In vitro characterization of cells derived from a patient with the GLA variant c.376A>G (p.S126G) highlights a non-pathogenic role in Fabry disease

Abstract

Fabry disease (FD) is a life-limiting disorder characterized by intracellular globotriaosylceramide (Gb3) accumulations. The underlying α-galactosidase A (α-GAL A) deficiency is caused by variants in the gene GLA. Variants of unknown significance (VUS) are frequently found in GLA and challenge clinical management. Here, we investigated a 49-year old man with cryptogenic lacunar cerebral stroke and the chance finding of the VUS S126G, who was sent to our center for diagnosis and initiation of a costly and life-long FD-specific treatment. We combined clinical examination with in vitro investigations of dermal fibroblasts (HDF), induced pluripotent stem cells (iPSC), and iPSC-derived sensory neurons. We analyzed α-GAL A activity in iPSC, Gb3 accumulation in all three cell types, and action potential firing in sensory neurons. Neurological examination and small nerve fiber assessment was normal except for reduced distal skin innervation. S126G iPSC showed normal α-GAL A activity compared to controls and no Gb3 deposits were found in all three cell types. Baseline electrophysiological characteristics of S126G neurons showed no difference compared to healthy controls as investigated by patch-clamp recordings. We pioneer multi-level cellular characterization of the VUS S126G using three cell types derived from a patient and provide further evidence for the benign nature of S126G in GLA, which is of great importance in the management of such cases in clinical practice.

Highlights

• •The GLA variant S126G is not associated with Fabry symptoms in the presented case
• •S126G has no effect on α-GAL A activity or Gb3 levels in this patient
• •S126G sensory neurons show no electrophysiological abnormalities

1. Introduction

Fabry disease (FD) is an X-linked lysosomal storage disorder caused by variants in the α-galactosidase A (α-GAL A) gene (GLA) [25]. With an estimated prevalence of 1/8000 [27], FD is the second most common inherited metabolic storage disease [62]. The underlying enzyme malfunction leads to intracellular accumulation of the glycosphingolipid substrate globotriaosylceramide (Gb3) in various organs and tissues. Patients present with diverse symptoms including cardiomyopathy, renal failure, cerebral stroke, and pain due to small fiber neuropathy (SFN) [20]. FD-associated pain is an early hallmark of the disease, with a prevalence of 60–70% in men [21,69]. Traditionally, FD phenotypes were categorized into “classic” with symptom manifestation already in childhood and “late-onset” with milder symptoms starting in adulthood [3,25]. The degree of enzymatic impairment and symptom severity corresponds to the location of the >900 [66] different gene variants known [53] differentiated into “benign” to “pathogenic” [52]. However, some variants remain of unknown significance (VUS). To properly decide on the initiation of FD-specific treatment which is costly, life-long, and may lead to potentially severe side effects [12,28], proper classification of these variants is crucial.

S126G (p.Ser126Gly, c.376 A > G) with a population frequency of 0.06% [11] is such a genetic variant considered “benign” and a VUS [37]. While in one study, end-stage renal disease was described in two patients [72], other studies did not show a pathogenic nature of S126G [15,39,51]. Since previous data was mostly acquired by clinical assessment of patients and enzyme activity measurements [2,15,39], basic science approaches on cellular level, going beyond the established measures, can help to answer the question of pathogenicity and strengthen patients` confidence in medical decisions. Here, we present data on the GLA variant S126G, linking thorough clinical examination with a patient-derived in vitro system to investigate morphology and functionality of skin cells, induced pluripotent stem cells (iPSC), and sensory neurons. By analyzing α-GAL A activity, Gb3 accumulation, and neuronal action potential firing, we tested the hypothesis that no molecular signs of FD are found in S126G cells.

2. Materials and methods

2.1. Clinical setting

The 49-year old man was seen at our Fabry Center for Interdisciplinary Therapy Würzburg (FAZIT). One year ago, he had experienced an episode of sensory impairment at the right hand together with an episodic reduction of fine motor skills lasting for few minutes. He also reported acute numbness at the right knee and tingling of the right face. Diagnostic work-up at the Neurological Department in his hometown revealed a punctual lacunar restriction on diffusion-weighted magnetic resonance imaging (MRI) of the brain in the left cerebral hemisphere. Since cardiovascular examinations were normal and the etiology remained obscure, FD diagnostics was initiated in the young patient, which revealed the genetic variant S126G in the GLA gene as a chance finding. In the following days, all symptoms resolved completely and the patient reported well-being. No such symptoms had occurred before and since then. Explicit interview also revealed that the patient had never experienced other symptoms characteristic for FD such as acral or other physical pain, hypohidrosis or hearing problems, and did not have angiokeratomas. α-GAL A activity in peripheral blood leukocytes (0.6 nmol/min/mg protein, normal range 0.4–1.0) and lyso-Gb3 (0.6 ng/ml, normal range < 0.9) levels were both normal. Also, family history was negative.

2.2. Clinical examination

The patient underwent complete neurological examination, nerve conduction studies of the right sural and tibial nerves, and a sympathetic skin response test. Further, the somatosensory evoked potentials were recorded upon stimulation of the tibial nerve. For cerebral blood flow, extra- and transcranial duplex sonography was performed and the patient underwent cranial MRI including time-of-flight sequences. For the assessment of small caliber nerve fibers, quantitative sensory testing (QST) was done at the right dorsal foot and a 6-mm skin punch biopsy was taken from the lower leg and the back paraspinally at dermatome level th5 for analysis of intraepidermal nerve fiber density (IENFD). The patient further underwent renal and cardiac assessment to detect potential nephropathy or cardiomyopathy. He had also undergone ophthalmological examination as an outpatient in his home town two years before and provided the medical report.

2.3. Skin biopsy and fibroblast cultivation

Human dermal fibroblasts (HDF) were obtained from the diagnostic skin punch biopsies following a published protocol [68]. For in vitro experiments, half of the skin biopsy from the lower leg was used. After mechanical separation of dermis and epidermis, dermal tissue was cultivated in fibroblast expansion medium (Dulbecco's modified eagle's medium/Nutrient Mixture F-12 (Thermo Fisher Scientific, Waltham, MA, USA), 10% fetal calf serum (FCS, Biochrom, Berlin, Germany), 100 U/ml penicillin/streptomycin (Pen/Strep, Thermo Fisher Scientific, Waltham, MA, USA)) under 5% CO₂ and 37 °C [34].

2.4. Generation and cultivation of iPSC

HDF obtained by skin punch biopsy were reprogrammed to iPSC with the StemRNA 3rd Gen Reprogramming Kit (ReproCell, Yokohama, Japan) as previously described [8]. In brief, non-modified RNA coding for reprogramming factors and immune evasion factors, and the microRNA cluster 302/367 were transiently transfected into HDF for four consecutive days using Lipofectamine™ RNAiMAX transfection reagent (Thermo Fisher Scientific, Waltham, MA, USA). Emerging iPSC colonies were isolated and expanded in StemMACS™ iPS-Brew XF, human medium (Miltenyi Biotec, Bergisch Gladbach, Germany) supplemented with 100 U/ml Pen/Strep on hESC qualified Matrigel- (Corning, Corning, NY, USA) coated 6-well plates (Greiner Bio-One, Kremsmünster, Austria) under 5% CO₂ and 37 °C. Routine culture required daily medium change and passaging twice a week using 2 mM EDTA (Thermo Fisher Scientific, Waltham, MA, USA) in phosphate-buffered saline (PBS, Sigma-Aldrich, St. Louis, MS, USA) for detaching. To inhibit apoptosis, 10 μM Y27632 (Miltenyi Biotec, Bergisch Gladbach, Germany) was added to the medium for the first 24 h after passaging. Expanded clones were analyzed for pluripotency factor expression, three-germ-layer-formation potential, normal karyotype, and GLA sequence ([8] and Fig. S1).

2.5. Sensory neuron differentiation

iPSC were differentiated to sensory neurons as previously described [36]. In brief, iPSC were seeded into growth factor-reduced Matrigel-coated (Corning, Corning, NY, USA) 6-well plates, cultured until confluence, and differentiated in a ten day protocol with daily change of defined differentiation medium. Until day two, neuralization of iPSC was induced via dual SMAD inhibition using the small-molecule inhibitors LDN-193189 and SB431542. Between day three and ten, three additional small-molecule inhibitors (CHIR99021, SU5402, DAPT) acting on WNT-, notch-receptor, and FGF-receptor signaling were applied to obtain nociceptive sensory like neurons via a neural crest cell intermediate. From day eleven on, cells were matured for at least five weeks with the addition of β-nerve growth factor (β-NGF), brain-derived neurotrophic factor (BDNF), and glia cell-derived neurotrophic factor (GDNF) to produce fully active post-mitotic sensory neurons.

2.6. Immunocytochemistry

For immunocytochemistry (ICC), iPSC and HDF were seeded on Matrigel- (Corning, Corning, NY, USA) or Cultrex-coated (Bio-Techne, Minneapolis, MN, USA) high-precision glass coverslips (Hartenstein, Würzburg, Germany) two days prior to immunoreaction and pre-seeded sensory neurons were stained five to seven weeks after differentiation. Cells were fixed with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA, USA) in PBS for 15 min at room temperature, washed with PBS, and unspecific binding was blocked with 5% fetal bovine serum (FBS, Merck, Darmstadt, Germany) in PBS. Primary antibodies were diluted in blocking solution, supplemented with 0.1% saponin for permeabilization, and incubated at 4 °C overnight. Primary antibodies included anti-peripherin (mouse anti-human IgG, conjugated to Alexa Fluor 488, 1:200, sc-377,093, Santa Cruz Biotechnology, Dallas, TX, USA) and anti-vimentin (rabbit anti-human IgG, 1:200, ab92547, Abcam, Cambridge, UK). Cells were washed with PBS and incubated with secondary antibodies diluted in PBS for two hours at room temperature. Secondary antibody used was 488-labeled donkey anti-rabbit IgG (1:100, Jackson ImmunoResearch Europe, Ely, UK). After washing with PBS, coverslips were mounted using Aqua-Poly/Mount mounting solution (Polysciences, Warrington, PA, USA). To label nuclei, 4′,6-diamidino-2-phenylindole (DAPI) was included in the first washing step before mounting. For visualization of intracellular Gb3, Shiga toxin subunit B (STxB, Sigma-Aldrich, St. Louis, MO, USA) was coupled to Alexa Fluor 555 dye (Thermo Fisher Scientific, Waltham, MA, USA) (STxB::555) as previously described [61], and added to the primary antibody solution. Images were acquired on a Zeiss Axio Imager 2 upright microscope (Zeiss, Oberkochen, Germany) controlled by ZEN (blue edition) software (Zeiss, Oberkochen, Germany) and processed with the Fiji distribution of ImageJ software [60].

2.7. α-GAL A activity assay

α-GAL A activity was measured in iPSC with the Alpha Galactosidase Activity Assay Kit (Abcam, Cambridge, UK). In brief, iPSC were lysed via ultrasonic pulses, centrifuged, and supernatant containing α-GAL A was collected for analysis. Substrate cleavage by α-GAL A was detected fluorometrically after two hours of incubation at 37 °C using a Tecan microplate reader (Tecan, Männedorf, Switzerland). For normalization, total protein concentration was determined using a BCA assay (Interchim, Montluçon, France).

2.8. Patch-clamp recordings

For electrophysiological recordings, sensory neurons were seeded on Matrigel- (Corning, Corning, NY, USA) or Cultrex-coated (Bio-Techne, Minneapolis, MN, USA) high-precision glass coverslips directly after differentiation and measured after five to ten weeks of maturation. On the day of analysis, neurons were transferred to a microscope chamber filled with physiological bath solution (180 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 10 mM glucose, 5 mM HEPES). To form the recording electrode, borosilicate capillaries (Kimble Chase Life Science and Research Products, Meiningen, Germany) were pulled with a P-1000 micropipette puller (Sutter Instrument, Novato, CA, USA) to a final input resistance of 3–6 MΩ, filled with intracellular pipette solution (170 mM KCl, 2 mM MgCl₂, 1 mM EGTA, 1 mM ATP, and 5 mM HEPES), and imposed on an Ag/AgCl electrode. A second Ag/AgCl reference grounding electrode was applied to the bath solution directly to close the electrical circuit. Both electrodes were connected to an EPC10 amplifier (HEKA, Ludwigshafen, Germany) controlled by the Patchmaster software (HEKA, Lambrecht, Germany) to inject and measure currents with a sample rate of 20 kHz. Neuronal somas were approached with the recording electrode using a micromanipulator (Luigs & Neumann, Ratingen, Germany). After gigaseal formation, fast pipette capacitance was compensated. Whole-cell (WC) configuration was achieved via short suction rupturing the membrane followed by slow membrane capacitance compensation. Electrical configurations and parameters were constantly assessed in Patchmaster via a 5-mV test pulse. To control for access integrity during WC configuration, measurements were filtered to an access resistance below 15 MΩ.

To record action potentials (AP), square-wave currents between 0 and 290 pA were injected into the neurons incrementally in 10-pA steps, and the first evoked AP was used to determine AP parameters with Stimfit software [29]. AP amplitude was considered the difference between resting membrane potential and peak depolarized potential, and AP duration was determined at half-maximum amplitude. The threshold potential was determined at the turning point of the AP depolarizing phase characterized by a slope of 5 mV/ms [14]. The minimum current needed to elicit an AP was considered the rheobase of the cell. To investigate firing behavior, APs were counted over 500 ms of stimulation with currents resembling 1×, 2×, 3×, and 4× multiples of the rheobase. Neurons that elicited multiple action potentials were included as nociceptors.

2.9. Statistics

Statistical analysis was performed with SPSS software version 28.0.1.0 (IBM, Ehningen, Germany). Data was analyzed for normal distribution using Shapiro-Wilk test. Normally distributed data were investigated using the Student's t-test and one-way ANOVA for two and more than two groups, respectively. Otherwise the nonparametric Mann-Whitney-U test was applied. Differences were considered significant at p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001). Graphs were generated in Graphpad PRISM software version 9.3.1 (GraphPad Software, Inc., La Jolla, CA, USA).

3. Results

3.1. Normal clinical assessment

Neurological examination was completely normal. Nerve conduction studies of the right sural and tibial nerves showed normal results. Sympathetic skin response at the feet after stimulation at the hand was present and the P40-latencies of the somatosensory evoked potentials were in the normal range. Extra- and transcranial duplex sonography showed normal blood flow. Cranial MRI including time-of-flight sequences was also unremarkable, particularly without white matter lesions or vascular ectasia. QST at the right foot revealed normal mechanic and thermal perception and pain thresholds compared with published normative data [43]. IENFD was reduced at the lower leg (2.1 fibers/mm), while the proximal skin innervation was normal (17 fibers/mm) compared with our laboratory normative values (lower leg: 9 +/-3 fibers/mm; back: 25 +/-8 fibers/mm). Further clinical examination revealed normal renal function with a glomerular filtration rate (e-GFR) of 103 ml/min /1.73 m² and no proteinuria. Cardiac function was also normal without cardiomyopathy. Also, no cornea verticillata had been found during an ophthalmological assessment as documented in the respective medical report the patient provided.

3.2. S126G iPSC show normal α-GAL A activity

To assess the effect of the genetic variant S126G on enzyme stability, we first determined the location of serine at position 126 within the α-GAL A homodimer (Fig. 1A). The affected serine resides on the protein surface distant to the active site of α-GAL A. Further, we measured α-GAL A activity in cell lysates of S126G and healthy control iPSC. α-GAL A activity was similar in both genotypes (Fig. 1B).

Fig. 1.

Variant site and α-GAL A activity in iPSC.

(A) Visualization of α-GAL A homodimer (grey, PDB identifier: 1R46, [24]) including the site affected by the genetic variant S126G (green) on the protein surface. (B) α-GAL A activity measured in cell lysates of a healthy control iPSC clone and two independent S126G clones. No difference in enzyme activity was found for control and S126G iPSC (one-way ANOVA, n_Control = 3, n_S126G = 3). Bar graph of means with standard deviations. Abbreviations: α-GAL A, α-galactosidase A; iPSC, induced pluripotent stem cells; PDB, protein data base.

3.3. No Gb3 deposits are detectable in S126G cells

To test for FD-specific intracellular Gb3 accumulations in S126G cells, fluorescently labeled STxB was used to visualize Gb3 in HDF, iPSC, and differentiated sensory neurons. As a positive control for STxB immunoreaction, we used HDF of a patient carrying the FD-related nonsense variant Q357X. While massive Gb3 deposits were visible in the positive control HDF (Fig. 2A), S126G HDF showed hardly any STxB signal similar to healthy control cells (Fig. 2B). The same was true for iPSC (Fig. 2C) and sensory neurons (Fig. 2D).

Fig. 2.

Visualization of intracellular Gb3 accumulations in healthy control and S126G cells.

(A) Representative photomicrograph of HDF carrying the variant Q357X (nonsense) in the GLA gene as positive control for STxB immunoreactivity. Q357X HDF showed massive intracellular STxB::555 signal corresponding to Gb3 deposits. (B) Representative photomicrographs of healthy control and S126G HDF immunoreacted with STxB::555. No Gb3 accumulations were detected in control and S126G HDF. (C) Representative photomicrographs of healthy control and S126G iPSC immunoreacted with STxB::555. No Gb3 accumulations were present in control and S126G iPSC. (D, left) Representative photomicrographs of control and S126G sensory neuron clusters immunoreacted with STxB::555 and the peripheral nervous system marker PRPH. No Gb3 accumulations were found in control and S126G sensory neuron clusters. (D, right) Representative photomicrographs of control and S126G single sensory neuron somas immunoreacted with STxB::555. No Gb3 accumulations were detected in control and S126G sensory neuron somas. Photomicrographs show overlay of fluorescence signal and phase contrast image. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; Gb3, globotriaosylceramide; GLA, α-galactosidase A gene; HDF, human dermal fibroblasts; iPSC, induced pluripotent stem cells; PhC, phase contrast microscopy; PRPH, peripherin; STxB, Shiga toxin subunit B.

3.4. AP firing of S126G sensory neurons equals control cells

To investigate effects of the genetic variant S126G on sensory neuron excitability, AP of iPSC-derived S126G and healthy control neurons were analyzed in current-clamp mode. Fig. 3A shows a representative AP recorded from a S126G neuron with illustration of analyzed parameters. Fig. 3B shows representative AP trains in control and S126G neurons elicited by current injection demonstrating the ability of repetitive firing in sensory neurons of both genotypes.

Fig. 3.

Depiction of AP parameters and repetitive firing.

(A) Representative AP recorded from a S126G sensory neuron. Resting membrane potential (V_Membrane), threshold potential (V_Threshold), AP duration (H_{Max width}), and amplitude are indicated. (B) Representative current clamp recordings of repetitive AP firing in control and S126 sensory neurons. AP trains were elicited by 500 ms current injection equivalent to 2× rheobase current. Abbreviations: AP, action potential.

First, we investigated the shape of the AP characterized by amplitude and duration measured at half maximum width. We found no difference in AP amplitude and duration (H_{Max width}) between S126G and control neurons, confirming normal shape of the AP in S126G neurons (Fig. 4A,B). In addition, the threshold potential, directly reflecting excitability of neurons [18], showed no difference between S126G and control neurons (Fig. 4C) while the same was true for the resting membrane potential (Fig. 4D). The minimum current needed to elicit an AP (rheobase) showed no difference between S126G and control neurons (Fig. 4E). To further assess neuronal excitability, AP were counted within 500 ms upon stimulation with multiples of the rheobase. While AP firing was lower in S126G neurons for an injected current reflecting 2× the rheobase (Fig. 4F, p < 0.05), we found no differences in AP firing between S126G and control neurons for injected currents reflecting 1×, 3×, and 4× the rheobase (Fig. 4F). Overall, no relevant pathological differences in AP shape and firing were found in S126G neurons compared to healthy control neurons.

Fig. 4.

Analysis of AP characteristics, membrane potential, and firing frequency in S126G and healthy control sensory neurons.

(A) AP amplitude (n_Control = 39, n_S126G = 39), (B) AP duration (H_{Max width}, n_Control = 39, n_S126G = 39), (C) threshold potential (V_Threshold, n_Control = 38, n_S126G = 38), and (D) resting membrane potential (V_Membrane, n_Control = 38, n_S126G = 38) showed no differences between S126G and healthy control neurons. (E) The minimum current to elicit an AP (Rheobase, n_Control = 39, n_S126G = 39) showed no difference between S126G neurons and healthy control neurons. (F) Neuronal firing frequency determined by AP counts within 500 ms of stimulation with multiples of the rheobase showed no difference between S126G and healthy control neurons for 1×, 3×, and 4× rheobase (n_Control = 30, n_S126G = 35), while firing frequency was reduced in S126G neurons for 2× rheobase (p < 0.05). Data was collected and pooled from two clones per iPSC line and a minimum of three independent differentiations. *p < 0.05. Box and whisker plots with min to max and single data points. Abbreviations: AP, action potential.

4. Discussion

FD is a life-threatening disease that requires permanent and costly treatment. Hence, correct diagnosis is critical to protect patients who carry genetic variants without relevant pathogenicity from potential adverse events, but also to determine patients in need of early FD-specific treatment by enzyme replacement [26] or chaperone therapy [74]. Besides infusion-associated symptoms such as fever [63], severe side effects including throat tightness [47], anaphylactic reaction [4,67], and atrial fibrillation [5] were reported.

VUS are a crucial challenge in the clinical management of patients with suspected FD and data available in literature as well as classification in databases is conflicting. Several genetic VUS are known in the GLA gene and are controversially discussed such as A143T and D313Y. While some studies report on cardiac involvement and reduced GLA activity in A143T [13,70], others did not observe FD-related organ involvement [57,64]. Similarly, D313Y was considered pathogenic due to reports of neurological and renal symptoms [16,76], while no organ manifestation was found in other studies [30,48]. As a result, classification of VUS in databases changes frequently and caregivers have to decide on FD-specific treatment based on the individual case. For S126G, data availability is even more limited. We report on a man who suffered from lacunar cerebral stroke and was diagnosed with FD based on the mere chance finding of the VUS S126G in the GLA gene. We have combined neurological examination with the study of patient-derived cells in vitro and provide multilevel evidence for a non-pathogenic status of S126G in this patient.

At our center, we diagnosed our patient not to have FD and no FD-specific treatment was recommended. He was instructed to get in touch if any symptoms suspicious of FD occur, which did not happen until the date of this report (i.e. in the last four years). In synopsis of all findings, we assume the lacunar cerebral stroke in our patient as etiologically cryptogenic, which is the case in 26% of all stroke patients [50]. Vice versa, screening studies for FD in young patients with cerebral stroke of unknown etiology reported a prevalence <1% based on α-GAL A activity and/or genetic testing [1,38,58]. However, considering VUS in such screening approaches can artificially inflate the prevalence up to 7% [17]. The finding of a GLA VUS in absence of FD-specific signs and symptoms, negative family history, normal lyso-Gb3, and α-GAL A activity should hence be interpreted with great caution. Particularly the latter is currently considered to exclude FD in a male person per definition.

The effect of a genetic variant on α-GAL A activity is determined by its impact on either integrity of the active site or overall conformational stability [23]. As shown by recent studies, variants residing aside from the active site and α-GAL A core are less prone to cause severe FD symptoms [53,54]. The location of serine 126 on the surface of the enzyme therefore suggests a non-pathogenic variant. In addition, we confirm in iPSC that S126G has no effect on α-GAL A activity (Fig. 1B) as was also reported before [6,49,59]. Assessing enzyme activity in vitro was considered the most accurate predictor of FD phenotype compared to plasma lyso-Gb3 and computational algorithms [42]. In contrast to α-GAL A measurement in plasma, cell-based activity assays are further insensitive to false detection of enzymatic pseudodeficiency [7,31]. However, a recent study suggested an impact of the S126G variant on enzyme trafficking towards lysosomes despite normal enzyme activity, questioning the significance of mere α-GAL A activity measurement for FD diagnosis [56]. Further investigation on both substrate and cellular function level was hence advised.

Gb3 is present in many tissues [10], however, investigating deposits in one cell type alone is not sufficient to assess the outcome of a variant. Although α-GAL A activity is determined by the variants position, translation to intracellular Gb3 deposits is likewise dependent on cell type-specific heterogeneous Gb3 metabolism. For instance, it was shown that plasma membrane expression of the receptor Gb3 is regulated by the cell cycle leading to expression differences between proliferative and non-proliferative cells and cell types [44]. We show the corresponding absence of Gb3 accumulation in HDF, epigenetically rejuvenated iPSC, and terminally differentiated sensory neurons (Fig. 2). In FD patient–derived HDF and sensory neurons, massive deposits were previously reported [32,36]. Further, genetic knockout of GLA caused intracellular accumulation in otherwise healthy iPSC [33]. Missing detection of STxB signal in all three cell types reinforces the notion of a non-pathogenic role of S126G.

Determination of IENFD with low inter- and intra-observer variability [65] is frequently used in patients with suspected small fiber neuropathy [35,40], however, reduction of skin innervation as a typical finding is not specific [19,71]. Given the reduced distal IENFD in our patient, we analyzed the in vitro excitability of sensory neurons carrying the S126G variant. We not only aimed to uncover potential abnormalities reflecting current nerve pathology, but also to predict neuronal dysfunction associated with a putative late-onset phenotype. Rheobase, threshold potential, and resting membrane potential as key parameters reflecting membrane excitability [18,41,46] were not different compared to controls (Fig. 4). Interestingly, we found a reduced AP firing frequency for S126G neurons when stimulated with 2× rheobase current, while no difference was observed for both lower and higher suprathreshold currents. In a study on gain-of-function variants of the voltage-gated sodium channel Na_v1.8, increase rather than decrease in firing frequency was directly linked to neuropathic pain [22]. Therefore, mere reduction of AP firing frequency without apparent influence on other AP parameters and only in one of the tested stimulus conditions is considered a finding without biological relevance. Together with the unremarkable neurological examination, our results add further evidence that determination of IENFD alone is not sufficient for diagnosing small fiber neuropathy and associated FD.

An intrinsic limitation of our study is the focus on one individual carrying the S126G variant. As reported before, phenotypic diversity occurs in patients sharing the same GLA variant [9]. In women, this might be ascribed to random X-inactivation and subsequent difference in cell-to-cell enzyme activity [45,73]. Since severe phenotypic variability was also found in men and even within one family [55], epigenetic modifications were proposed as an explanation [75]. However, phenotypic variability is mainly reported for variants abolishing α-GAL A activity [9,55,75], hence reflecting resilience to enzyme malfunction. Patient-to-patient variability is therefore expected to be a minor issue with variants allowing for normal enzyme activity.

We present translational data combining clinical examination and in vitro research to assess the effect of the VUS S126G in FD-associated GLA gene. Our findings suggest no pathological effects of this variant on enzyme activity and overall cellular fate and add to the evidence categorizing the S126G as benign.

Author contributions

Maximilian Breyer: Methodology, Formal analysis, Investigation, Writing – Original Draft, Visualization. Julia Grüner: Methodology, Investigation. Alexandra Klein: Methodology, Formal analysis, Investigation. Laura Finke: Investigation. Katharina Klug: Methodology. Markus Sauer: Project administration, Funding acquisition. Nurcan Üçeyler: Conceptualization, Resources, Writing – Original Draft, Supervision, Project administration, Funding acquisition.

Funding

This project was supported by the Interdisciplinary Center for Clinical Research Würzburg (Interdisziplinäres Zentrum für Klinische Forschung (IZKF); N-375) and by the German Research Foundation (Deutsche Forschungsgemeinschaft (DFG), Sonderforschungsbereich SFB1158). J.G. was supported by DFG (KFO5001 ResolvePAIN, project ID: 426503586). N.Ü. was supported by DFG (UE171/15-1).

Institutional review board statement

The study was approved by the Ethics Committee of the University of Würzburg Medical Faculty (#26/19) on 11 September 2019.

Informed consent statement

Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient to publish this paper.

CRediT authorship contribution statement

Maximilian Breyer: Conceptualization, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft. Julia Grüner: Formal analysis, Methodology, Validation, Writing – review & editing. Alexandra Klein: Formal analysis, Methodology, Writing – review & editing. Laura Finke: Formal analysis, Methodology, Writing – review & editing. Katharina Klug: Formal analysis, Methodology, Writing – review & editing. Markus Sauer: Conceptualization, Funding acquisition, Supervision, Writing – review & editing. Nurcan Üçeyler: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – original draft.

Author statement

Data is available upon request from the corresponding author.

Declaration of Competing Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.