Novel mutations in SLC6A5 with benign course in hyperekplexia

Infants suffering from life-threatening apnea, stridor, cyanosis, and increased muscle tone may often be misdiagnosed with infantile seizures and inappropriately treated because of lack and delay in genetic diagnosis. Here, we report a patient with increased muscle tone after birth and hypertonic attacks with life-threatening apnea but no epileptiform patterns in EEG recordings. We identified novel compound heterozygous variants in SLC6A5 (NM_004211.4:c.[1429T > C];[1430delC]) by trio whole-exome sequencing, containing a base deletion inherited by the asymptomatic mother leading to a frameshift (c.1430delC, p.Ser477PhefsTer9) and a de novo base exchange leading to an amino acid change (c.1429T > C, p.Ser477Pro). To date, there are four known disease-associated genes for primary hyperekplexia, all of which are involved in the functioning of glycinergic synapses. SLC6A5 encodes the sodium- and chloride-dependent glycine transporter 2 (GlyT2), which recaptures glycine, a major inhibitory transmitter in the brainstem and spinal cord. The diagnosis altered the patient's medical care to his benefit because SLC6A5 mutations with rather benign courses of hyperekplexia may be spared of needless pharmacotherapy. Symptoms eventually decreased in frequency until about once in 2 mo at 2 yr age. We present the first report of halting hyperekplexia episodes by maternal soothing in multiple instances. We highlight the importance of clarifying the genetic diagnosis by rapid next-generation sequencing techniques in this group of infantile apneic attacks with hyperekplexia due to the broad differential diagnoses.


INTRODUCTION
Patients with sudden apneic attacks often present with diffuse patient histories and challenge pediatric physicians. Infant apnea may stem from anatomical obstructive or various central causes, oftentimes puzzling on-call physicians. Environmental or acquired etiologies of apneic attacks such as infections, premature birth, body temperature, sleep status, body position, and nicotine exposure play a major role in the differential diagnosis (Gao et al. 2017;Oishi et al. 2018). However, physicians are also oftentimes confronted with inborn genetic errors of the central nervous system-for example, congenital central hypoventilation with PHO2XB or LBX1 mutations (Hernandez-Miranda et al. 2018;Zaidi et al. 2018) or congenital myasthenic syndromes with episodic apnea due to pathogenic variants in CHAT (Mallory et al. 2009). not observed any pathological reflexes, delay of motor development, nor radiological signs in transfontanellar sonography that would suggest a beginning cerebral palsy. Of note, clubfeet may be a sign of an increased muscle tone . In our patient, the clubfeet were treated with an orthopedic redression in plaster and left no remaining defect on feet posture or gait.

Genomic Analyses
To uncover the genetic cause of the attacks, we first performed Mendeliome sequencing (see Methods and Supplemental Table S4; Fazeli et al. 2016). After initial filtering, we had only observed one variant in SLC6A5 and initially no other conclusive results that could have explained the clinical features sufficiently. In search for the genetic diagnosis, we performed trio whole-exome sequencing (WES) (see Methods and Supplemental Tables S5-S7), which revealed a compound heterozygous variant in SLC6A5 (NM_004211.4:c.[1429T > C];[1430delC]), consisting of a deletion inherited by the mother (c.1430delC, p.Ser477PhefsTer9) and a (likely) de novo base exchange (c.1429T > C, p.Ser477Pro) in the patient (see Fig. 1). The patient has a novel variant, which was previously not found in ClinVar or gnomAD. However, the mother's variant has a reported frequency of 2/246158 in gnomAD. Both variants have been listed now in ClinVar (SCV000897641 and SCV000897642). The bioinformatic reanalysis and annotation of the Mendeliome next-generation sequencing data confirmed the results as well, which was found by the trio WES results in SLC6A5 as shown in Supplemental Tables S5-S7. To double-check the sequences and the phasing, we subcloned PCR products of the peripheral blood DNA of the index patient and his mother in order to split the SLC6A5 alleles into separate plasmids. We used TOPO-TA cloning and subsequently performed a plasmid purification and dideoxy sequencing of the alleles in the plasmid separately (see Methods). Figure 1. (A) Pedigree and chromatogram of the two variants in SLC6A5. The mother is a heterozygous carrier of a deletion (c.1430delC, p.Ser477PhefsTer9). The patient carries the deletion inherited by the mother and additionally showed another heterozygous de novo mutation (c.1429T > C, p.Ser477Pro). (B) Review of alignments in patient's compound heterozygous variant from whole-exome sequencing (WES) with the two mutations located next to each other (varbank; https://varbank.ccg.uni-koeln.de).

B A
Novel mutations in SLC6A5 with hyperekplexia The dideoxy sequencing results confirmed these variants in the patient and his mother (see Fig. 2). We also performed kinship analysis to confirm that the patient is indeed the biological offspring of these parents (see Supplemental Methods; Supplemental Table S3).
SLC6A5 encodes a sodium-and chloride-dependent glycine neurotransmitter transporter (GlyT2) consisting of 12 transmembrane regions. The mutations on position p.Ser477 are located in transmembrane region 6 (see Fig. 3), which has a role in transporter binding and transporter activity. Position p.Ser477 is expected to serve as one of eight sodium-binding sites by similarity to DAT (sodium-dependent dopamine transporter) according to a previously published analysis (Benito-Muñoz et al. 2018).
To elaborate on the importance of the residue p.Ser477 for metal binding, multiple sequence alignment was performed for SLC6A5 between the sodium-dependent dopamine transporter from Drosophila melanogaster (DAT_DROME) as well as members of the human SLC6 family of sodium-and chloride-dependent neurotransmitter transporters (see Supplemental Fig. S1). The alignment was generated using the NCBI HomoloGene Protein Multiple Alignment platform, analyzed with the MUSCLE algorithm (Edgar 2004), provided by the Jalview web service, and visualized with the Jalview online tool (version 2.11.0) (Waterhouse et al. 2009). The secondary structure was predicted with the JPred Secondary Structure Prediction tool (Drozdetskiy and Cole 2015). Amino acid color labels were selected for the block substitution matrix 62.
In addition, a copy-number variant (CNV) analysis was performed to ensure we have not overlooked any other pathogenic variant. However, none of the detected CNVs shows any relevance to the phenotype (see Supplemental Tables S8-S10).

DISCUSSION
We report a 2-yr-old boy from non-consanguineous German parents with tonic-jittery attacks with tachycardia and intermittently aggravated hypertonic phases in extremities that led to life-threatening apnea with cyanosis immediately after birth. In deciphering the underlying etiology, we identified novel compound heterozygous variants in SLC6A5 (NM_004211.4: , containing a deletion inherited by the asymptomatic mother (c.1430delC, p.Ser477PhefsTer9) and a likely de novo base exchange (c.1429T > C, p.Ser477Pro). A plausible alternative mode of inheritance may be germline mosaicism of the patient's father.
Congenital hyperekplexia is a rare, potentially treatable neurogenetic disorder, and the diagnosis has been largely based on clinical findings (Thomas 2015). Affected individuals typically show generalized stiffness immediately or soon after birth. An excessive startle reflex to unexpected stimuli is the second main feature in hyperekplexia. Startle episodes are reported in some cases before birth (Thomas 2015). Although the symptoms are clearly defined, hyperekplexia can be confused with neonatal epilepsy, thus delaying diagnosis. Patients with hyperekplexia stay conscious during the tonic-jittery attacks, which distinguishes the disease from epileptic seizures. Although the symptoms often diminish during the first years of life, the excessive startle response can persist well into adulthood, leading to serious injuries from unprotected falls (Bode and Lynch 2014).
Although no epileptiform discharges were observed, the patient was treated with antiepileptic medication (levetiracetam) for 6 mo because the use of levetiracetam in a child with hyperekplexia was reported to have dramatically decreased the frequency of attacks Figure 3. Visualization of SLC6A5 as a transmembrane protein with 12 transmembrane regions, and the mutation in our patient on position p. Ser477 in region 6 is marked in red with a red arrow, as well as the missense mutations (p.Trp482Arg, p.Asn509Ser), nonsense mutation (p.Tyr377Ter), and frameshift mutation (p. Val432PhefsTer99) (Rees et al. 2001;Benito-Muñoz et al. 2018). Position 477 is known as a Na + -binding site. Each amino acid is abbreviated by its first-letter code.
Novel mutations in SLC6A5 with hyperekplexia  (Hussain et al. 2013). After the introduction of levetiracetam in our patient, we have observed a relaxation of initially increased muscle tone, but there were no changes in the hyperekplexia attacks. Thus, we suggest that a rather benign course of hyperekplexia with SLC6A5 mutations may render a continuous pharmacotherapeutic management dispensable. However, it should be debated on the basis of the individual case if the patient may benefit from a preventive pharmacotherapy in the first 3 mo of life because of severe apnea attacks or a pro re nata (PRN) medication (e.g., buccal midazolam). Of note, we present the first report of halting hyperekplexia attacks with maternal soothing and avoiding surrounding auditory stimuli.
In GlyT2-KO mice, glycinergic inhibition, such as in hyperekplexia, has proven to be mainly relevant in feedback regulation of respiratory reflexes. GlyT2 inhibition leads to depletion of intracellular glycine storage and limitation of glycine accumulation in synaptic vesicles (Morrow et al. 1998). In recurring cases without clinical clues, genetic workup might be useful to shed some light on the nature and course of the disease. For instance, if patients present with startle reflexes, it may allude to a rather benign variant such as hyperekplexia. A forced flexion of head and legs over the trunk ("Vigevano maneuver") has been reported to stop sudden attacks of hyperekplexia in infancy (Vigevano et al. 1989). Therapeutic management in hyperekplexia might include medication with an allosteric potentiator of the inhibitory GABA A receptor clonazepam in patients with variants in GLRA1 (Tijssen et al. 1997) and SLC6A5 (Bakker et al. 2009;Thomas 2015). The stimulation of P2X purinergic receptors with βγ-methylene adenosine 5 ′ -triphosphate has also been shown to induce the up-regulation of GlyT2 transport activity by increasing total and plasma membrane expression and reducing transporter ubiquitination (Villarejo-López et al. 2017). As a nonharmful therapeutic strategy in our case, maternal soothing was observed as halting hyperekplexia attacks in repeated instances.
We report two variants, which are located directly next to each other (see Fig. 1), and confirmed them by subcloning and subsequent plasmid sequencing.
Our results confirm that the healthy mother is a carrier of a deletion on position c.1430delC-a heterozygous frameshift variant that has been predicted to be subject to nonsense-mediated decay (NMD) by the NMDEsc Predictor. Because the mother remains asymptomatic, we may argue that the deficient SLC6A5 is partially cleared away by the NMD pathway or does not cause any other dominant negative effect. In a previous report, asymptomatic parents of multiple patients with SLC6A5-related hyperekplexia have been observed with truncating variants (Rees et al. 2006), which is in line with our observation.
In addition to this frameshift variant, the patient had a missense mutation (p.Ser477Pro) with a change from serine to proline. Please see Figure 3 for a comparison of our patient's mutation site as well as the missense, nonsense, and frameshift mutations that are discussed in the following. In a previous report of a severely affected individual (Rees et al. 2006), confocal microscopy of transfected HEK293 cells showed that nonmutated EGFP-hGlyT2 was readily expressed at the cell surface, whereas the mutants (p.Tyr377Ter and p.Val432PhefsTer99 among others) appeared to be cytoplasmic and showed no distinct expression at the cell surface with reduced [ 3 H]glycine uptake. Furthermore, Rees and colleagues have also shown in two-electrode voltage clamp analysis in Xenopus oocytes that a mutation in p.Trp482Arg did not respond to glycine (up to 10 mM), but was present at the cell surface as demonstrated by sodium-dependent and glycine-insensitive transient currents (see Fig. 3)-that is, the mutants p.Trp482Arg (transmembrane region 6) were functionally inert. In another report on the sodium-binding activity in SLC6A5, the mutational site in our patient, p.Ser477 in transmembrane region 6, was observed to serve as one of eight sodium-binding sites based on an homology model of GlyT2 dDAT (sodium-dependent dopamine transporter, Q7K4Y6, DAT_DROME) and was confirmed experimentally by electrophysiologcal examinations (Benito-Muñoz et al. 2018). In Supplemental Figure S1, we show a multiple sequence alignment of human SLC6A5 between the sodium-dependent dopamine transporter from Drosophila melanogaster (DAT_DROME) and paralog members of the human SLC6 family. This alignment indicates that the residue p.Ser477 is strictly conserved through sodium-and chloride-dependent neurotransmitter transporters, highlighting its importance in sodium binding. Overall secondary structure prediction (jnetpred) demonstrates the location of p.Ser477 in an α-helical section, consistent with its transmembrane location. As expected, no coiled-coil structure was predicted for this section (Lupas et al. 1991). Prediction of solvent accessibility (Jnet Burial) shows a medium exposure of residue p.Ser477. Thus, we conjecture that the missense variant in our patient (c.1429C > T, p.Ser477Pro) serves as a pathogenic mutation on the protein level because of the strict conservation throughout species, the damaging effect of a proline introduction in the α-helix, and the aforementioned electrophysiological observations of deficient sodium binding (Benito-Muñoz et al. 2018). On the basis of these findings and the experimental observations by Rees and colleagues, we hypothesize that the missense mutation (p.Ser477Pro) may lead to deficient metal binding in GlyT2 at transmembrane region 6, which renders it functionally inert. and to reduced expression levels because of the disrupting effect of proline onto the α-helix.
The patient's life-threatening apneic attacks first raised suspicions toward congenital hypoventilation as a severe differential diagnosis to hyperekplexia. LBX1 and PHOX2B mutations have previously been shown to impair the development of a small subpopulation of neurons in the medulla oblongata that are essential for respiratory control (Hernandez-Miranda et al. 2018). Other differential diagnoses of neuromuscular symptoms with an infantile onset may stem from "channelopathies" (i.e., a heterogeneous group of disorders resulting from the dysfunction of transmembrane ion channels). These more severe cases comprise patients with mutations in genes that are components of the nonselective sodium leak channel complex (NALCN channelosome) and-depending on the inheritance pattern -either present with muscular hypertonia and distal contractures  or hypotonia, psychomotor retardation, and dysmorphic features (Bramswig et al. 2018), as well as patients with mutations in the sodium voltage-gated channel α subunit 4 gene (SCN4A), which may present with congenital myopathy or as congenital myasthenic syndrome (Sloth et al. 2018;Elia et al. 2019).
Hyperekplexia patients with a benign phenotype and variants in SLC6A5 are significantly less likely to have recurrent infantile apnea than those with GLRA1 variants (Thomas 2015). Moreover, patients with variants in GLRB and SLC6A5 are more likely to have a developmental delay than those with GLRA1 variants. Thus, an early genetic workup helps in recognizing the patient's symptoms, providing parents with genetic counseling, and avoiding unnecessary medication and its accompanying adverse effects in early postnatal development. Nextgeneration sequencing (NGS) to uncover the underlying cause is indicated because of the broad spectrum of genetic differential diagnosis for apneic attacks and epilepsy-like clinical presentations.
Diagnosing rare diseases with NGS in perinatal settings has become highly rapid, economical, and efficient, but it comes with careful consideration of parental consent, ethical framework, and sparing trauma for patients and parents (Daoud et al. 2016;Fazeli et al. 2016;Poulsen et al. 2016;Borghesi et al. 2017;Kuehne et al. 2019). Importantly, receiving a genetic diagnosis might enable physicians to administer specific therapy or at least deter unnecessary drug exposure.
In conclusion, we report a novel compound heterozygous variant in SLC6A5 with already well-established symptoms that may have been overlooked initially because of the broad differential diagnoses of apneic attacks. If-as in this case-a variant in SLC6A5 is revealed to cause the disease, the patient is expected to have a benign form of hyperekplexia. We report the first instances of stopping hyperekplexia attacks with maternal soothing, thus the patient could be spared anti-epileptic medication. Because the patients are conscious during attacks, the family members could be informed of strategies to halt the attacks without using medication, such as soothing the patient. Lastly, we may offer genetic counseling to the patient's family regarding the expected ordinary thriving and sensomotoric development in this benign disease course.

METHODS
Written informed consent was obtained from the parents for genetic investigations and recording and publishing of the disease-related information. The study was approved by the institutional review board of the Ethics Committee of the University Hospital of Cologne.
To uncover the genetic cause in this family, we performed Mendeliome sequencing, a commercial gene panel (Illumina TruSight One, Illumina) including 4.813 genes responsible for rare diseases (see Supplemental Tables S1 and S4; Fazeli et al. 2016;Alawbathani et al. 2018). The sequencing was performed on a MiSeq sequencer (Illumina) using the TruSight One chemistry for target extraction (Illumina). Because it was inconclusive in the first analysis, we performed trio whole-exome sequencing, which led to the diagnosis. Reanalysis of the Mendeliome confirmed the results. Genomic DNA samples isolated from peripheral blood of the index patient and his parents were enriched with the NimbleGen SeqCap EZ Human Exome Library v2.0 (Roche) following the manufacturer's instructions. The trio was sequenced on a HiSeq 2000 sequencer (Illumina) with 2 × 101-bp reads, producing a mean coverage of the target regions of 94× for the index patient, 89× for the father, and 147× for the mother (see Supplemental Tables S1 and S5-S7). To confirm the variants in SLC6A5 and to validate the cosegregation within the family, we performed dideoxy sequencing.
Using early versions of the Cologne Center for Genomics exome pipeline, the sequencing data of the Mendeliome sequencing was analyzed with version 2.10, and the WES trio and the reanalysis of Mendeliome sequencing data were analyzed with version 2.14, only with differences in technical fixes-for example, activating Ion Torrent and Illumina gene panels, disabling downsampling in variant callers, and various bug fixes in parameter parsing or disk space usage (Kawalia et al. 2015). For further bioinformatics analysis of NGS data, refer to the Methods section in our Supplemental Material.
The variants were filtered for a de novo and compound heterozygous inheritance model without consanguine familiar background; with an allele read frequency window of 25%-75%. Variants were considered with a minor allele frequency of 0.1% or less. From a total number of 16 rare functional variants (see Supplemental Table S2), we checked the variants for a quality of >100, nonsynonymous coding, polymorphism predictions. We classified the remaining variants according to the American College of Medical Genetics and Genomics-Association for Molecular Pathology (ACMG-AMP) guidelines and refined Sherloc criteria (Richards et al. 2015;Nykamp et al. 2017). Solely the mentioned variants in SLC6A5 fitted to all of the filter criteria, were classified as pathogenic in both variant classifications, and could be matched to the phenotype of our patient. Table 1 shows the gene variants, according to the ACMG-AMP criteria, and refined Sherloc criteria (Nykamp et al. 2017) in the  Richards et al. 2015;Nykamp et al. 2017.
The ACMG and Sherloc scores were called for each variant, the ACMG-AMP and Sherloc classification was called together for compound heterozygous variants.
Only the SLC6A5 variant was clearly classified as pathogenic. The variants in ABCC6 locate to the nucleotide-binding fold 1 (p.Val787Ile) and the seventh cytoplasmic loop (p.Arg1030Gln), whereas p.Val787Ile has been published in a patient with Pseudoxanthoma elasticum (PXE) leading to strokes due to vascular mineralization. However, ACMG criteria imply uncertain significance for variants and there were no other neurological manifestations, which renders the ABCC6 variants highly unlikely to be the primary cause of the patient's disease. Variants in KIAA0513 have been associated with schizophrenia; however, as they were not reported in combination with seizure-like episodes, a modifying role cannot be excluded. For SLC6A5, the identified homozygous frameshift variant explains the phenotype of hyperekplexia adequately. remaining variants, including four compound-heterozygous variants of uncertain significance in the ATP Binding Cassette Subfamily C Member 6 (ABCC6) gene and KIAA0513 gene, and both SLC6A5 variants we present here, which are highlighted in Supplemental Table S11. To evaluate if a variant was subjected to NMD, we used the NMDEsc Predictor online tool (https ://nmdprediction.shinyapps.io/nmdescpredictor/). In addition to the above-mentioned resources, the variants were also checked in gnomAD (https://gnomad.broadinstitute.org/) for reports in exome-or genome-wide population studies and ClinVar (https://www.ncbi .nlm.nih.gov/clinvar/) for supporting evidence and clinical significance.
To confirm the relations of patient and parents, we also performed kinship analysis with varbank by analyzing the proportion of shared rare alleles (see Supplemental Methods and Supplemental Table S3 for further details).
The pipeline performed CNV calling for the patient, father, and mother individually, using three different callers: ExomeDepth (Plagnol et al. 2012), XHMM (Fromer et al. 2012), and CoNIFER (Krumm et al. 2012). For more details on CNV calling, please see Supplemental Tables S8-S10 for called de novo CNVs and details about callers.
In a next step, we wanted to confirm the allele specificity of the SLC6A5 variants. After a standard PCR procedure to child and mother's samples with HotStar HiFidelity DNA polymerase (QIAGEN) with a proofreading 3 ′ to 5 ′ exonuclease activity, the PCR products were purified to remove proofreading enyzme with the PCR clean-up Gel extraction kit (Macherey-Nagel) according to the protocol.
After purification of the PCR product, a 3 ′ A-overhang was added by incubation of the PCR product with a nonproofreading MyTaq DNA polymerase, dNTPs, and MyTaq 1× PCR buffer (BIOLINE) for 10 min at 72°C. The product was ligated into the pCR4-TOPO TA Vector (Invitrogen) according to the manufacturer's protocol for the "TOPO TA Cloning Kit for sequencing." The ligation product was transformed into "One Shot Mach1 Phage-Resistant" Chemically Competent E. coli (Thermo Fisher Scientific). Cells were plated and incubated overnight at 37°C. Several colonies were picked to inoculate 5-mL cultures and incubated overnight at 37°C. Plasmids were purified with the NucleoSpin Plasmid kit (Machery-Nagel). Subsequent dideoxy sequencing of the plasmids insert confirmed the above-mentioned variants in the index patient and his mother (see Fig. 2).

ADDITIONAL INFORMATION Data Deposition and Access
The pathogenic SLC6A5 variants have been submitted to ClinVar (http://www.ncbi.nlm.nih .gov/clinvar/) and can be found under accession numbers SCV000897641 and SCV000897642. We have no further data to be deposited, because we are not allowed to publish the full exome variant data set based on IRB approval and patient consent.

Ethics Statement
Informed consent was obtained from the patient and parents for genetic investigations and recording and publishing of the disease-related information. The study was approved by the institutional review board of the Ethics Committee of the University Hospital of Cologne.
the Gerok program of the Faculty of Medicine, University of Cologne. We also thank the Regional Computing Center of the University of Cologne (RRZK) for providing computing time for the bioinformatics analyses on the DFG-funded High Performance Computing (HPC) system CHEOPS, as well as for support.

Author Contributions
H.S.D. analyzed clinical, genetic, and bioinformatic data and wrote the manuscript. A.M. summarized clinical findings. P.H. contributed to the clinical diagnosis, description, and management of the patient. M.K. contributed to the genetic workup and data analysis and revised the manuscript. A.K., R.S., A.M., P.N., S.M., and H.T. contributed the genetic and bioinformatic analysis of the data and revised the manuscript. S.C. obtained funding, analyzed the data, coordinated and supervised the work, and revised the manuscript. All authors approved the manuscript before submission.