2.1 Patient consents and kinship analysis

All human study procedures and protocols complied with Yale University's Human Investigation Committee and Human Research Protection Program. Written informed consent for genetic studies was obtained from all participants. Relationship between proband, siblings, and parents in the Family NG1917 was estimated using the pairwise identity-by-descent calculation in PLINK.¹⁶ The identity-by-descent sharing between the proband, siblings and parents is between 45% and 55%. For principal component analysis, mapping and variant calling, de novo and inherited variant analysis and filtering, see the supplementary material.

2.2 Animal compliance

Animal procedures were approved by the Vanderbilt University Medical Center (VUMC) Institutional Animal Care and Use Committee (IACUC). This includes the generation of the Prkd1 mutant mouse model, kainic acid administration, tissue collection, behavioral experiments in mice, and the handling and use of Xenopus laevis frogs for oocyte collection. The description of the Prkd1 mutant mouse generation is provided in the supplementary material.

2.3 Transmission electron microscopy

Sural nerves were dissected from adult mice, then fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate for 1 h at room temperature (RT) and then at 4°C overnight. The Vanderbilt Electron Microscopy core further processed the sural nerve samples by washing and fixing them in 1% osmium tetroxide solution for 1 h at RT and then with 0.5% OsO₄ for 24 h. Then, the tissue samples underwent a series of ethanol dehydration steps (50% for 5 min, 75% for 15 min, 95% twice for 15 min each, and 100% thrice, 20 min each) before they were embedded in Spurr resin at 60°C for 24 to 48 h. Semi-thin sections (500 nm) were stained with toluidine blue and examined for positioning. Ultrathin sections (80 nm) were then cut and stained with uranyl acetate and lead citrate and placed on copper grids. Images were observed using a Philips/FEI T-12 transmission electron microscope and captured with the Philips/FEI T-12 Transmission Electron Microscope camera.¹⁷

2.4 Accelerated rotarod assay

A neuromotor coordination task was performed using an accelerating rotating cylinder (model 47 600; Ugo Basile, S.R. Biological Research Apparatus). Both controls (wild-types) Prkd1^+/+ and heterozygous Prkd1^+/E77X mice were tested. The cylinder was 3 cm in diameter and was covered with scored plastic. Mice were confined to a 4-cm-long section of the cylinder by gray Plexiglass dividers. Two to five mice were placed on the cylinder at once. The rotation rate of the cylinder increased over a 4-min period from 4 to 40 rpm. The latency of each mouse to fall off the rotating cylinder was automatically recorded by the device. Mice that remained on the rotarod after the 300 s trial period were removed and given a score of 300 s. The test was performed as three trials daily for three consecutive days with an inter-trial interval of at least 30 min.¹⁸

2.5 Balance beam assay

To assess fine motor coordination and balance, we used a 1 m-long steel balance beams of either 12 mm or 6 mm thickness. The beams were placed about 50 cm from the ground and positioned between two pillars. At the start of the test, mice began on an open, square platform and ended in an enclosed black box with bedding as motivation for the mice to cross. Mice were trained for 2 days (three trials per day for each beam) beginning with the thicker beam (12 mm) and progressing to the thinner beam (6 mm). The mice were tested consecutively on each beam with 10-min relief periods between each trial. The third day was used as the test day with three trials for each beam. The mice had about 60 s to traverse the beam and were scored on the neurological scoring system for beam walking adapted from Feeney and colleagues.¹⁹ This scoring system is based on the ability of the mouse to cross the beam and accounts for the number of paw slips. The mice received a score ranging from 1 to 7, based on their ability to complete the task, to place affected limbs on beam, and on the number of paw slips. This neurological scoring system considers a high score of 7 to be indicative of a wild-type mouse phenotype with no coordination deficits, and a low score of 1 indicative of severe motor defects.²⁰

2.6 Kainic acid administration and behavioral scoring of seizures

Behavioral observations began immediately following intraperitoneal administration of 20 mg/Kg kainic acid per mouse and continued for up to 1 h, after which, mice were returned to their home cage or sacrificed. Cages were not returned to the vivarium until at least 3 h after drug administration when it was confirmed that no further seizure activities were observed. Mice were scored live at the time of the treatment and videotaped for additional coding of activity levels by a second experimenter, who was fully blinded to experimental conditions. Mice were rated for immobility time across three continuous 10-min time bins. Kainic acid was scored according to a modified Racine scale in which stage 3 head bobs represents myoclonic jerks. The six stages in the Racine scale are: stage 1, immobility/flattening; stage 2, extension of forelimb and/or tail extension, rigid posture; stage 3, tics, repetitive movements, head bobs; stage 4, rearing and falling; stage 5, continuous rearing and falling, barrel rolling; and stage 6, severe tonic–clonic seizures.²¹

2.7 K⁺ influx measurements in Xenopus laevis oocytes

All frog handling and procedures were approved by the VUMC Institutional Animal Care and Use Committee (IACUC). Detailed methods on oocyte collection, cDNA clones, RNA transcription, and injection are included in the supplementary material and described in previous work.²² For the fluxes, oocytes were pre-incubated for 10 min in an isosmotic solution containing 96 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgSO4, 5 mM Hepes (pH 7.4), and 1 mM glucose, or in hypotonic solution with a NaCl content reduced to 48 mM. At the end of the preincubation period, the medium was aspirated and replaced with identical solutions containing 100 μM ouabain, and ⁸⁶Rb (5 μCi/ml). After a 1-h ⁸⁶Rb uptake, the oocytes were washed three times with ice-cold buffer, individually placed in liquid scintillation vials with 200 μl of 0.25 N NaOH for 1 h, and neutralized with 250 μl of acetic acid glacial. Liquid scintillation fluid (5 ml, Biosafe II) was added to each vial, and radioactivity was counted using a Packard Tri-Carb β-scintillation counter. K⁺ flux was calculated from ⁸⁶Rb counts and expressed in nanomoles K⁺ per oocyte per h. Calculation was based on measuring and averaging the counts (cpm) of 5 μl of aliquots of radioactive extracellular solution and relating these counts to the amount of K⁺ contained in these aliquots (e.g., 1 cpm = 2.5 pmol K⁺).

2.8 Cell culture and surface biotinylation

Human embryonic kidney (HEK) cells transfected with GFP-KCC3 and His-6-strep II-tagged PKD1 were grown in 10-cm Corning culture dishes at 37°C, air/5% CO₂. For biotinylation, cells were washed in ice cold CaCl₂ and MgCl₂ containing HBSS (HBSS²⁺), then incubated with freshly prepared Sulfo-NHS-biotin (Thermo Scientific, Waltham, MA, 0.5 mg/ml in cold HBSS²⁺ for 30 min at 4°C. After two washes with ice cold HBSS²⁺, 1 ml of lysis buffer (150 mM NaCl, 50 mM Tris–HCl pH 7.5, 1% Triton X-100 with proteases inhibitors) was added to the plate and cells lysate was collected. Biotinylated (cell surface) proteins were isolated by immunoprecipitation with streptavidin agarose beads (Thermo Scientific). Streptavidin immuno-precipitated proteins and whole cell lysate were used to detect cell surface and total cellular KCC3 by immunoblotting.

2.9 Immunofluorescence

HEK cells transfected with GFP-KCC3 (KCC3 tagged with GFP epitope) and His-6-strep II-tagged were cultured on glass coverslips until they reached 100% confluency. Cells were fixed with cold (−20°C) methanol, washed with 3 × 5 min with HBSS²⁺, and followed by incubation in blocking buffer (5% BSA, 0.1% Triton X-100, and 0.1% Tween 20 in HBSS²⁺) for 2 h and incubation with primary antibody overnight. Slides were washed 3 × 5 min with HBSS²⁺, and incubated with fluorophore-conjugated (Cy3 or FITC) secondary antibodies for 1 h. Chambers were removed from the microscope slide and mounted with ProLong Gold antifade reagent with DAPI (Invitrogen). Slides were imaged on a Zeiss LSM 880 laser scanning confocal microscope. Samples were scanned using a × 63 oil objective. The images were exported as TIFF files using Zeiss ZEN Lite 2012 software.

2.10 Immunoblotting

Equal volumes (40 μl) of cell lysates in sample buffer was subjected to gradient (4%–20%) SDS-polyacrylamide gel electrophoresis (PAGE) and proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Thermo Fisher Scientific). Membranes were probed with primary antibodies overnight at 4°C (KCC3, His-6 [Millipore], and Ezrin [Millipore]), followed by horseradish peroxidase (HRP)-conjugated secondary anti-mouse (Jackson ImmunoResearch) or anti-rabbit antibodies (Sigma-Aldrich), and revealed using enhanced chemiluminescence (PerkinElmer).

2.11 Quantitative PCR

Sciatic nerves were isolated from WT or Prkd1^+/E77X mice, lysed, and processed using the RNeasy mini kit (Qiagen). RNA quality and quantity were assessed by measuring absorbance at 260, 280, and 320 nm. Reverse-Transcription was performed by incubating 1 μg RNA with random hexamers, dNTPs, and SuperScript II (Invitrogen), for 1 h at 37°C, followed by denaturation for 5 min at 95°C. Quantitative PCR reactions contained 12.5 μl SYBR Green PCR master mix (Applied Biosystems, Foster City, CA), 1 μl each primer (1 μM), 9.5 μl water, and 1 μl cDNA. Relative mRNA expression levels were calculated by the ΔΔCt method. Statistical analysis was performed using one-way analysis of variance (ANOVA).

3.1 Clinical presentation of peripheral motor neuropathy and seizures

The PRKD1 variant was first discovered in a young female patient (Table S1) who developed an unusual and progressive seizure disorder. The patient presented at age 7 with a first-time seizure, during which she experienced an episode of whole body shaking and intermittent leg weakness. She continued to have similar episodes occurring multiple times during sleep, which typically occurred 2–3 h after sleep onset, lasting 30 s to 2 min. At the time of first presentation, her past medical history and family history were negative. Her exam was normal, and after initial workup, she was diagnosed with idiopathic generalized epilepsy syndrome after multifocal myoclonic seizures were captured on EEG. After further EEG studies and clinical evaluation, she was diagnosed with juvenile myoclonic epilepsy (JME) and her symptoms were partially responsive to valproic acid (VPA) and clonazepam. However, over the following couple of years her symptoms progressed despite treatment. She began having daytime seizures as well as progressive unsteadiness of her feet, gait, and fine motor dyspraxia, worsening action/intention tremor of her hands and feet with development of a wide based and tremulous gate, and hyperreflexia. Given the progressive nature and unclear etiology of her condition, she underwent further diagnostics, including an MRI of the brain as well as sural nerve and muscle biopsies.

Initial workup involved MRI of the brain, which revealed normal morphology, and a follow up MRI 9 years later continued to show unremarkable findings (Figure 1(A)). The patient underwent multiple MRIs since presentation with no findings to explain clinical course and disease. A dedicated 3 Tesla MRI brain under an epilepsy protocol, was performed when the patient was 21 years old. MRI imaging consistently revealed normal structure and signal of the hippocampi bilaterally, no heterotopic gray matter or focal cortical dysplasia, no polymicrogyria, and no abnormality in the brainstem. Note was made of a 1 mm focal area of signal intensity in the inferior right temporal gray matter, which cannot exclude an encephalocele; however, this finding was not re-demonstrated on any prior or subsequent scans. Of note, the patient has recently undergone repeat MRI with findings of a 4 mm T2 hyperintense lesion in the right pituitary gland, which likely represents a cystic microadenoma. A muscle biopsy showed well-preserved fascicles and normal staining (Figure 1(B)). Additional biopsies performed, however, suggested unusual abnormalities in myelination. The sural nerve biopsy; however, demonstrated an abnormal degree and pattern of myelination. Electron microscopy of the nerve biopsy was then performed which demonstrated nonspecific nerve sheath degeneration and abnormally myelinated Schwann cells with abnormal thickness and folding pattern (Figure 1(C),(D)). Metabolic and genetic testing performed initially did not uncover any known variants, therefore, whole exome sequencing was performed, which uncovered the variant in PRKD1.

3.2 Identification of the PRKD1 variant in the clinical case

To detect potential disease-causing variants with comprehensive genetic information, whole exome sequencing was performed on genomic DNA from the proband, both parents, and two siblings as described in the Section 2. The sequencing run achieved an average of 97.7% and 94.5% bases with at least 8x and 20x coverage, respectively (Table S1). Principal component analysis showed that these subjects were of Mexican ancestry. The identity-by-descent shared between parents and offspring was ~50%, which confirms the parent-offspring relationship (Table S2). Variants were called using the GATK Best Practices pipeline^{23, 24} and annotated using ANNOVAR.²⁵ The deleterious impact of missense variant was inferred using the MetaSVM algorithm.²⁶ Variants in genes of interest were validated using direct Sanger sequencing.

As parents did not have a history of seizures, we suspected that the disease in the proband may be caused by either de novo or recessive variants. To identify de novo variants, the TrioDeNovo program²⁷ was used and high-stringency filtering criteria were applied as described in the Section 2. We identified two de novo variants in the proband NG1917-1. The first is a novel stop-gain variant (p.E79X) in PRKD1 (Figure 1(E),(F)) and the second is a MetaSVM-tolerant missense (D-Mis) variant (p.R56C) in HIST1H4E (Table S3). Only p.E79X in PRKD1 is absent among >2.8 x 10⁵ alleles in gnomAD.²⁸ HIST1H4E, which encodes histone 1H4e, is important for functional information on H4 histones. In silico algorithms that predict the functional impact and biological relevance suggests that p.E79X in PRKD1 is most likely to contribute to the phenotype. Further, PRKD1 is expressed in several brain regions and tibial nerve according to the Genotype-Tissue Expression (GTEx) database²⁹; in contrast, HIST1H4E is barely expressed in any brain regions or tibial nerve per GTEx. For the recessive variants, we did not identify any genes that harbor homozygous or compound heterozygotes in the proband NG1917-1 using an allele frequency cutoff (MAF) of 10⁻³. Of note, one shared compound heterozygous variant in KCNH6 was identified in two siblings (Table S3). We also sought to identify rare (MAF < 2x10⁻⁵ in gnomAD) transmitted loss-of-function (LoF) and D-Mis heterozygous variants with incomplete penetrance but none of them are likely to be pathogenically significant (Table S3).

3.3 Generation and characterization of PKD1-E77X mice

Using CRISPR/Cas9 gene editing, we created a mouse that reproduces the PKD1-p.E79X variant found in the human patient. PKD1 is encoded by the PRKD1 gene in humans, and the Prkd1 gene in mice. Since the human E79 residue is located at position 77 in the mouse, we created a PKD1-E77X mouse line (Figure 2(A),(B)). After backcrossing a founder line to C57BL/6J mice for four generations, we crossed heterozygote mice to produce wild-type, heterozygous, and homozygous littermates. As seen in Figure 2(D), there was a significant deficit in the number of homozygous animals generated. Out of 125 pups produced in 16 litters, only three homozygote animals were produced. Thus, in contrast to the sex of the mice which was distributed according to a Mendelian distribution (χ² = 3.53, p = 0.060), the mutant Prkd1 allele was not properly distributed (χ² = 34.186, p < 0.001).

To assess nerve pathology in the PKD1 mutant mice, we isolated the sural nerves of WT, Prkd1^+/E77X, and Prkd1^E77X/E77X mice and processed them for electron microscopy (Figure 3(A),(D)). While the morphology of the nerve fibers was overall normal (Figure 3(A),(B)), several nerve tomacula pathologies were identified in sural nerves of Prkd1^E77X/E77X (Figure 3(C),(D)) similar to the pathology seen in the patient (Figure 1(C),(D)). We then assessed sural nerve fiber thickness, by measuring g-ratios on sural nerves isolated from WT, Prkd1^+/E77X, and Prkd1^E77X/E77X mice. Prkd1^E77X/E77X mice nerve fiber g-ratios were significantly lower than in wild-types and heterozygote mice (p < 0.01) (Figure 3(E)), possibly indicating larger fibers. To determine if the decrease in g-ratios was due to myelin thickness and/or axon diameter, we individually measured these two parameters on a large number of fibers, and could assign the increase in g-ratio solely to an increase in myelin thickness (Figure 3(F)). There was no measurable difference in axon diameter (Figure 3(G),(H)).

To determine if the nerve pathology affected motor coordination, balance, and/or fine motor movement in the mice, we performed several neurobehavioral tests on wild-type and heterozygote mice, as unfortunately, the breeding did not produce homozygote mice for these experiments. Male and female mice of age P60-P70 were subjected to the accelerated rotarod (Figure 4(A)), balance beam (Figure 4(B),(C)), open-field (Figure S1A-B), Porsolt forced swim (Figure S1C), and gait (Figure S2A-D) tests. There was no observed abnormal locomotor phenotype in the mutant Prkd1^+/E77X mice, compared to their wild-type counterpart. As the sural nerve has a purely sensory function, we also assessed sensory responses via measuring responses to noxious mechanical and heat stimuli through Von Frey (Figure S2E), and hotplate (Figure S2F) assays. The force needed for paw withdrawal upon filament pressure on the foot and the time needed for response to heat stimulus on a hotplate were similar between the two genotypes. Thus, there was no motor or sensory deficit measurable in the PKD1 mutant mice.

Because of the patient history of epileptic seizures and the possible involvement of KCC3 in seizure susceptibility,³⁰ we utilized a kainate-induced seizure model in PKD1 mutant mice compared to wild-type littermates (Figure 5). After intra-peritoneal injection of 20 mg/Kg kainic acid, mice were observed for 60 min and the latency to stage 3 myoclonic seizures (see methods) was measured. Prkd1^+/E77X mice had a clear increase in seizure susceptibility and severity compared to wild-type mice (Figure 5(A),(B)), where the latency of stage 3 seizures was significantly reduced in Prkd1^+/E77X (p < 0.001). Also, Prkd1^+/E77X mice experienced more severe seizures than WT mice. In addition, Prkd1^+/E77X mice exhibited a significantly larger number of myoclonic jerks than their wild-type counterparts (Figure 5(C), p < 0.01). Since changes in corpus callosum are associated with changes in brain weight³¹ and KCC3 loss of function is linked to ACCPN¹³ and increased seizure susceptibility,¹⁴ we were interested in measuring brain weight of Prkd1^+/E77X mice. There was no observed difference in averaged brain weights among WT and Prkd1^+/E77X mice (Figure 5(D)), concluding that the effect of PKD1 on seizures might act independently of the relationship between KCC3 and brain cell volume.

3.4 Effect of PKD1 variant on KCC3 function

PKD1 was identified as a hit in a kinome-wide RNAi screen carried out in HEK293 cells expressing KCC3.¹⁵ The screen was designed to recognize genes required for KCC3 phosphorylation at residue Thr⁹⁹¹.¹⁵ In addition, this study demonstrated that PKD1 knockdown in HEK293 cells resulted in decreased KCC3 phosphorylation at residues Thr⁹⁹¹ and Thr¹⁰⁴⁸. Given the importance of residue Thr⁹⁹¹ phosphorylation in swelling-induced stimulation of KCC3 activity,^{32, 33} the swelling-regulated function of KCC3 was assessed in Xenopus laevis oocytes co-injected with wild-type or catalytically inactive PKD1 (Figure 6(A)). Oocytes expressing catalytically inactive PKD1 demonstrated a small but significant increase in K⁺ influx, consistent with decreased phosphorylation of the cotransporter. To assess whether the increase was related to increased membrane expression, we co-transfected HEK2 cells with KCC3 and wild-type or mutant PKD1 kinase. Immunoblot assays (Figure 6(C),(D)) revealed no change in the level of KCC3 expression at the plasma membrane, irrespective of PKD1 activity. Note that the kinase was shown to co-localize with the cotransporter (Figure 6(B)), indicating that the proteins are likely to be part of a multi-protein complex. These results indicate that an increase in KCC3 mediated K⁺ flux is due to the increase in KCC3 activity and not the expression at the plasma membrane.