Developmental and epileptic encephalopathies (DEE) are a devastating group of diseases, often with poor response to pharmacological treatment, resulting in a lifelong burden of seizures, developmental delay, and intellectual disability. Since the original discovery over 20 years ago that de novo mutations in voltage-gated ion channels can directly cause early childhood epilepsy syndromes (Singh et al., 1998), the number of genes and gene variants associated with DEE has grown exponentially. Current estimates indicate that ~30% of DEE patients carry at least one pathogenic variant in the top 100 known gene candidates for the disease, with about a third of the affected genes falling into the category of voltage- and ligand-gated ion channels (Noebels, 2017; Oyrer et al., 2018; Wang and Frankel, 2021). Despite the wealth of genetic data available and recent efforts to model the effects of the pathogenic variants in genetically modified mice, we understand very little about the mechanisms that underlie the brain and neuronal circuit alterations responsible for seizures or the reasons for their drug resistance.

Here, we focus on seizures and pharmacological response profile associated with mutations in the HCN1 gene, which encodes a hyperpolarization-activated cyclic nucleotide-regulated non-selective cation channel expressed prominently in the brain (Santoro and Shah, 2020). To date, a total of 40 different missense variants in the HCN1 gene have been reported in patients affected by epilepsy and/or neurodevelopmental disorders (source: HGMD Professional 2020.4). Among these, at least 12 different de novo HCN1 variants have been linked to DEE (Nava et al., 2014; Butler et al., 2017; Marini et al., 2018; Wang et al., 2019).

The functional role of HCN1 channels is tied closely to their distinct subcellular localization in the two classes of neurons where HCN1 protein is primarily expressed, pyramidal neurons in neocortex and hippocampus, and parvalbumin-positive (PV+) interneurons across the brain (Notomi and Shigemoto, 2004). Within these neuronal classes, the subcellular localization of HCN1 channel subunits is tightly regulated. In excitatory (pyramidal) neurons, the channel is targeted to the distal portion of the apical dendrites, where it constrains the dendritic integration of excitatory postsynaptic potentials, limiting excitability (Tsay et al., 2007; George et al., 2009). In inhibitory (PV+) neurons, HCN1 channels are localized exclusively to axonal terminals, where they facilitate rapid action potential (AP) propagation (Roth and Hu, 2020) and may regulate GABA release (Southan et al., 2000; Aponte et al., 2006). Decreased expression or activity of HCN1 channels, as observed in several rodent models of acquired epilepsy, has been thus posited to increase overall network excitability and contribute to epileptogenesis and seizures (Jung et al., 2011; McClelland et al., 2011; Brennan et al., 2016).

In this study, we examine two new genetic mouse models that reproduce de novo HCN1 variants previously shown to be associated with severe forms of neonatal- and childhood-onset epilepsy (Parrini et al., 2017; Marini et al., 2018; Fernández-Marmiesse et al., 2019). A prior attempt at modeling HCN1-linked DEE in mice (Bleakley et al., 2021) resulted in a relatively mild phenotype, wherein spontaneous seizures could not be systematically documented, despite the presence of epileptiform spikes on electrocorticogram recordings (ECoG). Furthermore, it has been reported that mice with either global or forebrain-restricted deletion of HCN1 also lack spontaneous seizures (Nolan et al., 2003; Nolan et al., 2004), although seizure susceptibility is increased (Huang et al., 2009; Santoro et al., 2010). This prompted the question whether there may be some general limitation to modeling HCN1-linked DEE in mice. To examine this question and probe further whether and how HCN1 sequence variations may give rise to DEE, we generated mouse models for two other HCN1 variants linked to DEE in humans.

We examined variants c.1172G>A/p.Gly391Asp and c.459G>C/p.Met153Ile based on three criteria: severity of disease; occurrence in at least two independent patients with similar epilepsy phenotypes; and differing biophysical effects on channel function when tested in heterologous expression systems (Marini et al., 2018). Moreover, clinical reports suggest that seizures are poorly controlled with standard antiseizure medications and that certain anticonvulsants in fact exacerbate seizures in these patients (Marini et al., 2018). Our results show that both de novo variants lead to spontaneous seizures in mice and reproduce the paradoxical pharmacological responses seen in patients, providing new insights into the mechanistic basis for the limitations of current drug therapies.

Pathogenic sequence variants orthologous to human HCN1 variants p.G391D and p.M153I were generated using the CRISPR/Cas9 approach in the Hcn1 gene of inbred C57BL/6J mice (see ‘Materials and methods’; Tröder et al., 2018). The resulting lines, Hcn1^G380D/+ and Hcn1^M142I/+, were maintained using a male heterozygote × female wildtype (WT) breeding scheme and heterozygous animals were compared to WT littermate controls. For simplicity, the equivalent variants in mice and humans will be hereafter cited using the appropriate protein and gene nomenclature formatting, however, omitting positional information numbers (e.g., HCN1-GD or HCN1-MI, and Hcn1^GD/+ or Hcn1^MI/+).

General phenotypic features and gross brain anatomy

No overt differences were noted in pup development, except for consistently smaller body weights at weaning in Hcn1^GD/+ mice of both sexes (Figure 1A). Smaller body weights persisted for the life of Hcn1^GD/+ animals and were accompanied by an ~12% reduction in adult brain weight with gross brain anatomy otherwise normal (Figure 1B). Brain area measurements showed a similar reduction. Thus, both cerebellum and brainstem areas were significantly smaller in Hcn1^GD/+ mice compared to WT (Figure 1—figure supplement 1), with the cerebellum more affected than the brainstem, consistent with the higher expression of HCN1 protein in this region (Notomi and Shigemoto, 2004). Reduced mean body weight was also observed in Hcn1^MI/+ males, but no decrease in brain size was noted in either sex for this line (Figure 1A and B).

Figure 1 with 1 supplement see all

Download asset Open asset

Survival curves revealed a significant number of premature deaths in both Hcn1^GD/+ and Hcn1^MI/+ animals (Figure 1C). In many instances, animals were found dead in the cage without having shown prior signs of distress, suggesting sudden death. The survival curves revealed some notable differences in the pattern of death observed between the two lines, both with regards to timing and sex specificity. Most deaths among Hcn1^GD/+ animals occurred before the age of 3 months (Figure 1C, left), with no difference between the sexes (survival rate at 30 weeks: 78% male versus 75% female Hcn1^GD/+ mice, p=0.369, Fisher’s exact probability test). In contrast, for Hcn1^MI/+ animals, females were more severely affected than males (survival rate at 30 weeks: 86% male versus 55% female Hcn1^MI/+ mice, p<0.001, Fisher’s exact probability test), with most deaths in either sex occurring at an age greater than 3 months (Figure 1C, right). These findings suggest important differences in the effects of the two variants on brain function and physiology, perhaps reflective of the divergent effects of the mutations on channel biophysical properties (Marini et al., 2018) and HCN1 protein expression (see below).

Given the high expression of HCN1 subunits in cerebellum, both in Purkinje neurons and basket cell axon terminals (Southan et al., 2000; Luján et al., 2005; Rinaldi et al., 2013), as well as the motor phenotype observed in HCN1 global knockout animals (Nolan et al., 2003; Massella et al., 2009), we quantified basic locomotion in an open-field setting (Figure 2A), assessed gait parameters (Figure 2D), and tested for motor coordination, learning, and strength using a vertical pole paradigm (Figure 2G). Hcn1^GD/+ and Hcn1^MI/+ animals showed locomotor hyperactivity in the open field, with significant increases in distance moved and running speed, but a reduction in the center versus border zone occupancy ratio, potentially indicative of increased anxiety-like behavior in both lines (Figure 2B and C, Figure 2—figure supplement 1). Gait analysis reflected the overall higher running speed of Hcn1^GD/+ and Hcn1^MI/+ animals compared to WT littermates (Figure 2E). When runs of similar speed were compared, significant changes in the base of support (BOS) were observed in Hcn1^GD/+ animals, with the front paws’ BOS being wider at both speed ranges (Figure 2F), and the hind paws’ BOS narrower at higher running speeds (Figure 2—figure supplement 2). No BOS alterations were noted in Hcn1^MI/+ animals or any of the other gait parameters measured (Figure 2F, Figure 2—figure supplement 3). When comparing the ability to climb down a vertical pole, Hcn1^GD/+ animals displayed significant difficulties, with only 1 of 18 animals able to successfully complete the test during trials 1 and 2, and 2 of 18 animals succeeding in trial 3 (Figure 2H, Figure 2—video 1 and Figure 2—video 2). In contrast, Hcn1^MI/+ animals performed as well as their WT littermates, with all mice successfully completing the task and their average latencies to reach the floor comparable between the two groups (Figure 2I, Figure 2—video 3, Figure 2—video 4).

Figure 2 with 7 supplements see all

Download asset Open asset

Figure 2—source data 1

Open-field parameters for Hcn1^GD/+ and Hcn1^MI/+ heterozygous mice in comparison to WT littermates.

Number of animals is indicated in parentheses. Times in border and center zones, respectively, are expressed as percentage of total trial duration. *Data was analyzed with a Mann–Whitney U test; ^#data was analyzed with a Student’s t-test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig2-data1-v2.docx

Download elife-70826-fig2-data1-v2.docx

Figure 2—source data 2

Catwalk gait analysis parameters for Hcn1^GD/+ and Hcn1^MI/+ heterozygous mice in comparison to WT littermates.

Parameters listed correspond to the right front paw. Number of animals is shown in parentheses. *Data was analyzed with a Mann–Whitney U test; ^#data was analyzed with a Student’s t-test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig2-data2-v2.docx

Download elife-70826-fig2-data2-v2.docx

Figure 2—source data 3

Vertical pole test parameters for Hcn1^GD/+ and Hcn1^MI/+ heterozygous mice in comparison to WT littermates.

Number of animals is indicated in parentheses. Latency data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig2-data3-v2.docx

Download elife-70826-fig2-data3-v2.docx

DEEs are characterized by the presence of significant comorbidities, including developmental delay and cognitive impairments; and patients carrying the HCN1 p.G391D and p.M153I variants were reported to display severe and mild intellectual disability, respectively (Marini et al., 2018). Since HCN1 expression is also very high both in pyramidal neuron dendrites and basket cell terminals throughout the forebrain cortex (Notomi and Shigemoto, 2004), we next sought to test the performance of Hcn1^GD/+ and Hcn1^MI/+ mice on behaviors that require forebrain cortical function. Spontaneous alternation is a naturalistic behavior in rodents, which is strongly dependent on intact spatial working memory and involves the activity of several brain regions, including hippocampus and neocortical structures (Lalonde, 2002). When tested in a Y maze-based spontaneous alternation task (Figure 3A), Hcn1^GD/+ animals failed to display normal alternation behavior, with alternation rates not exceeding the chance level of 50% (Figure 3B). In contrast, Hcn1^MI/+ animals showed equal performance compared to WT littermates. Both Hcn1^GD/+ and Hcn1^MI/+ animals needed significantly less time to complete the required 24 arm transitions (Figure 3C), consistent with the locomotor hyperactivity observed in these mice. We next tested the mice in an object recognition memory task using a 24 hr interval between training and testing (Figure 3D), a paradigm that assesses long-term memory and also requires the activity of multiple cortical regions (Barker and Warburton, 2011; Haettig et al., 2011). We found that mutant mice from both lines were able to distinguish the novel object from the familiar object, as reflected in positive discrimination indices (D.I.) (Figure 3E). Hcn1^GD/+ animals also spent a significantly higher time exploring the novel object, compared to their WT littermates, although the meaning of this finding is unclear.

Figure 3

Download asset Open asset

Figure 3—source data 1

Spontaneous alternation in the Y maze.

Alternation rates and time to complete 24 transitions are expressed as the mean of 2 days. Number of animals is indicated in parentheses. *Data was analyzed using a Mann–Whitney U test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig3-data1-v2.docx

Download elife-70826-fig3-data1-v2.docx

Figure 3—source data 2

Object recognition memory test.

Number of animals is indicated in parentheses. *Data was analyzed using a Mann–Whitney U test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig3-data2-v2.docx

Download elife-70826-fig3-data2-v2.docx

In summary, our results reveal a complex effect of the two variants on phenotype, with an overall more severe impact on the general fitness of Hcn1^GD/+ compared to Hcn1^MI/+ animals, reflected in smaller body size, reduced brain size, as well as altered gait and locomotion, reduced motor coordination and impairment in certain higher cognitive abilities.

Hcn1^GD/+ and Hcn1^MI/+ mice show spontaneous convulsive seizures

The sudden death phenotype observed in the two mouse lines, along with the epilepsy syndromes seen in patients carrying HCN1 p.G391D and p.M153I variants, suggests the animals may be undergoing generalized tonic–clonic seizures (GTCS), which in mice often result in death when escalating into tonic hindlimb extension. One such death event was indeed observed in a female from the Hcn1^MI/+ line, and several convulsive seizure events were witnessed in mice from both lines during routine cage inspections. We therefore set out to rigorously quantify the occurrence, frequency, and severity of seizures in Hcn1^GD/+ and Hcn1^MI/+ mice by using chronic video ECoG recordings for periods of 1–2 weeks at a time (Figure 4). Adult mice from both lines and sexes were found to display spontaneous GTCSs with distinct electrographic signature on ECoG traces (Figure 4A and C) and behavioral manifestation on video recordings (Figure 4—video 1, Figure 4—video 2).

Figure 4 with 3 supplements see all

Download asset Open asset

We recorded from a total of n = 8 Hcn1^GD/+ animals (two males, six females) 56–91 days old to avoid selecting against animals with early death; and a total of n = 12 Hcn1^MI/+ animals (six males, six females) 80–130 days old to capture the period during which sudden death is most frequently seen in females. Notably, despite the higher occurrence of death in females, seizures were observed with similar frequency in males and females from the Hcn1^MI/+ line (Figure 4D). Seizure events were comparatively infrequent, ranging from 0.6 to 4.1 seizures per week in Hcn1^GD/+ mice (with 2/8 individuals failing to show seizures during the time recorded) and from 0.6 to 5.1 seizures per week in Hcn1^MI/+ mice (with 2/12 individuals displaying no seizures during the recorded period). Despite the overall more severe phenotype of Hcn1^GD/+ animals, based on body weight, brain size, behavioral performance, and general appearance, when rated on a modified Racine scale (see ‘Materials and methods’) we found that animals from the Hcn1^GD/+ line had on average lower maximum grade seizures compared to Hcn1^MI/+ animals (Figure 4B and D).

As GTCS in mice usually reflect involvement of the limbic system, including the hippocampus, we sought to additionally characterize the epilepsy phenotype of Hcn1^GD/+ and Hcn1^MI/+ mice by performing morphological analysis and immunohistochemical labeling for known hippocampal markers of epilepsy. In line with our video ECoG findings, we observed upregulation of neuropeptide Y in dentate gyrus mossy fibers, alterations in granule cell layer morphology and extracellular matrix deposition in the dentate gyrus, as well as increased hippocampal gliosis in a majority of mutant animals from both lines (Figure 4—figure supplement 1). Such findings are consistent with the presence of spontaneous seizures and altered excitability within the hippocampal circuit (Houser, 1990; Marksteiner et al., 1990; Stringer, 1996; Magagna-Poveda et al., 2017). While the site of seizure onset may differ between species, with a more prevalent role of neocortex in human compared to mouse, these results confirm that Hcn1^GD/+ and Hcn1^MI/+ mice are an overall appropriate model for human HCN1-associated DEE, and that the two variants are causal to the epilepsy phenotype.

Effects of mutations on HCN1 expression and the intrinsic properties of hippocampal CA1 pyramidal neurons

Prior in vitro studies of the human p.G391D and p.M153I variants using heterologous expression systems have demonstrated the functional impact of these mutations on HCN1 channels and the mixed hyperpolarization-activated Na⁺/K⁺ current they generate, known as I_h (Marini et al., 2018). HCN1 subunits with the G391D mutation fail to express current on their own, but when combined with WT subunits (as occurs in patients and heterozygote Hcn1^GD/+ mice) yield, in the ~20% fraction of cells that show measurable currents, a ‘leaky’ channel with significantly decreased current density but a greatly increased voltage- and time-independent inward, depolarizing current component. In contrast, the M153I mutation only slightly reduces current density, but accelerates the opening kinetics and shifts the voltage dependence of the HCN1 channel to more depolarized potentials (midpoint of activation is shifted by +36 mV for homomeric mutant channels and by +12 mV for heteromeric M153I/WT channels). However, it is not known how these mutations affect HCN1 expression and neuronal physiology in a native environment, which includes both the HCN channel auxiliary subunit TRIP8b (Lewis et al., 2009; Santoro et al., 2009) and other channel modulators. To address this question, we performed immunohistochemical labeling of HCN1 protein and patch-clamp recordings from CA1 pyramidal neurons in Hcn1^GD/+ and Hcn1^MI/+ mice and compared their properties with WT littermate controls.

HCN1 antibody labeling of the hippocampus in WT animals showed the typical distribution of HCN1 subunits in CA1 pyramidal neurons (Figure 5A), where the channel is present in a gradient of increasing expression along the somatodendritic axis of the apical dendrite (Lörincz et al., 2002). This gradient of expression was preserved in brains from both lines, although labeling revealed a substantial decrease in overall HCN1 protein levels in Hcn1^GD/+ brains, with a smaller but still significant decrease observed in Hcn1^MI/+ brains (Figure 5B). In whole-cell patch-clamp recordings, such results were matched by a substantial reduction in the voltage sag observed in response to hyperpolarizing current injections, a hallmark of I_h activity in neurons (Figure 5C and D). The voltage sag was absent in Hcn1^GD/+ CA1 pyramidal neurons and significantly reduced in Hcn1^MI/+. The lack of voltage sag in Hcn1^GD/+ neurons could be due to decreased HCN1 protein expression, a dominant negative effect of HCN1-GD subunits on channel function and/or the increased instantaneous component of HCN channels containing HCN1-GD subunits, which would cause the channel to conduct an inward Na⁺ ‘leak’ current with no voltage dependence (Marini et al., 2018).

Figure 5 with 1 supplement see all

Download asset Open asset

Figure 5—source data 1

Action potential (AP) properties of Hcn1^GD/+ and Hcn1^MI/+ pyramidal neurons compared to respective WT littermate controls.

Number of cells is shown in parenthesis. Number of animals used for GD: WT n = 6 mice, GD n = 5 mice; and for MI: WT n = 3 mice, MI n = 3 mice. *Data was analyzed using a Mann–Whitney U test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig5-data1-v2.docx

Download elife-70826-fig5-data1-v2.docx

As wildtype HCN1 channels are partially open at the resting membrane potential (RMP) of most cells, including CA1 pyramidal neurons, they exert a well-characterized action to shift the RMP to more positive potentials and decrease the input resistance. Surprisingly, despite the loss of HCN1 protein expression, CA1 pyramidal neurons of Hcn1^GD/+ mice displayed a significant depolarization in the RMP; in contrast, there was no change in RMP in Hcn1^MI/+ neurons (Figure 5E). The more positive RMP of Hcn1^GD/+ mice was unexpected considering the strong reduction in HCN1 protein expression (Figure 5B) and hyperpolarized RMP in mice lacking HCN1 channels (Nolan et al., 2004). This may reflect the depolarizing action of the increased voltage-independent ‘leak’ current seen with certain HCN1 channel mutations, including p.G391D and p.M305L (Bleakley et al., 2021), and/or secondary changes in the function of other neuronal conductances. Input resistance was unaltered in both lines (GD: WT = 137.8 ± 7.1 MΩ, n = 20 cells, versus Hcn1^GD/+ = 153.7 ± 7.8 MΩ, n = 21; MI: WT = 128.4 ± 9.5 MΩ, n = 25, versus Hcn1^MI/+ = 125.5 ± 7.9 MΩ, n = 38). As input resistance in CA1 pyramidal neurons is typically increased in mice lacking HCN1 channels (Nolan et al., 2003; Nolan et al., 2004), in the case of Hcn1^GD/+ mice this may again reflect the offsetting actions of increased ‘leak’ current and decreased channel expression, or alternatively, offsetting effects due to alterations in the expression of other types of channels. It should be furthermore noted that recordings at the soma might not reveal changes in the input resistance occurring in distal dendrites, where HCN1 expression is highest (Figure 5A and B).

The possibility that the expression of mutant HCN1 channels may cause secondary changes in CA1 pyramidal neuron function is consistent with our finding that CA1 cells in both mouse lines show an impaired ability to fire APs, a phenotype not observed with acute application of the HCN channel-specific blocker ZD7288 (Gasparini and DiFrancesco, 1997; Figure 5—figure supplement 1). Thus, application of positive current steps from a common holding potential of –70 mV revealed an abnormal firing pattern in CA1 neurons from both lines (Figure 5C), wherein a majority of Hcn1^GD/+ neurons and a significant fraction of Hcn1^MI/+ neurons showed a reduction in the number of APs fired during 1-s-long depolarizing current steps (fraction of neurons with abnormal firing pattern: 16/21 cells = 76% Hcn1^GD/+ versus 13/35 cells = 37% Hcn1^MI/+, p=0.006, Fisher’s exact probability test showing a larger effect in the former cohort; 1/44 cells = 2% in the combined set of WT littermate neurons). In contrast, rheobase was unaltered in both lines when measured from a holding potential of –70 mV (GD: WT = 100 ± 9.6 pA, n = 20, versus Hcn1^GD/+ = 114.3 ± 8.0 pA, n = 21; MI: WT = 101.0 ± 6.8 pA, n = 25 versus Hcn1^MI/+ = 98.6 ± 6.4 pA, n = 38), and no changes were observed in AP threshold, AP peak or AP width (Figure 5—figure supplement 1). These findings suggest a lack of secondary changes in voltage-gated Na⁺ channel function, at least in pyramidal neurons, unlike what has been reported in the case of the p.M305L variant (Bleakley et al., 2021). However, alterations in the activity of other channel types (or alterations of Na⁺ channels in other neuronal classes) remain certainly possible. A microarray-based gene expression screening from WT, Hcn1^GD/+, and Hcn1^MI/+ total hippocampal lysates also did not reveal any changes in Na⁺ channel transcript levels, similar to those observed in Bleakley et al., 2021 (Source data 1; the only candidate with a p-value<0.001 being a K⁺ channel), although alterations in protein levels or post-translational modulation cannot be ruled out by this approach. Thus, further investigations will be needed to determine how altered conductance in Na⁺, K⁺, or other ion channels may potentially contribute to modified neuronal activity in Hcn1^GD/+ and Hcn1^MI/+ animals. Finally, no differences were observed between males and females for any of the measures assessed in either mouse line; hence, all reported data were pooled across sexes.

Overall, the observed changes reflect subtle alterations in the intrinsic properties of pyramidal neurons, which do not immediately explain the hyperexcitability and epilepsy phenotype observed in either mouse line. The reduced ability of Hcn1^GD/+ pyramidal neurons to maintain sustained firing may, in fact, act to limit circuit hyperexcitability in these animals – as reflected in the lower average maximum seizure grade reached by Hcn1^GD/+ animals both during spontaneous and drug-induced seizure events (Figure 4B and D; and see Figure 8C), compared to Hcn1^MI/+ animals. These results suggest that altered activity in non-pyramidal neurons may also play a role in the epilepsy phenotype.

Severe impairment in the axonal localization of HCN1 protein in Hcn1^GD/+ PV+ interneurons

As HCN1 subunits are highly expressed in PV+ interneurons in the brain, in addition to their expression in pyramidal neurons, we examined whether this expression is altered by the seizure-causing mutations. Whereas the highest site of expression of HCN1 subunits is at the distal portion of apical dendrites of pyramidal neurons, as detailed above, HCN1 is most strongly enriched in the axonal terminals of PV+ interneurons. The exquisite targeting of HCN1 protein to axonal terminals in PV+ interneurons can be particularly appreciated in basket cells from the cerebellar cortex (Figure 6). Here, HCN1 protein accumulates in a specialized structure, named the ‘pinceau,’ representing the axonal termination of basket cells onto the axon initial segment of Purkinje neurons. Surprisingly, at this location, we found a complete loss of HCN1 protein staining in Hcn1^GD/+ animals, while the distribution in Hcn1^MI/+ appeared relatively normal (Figure 6, top row). This result is unexpected as levels of HCN1 protein expressed from the WT allele should be normal, providing at least a half dosage in Hcn1^GD/+ cells, and suggests that the mutation may exert a dominant negative effect by co-assembling with WT HCN1 subunits, resulting in improper trafficking and/or premature degradation.

Figure 6 with 1 supplement see all

Download asset Open asset

Accumulation of misfolded protein in the endoplasmic reticulum or Golgi apparatus may also impair the trafficking of proteins other than HCN1. As several other proteins are known to specifically localize to the cerebellar pinceau (Kole et al., 2015), we used labeling for K_V1.2 to assess whether the expression of other channels normally targeted to the pinceau may also be disrupted in Hcn1^GD/+ brains. As shown in Figure 6 (middle row), K_V1.2 labeling of the pinceau was completely lost, similar to HCN1 protein labeling, in Hcn1^GD/+ but not Hcn1^MI/+ brains. The downregulation of K_V1.2 does not appear to be due to co-regulation with HCN1, as normal K_V1.2 labeling is still present in the cerebellum of HCN1 knockout mice (Figure 6—figure supplement 1) and the expression of the two channels at the cerebellar pinceau was shown to be independently regulated (Kole et al., 2015). As HCN1 channels in PV+ axonal terminals are thought to functionally interact with the Na⁺/K⁺ pump (Roth and Hu, 2020), we further sought to test for the presence of ATP1a3 expression at the pinceau. As shown in Figure 6 (bottom row), similar to HCN1 and K_V1.2, there was a complete loss of ATP1a3 labeling in Hcn1^GD/+ but not Hcn1^MI/+ animals compared to WT controls.

HCN1 channels are expressed in basket cell axons not just in the cerebellum but also throughout the forebrain cortex. Therefore, we next assessed protein labeling in hippocampus, particularly as this region appears to be involved in the mice epileptic phenotype (Figure 4—figure supplement 1). Similar to what was observed in the apical dendrites of CA1 pyramidal neurons (Figure 5A and B), immunolabeling of HCN1 protein in the axon terminals of hippocampal PV+ interneurons was strongly reduced in Hcn1^GD/+ animals, although not completely absent (Figure 7A). Since endogenous protein labeling is unable to distinguish between WT and mutant subunits, we sought to further probe for the trafficking of mutant HCN1 subunits in hippocampal PV+ interneurons using a viral transduction approach. To this end, we used mice that express Cre recombinase specifically in PV+ interneurons (Pvalb-Cre line) and stereotaxically injected into the hippocampal CA1 area an adeno-associated virus (AAV) driving Cre-dependent expression of hemagglutinin (HA)-tagged mouse HCN1 subunits (HA-HCN1). As shown in Figure 7B, following injection of either HA-HCN1-WT or HA-HCN1-MI, staining with anti-HA antibodies revealed efficient targeting and expression of virus-driven protein in the axonal terminals of PV+ interneurons. In stark contrast, injection of HA-HCN1-GD did not result in detectable levels of HA-tagged protein in axonal terminals, while protein accumulation was observed in the soma of targeted neurons (Figure 7B, top row).

Figure 7 with 2 supplements see all

Download asset Open asset

Quantitative analysis of HA-tagged protein expression within the pyramidal layer of virus-injected areas confirmed that the virtual absence of axonal staining for HA-HCN1-GD was consistent across multiple animals and injection sites, while HA-HCN1-WT and HA-HCN1-MI expression was comparable (all measurements performed excluding cell somas, see Figure 7—figure supplement 1 and ‘Materials and methods’ for details). Thus, the level of signal in sections from HA-WT-injected mice (63 ± 3 arbitrary units [au], n = 16 slices from n = 4 injection sites) was much greater than that observed in sections from HA-GD mice (10 ± 3 au, n = 12 slices from n = injection sites) and similar to levels in sections from HA-MI mice (64 ± 8 au, n = 18 slices from n = 4 injection sites). The signal in HA-GD mice differed significantly from that in HA-WT mice, whereas the signal in HA-MI mice did not differ significantly from that in HA-WT mice (one-way ANOVA with post hoc Holm–Šídák’s multiple-comparisons test: effect of Group, p<0.001, HA-WT versus HA-GD mice, p<0.001; HA-WT versus HA-MI mice, p=0.837). Co-staining with anti-parvalbumin antibodies furthermore revealed a striking decrease of axonal parvalbumin protein labeling in the targeted CA1 area from brains injected with virus expressing HA-HCN1-GD but not with virus expressing HA-HCN1-WT or HA-HCN1-MI (Figure 7B, middle row, and Figure 7—figure supplements 1 and 2). Quantitative analysis of parvalbumin protein expression within virus-injected areas (measured excluding cell somas, as described above) again demonstrated similar values in HA-WT mice (75 ± 3 au, n = 6 slices from n = 2 injection sites) and HA-MI mice (83 ± 7 au, n = 6 slices from n = 2 injection sites), but significantly reduced signal intensity in HA-GD-injected animals (48 ± 3 au, n = 12 slices from n = 4 injection sites; one-way ANOVA with post hoc Holm–Šídák’s multiple-comparisons test: effect of Group, p<0.001, HA-WT versus HA-GD mice, p=0.003; HA-WT versus HA-MI mice, p=0.418).

While the observed loss of parvalbumin, K_V1.2 or ATP1a3 labeling in axonal terminals does not distinguish between loss of protein expression and loss of axonal structure or morphology per se, these results underscore the markedly adverse effect of the HCN1-GD variant on PV+ interneuron cellular biology. Overall, our combined findings on native protein expression in the cerebellar pinceau and virus-transduced protein expression in the axonal terminals of hippocampal basket cells suggest that disruption of PV+ interneuron function, due to HCN1 protein misfolding or altered trafficking, may represent an important component of neuronal circuit dysfunction in HCN1-linked epilepsy.

Paradoxical response of Hcn1^GD/+ and Hcn1^MI/+ mice to antiseizure medications

Clinical reports from at least four different patients with HCN1-linked DEE (carrying variants p.M305L, p.G391D, and p.I380F; Marini et al., 2018; Bleakley et al., 2021; C. Marini, personal communication, March 2021) consistently revealed worsening of seizures after administration of either lamotrigine (LTG) or phenytoin. A particularly striking example is provided by the p.G391D variant, identified in two unrelated patients with similar course of disease. Both patients had neonatal seizure onset, characterized by daily asymmetric tonic seizures with apnea and cyanosis, severe developmental delay, and died between 14 and 15 months of age. Both patients also showed a paradoxical response to phenytoin, which caused the induction of status epilepticus. Given this precedent, we hypothesized that there may be systematic similarities between the adverse pharmacological response profile of patients with p.G391D and p.M153I variants and Hcn1^GD/+ and Hcn1^MI/+ mice, which would further endorse their use as models with construct, face, and potentially predictive validity.

Consistent with our hypothesis, we found that administration of LTG (23 mg/kg, i.p.) induced convulsive seizures, defined as grade 3 or higher on a modified Racine scale (see ‘Materials and Methods’), within 90 min from injection in 11/15 HCN1^GD/+ animals (Figure 8A and C, left). Several animals had multiple seizure bouts, ranging from rearing with forelimb clonus to wild running and jumping seizures, during the period of observation. Video ECoG recordings were performed on a subset of animals (2/15 mutant and 2 WT control animals) for a total of 24 hr after drug injection, confirming the presence of GTCS-like activity on ECoG in both animals (Figure 8A) but no seizure activity in control littermates (data not shown). Similar outcomes were seen in Hcn1^MI/+ animals, where convulsive seizures were observed in 6/9 mice in response to LTG (Figure 8A and C, right). Of note, four of the six Hcn1^MI/+ animals with adverse response had grade 6 seizures, again higher than the maximum seizure grade observed in Hcn1^GD/+ mice (Figure 8C). Video ECoG recordings performed on three Hcn1^MI/+ animals confirmed the presence of GTCS electrographic activity in 2/3 animals (Figure 8A, right). In both lines, all seizure events recorded through video ECoG were reflected in behavioral manifestations, implying that data collected through observation in non-implanted animals are likely to have captured all seizure events generated in response to LTG injection.

Figure 8

Download asset Open asset

Results with phenytoin (30 mg/kg PE, i.p.) were even more striking as 14/18 animals tested from the Hcn1^GD/+ line developed convulsive seizures within 2.5 hr of injection (this number includes 11/14 animals with no prior drug exposure and 3/4 animals previously exposed to LTG; Figure 8B and C, left). Latency to first seizure after injection was longer than observed with LTG, in line with the slower time to peak plasma concentration for phenytoin. In addition, multiple seizures were observed during a period up to 6 hr after injection, again consistent with the longer half-life of phenytoin compared to LTG (Markowitz et al., 2010; Hawkins et al., 2017). Among Hcn1^MI/+ animals, we observed seizures in 10/15 mice following phenytoin injection, with similar temporal pattern and ECoG signatures compared to Hcn1^GD/+ mice (this number includes 6/8 animals without prior drug exposure and 4/7 animals previously exposed to LTG, Figure 8B and C, right).

All the Hcn1^GD/+ or Hcn1^MI/+ animals tested, either in response to LTG or phenytoin, were randomized to receive a control vehicle injection either 1 week before or 1 week after the drug administration session. None had convulsive seizures in response to vehicle injection during the period of observation (4 hr for behavioral experiments and 24 hr for video ECoG experiments), further confirming that the seizures observed in the period following drug administration were a direct consequence of the pharmacological challenge (Figure 8C). As an additional control, we tested the effects of a third anticonvulsant, namely, sodium valproate (VPA), which has not been reported to worsen seizures in any of the patients with pathogenic HCN1 sequence variations in our database. Administration of sodium valproate (250 mg/kg, i.p.) did not result in the occurrence of convulsive seizures in any of the Hcn1^GD/+ or Hcn1^MI/+ animals tested (Figure 8C). However, use of 250 mg/kg VPA resulted in profound sedation of Hcn1^GD/+ mice, which was not seen in Hcn1^MI/+ mice or WT littermate controls, with 9/10 Hcn1^GD/+ animals failing to respond to gentle touch or passive moving of the tail 45 min after drug injection versus 2/8 Hcn1^MI/+ and 0/8 WT animals (note that tail pinching still elicited a response in all mice tested).

What may be the mechanism by which LTG and phenytoin induce seizures in mice with pathogenic HCN1 variants? Although both drugs act as Na⁺ channel blockers, LTG is also widely considered an HCN channel activator, a property thought to contribute to its antiepileptic effects (Poolos et al., 2002). Despite multiple reports on its action on I_h in neurons, showing somewhat inconsistent outcomes (Peng et al., 2010; Huang et al., 2016), no studies have been conducted to directly test the effect of LTG on isolated HCN channels expressed in heterologous systems. To obtain further clarification on the mechanism of action of LTG, we tested the drug on wildtype HCN1 and HCN2 channels after transient transfection in HEK293T cells (Figure 9). Surprisingly, we found that LTG had no effects on either the midpoint voltage of activation or maximal tail current density of either HCN1 or HCN2 channels; LTG also had no effects when HCN1 was coexpressed with its auxiliary subunit TRIP8b (Figure 9B). In contrast, under the same conditions, LTG exerted its expected inhibitory effect on Na_V1.5 sodium channel currents (Figure 9—figure supplement 1; Qiao et al., 2014). These results clearly show that LTG is not a direct activator of HCN channels and suggest that any effects reported in neurons are the result of indirect regulation of I_h, perhaps secondary to Na⁺ channel block, and likely contingent on cell-specific, state-dependent signal transduction pathways. Although further experiments in mouse brain slices are needed to fully rule out an effect of this drug on native HCN channels with pathogenic sequence variants, it seems likely the effects of LTG on Hcn1^GD/+ and Hcn1^MI/+ animals are due to its action to directly block Na⁺ channels – similar to the effects of phenytoin – rather than an enhancement of HCN currents.

Figure 9 with 1 supplement see all

Download asset Open asset

Figure 9—source data 1

Lamotrigine has no direct effect on HCN1 or HCN2 channel activity.

Number of cells is indicated in parenthesis. ^#Data was analyzed with a Student’s t-test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig9-data1-v2.docx

Download elife-70826-fig9-data1-v2.docx

While we have at present no available indication that either of the two patients carrying the HCN1 p.M153I variant showed adverse effects to LTG or phenytoin treatment, our results in mice suggest that despite the clear differences in their phenotypes, there may be some fundamental similarities in the disease mechanisms at play both in patients and mouse models for HCN1-linked DEE.

Reduced efficacy of HCN/I_h-blocking compounds on HCN1-G391D channels and Hcn1^GD/+ neurons

The observation that several of the known HCN1 variants, when tested for their properties in vitro (Nava et al., 2014; Marini et al., 2018; Porro et al., 2021) or in vivo (Bleakley et al., 2021), exhibit an increase in ‘leak’ current suggests that use of HCN channel blockers may be beneficial in the treatment of patients with certain HCN1 sequence variants. This seems especially warranted as the ‘leak’ component has been directly linked to hyperexcitability in cortical pyramidal neurons (Bleakley et al., 2021). Several such blockers exist, although at present none is able to efficiently cross the blood–brain barrier. More importantly, all currently available inhibitors that are specific for HCN channels, such as ZD7288, ivabradine, zatebradine, cilobradine, and their derivatives (Melchiorre et al., 2010; Del Lungo et al., 2012), are thought to act through the same mechanism, namely, as pore blockers by interacting with residues located within the pore cavity. At the same time, many of the most severe HCN1 variants, including p.G391D, are located in the S6 transmembrane domain of the channel, which forms the inner lining of the pore. Such observations raise the question whether pathogenic mutations affecting pore structure may decrease the efficacy of currently available HCN-blocking compounds.

The structural models presented in Figure 10F illustrate the proximity of the G391D mutation to the known binding site of ZD7288 in the HCN pore cavity, wherein the residue immediately preceding G391 (V390 in human HCN1, corresponding to V379 in mouse) was shown to critically affect the sensitivity of the channel to the drug (Cheng et al., 2007). The presence of charged aspartate residues in G391D-containing subunits leads to a disruption in the symmetry of the pore in heteromeric channels containing at least one WT subunit, and consequently in the relative distance between the V390 side chains facing the cavity. To test whether this predicted disruption in the ZD7288 binding site leads to a reduced efficacy of the drug, we expressed heteromeric HCN1 channels containing WT and G391D subunits in HEK cells. As illustrated in Figure 10A and B, compared to the effect of ZD7288 to reduce the current expressed by WT HCN1 channels by ~67%, the drug was minimally effective in blocking either the instantaneous or the time-dependent current component generated by mutant WT/G391D channels (~15% inhibition). Addition of Cs⁺, another HCN channel blocker (albeit a less specific one, compared to ZD7288) with a different mechanism of action and binding site (DiFrancesco, 1982), led to a strong reduction in the current expressed by both WT and G391D channels (Figure 10A), confirming that the observed currents were indeed generated by HCN1 WT/G391D heteromeric channels.

Figure 10

Download asset Open asset

Figure 10—source data 1

Lack of effect of ZD7288 on Hcn1^GD/+ neurons.

Number of cells is shown in parenthesis. *Data was analyzed with a paired t-test. ^#Data was analyzed with a Wilcoxon matched-pairs signed-rank test. Data represent mean ± SEM.

https://cdn.elifesciences.org/articles/70826/elife-70826-fig10-data1-v2.docx

Download elife-70826-fig10-data1-v2.docx

Parallel experiments in brain slices, where bath application of ZD7288 is the most common method used to isolate the contribution of I_h to neuronal physiology, demonstrate the additional difficulties posed by the limited efficacy of the drug in Hcn1^GD/+ neurons. As shown in Figure 10C–E, application of the compound did not modify either the RMP or input resistance in Hcn1^GD/+ neurons, while appropriately eliminating voltage sag, hyperpolarizing the RMP, and increasing input resistance in WT neurons. These results leave unanswered the question whether the positive shift in RMP in Hcn1^GD/+ neurons is directly due to ‘leaky’ HCN1 channels or is a secondary response to changes in other conductances. The development of new HCN-blocking compounds with different mechanisms of action will be important both in allowing a more targeted investigation of the role of the ‘leak’ component in the physiology of affected neurons, as well as devising tailored therapeutic interventions in patients carrying HCN1 pathogenic sequence variants with significant loss of channel voltage dependence.

The speed and precision of next-generation sequencing (NGS) allow the efficient genetic assessment of children that present with early-onset epilepsies. The overarching goal of such precision medicine is to identify the specific molecular and network mechanisms that link distinct epilepsy variants in individual genes to particular disease profiles, resulting in more effective gene- or variant-specific therapeutic strategies. Despite the fact that dysfunction in ion channels carrying missense mutations can be promptly modeled in heterologous expression systems, such assessments do not sufficiently reflect the dysfunction caused by the sequence variations in vivo, preventing the direct translation of such datasets into applicable clinical strategies. In our study, we set out to fill this information gap by designing new preclinical models, which reproduce key disease features of HCN1-linked DEE. The Hcn1^GD/+ and Hcn1^MI/+ knock-in mice we generated display not only an appropriate and robust epilepsy phenotype, along with comorbid behavioral abnormalities, but illustrate the difficulties intrinsic to the treatment of genetic developmental epilepsies, including the adverse response of individuals with HCN1 variants to conventional antiseizure medications, and the potential for altered drug sensitivity of mutant channels. These results underscore how mutations in voltage-gated ion channels can fundamentally alter network properties and the response of neuronal circuits to pharmacological challenge. Some of these alterations may be acute, while others may result from HCN1 channel dysfunction impacting brain development, as multiple studies have demonstrated the influence of HCN channels on cortical maturation (Schlusche et al., 2021) and the complex interplay between HCN isoforms in shaping developing neuronal networks (Surges et al., 2006; Bender and Baram, 2008; Stoenica et al., 2013). In this scenario, our models provide a useful system for testing mechanistic hypotheses and developing new therapeutic approaches.

A notable feature of our models and patients with HCN1 sequence variations is the paradoxical induction of seizures by LTG and phenytoin. A similar phenomenon has been reported both in Dravet syndrome patients carrying SCN1A pathogenic variants (Perucca and Perucca, 2019) and in mice modeling Scn1a haploinsufficiency (Scn1a^+/-; Hawkins et al., 2017). In Dravet syndrome, loss-of-function mutations in the SCN1A sodium channel gene are thought to promote seizures by exerting a net effect to reduce the excitability of inhibitory interneurons (Catterall, 2018). This has led to the suggestion that certain anticonvulsant drugs acting as Na⁺ channel blockers paradoxically exacerbate seizures in this syndrome as a result of a synergistic action to further suppress inhibition. Interestingly, the axon-specific expression of HCN1 channels in PV+ interneurons matches the similarly polarized subcellular distribution of voltage-gated Na⁺ channels, wherein Na_V1.1 (encoded by SCN1A) is the predominant subtype expressed in the axon of PV+ interneurons (Ogiwara et al., 2007; Ogiwara et al., 2013; Dutton et al., 2013; Hedrich et al., 2014; Hu and Jonas, 2014). To maintain ionic homeostasis during repetitive firing, PV+ interneuron axons also strongly express Na⁺/K⁺-ATPases (Peng et al., 1997), which will activate in response to Na⁺ entry, leading to membrane hyperpolarization due to electrogenic pump activity. Recent pharmacological studies in vitro have proposed that axonal expression of HCN channels can counteract the hyperpolarizing Na⁺ pump current during repetitive firing, allowing for fast AP firing and propagation (Roth and Hu, 2020). Thus, blockade of HCN channel activity in PV+ interneuron axon terminals results in a failure in AP propagation and reduced perisomatic inhibition onto target excitatory neurons (Southan et al., 2000; Aponte et al., 2006; Roth and Hu, 2020). Since the only HCN subtype expressed in PV+ interneuron terminals is HCN1, diminished GABAergic input onto excitatory neurons as a result of HCN1 protein dysfunction could contribute to excitation/inhibition imbalance and epilepsy. Such an effect could be exacerbated by further inhibition of Na⁺ channel activity by antiseizure medications, similar to what is thought to occur in Dravet syndrome. HCN1 subunits also substantially contribute to regulating the excitability of a second class of inhibitory neurons, namely, somatostatin-positive interneurons, which target the dendrites of pyramidal cells (Matt et al., 2011). Here, HCN1 acts along with HCN2 to set the somatic resting potential and facilitate the spontaneous activity of this class of interneurons.

While our histological analysis provides some evidence as to the deleterious effects of the HCN1-GD variant on axonal function, there are some important limitations to the interpretation of this data in the absence of a direct physiological evaluation of the intrinsic and synaptic properties of PV+ interneurons. Future studies aimed at investigating the physiological effects of HCN1 mutations in genetically identified interneuron populations, along with conditional knock-in mouse lines that limit Hcn1^GD or Hcn1^MI allele expression to select classes of inhibitory neurons, will provide a more definitive answer to the question of how different variants may affect perisomatic and/or dendritic inhibition in HCN1-linked epilepsy.

When tested in heterologous expression systems, epilepsy-linked HCN1 mutations have been shown to affect the channel in several different ways. These actions include loss of function due to decreased HCN1 channel activation, decreased protein levels, and/or impaired targeting to the plasma membrane (Porro et al., 2021). Some of the variants exert a dominant negative effect when a mutant subunit assembles into a heteromeric channel with WT subunits. Conversely, gain-of-function effects include accelerated activation kinetics, depolarized midpoint voltage of activation (as in M153I), and an increased voltage-independent component of the HCN current, that is, generation of a depolarizing ‘leak’ current (as in G391D). Why do HCN1 mutations that in heterologous expression systems have divergent effects on channel function equally result in epilepsy, and a similar paradoxical response to certain Na⁺ channel blockers? Our results in Hcn1^GD/+ and Hcn1^MI/+ mice show that both these variants ultimately led to a net decrease in I_h-dependent voltage sag, and a significant decrease in protein expression levels (Figure 5), despite the p.M153I variant yielding an apparent gain-of-function channel upon heterologous expression (Marini et al., 2018). This suggests that misfolding and impaired biosynthesis of mutant HCN1 subunits may be a prevalent consequence likely to affect the function of neurons in multiple ways. An increase in cellular stress, for example, would be expected to affect PV+ neurons in particular, as they are known to be extremely sensitive to metabolic and oxidative stress (Kann, 2016). Such a mechanism may explain why two variants with seemingly divergent biophysical alterations in channel functional properties ultimately cause overlapping phenotypes.

The divergent characteristics of the two variants do, however, lead to distinguishable phenotypes. This is demonstrated by the overall greater severity of the Hcn1^GD/+ phenotype (reduced brain size, altered gait, impaired pole climbing, lack of spontaneous alternation, and greater reductions in HCN1 protein and AP firing), as well as the remarkable sex dependence of mortality in Hcn1^MI/+ mice. This observation suggests that, while some fundamental mechanisms are affected in similar ways, there are distinctions in the manner in which physiological networks are perturbed by different HCN1 variants. Moreover, similar to other ion channel-linked syndromes, a variety of disease phenotypes are observed in patients carrying HCN1 pathogenic variants, with DEEs seen in ~20% of cases (Marini et al., 2018). Overall, if both familial and sporadic patients are considered, the predominant phenotypes are milder and include genetic generalized epilepsies (Bonzanni et al., 2018), as well as the spectrum of genetic epilepsy with febrile seizures plus (GEFS+) (Marini et al., 2018). In addition, some variants do not show full penetrance. An analogous picture emerges from the study of a third HCN1 variant recently modeled in mice (Hcn1^M294L, corresponding to human p.Met305Leu; Bleakley et al., 2021). Similar to HCN1-G391D, this variant generates a channel with a significant ‘leak’ or voltage-independent current component, although protein levels in the Hcn1^M294L mouse brain are less dramatically affected. The epileptic phenotype of Hcn1^M294L mice is considerably milder than either Hcn1^GD/+ or Hcn1^MI/+ animals. Intriguingly, however, Hcn1^M294L mice are also adversely affected by LTG, which caused the induction of seizures, similar to heterozygous animals from the two lines presented here. From the therapeutic point of view, the observed mechanistic convergence does offer some hope that patients with genetic HCN1 channel dysfunction can be treated following similar principles or strategies across different variants. A larger convergence in treatment strategies, centered on the degree of interneuron dysfunction across multiple epilepsies, may also be envisioned to emerge in the future.

Finally, what may such strategies look like? Our data once again stress the difficulty of devising targeted therapeutic approaches for ion channel-associated DEE syndromes using existing tools. As stated above, several of the identified epilepsy-associated HCN1 variants generate channels with what have been deemed ‘gain of aberrant’ function properties (Bleakley et al., 2021), implying that HCN channel blockers may be potentially helpful. At the same time, several of the most severe pathogenic HCN1 variants are located in the S6 segment, which forms the lining of the pore cavity, or in the immediate vicinity to the channel’s intracellular gate. These include G391D (Figure 10) and two more variants at the same position, G391C and G391S, as well as M379R, I380F, A387S, F389S, I397L, S399P, and D401H (Nava et al., 2014; Lucariello et al., 2016; Marini et al., 2018; Wang et al., 2019). While we do not have complete information available about the effects of known HCN inhibitors on the other variants, results presented here using ZD7288 on G391D-containing HCN1 channels strongly suggest that the mutations may decrease the efficacy of the pore-blocking compounds, reducing their ability to suppress both the time-dependent and time-independent current component generated by the mutant channels (Marini et al., 2018; Porro et al., 2021). There is therefore a need to develop new small-molecule compounds that target the HCN1 channels at sites outside the pore cavity.

The available 3D structures of HCN1, combined with molecular dynamics simulations, coarse-grained modeling, and molecular docking, suggest that there are many opportunities for such a pharmacological-targeting approach to succeed (Lee and MacKinnon, 2017; Gross et al., 2018; Porro et al., 2019). The structures have indeed revealed unique intracellular regulatory modules, including both the HCN domain and C-linker/CNBD, which may allow for the targeting of HCN channels through allosteric modulation of pore gating properties. Furthermore, the recently determined structures of HCN4, the main cardiac isoform of HCN channels, have demonstrated differences in the relative arrangement of these regulatory modules (Saponaro et al., 2021), which may be exploited towards the development of subtype-specific drugs. Compound library screenings for small molecules with HCN1-isoform specificity have already yielded some initial and promising results (McClure et al., 2011; Harde et al., 2019). Other approaches include the design of cell-penetrant peptides, which interfere with the interaction between HCN channels and their regulatory auxiliary subunit, TRIP8b (Han et al., 2015; Saponaro et al., 2018), as well as antisense oligonucleotide-based strategies aimed at HCN1 mRNA downregulation. Of note, the latter strategy may be particularly effective in light of the hypothesis that protein misfolding may be an important component of the pathology in HCN1-linked DEE, together with the observation that Hcn1 null mice do not have spontaneous seizures (Huang et al., 2009; Santoro et al., 2010).

In conclusion, the availability of strong preclinical models, such as the Hcn1^GD/+ and Hcn1^MI/+ mouse lines presented here, provides an ideal platform for further investigation of mechanisms underlying epileptic encephalopathies and testing new and paradigm-shifting therapeutic strategies.

All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for Figures 2, 3, 5, 9 and 10. Microarray expression data have been deposited in GEO under accession code GSE209630.

1. 1. Merseburg A
   2. Stockebrand M
   3. Isbrandt D
   4. Santoro B

   (2022) NCBI Gene Expression Omnibus

   ID GSE209630. Expression data from the hippocampus of epileptic mice carrying two different pathogenic de novo mutations in the hyperpolarization-activated, cyclic nucleotide-gated, non-selective cation channel gene Hcn1 in comparison to wildtype controls.

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE209630

1. 1. Aponte Y
   2. Lien CC
   3. Reisinger E
   4. Jonas P

   (2006) Hyperpolarization-activated cation channels in fast-spiking interneurons of rat hippocampus

   The Journal of Physiology 574:229–243.

   https://doi.org/10.1113/jphysiol.2005.104042

   • PubMed
   • Google Scholar
2. 1. Barker GRI
   2. Warburton EC

   (2011) When is the hippocampus involved in recognition memory?

   The Journal of Neuroscience 31:10721–10731.

   https://doi.org/10.1523/JNEUROSCI.6413-10.2011

   • PubMed
   • Google Scholar
3. 1. Bender RA
   2. Baram TZ

   (2008) Hyperpolarization activated cyclic-nucleotide gated (HCN) channels in developing neuronal networks

   Progress in Neurobiology 86:129–140.

   https://doi.org/10.1016/j.pneurobio.2008.09.007

   • PubMed
   • Google Scholar
4. 1. Bleakley LE
   2. McKenzie CE
   3. Soh MS
   4. Forster IC
   5. Pinares-Garcia P
   6. Sedo A
   7. Kathirvel A
   8. Churilov L
   9. Jancovski N
   10. Maljevic S
   11. Berkovic SF
   12. Scheffer IE
   13. Petrou S
   14. Santoro B
   15. Reid CA

   (2021) Cation leak underlies neuronal excitability in an HCN1 developmental and epileptic encephalopathy

   Brain : A Journal of Neurology 144:2060–2073.

   https://doi.org/10.1093/brain/awab145

   • PubMed
   • Google Scholar
5. 1. Bonzanni M
   2. DiFrancesco JC
   3. Milanesi R
   4. Campostrini G
   5. Castellotti B
   6. Bucchi A
   7. Baruscotti M
   8. Ferrarese C
   9. Franceschetti S
   10. Canafoglia L
   11. Ragona F
   12. Freri E
   13. Labate A
   14. Gambardella A
   15. Costa C
   16. Rivolta I
   17. Gellera C
   18. Granata T
   19. Barbuti A
   20. DiFrancesco D

   (2018) A novel de novo HCN1 loss-of-function mutation in genetic generalized epilepsy causing increased neuronal excitability

   Neurobiology of Disease 118:55–63.

   https://doi.org/10.1016/j.nbd.2018.06.012

   • PubMed
   • Google Scholar
6. 1. Brennan GP
   2. Baram TZ
   3. Poolos NP

   (2016) Hyperpolarization-activated cyclic nucleotide-gated (HCN) channels in epilepsy

   Cold Spring Harbor Perspectives in Medicine 6:a022384.

   https://doi.org/10.1101/cshperspect.a022384

   • PubMed
   • Google Scholar
7. 1. Butler KM
   2. da Silva C
   3. Alexander JJ
   4. Hegde M
   5. Escayg A

   (2017) Diagnostic yield from 339 epilepsy patients screened on a clinical gene panel

   Pediatric Neurology 77:61–66.

   https://doi.org/10.1016/j.pediatrneurol.2017.09.003

   • PubMed
   • Google Scholar
8. 1. Catterall WA

   (2018) Dravet syndrome: A sodium channel interneuronopathy

   Current Opinion in Physiology 2:42–50.

   https://doi.org/10.1016/j.cophys.2017.12.007

   • PubMed
   • Google Scholar
9. 1. Cheng L
   2. Kinard K
   3. Rajamani R
   4. Sanguinetti MC

   (2007) Molecular mapping of the binding site for a blocker of hyperpolarization-activated, cyclic nucleotide-modulated pacemaker channels

   The Journal of Pharmacology and Experimental Therapeutics 322:931–939.

   https://doi.org/10.1124/jpet.107.121467

   • PubMed
   • Google Scholar
10. 1. Del Lungo M
    2. Melchiorre M
    3. Guandalini L
    4. Sartiani L
    5. Mugelli A
    6. Koncz I
    7. Szel T
    8. Varro A
    9. Romanelli MN
    10. Cerbai E

    (2012) Novel blockers of hyperpolarization-activated current with isoform selectivity in recombinant cells and native tissue

    British Journal of Pharmacology 166:602–616.

    https://doi.org/10.1111/j.1476-5381.2011.01782.x

    • PubMed
    • Google Scholar
11. 1. DiFrancesco D

    (1982) Block and activation of the pace-maker channel in calf purkinje fibres: effects of potassium, caesium and rubidium

    The Journal of Physiology 329:485–507.

    https://doi.org/10.1113/jphysiol.1982.sp014315

    • PubMed
    • Google Scholar
12. 1. Dutton SB
    2. Makinson CD
    3. Papale LA
    4. Shankar A
    5. Balakrishnan B
    6. Nakazawa K
    7. Escayg A

    (2013) Preferential inactivation of scn1a in parvalbumin interneurons increases seizure susceptibility

    Neurobiology of Disease 49:211–220.

    https://doi.org/10.1016/j.nbd.2012.08.012

    • PubMed
    • Google Scholar
13. 1. Escartin C
    2. Galea E
    3. Lakatos A
    4. O’Callaghan JP
    5. Petzold GC
    6. Serrano-Pozo A
    7. Steinhäuser C
    8. Volterra A
    9. Carmignoto G
    10. Agarwal A
    11. Allen NJ
    12. Araque A
    13. Barbeito L
    14. Barzilai A
    15. Bergles DE
    16. Bonvento G
    17. Butt AM
    18. Chen W-T
    19. Cohen-Salmon M
    20. Cunningham C
    21. Deneen B
    22. De Strooper B
    23. Díaz-Castro B
    24. Farina C
    25. Freeman M
    26. Gallo V
    27. Goldman JE
    28. Goldman SA
    29. Götz M
    30. Gutiérrez A
    31. Haydon PG
    32. Heiland DH
    33. Hol EM
    34. Holt MG
    35. Iino M
    36. Kastanenka KV
    37. Kettenmann H
    38. Khakh BS
    39. Koizumi S
    40. Lee CJ
    41. Liddelow SA
    42. MacVicar BA
    43. Magistretti P
    44. Messing A
    45. Mishra A
    46. Molofsky AV
    47. Murai KK
    48. Norris CM
    49. Okada S
    50. Oliet SHR
    51. Oliveira JF
    52. Panatier A
    53. Parpura V
    54. Pekna M
    55. Pekny M
    56. Pellerin L
    57. Perea G
    58. Pérez-Nievas BG
    59. Pfrieger FW
    60. Poskanzer KE
    61. Quintana FJ
    62. Ransohoff RM
    63. Riquelme-Perez M
    64. Robel S
    65. Rose CR
    66. Rothstein JD
    67. Rouach N
    68. Rowitch DH
    69. Semyanov A
    70. Sirko S
    71. Sontheimer H
    72. Swanson RA
    73. Vitorica J
    74. Wanner I-B
    75. Wood LB
    76. Wu J
    77. Zheng B
    78. Zimmer ER
    79. Zorec R
    80. Sofroniew MV
    81. Verkhratsky A

    (2021) Reactive astrocyte nomenclature, definitions, and future directions

    Nature Neuroscience 24:312–325.

    https://doi.org/10.1038/s41593-020-00783-4

    • PubMed
    • Google Scholar
14. 1. Fernández-Marmiesse A
    2. Roca I
    3. Díaz-Flores F
    4. Cantarín V
    5. Pérez-Poyato MS
    6. Fontalba A
    7. Laranjeira F
    8. Quintans S
    9. Moldovan O
    10. Felgueroso B
    11. Rodríguez-Pedreira M
    12. Simón R
    13. Camacho A
    14. Quijada P
    15. Ibanez-Mico S
    16. Domingno MR
    17. Benito C
    18. Calvo R
    19. Pérez-Cejas A
    20. Carrasco ML
    21. Ramos F
    22. Couce ML
    23. Ruiz-Falcó ML
    24. Gutierrez-Solana L
    25. Martínez-Atienza M

    (2019) Rare variants in 48 genes account for 42% of cases of epilepsy with or without neurodevelopmental delay in 246 pediatric patients

    Frontiers in Neuroscience 13:1135.

    https://doi.org/10.3389/fnins.2019.01135

    • PubMed
    • Google Scholar
15. 1. Festing MF

    (2018) On determining sample size in experiments involving laboratory animals

    Laboratory Animals 52:341–350.

    https://doi.org/10.1177/0023677217738268

    • PubMed
    • Google Scholar
16. 1. Freitag S
    2. Schachner M
    3. Morellini F

    (2003) Behavioral alterations in mice deficient for the extracellular matrix glycoprotein tenascin-R

    Behavioural Brain Research 145:189–207.

    https://doi.org/10.1016/s0166-4328(03)00109-8

    • PubMed
    • Google Scholar
17. 1. Gasparini S
    2. DiFrancesco D

    (1997) Action of the hyperpolarization-activated current (ih) blocker ZD 7288 in hippocampal CA1 neurons

    Pflugers Archiv 435:99–106.

    https://doi.org/10.1007/s004240050488

    • PubMed
    • Google Scholar
18. 1. George MS
    2. Abbott LF
    3. Siegelbaum SA

    (2009) HCN hyperpolarization-activated cation channels inhibit epsps by interactions with M-type K(+) channels

    Nature Neuroscience 12:577–584.

    https://doi.org/10.1038/nn.2307

    • PubMed
    • Google Scholar
19. 1. Gross C
    2. Saponaro A
    3. Santoro B
    4. Moroni A
    5. Thiel G
    6. Hamacher K

    (2018) Mechanical transduction of cytoplasmic-to-transmembrane-domain movements in a hyperpolarization-activated cyclic nucleotide-gated cation channel

    The Journal of Biological Chemistry 293:12908–12918.

    https://doi.org/10.1074/jbc.RA118.002139

    • PubMed
    • Google Scholar
20. 1. Haettig J
    2. Stefanko DP
    3. Multani ML
    4. Figueroa DX
    5. McQuown SC
    6. Wood MA

    (2011) HDAC inhibition modulates hippocampus-dependent long-term memory for object location in a CBP-dependent manner

    Learning & Memory 18:71–79.

    https://doi.org/10.1101/lm.1986911

    • PubMed
    • Google Scholar
21. 1. Haeussler M
    2. Schönig K
    3. Eckert H
    4. Eschstruth A
    5. Mianné J
    6. Renaud JB
    7. Schneider-Maunoury S
    8. Shkumatava A
    9. Teboul L
    10. Kent J
    11. Joly JS
    12. Concordet JP

    (2016) Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA selection tool CRISPOR

    Genome Biology 17:148.

    https://doi.org/10.1186/s13059-016-1012-2

    • PubMed
    • Google Scholar
22. 1. Han Y
    2. Lyman K
    3. Clutter M
    4. Schiltz GE
    5. Ismail Q-A
    6. Prados DB
    7. Luan C-H
    8. Chetkovich DM

    (2015) Identification of small-molecule inhibitors of hyperpolarization-activated cyclic nucleotide-gated channels

    Journal of Biomolecular Screening 20:1124–1131.

    https://doi.org/10.1177/1087057115589590

    • PubMed
    • Google Scholar
23. Conference

    1. Harde E
    2. Weber M
    3. Wellbourne-Wood J
    4. Schoenenberger P
    5. Paina M
    6. Rolland JF
    7. Chernysheva M
    8. Koc B
    9. Redondo R
    10. Wamsteeker Cusulin J
    11. Helmchen F
    12. O’Connor E
    13. Hall BJ

    (2019)

    Linking selective pharmacological suppression of HCN1 to synaptic integration, prefrontal cortical function and working memory

    Presented at the Society for Neuroscience.

    • Google Scholar
24. 1. Hawkins NA
    2. Anderson LL
    3. Gertler TS
    4. Laux L
    5. George AL
    6. Kearney JA

    (2017) Screening of conventional anticonvulsants in a genetic mouse model of epilepsy

    Annals of Clinical and Translational Neurology 4:326–339.

    https://doi.org/10.1002/acn3.413

    • PubMed
    • Google Scholar
25. 1. Hedrich UBS
    2. Liautard C
    3. Kirschenbaum D
    4. Pofahl M
    5. Lavigne J
    6. Liu Y
    7. Theiss S
    8. Slotta J
    9. Escayg A
    10. Dihné M
    11. Beck H
    12. Mantegazza M
    13. Lerche H

    (2014) Impaired action potential initiation in gabaergic interneurons causes hyperexcitable networks in an epileptic mouse model carrying a human na(V)1.1 mutation

    The Journal of Neuroscience 34:14874–14889.

    https://doi.org/10.1523/JNEUROSCI.0721-14.2014

    • PubMed
    • Google Scholar
26. 1. Houser CR

    (1990) Granule cell dispersion in the dentate gyrus of humans with temporal lobe epilepsy

    Brain Research 535:195–204.

    https://doi.org/10.1016/0006-8993(90)91601-c

    • PubMed
    • Google Scholar
27. 1. Hu H
    2. Jonas P

    (2014) A supercritical density of na(+) channels ensures fast signaling in gabaergic interneuron axons

    Nature Neuroscience 17:686–693.

    https://doi.org/10.1038/nn.3678

    • PubMed
    • Google Scholar
28. 1. Huang Z
    2. Walker MC
    3. Shah MM

    (2009) Loss of dendritic HCN1 subunits enhances cortical excitability and epileptogenesis

    The Journal of Neuroscience 29:10979–10988.

    https://doi.org/10.1523/JNEUROSCI.1531-09.2009

    • PubMed
    • Google Scholar
29. 1. Huang YY
    2. Liu YC
    3. Lee CT
    4. Lin YC
    5. Wang ML
    6. Yang YP
    7. Chang KY
    8. Chiou SH

    (2016) Revisiting the lamotrigine-mediated effect on hippocampal gabaergic transmission

    International Journal of Molecular Sciences 17:E1191.

    https://doi.org/10.3390/ijms17071191

    • PubMed
    • Google Scholar
30. 1. Jung S
    2. Warner LN
    3. Pitsch J
    4. Becker AJ
    5. Poolos NP

    (2011) Rapid loss of dendritic HCN channel expression in hippocampal pyramidal neurons following status epilepticus

    The Journal of Neuroscience 31:14291–14295.

    https://doi.org/10.1523/JNEUROSCI.1148-11.2011

    • PubMed
    • Google Scholar
31. 1. Kann O

    (2016) The interneuron energy hypothesis: implications for brain disease

    Neurobiology of Disease 90:75–85.

    https://doi.org/10.1016/j.nbd.2015.08.005

    • PubMed
    • Google Scholar
32. 1. Kitanaka J
    2. Kitanaka N
    3. Hall FS
    4. Fujii M
    5. Goto A
    6. Kanda Y
    7. Koizumi A
    8. Kuroiwa H
    9. Mibayashi S
    10. Muranishi Y
    11. Otaki S
    12. Sumikawa M
    13. Tanaka KI
    14. Nishiyama N
    15. Uhl GR
    16. Takemura M

    (2015) Memory impairment and reduced exploratory behavior in mice after administration of systemic morphine

    Journal of Experimental Neuroscience 9:27–35.

    https://doi.org/10.4137/JEN.S25057

    • PubMed
    • Google Scholar
33. 1. Kole MJ
    2. Qian J
    3. Waase MP
    4. Klassen TL
    5. Chen TT
    6. Augustine GJ
    7. Noebels JL

    (2015) Selective loss of presynaptic potassium channel clusters at the cerebellar basket cell terminal pinceau in adam11 mutants reveals their role in ephaptic control of purkinje cell firing

    The Journal of Neuroscience 35:11433–11444.

    https://doi.org/10.1523/JNEUROSCI.1346-15.2015

    • PubMed
    • Google Scholar
34. 1. Lalonde R

    (2002) The neurobiological basis of spontaneous alternation

    Neuroscience and Biobehavioral Reviews 26:91–104.

    https://doi.org/10.1016/s0149-7634(01)00041-0

    • PubMed
    • Google Scholar
35. 1. Lee CH
    2. MacKinnon R

    (2017) Structures of the human HCN1 hyperpolarization-activated channel

    Cell 168:111–120.

    https://doi.org/10.1016/j.cell.2016.12.023

    • PubMed
    • Google Scholar
36. 1. Lewis AS
    2. Schwartz E
    3. Chan CS
    4. Noam Y
    5. Shin M
    6. Wadman WJ
    7. Surmeier DJ
    8. Baram TZ
    9. Macdonald RL
    10. Chetkovich DM

    (2009) Alternatively spliced isoforms of trip8b differentially control h channel trafficking and function

    The Journal of Neuroscience 29:6250–6265.

    https://doi.org/10.1523/JNEUROSCI.0856-09.2009

    • PubMed
    • Google Scholar
37. 1. Lörincz A
    2. Notomi T
    3. Tamás G
    4. Shigemoto R
    5. Nusser Z

    (2002) Polarized and compartment-dependent distribution of HCN1 in pyramidal cell dendrites

    Nature Neuroscience 5:1185–1193.

    https://doi.org/10.1038/nn962

    • PubMed
    • Google Scholar
38. 1. Lucariello M
    2. Vidal E
    3. Vidal S
    4. Saez M
    5. Roa L
    6. Huertas D
    7. Pineda M
    8. Dalfó E
    9. Dopazo J
    10. Jurado P
    11. Armstrong J
    12. Esteller M

    (2016) Whole exome sequencing of rett syndrome-like patients reveals the mutational diversity of the clinical phenotype

    Human Genetics 135:1343–1354.

    https://doi.org/10.1007/s00439-016-1721-3

    • PubMed
    • Google Scholar
39. 1. Luján R
    2. Albasanz JL
    3. Shigemoto R
    4. Juiz JM

    (2005) Preferential localization of the hyperpolarization-activated cyclic nucleotide-gated cation channel subunit HCN1 in basket cell terminals of the rat cerebellum

    The European Journal of Neuroscience 21:2073–2082.

    https://doi.org/10.1111/j.1460-9568.2005.04043.x

    • PubMed
    • Google Scholar
40. 1. Magagna-Poveda A
    2. Moretto JN
    3. Scharfman HE

    (2017) Increased gyrification and aberrant adult neurogenesis of the dentate gyrus in adult rats

    Brain Structure & Function 222:4219–4237.

    https://doi.org/10.1007/s00429-017-1457-4

    • PubMed
    • Google Scholar
41. 1. Marini C
    2. Porro A
    3. Rastetter A
    4. Dalle C
    5. Rivolta I
    6. Bauer D
    7. Oegema R
    8. Nava C
    9. Parrini E
    10. Mei D
    11. Mercer C
    12. Dhamija R
    13. Chambers C
    14. Coubes C
    15. Thévenon J
    16. Kuentz P
    17. Julia S
    18. Pasquier L
    19. Dubourg C
    20. Carré W
    21. Rosati A
    22. Melani F
    23. Pisano T
    24. Giardino M
    25. Innes AM
    26. Alembik Y
    27. Scheidecker S
    28. Santos M
    29. Figueiroa S
    30. Garrido C
    31. Fusco C
    32. Frattini D
    33. Spagnoli C
    34. Binda A
    35. Granata T
    36. Ragona F
    37. Freri E
    38. Franceschetti S
    39. Canafoglia L
    40. Castellotti B
    41. Gellera C
    42. Milanesi R
    43. Mancardi MM
    44. Clark DR
    45. Kok F
    46. Helbig KL
    47. Ichikawa S
    48. Sadler L
    49. Neupauerová J
    50. Laššuthova P
    51. Šterbová K
    52. Laridon A
    53. Brilstra E
    54. Koeleman B
    55. Lemke JR
    56. Zara F
    57. Striano P
    58. Soblet J
    59. Smits G
    60. Deconinck N
    61. Barbuti A
    62. DiFrancesco D
    63. LeGuern E
    64. Guerrini R
    65. Santoro B
    66. Hamacher K
    67. Thiel G
    68. Moroni A
    69. DiFrancesco JC
    70. Depienne C

    (2018) HCN1 mutation spectrum: from neonatal epileptic encephalopathy to benign generalized epilepsy and beyond

    Brain 141:3160–3178.

    https://doi.org/10.1093/brain/awy263

    • PubMed
    • Google Scholar
42. 1. Markowitz GJ
    2. Kadam SD
    3. Boothe DM
    4. Irving ND
    5. Comi AM

    (2010) The pharmacokinetics of commonly used antiepileptic drugs in immature CD1 mice

    Neuroreport 21:452–456.

    https://doi.org/10.1097/wnr.0b013e328338ba18

    • PubMed
    • Google Scholar
43. 1. Marksteiner J
    2. Ortler M
    3. Bellmann R
    4. Sperk G

    (1990) Neuropeptide Y biosynthesis is markedly induced in mossy fibers during temporal lobe epilepsy of the rat

    Neuroscience Letters 112:143–148.

    https://doi.org/10.1016/0304-3940(90)90193-d

    • PubMed
    • Google Scholar
44. 1. Massella A
    2. Gusciglio M
    3. D’Intino G
    4. Sivilia S
    5. Ferraro L
    6. Calzà L
    7. Giardino L

    (2009) Gabapentin treatment improves motor coordination in a mice model of progressive ataxia

    Brain Research 1301:135–142.

    https://doi.org/10.1016/j.brainres.2009.09.004

    • PubMed
    • Google Scholar
45. 1. Matt L
    2. Michalakis S
    3. Hofmann F
    4. Hammelmann V
    5. Ludwig A
    6. Biel M
    7. Kleppisch T

    (2011) HCN2 channels in local inhibitory interneurons constrain LTP in the hippocampal direct perforant path

    Cellular and Molecular Life Sciences 68:125–137.

    https://doi.org/10.1007/s00018-010-0446-z

    • PubMed
    • Google Scholar
46. 1. McClelland S
    2. Flynn C
    3. Dubé C
    4. Richichi C
    5. Zha Q
    6. Ghestem A
    7. Esclapez M
    8. Bernard C
    9. Baram TZ

    (2011) Neuron-restrictive silencer factor-mediated hyperpolarization-activated cyclic nucleotide gated channelopathy in experimental temporal lobe epilepsy

    Annals of Neurology 70:454–464.

    https://doi.org/10.1002/ana.22479

    • PubMed
    • Google Scholar
47. 1. McClure KJ
    2. Maher M
    3. Wu N
    4. Chaplan SR
    5. Eckert WA
    6. Lee DH
    7. Wickenden AD
    8. Hermann M
    9. Allison B
    10. Hawryluk N
    11. Breitenbucher JG
    12. Grice CA

    (2011) Discovery of a novel series of selective HCN1 blockers

    Bioorganic & Medicinal Chemistry Letters 21:5197–5201.

    https://doi.org/10.1016/j.bmcl.2011.07.051

    • PubMed
    • Google Scholar
48. 1. Melchiorre M
    2. Del Lungo M
    3. Guandalini L
    4. Martini E
    5. Dei S
    6. Manetti D
    7. Scapecchi S
    8. Teodori E
    9. Sartiani L
    10. Mugelli A
    11. Cerbai E
    12. Romanelli MN

    (2010) Design, synthesis, and preliminary biological evaluation of new isoform-selective f-current blockers

    Journal of Medicinal Chemistry 53:6773–6777.

    https://doi.org/10.1021/jm1006758

    • PubMed
    • Google Scholar
49. 1. Nava C
    2. Dalle C
    3. Rastetter A
    4. Striano P
    5. Kovel CGF
    6. Nabbout R
    7. Cancès C
    8. Ville D
    9. Brilstra EH
    10. Gobbi G
    11. Raffo E
    12. Bouteiller D
    13. Marie Y
    14. Trouillard O
    15. Robbiano A
    16. Keren B
    17. Agher D
    18. Roze E
    19. Lesage S
    20. Nicolas A
    21. Brice A
    22. Baulac M
    23. Vogt C
    24. El Hajj N
    25. Schneider E
    26. Suls A
    27. Weckhuysen S
    28. Gormley P
    29. Lehesjoki AE

    (2014) De novo mutations in HCN1 cause early infantile epileptic encephalopathy

    Nature Genetics 46:640–645.

    https://doi.org/10.1038/ng.2952

    • Google Scholar
50. 1. Noebels J

    (2017) Precision physiology and rescue of brain ion channel disorders

    The Journal of General Physiology 149:533–546.

    https://doi.org/10.1085/jgp.201711759

    • PubMed
    • Google Scholar
51. 1. Nolan MF
    2. Malleret G
    3. Lee KH
    4. Gibbs E
    5. Dudman JT
    6. Santoro B
    7. Yin D
    8. Thompson RF
    9. Siegelbaum SA
    10. Kandel ER
    11. Morozov A

    (2003) The hyperpolarization-activated HCN1 channel is important for motor learning and neuronal integration by cerebellar purkinje cells

    Cell 115:551–564.

    https://doi.org/10.1016/s0092-8674(03)00884-5

    • PubMed
    • Google Scholar
52. 1. Nolan MF
    2. Malleret G
    3. Dudman JT
    4. Buhl DL
    5. Santoro B
    6. Gibbs E
    7. Vronskaya S
    8. Buzsáki G
    9. Siegelbaum SA
    10. Kandel ER
    11. Morozov A

    (2004) A behavioral role for dendritic integration: HCN1 channels constrain spatial memory and plasticity at inputs to distal dendrites of CA1 pyramidal neurons

    Cell 119:719–732.

    https://doi.org/10.1016/j.cell.2004.11.020

    • PubMed
    • Google Scholar
53. 1. Notomi T
    2. Shigemoto R

    (2004) Immunohistochemical localization of ih channel subunits, HCN1-4, in the rat brain

    The Journal of Comparative Neurology 471:241–276.

    https://doi.org/10.1002/cne.11039

    • PubMed
    • Google Scholar
54. 1. Ogiwara I
    2. Miyamoto H
    3. Morita N
    4. Atapour N
    5. Mazaki E
    6. Inoue I
    7. Takeuchi T
    8. Itohara S
    9. Yanagawa Y
    10. Obata K
    11. Furuichi T
    12. Hensch TK
    13. Yamakawa K

    (2007) Nav1.1 localizes to axons of parvalbumin-positive inhibitory interneurons: a circuit basis for epileptic seizures in mice carrying an scn1a gene mutation

    The Journal of Neuroscience 27:5903–5914.

    https://doi.org/10.1523/JNEUROSCI.5270-06.2007

    • PubMed
    • Google Scholar
55. 1. Ogiwara I
    2. Iwasato T
    3. Miyamoto H
    4. Iwata R
    5. Yamagata T
    6. Mazaki E
    7. Yanagawa Y
    8. Tamamaki N
    9. Hensch TK
    10. Itohara S
    11. Yamakawa K

    (2013) Nav1.1 haploinsufficiency in excitatory neurons ameliorates seizure-associated sudden death in a mouse model of dravet syndrome

    Human Molecular Genetics 22:4784–4804.

    https://doi.org/10.1093/hmg/ddt331

    • PubMed
    • Google Scholar
56. 1. Oyrer J
    2. Maljevic S
    3. Scheffer IE
    4. Berkovic SF
    5. Petrou S
    6. Reid CA

    (2018) Ion channels in genetic epilepsy: from genes and mechanisms to disease-targeted therapies

    Pharmacological Reviews 70:142–173.

    https://doi.org/10.1124/pr.117.014456

    • PubMed
    • Google Scholar
57. 1. Parrini E
    2. Marini C
    3. Mei D
    4. Galuppi A
    5. Cellini E
    6. Pucatti D
    7. Chiti L
    8. Rutigliano D
    9. Bianchini C
    10. Virdò S
    11. De Vita D
    12. Bigoni S
    13. Barba C
    14. Mari F
    15. Montomoli M
    16. Pisano T
    17. Rosati A
    18. Group CS
    19. Guerrini R

    (2017) Diagnostic targeted resequencing in 349 patients with drug-resistant pediatric epilepsies identifies causative mutations in 30 different genes

    Human Mutation 38:216–225.

    https://doi.org/10.1002/humu.23149

    • PubMed
    • Google Scholar
58. 1. Peng L
    2. Martin-Vasallo P
    3. Sweadner KJ

    (1997)

    Isoforms of na,K-atpase alpha and beta subunits in the rat cerebellum and in granule cell cultures

    The Journal of Neuroscience 17:3488–3502.

    • PubMed
    • Google Scholar
59. 1. Peng BW
    2. Justice JA
    3. Zhang K
    4. He XH
    5. Sanchez RM

    (2010) Increased basal synaptic inhibition of hippocampal area CA1 pyramidal neurons by an antiepileptic drug that enhances I(H)

    Neuropsychopharmacology 35:464–472.

    https://doi.org/10.1038/npp.2009.150

    • PubMed
    • Google Scholar
60. 1. Perucca P
    2. Perucca E

    (2019) Identifying mutations in epilepsy genes: impact on treatment selection

    Epilepsy Research 152:18–30.

    https://doi.org/10.1016/j.eplepsyres.2019.03.001

    • PubMed
    • Google Scholar
61. 1. Poolos NP
    2. Migliore M
    3. Johnston D

    (2002) Pharmacological upregulation of h-channels reduces the excitability of pyramidal neuron dendrites

    Nature Neuroscience 5:767–774.

    https://doi.org/10.1038/nn891

    • PubMed
    • Google Scholar
62. 1. Porro A
    2. Saponaro A
    3. Gasparri F
    4. Bauer D
    5. Gross C
    6. Pisoni M
    7. Abbandonato G
    8. Hamacher K
    9. Santoro B
    10. Thiel G
    11. Moroni A

    (2019) The HCN domain couples voltage gating and camp response in hyperpolarization-activated cyclic nucleotide-gated channels

    eLife 8:e49672.

    https://doi.org/10.7554/eLife.49672

    • Google Scholar
63. 1. Porro A
    2. Abbandonato G
    3. Veronesi V
    4. Russo A
    5. Binda A
    6. Antolini L
    7. Granata T
    8. Castellotti B
    9. Marini C
    10. Moroni A
    11. DiFrancesco JC
    12. Rivolta I

    (2021) Do the functional properties of HCN1 mutants correlate with the clinical features in epileptic patients?

    Progress in Biophysics and Molecular Biology 166:147–155.

    https://doi.org/10.1016/j.pbiomolbio.2021.07.008

    • PubMed
    • Google Scholar
64. 1. Qiao X
    2. Sun G
    3. Clare JJ
    4. Werkman TR
    5. Wadman WJ

    (2014) Properties of human brain sodium channel α-subunits expressed in HEK293 cells and their modulation by carbamazepine, phenytoin and lamotrigine

    British Journal of Pharmacology 171:1054–1067.

    https://doi.org/10.1111/bph.12534

    • PubMed
    • Google Scholar
65. 1. Rinaldi A
    2. Defterali C
    3. Mialot A
    4. Garden DLF
    5. Beraneck M
    6. Nolan MF

    (2013) HCN1 channels in cerebellar purkinje cells promote late stages of learning and constrain synaptic inhibition

    The Journal of Physiology 591:5691–5709.

    https://doi.org/10.1113/jphysiol.2013.259499

    • PubMed
    • Google Scholar
66. 1. Roth FC
    2. Hu H

    (2020) An axon-specific expression of HCN channels catalyzes fast action potential signaling in gabaergic interneurons

    Nature Communications 11:2248.

    https://doi.org/10.1038/s41467-020-15791-y

    • Google Scholar
67. 1. Santoro B
    2. Piskorowski RA
    3. Pian P
    4. Hu L
    5. Liu H
    6. Siegelbaum SA

    (2009) TRIP8b splice variants form a family of auxiliary subunits that regulate gating and trafficking of HCN channels in the brain

    Neuron 62:802–813.

    https://doi.org/10.1016/j.neuron.2009.05.009

    • PubMed
    • Google Scholar
68. 1. Santoro B
    2. Lee JY
    3. Englot DJ
    4. Gildersleeve S
    5. Piskorowski RA
    6. Siegelbaum SA
    7. Winawer MR
    8. Blumenfeld H

    (2010) Increased seizure severity and seizure-related death in mice lacking HCN1 channels

    Epilepsia 51:1624–1627.

    https://doi.org/10.1111/j.1528-1167.2010.02554.x

    • PubMed
    • Google Scholar
69. 1. Santoro B
    2. Shah MM

    (2020) Hyperpolarization-activated cyclic nucleotide-gated channels as drug targets for neurological disorders

    Annual Review of Pharmacology and Toxicology 60:109–131.

    https://doi.org/10.1146/annurev-pharmtox-010919-023356

    • PubMed
    • Google Scholar
70. 1. Saponaro A
    2. Cantini F
    3. Porro A
    4. Bucchi A
    5. DiFrancesco D
    6. Maione V
    7. Donadoni C
    8. Introini B
    9. Mesirca P
    10. Mangoni ME
    11. Thiel G
    12. Banci L
    13. Santoro B
    14. Moroni A

    (2018) A synthetic peptide that prevents camp regulation in mammalian hyperpolarization-activated cyclic nucleotide-gated (HCN) channels

    eLife 7:e35753.

    https://doi.org/10.7554/eLife.35753

    • PubMed
    • Google Scholar
71. 1. Saponaro A
    2. Bauer D
    3. Giese MH
    4. Swuec P
    5. Porro A
    6. Gasparri F
    7. Sharifzadeh AS
    8. Chaves-Sanjuan A
    9. Alberio L
    10. Parisi G
    11. Cerutti G
    12. Clarke OB
    13. Hamacher K
    14. Colecraft HM
    15. Mancia F
    16. Hendrickson WA
    17. Siegelbaum SA
    18. DiFrancesco D
    19. Bolognesi M
    20. Thiel G
    21. Santoro B
    22. Moroni A

    (2021) Gating movements and ion permeation in HCN4 pacemaker channels

    Molecular Cell 81:2929–2943.

    https://doi.org/10.1016/j.molcel.2021.05.033

    • PubMed
    • Google Scholar
72. 1. Schlusche AK
    2. Vay SU
    3. Kleinenkuhnen N
    4. Sandke S
    5. Campos-Martín R
    6. Florio M
    7. Huttner W
    8. Tresch A
    9. Roeper J
    10. Rueger MA
    11. Jakovcevski I
    12. Stockebrand M
    13. Isbrandt D

    (2021) Developmental HCN channelopathy results in decreased neural progenitor proliferation and microcephaly in mice

    PNAS 118:e2009393118.

    https://doi.org/10.1073/pnas.2009393118

    • PubMed
    • Google Scholar
73. 1. Singh NA
    2. Charlier C
    3. Stauffer D
    4. DuPont BR
    5. Leach RJ
    6. Melis R
    7. Ronen GM
    8. Bjerre I
    9. Quattlebaum T
    10. Murphy JV
    11. McHarg ML
    12. Gagnon D
    13. Rosales TO
    14. Peiffer A
    15. Anderson VE
    16. Leppert M

    (1998) A novel potassium channel gene, KCNQ2, is mutated in an inherited epilepsy of newborns

    Nature Genetics 18:25–29.

    https://doi.org/10.1038/ng0198-25

    • PubMed
    • Google Scholar
74. 1. Southan AP
    2. Morris NP
    3. Stephens GJ
    4. Robertson B

    (2000) Hyperpolarization-activated currents in presynaptic terminals of mouse cerebellar basket cells

    The Journal of Physiology 526 Pt 1:91–97.

    https://doi.org/10.1111/j.1469-7793.2000.t01-1-00091.x

    • PubMed
    • Google Scholar
75. 1. Stoenica L
    2. Wilkars W
    3. Battefeld A
    4. Stadler K
    5. Bender R
    6. Strauss U

    (2013) HCN1 subunits contribute to the kinetics of I(h) in neonatal cortical plate neurons

    Developmental Neurobiology 73:785–797.

    https://doi.org/10.1002/dneu.22104

    • PubMed
    • Google Scholar
76. 1. Stringer JL

    (1996) Repeated seizures increase GFAP and vimentin in the hippocampus

    Brain Research 717:147–153.

    https://doi.org/10.1016/0006-8993(96)00059-5

    • PubMed
    • Google Scholar
77. 1. Surges R
    2. Brewster AL
    3. Bender RA
    4. Beck H
    5. Feuerstein TJ
    6. Baram TZ

    (2006) Regulated expression of HCN channels and camp levels shape the properties of the h current in developing rat hippocampus

    The European Journal of Neuroscience 24:94–104.

    https://doi.org/10.1111/j.1460-9568.2006.04880.x

    • PubMed
    • Google Scholar
78. 1. Tröder SE
    2. Ebert LK
    3. Butt L
    4. Assenmacher S
    5. Schermer B
    6. Zevnik B

    (2018) An optimized electroporation approach for efficient CRISPR/cas9 genome editing in murine zygotes

    PLOS ONE 13:e0196891.

    https://doi.org/10.1371/journal.pone.0196891

    • PubMed
    • Google Scholar
79. 1. Tsay D
    2. Dudman JT
    3. Siegelbaum SA

    (2007) HCN1 channels constrain synaptically evoked ca2+ spikes in distal dendrites of CA1 pyramidal neurons

    Neuron 56:1076–1089.

    https://doi.org/10.1016/j.neuron.2007.11.015

    • PubMed
    • Google Scholar
80. 1. Van Erum J
    2. Van Dam D
    3. De Deyn PP

    (2019) PTZ-induced seizures in mice require a revised racine scale

    Epilepsy & Behavior 95:51–55.

    https://doi.org/10.1016/j.yebeh.2019.02.029

    • PubMed
    • Google Scholar
81. 1. Wang J
    2. Wen Y
    3. Zhang Q
    4. Yu S
    5. Chen Y
    6. Wu X
    7. Zhang Y
    8. Bao X

    (2019) Gene mutational analysis in a cohort of chinese children with unexplained epilepsy: identification of a new KCND3 phenotype and novel genes causing dravet syndrome

    Seizure 66:26–30.

    https://doi.org/10.1016/j.seizure.2019.01.025

    • PubMed
    • Google Scholar
82. 1. Wang W
    2. Frankel WN

    (2021) Overlaps, gaps, and complexities of mouse models of developmental and epileptic encephalopathy

    Neurobiology of Disease 148:105220.

    https://doi.org/10.1016/j.nbd.2020.105220

    • PubMed
    • Google Scholar

Author details

1. Andrea Merseburg

   1. German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany
   2. University of Cologne, Institute for Molecular and Behavioral Neuroscience, Cologne, Germany
   3. University of Cologne, Center for Molecular Medicine Cologne, Cologne, Germany

   Contribution

   Conceptualization, Funding acquisition, Investigation, Project administration, Supervision, Validation, Visualization, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0630-6564

2. Jacquelin Kasemir

   1. German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany
   2. University of Cologne, Institute for Molecular and Behavioral Neuroscience, Cologne, Germany

   Contribution

   Investigation, Visualization, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9176-5241

3. Eric W Buss

   Department of Neuroscience, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University, New York, United States

   Contribution

   Data curation, Investigation, Validation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0473-4717

4. Felix Leroy

   Department of Neuroscience, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University, New York, United States

   Contribution

   Data curation, Investigation, Validation, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1715-3233

5. Tobias Bock

   Department of Neuroscience, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University, New York, United States

   Contribution

   Data curation, Investigation, Validation, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7734-1183

6. Alessandro Porro

   Department of Biosciences, University of Milan, Milan, Italy

   Contribution

   Data curation, Investigation, Validation, Visualization, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4845-6165

7. Anastasia Barnett

   Department of Neuroscience, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University, New York, United States

   Contribution

   Data curation, Investigation

   Competing interests

   No competing interests declared

8. Simon E Tröder

   In vivo Research Facility, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany

   Contribution

   Methodology, Resources, Validation, Writing – review and editing

   Competing interests

   No competing interests declared

9. Birgit Engeland

   1. German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany
   2. University of Cologne, Institute for Molecular and Behavioral Neuroscience, Cologne, Germany

   Contribution

   Investigation, Project administration, Supervision, Validation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9104-7194

10. Malte Stockebrand

    1. German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany
    2. University of Cologne, Institute for Molecular and Behavioral Neuroscience, Cologne, Germany

    Contribution

    Investigation, Project administration, Validation

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-9009-137X

11. Anna Moroni

    Department of Biosciences, University of Milan, Milan, Italy

    Contribution

    Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-1860-406X

12. Steven A Siegelbaum

    Department of Neuroscience, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University, New York, United States

    Contribution

    Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-0242-7505

13. Dirk Isbrandt

    1. German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany
    2. University of Cologne, Institute for Molecular and Behavioral Neuroscience, Cologne, Germany
    3. University of Cologne, Center for Molecular Medicine Cologne, Cologne, Germany

    Contribution

    Conceptualization, Data curation, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing

    For correspondence

    dirk.isbrandt@dzne.de

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-4720-1016

14. Bina Santoro

    Department of Neuroscience, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University, New York, United States

    Contribution

    Conceptualization, Funding acquisition, Investigation, Project administration, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing

    For correspondence

    bs73@columbia.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-4277-1992

• 1,839

  views

• 379

  downloads

• 9

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.