Introduction

A molecular diagnosis in the early-onset epileptic encephalopathies (EOEEs) often has implications for management and prognosis.1, 2 Mutations in Kv7.2 (KCNQ2) and Kv7.3 (KCNQ3) genes, encoding for voltage-gated potassium channel subunits underlying the neuronal M-current (IKM),3 are responsible for early-onset epileptic diseases with a widely diverging phenotypic presentation.4 The M channel is a slowly activating and deactivating potassium conductance that has a critical role in determining the subthreshold electroexcitability of neurons as well as the responsiveness to synaptic inputs. Dominant mutations in this channel were first described in 1998 associated to benign familial neonatal epilepsy (BFNS; OMIM:121200) with a favorable prognosis.4, 5, 6 In stark clinical contrast, KCNQ2 mutations are also associated with EOEE, which has been called KCNQ2 encephalopathy (OMIM 613720).7 Patients with KCNQ2 encephalopathy display tonic seizures 2–3 days after birth with EEG characterized by suppression–burst patterns or multifocal epileptic activity and have poor neurological outcome.8, 9, 10 After the first EOEE-associated KCNQ2 mutation reported,11 more mutations in KCNQ2 had been described subsequently, most of them de novo. Now, it has been shown as one of the most common causes of EOEE including the Ohtahara syndrome.7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18

In this work, we aim to show the clinical, electrophysiological and genetic features from 13 KCNQ2-positive Spanish patients.

Materials and methods

This study has been approved by the ethics committee from Hospital Universitario Niño Jesús (Madrid, Spain). Detailed familiar and clinical information including age of onset, perinatal data, familiar history, initial and evolutionary seizures semiology, current seizure status, initial and evolutionary psychomotor development, ictal EEG and MRI findings, treatment over time and response to this, were collected for 13 patients with 13 different KCNQ2 mutations revealed through a targeted NGS-based panel for epilepsy genes. Informed consent was obtained from the parents of all the patients included in the study.

This cohort of KCNQ2 patients were extracted from 80 probands recruited from different Spanish Neuropediatric Units who experienced seizures of unknown etiology and studied through a custom-targeted NGS panel encompassing 155 epilepsy-associated genes to achieve a molecular diagnosis. A custom Sure Select oligonucleotide probe library was designed to capture the coding regions of these epilepsy-associated genes and requested to Agilent Technologies (Santa Clara, CA, USA). Captured fragments were sequenced in pair-end 100-base mode using Miseq Illumina platform from Illumina (San Diego, CA, USA). Reads were aligned to the reference genome GRCh37 with BWA v0.7.5a.19 Varscan and SAMtools v0.1.19 were the variant detection software used,20, 21 and Annovar for annotation.22 Prioritization was based on frequency of variants detected given in 1000 genomes, dbSNP and our in-house population database (collected sequencing data from 80 patients and 120 controls from Spanish origin). Sanger dideoxy sequencing method was used to perform familiar studies of KCNQ2 mutations detected in the patients. Missense mutations were evaluated with CONDEL (http://bg.upf.edu/fannsdb/).23 This tool integrates the output of five computational tools (SIFT, Polyphen2, MAPP, LogR Pfam E-value and MutationAssessor) aimed at assessing the impact of non-synonymous SNVs on protein function. To do this, it computes a weighted average of the scores of these tools.

The novel KCNQ2 mutations have been submitted to the ClinVar (https://submit.ncbi.nlm.nih.gov/subs/variation_submission/SUB1205799/overview) and to the Locus Specific (http://databases.lovd.nl/shared/users/01470) databases.

Results

Clinical findings

From the 13 patients studied, nine of them (P1–P9) presented a severe electroclinical phenotype (EOEE), whereas the remaining four patients (P10–P13) were cataloged as benign neonatal seizures (BNS; Table 1). None of the cases, except P1, showed a prominent perinatal history. In this patient, it was observed perinatal asphyxia data (presence of meconium and the need of pocket resuscitation mask) without showing hypoxic-ischemic encephalopathy in the following hours.

Table 1 Clinical features of 13 Spanish patients positive for KCNQ2 mutations
Full size table

Five patients had positive family history, until third-degree relatives, of epilepsy. All children had seizures onset in the neonatal period. The successive treatments applied are summarized in Table 2.

Table 2 Antiepileptic drugs administrated to KCNQ2-positive patients
Full size table

From the nine patients with KCNQ2-EOEE, seven debuted with tonic seizures, in six patients, seizures were accompanied by autonomic features characterized by brief respiratory pauses (apnea episodes). In the evolution, four presented epileptic spasms, and five showed exaggerated startle attacks. Five patients presented suppression–burst pattern in EEG, with four of them cataloged as Ohtahara syndrome and one with early myoclonic epileptic encephalopathy of Aicardi. Four patients developed EEG compatible with hypsarrythmia, two of them with an initial suppression–burst pattern. All the patients developed early psychomotor delay and presented intercritical multifocal epileptic activity during evolution. In two patients, MRI showed an altered pattern: P2, a cortico-subcortical atrophy; and P4, a slight delay in the myelination. Seven patients showed hypotonia. The mean follow-up period of the patients was 1–10 years and five are currently seizure free.

All BNS cases, with a mean follow-up period of 2–4 years, showed normal psychomotor development. P10 showed hyperechoic peripheral white matter lesions in the thalamus, presenting later a normal MRI. Three of them showed focal seizures at the onset and two showed epileptiform discharges in initial EEG and gradually returned toward normal subsequently.

Genetic findings

Thirteen different KCNQ2 mutations, 12 missense and one frameshift, were found in the 13 patients recruited in this work, nine EOEE-associated, two BNS-associated and two BFNS-associated (Table 3). Familiar genetic study was performed in all the cases. All the mutations found in EOEE cases and two BNS-associated mutations were de novo or the result of parental mosaicism, as it was not detected in DNA extracted from the whole blood of their parents. However, it is important to remark that in three of the EOEE patients, a positive family history for epilepsy was found. In two of the four BNS patients, the mutation was found in different family members with BNS showing cosegregation. Ten mutations were found to be novel, and five caused alteration of amino acid residues at which different missense changes had been reported previously in patients with BFNS and EOEE (Table 3). Mutations p.A265T and p.R553W caused alteration of amino acid at which two different missense changes have been previously described, one of them BFNS-associated and the other EOEE-associated. The mutation c.388G>A found in P8 takes place in the first nucleotide of exon 3, but this variant do not seem to cancel the consensus sequence required for the adequate splicing of such exon, according to the prediction made with human splicing program finder (HSF v.3.0). All the mutations have CONDEL values between 0.57 and 0.74, and all are predicted to be pathogenic. Amino acids 130, 201, 265, 268, 274, 284, 294, 306, 553 are highly conserved during evolution as it is shown by a SIFT value of zero, and the amino acids changes considered more drastic are R201C, T274M, R553W, R553L, R553Q.

Table 3 Summary of the variants identified in the KCNQ2 gene in 13 Spanish patients
Full size table

Discussion

Ages of onset and EEG findings are similar to those previously described.7, 8, 10, 14, 15, 17, 24 Nevertheless, we noted that five patients (38%) showed exaggerated startle attacks (nonepileptic tonic postures that appear as a result of different sensorial stimulus like sound or touch), which had never been published before in these patients. Only Allen et al.14 described nonepileptic extensor jerks in one case. Although seizures may be similar in both phenotypes,10, 14 neurological status and abnormal EEG pattern allow to distinguish them (EEOE vs BFNS).7, 11, 14 These electroclinical features may suggest the diagnosis of KCNQ2 encephalopathy.

From the nine patients with KCNQ2 encephalopathy, five remain seizure free but with severe impairment in psychomotor development, similar to that observed in other series.14, 15, 24 All these patients presented initial refractory epilepsy, but five became seizure free when they took inhibitors of sodium channels as it had been previously reported.7, 15, 24 Numis et al.24 reported that this beneficial effect seems to be due to the fact that both sodium and potassium channels are linked and located at critical points in the neuronal membrane. Thus, modulation of one of them, can significantly improve the function of the whole channel system. Retigabine has been postulated as an even more effective drug as it increases the potassium current through KCNQ channels;18, 24 however, it was not used in our cohort because of adverse effects.

Mutations in the KCNQ2 gene was initially discovered in patients with BFNS;5, 6 however, many cases associated with EOEE7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 were subsequently recognized. In the literature, it has been estimated that between 5 and 23% of all EOEE cases are secondary to mutations in KCNQ2.7, 15, 17 In our cohort, mutations in KCNQ2 accounts for about 16% of pediatric patients with seizures of unknown etiology. The reason why KCNQ2 mutations can lead to these two markedly different phenotypes is currently unknown. From the occurrence of KCNQ2 deletions in BFNS, it is known that a total loss of function of one allele leads to the benign phenotype only; therefore, haploinsufficiency cannot explain the more severe KCNQ2-EOEE. Even more, all KCNQ2-EOEE mutations published so far are missense mutations, but never loss-of-function mutations (always BFNS-associated). A possible explanation for this finding would be that haploinsufficiency may be physiologically reverted as it is shown in BFNS patients meanwhile alleles with missense mutations can be used to form functional channels but with different sensitivities to voltage promoted by the amino acid change. This finding could lead to the development of personalized treatment for KCNQ2-EOEE. Therapies using antisense oligonucleotides are being specifically developed for diseases with known dominant-negative mutations. Specifically inhibiting transcription or translation of the mutated KCNQ2 allele would lead to a loss-of-function situation, mimicking a KCNQ2 haploinsufficiency, which in turn, is known to lead to the milder BFNS phenotype. Therefore such strategy has the potential to turn a severe EOEE into a benign neonatal epilepsy syndrome.15

There has been reported several studies trying to explain the quite different phenotypes associated to KCNQ2 mutations. First explanation is the extent of the mutation-induced functional K+ channel impairment. Functional studies have revealed that KCNQ2-BFNS causing mutations decrease IKM conductance, and a 25% reduction in Kv7.2 currents appears sufficient to increase neuronal excitability to epileptogenic levels in early infancy;25 animal models appeared to confirm such conclusion.26 More dramatic functional deficits have been found in channels carrying mutations associated with more severe epileptic phenotypes.18, 27 With regard to this, it has been shown that two mutations affecting the same positively charged residue in the S4 domain of KV7.2 were found in children affected with BFNS (R213W) or with EOEE (R213Q). Functional studies revealed that both mutations increased cell firing frequency,27 with the R213Q mutation prompting more dramatic functional changes compared with the R213W mutation. These results suggested that the clinical disease severity may be related to the extent of the mutation-induced functional K+ channel impairment, and set the preclinical basis for the potential use of Kv7 openers as a targeted anticonvulsant therapy to improve developmental outcome in neonates with KV7.2 encephalopathy. Nevertheless, a recent work28 showed that three mutations placed in the voltage-sensing domain (R144Q, R201C and R201H) found in patients with EOEE stabilized the activated state of the channel, suggesting that a gain-of-function mechanism may also cause KCNQ2-EOEE. In brief, depending on the position and the features of the amino acid change, its influence in the voltage sensitivity of such channel will be variable. The change could promote increase and decrease of channel activity and with different intensity levels.

Three patients with supposed de novo mutations had positive family history for epilepsy in our cohort. It is important to highlight that a different pattern for heredity must be considered given that mosaicism may have an important role in the genetic burden of neurologic disease.29, 30 Electroclinical pattern and seizures response with inhibitors of sodium channels could be essential for the early diagnosis of KCNQ2 encephalopathy helping a more effective and targeted treatment.

To conclude, it should be noted that the presence of tonic seizures in the neonatal period accompanied by autonomic features, suppression–burst pattern in EEG and exaggerated startle attacks could be essential for suspected diagnosis of KCNQ2-EOEE and begin early treatment with sodium channels inhibitors until genetic confirmation. This work also shows that KCNQ2 mutations contribute to the development of about 16% of total pediatric Spanish patients with seizures of unknown etiology. It also confirms (1) the broader genetic heterogeneity of this gene with 10 novel KCNQ2 mutations found and (2) the two different phenotypes promoted depending of the amino acid change involved, even for the same protein position.

All procedures performed in the studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.