT2 is why knockdown in primary hippocampal neurons resulted in a severe impairment of synaptic transmission that was not seen in the slice recordings from knockout animals. This could be due to the interesting finding that the expression of PRRT2 in the brain has a very restricted distribution, suggesting that the population of synapses recorded from in slice contained only a minority of knockout synapses. It is also possible that the intact brain can partially compensate for impairments in PRRT2 function over time, which could explain why seizures in humans remit around 2 years of age. Understanding these relationships between cellular, network, and behavioral phenotypes—where they converge and why they sometimes break down—is key to ultimately understanding neurological disease.