HCN1 Gain-Of-Function Mutations – a New Cause of Epileptic Encephalopathy

De Novo Mutations in HCN1 Cause Early Infantile Epileptic Encephalopathy.

Nava C, Dalle C, Rastetter A, Striano P, de Kovel CG, Nabbout R, Cancès C, Ville D, Brilstra EH, Gobbi G, Raffo E, Bouteiller D, Marie Y, Trouillard O, Robbiano A, Keren B, Agher D, Roze E, Lesage S, Nicolas A, Brice A, Baulac M, Vogt C, El Hajj N, Schneider E, Suls A, Weckhuysen S, Gormley P, Lehesjoki AE, De Jonghe P, Helbig I, Baulac S, Zara F, Koeleman BP; EuroEPINOMICS RES Consortium. Nat Genet 2014;46:640–645.

Hyperpolarization-activated, cyclic nucleotide–gated (HCN) channels contribute to cationic Ih current in neurons and regulate the excitability of neuronal networks. Studies in rat models have shown that the Hcn1 gene has a key role in epilepsy, but clinical evidence implicating HCN1 mutations in human epilepsy is lacking. We carried out exome sequencing for parent-offspring trios with fever-sensitive, intractable epileptic encephalopathy, leading to the discovery of two de novo missense HCN1 mutations. Screening of follow-up cohorts comprising 157 cases in total identified 4 additional amino acid substitutions. Patch-clamp recordings of Ih currents in cells expressing wild-type or mutant human HCN1 channels showed that the mutations had striking but divergent effects on homomeric channels. Individuals with mutations had clinical features resembling those of Dravet syndrome with progression toward atypical absences, intellectual disability and autistic traits. These findings provide clear evidence that de novo HCN1 point mutations cause a recognizable early-onset epileptic encephalopathy in humans.

Commentary

Early infantile epileptic encephalopathy (EIEE) refers to severe early onset epilepsy with associated neurodevelopmental abnormalities. The term was used initially to reflect EIEE with suppression-burst pattern on EEG, or Ohtahara syndrome, but it has been applied more widely to a larger and growing group of epileptic encephalopathies; patients with mutations in the EIEE genes display a range of epilepsy syndrome phenotypes, including infantile spasms, Lennox-Gastaut syndrome, Dravet syndrome, and migrating partial seizures of infancy. The formally designated EIEE genes in the Online Mendelian Inheritance in Man (OMIM) database now number 25, and many additional genes involved in severe early onset epilepsy have been identified in the era of next-generation sequencing (1–3). One such gene is HCN1 (EIEE25), encoding the hyperpolalization-activated, cyclic nucleotide-gated (HCN) channel (4). A group of four HCN channels mediates a cationic current (Ih). HCN channels possess an inherent negative feedback property – hyperpolarization activates HCN channels, whereas depolarization deactivates them. The net effect of HCN channel activation is a decrease in membrane input resistance. Therefore, Ih tends to stabilize the neuronal membrane potential toward resting potential. HCN1 has been localized to neuronal dendrites, where it plays a major role in membrane potential stabilization (4). Thus, HCN channels might be expected to play a role in epilepsy (5).

Nava and colleagues identify a novel gene for early onset epilepsy using a logical approach to patients with a known epilepsy syndrome. Their initial cohort of 39 cases was described as having many of the core features of Dravet syndrome, characterized by epilepsy with onset in infancy, an association of fever as a seizure trigger, a combination of seizure types typically involving GTCs and myoclonic seizures but also other generalized and sometimes focal seizure types, intractability, and intellectual impairment. The majority of patients with classic features of Dravet syndrome have mutations in the SCN1A gene that encodes the alpha subunit of the voltage-gated sodium channel (6). In girls with Dravet syndrome who do not have mutations (sequence changes or deletions), the current genetic differential diagnosis includes PCDH19, a gene that encodes protocadherin 19 and is associated with epilepsy in females with intellectual impairment (7). These two genes were evaluated in the first 39 cases using a clinical syndrome-guided approach to genetic diagnosis. When the results were negative, the authors performed whole exome sequencing of their cohort using a now well-established “trio” approach (2). Two cases harbored apparent HCN1 mutations—rare missense variants not present in control population data and predicted to disrupt protein function. At the time, HCNs were not clinically testable. The finding of the two apparently new mutations prompted the group to assess an additional 157 cases, and an additional four variants were identified, three confirmed to be de novo.

A major strength of the report by this international collaborative group is the functional assessment undertaken to assess pathogenicity for each variant and provide insight into the mechanisms of HCN dysfunction. The Chinese hamster ovary (CHO-K1) cell system was used to perform patch-clamp recordings, comparing cells expressing wild type (WT) HCN1 to cells expressing mutant forms of the protein. Only three of the six variant forms of the protein—S100P, H279Y, and D401H—displayed voltage-dependent inward currents consistent with the presence of functioning channels. These mutant forms of HCN1 showed slower deactivation as well as other features suggesting that the mutations conferred a gain of function (excess current). In contrast, the HCN1 R297T and S272P variant proteins did not display a measurable current and were present in less abundance in cell lysates compared to WT HCN1 lysates, suggesting that these variants led to lack of stability and thus degradation of the mutant HCN1 protein and loss of function. Indeed, even when the authors co-expressed these mutant forms of the HCN1 protein with WT HCN1, normal function was not restored. These data support the assertion that the mutations exert a “dominant negative” effect: the presence of a heterozygous mutation resulting in some amount of mutant protein, even in the presence of normal protein, disrupts the normal HCN1-mediated current.

The final variant, G47V, affected a strongly but not completely conserved amino acid and could not be confirmed as de novo due to lack of availability of one of the parents. Interestingly, despite this limitation and lack of functional corroboration of pathogenicity of this variant, the Dravet-like epilepsy phenotype of the G47V variant case is fully consistent with the other cases. The only different features are the absence of language and ataxia in this case, whereas the other cases all had behavioral disturbances and autism; only one case had truncal ataxia. The G47V variant would thus be considered highly suspect though not definitive, as the authors suggest. This degree of uncertainty of significance of a rare variant in the presence of a reasonably consistent clinical phenotype is one that will continue to be an issue as clinical sequencing of panels of genes and exome sequencing continues in patients with epilepsy.

The report from Nava and colleagues highlights the need to start with a phenotype-based approach to genetic testing—but when that is unsuccessful, to move on to a genotype-first approach. Their approach parallels the approach undertaken now in the clinic, which is far from perfect but offers the possibility of genetic diagnosis to an increasing number of children with severe early onset epilepsy (8). Their work nicely demonstrates the value of collaborative sequencing efforts and functional validation of suspected variants. The impact of each mutation—gain or loss of function—could not have been predicted a priori by examining what is currently known about the protein structure of HCN1. The importance of understanding whether epilepsy results from gain or loss of function of a channel is important scientifically as we begin to understand the role of this protein in clinical epilepsy. This type of analysis also has tremendous potential clinical implications as basic and clinical researchers are undertaking rational trials of drugs in vitro and in patients based on the known or presumed mechanism of each patient's mutation. For example, some patients with mutations in the EIEE gene KCNT1 may be amenable to treatment with quinidine, which reduces the excess potassium currents produced by gain of function mutations (9, 10). However, patients with loss of function mutations could be worsened by administration of such a treatment. With the current opportunity to target treatment for HCN1-related epilepsy comes the responsibility to understand each patient's mutation well. Nava and colleagues have done so in an exemplary fashion, providing not only a new gene for EIEE but also a model of how such findings should be reported.