Variants in NGLY1 lead to intellectual disability, myoclonus epilepsy, sensorimotor axonal polyneuropathy and mitochondrial dysfunction

Abstract NGLY1 encodes the enzyme N‐glycanase that is involved in the degradation of glycoproteins as part of the endoplasmatic reticulum‐associated degradation pathway. Variants in this gene have been described to cause a multisystem disease characterized by neuromotor impairment, neuropathy, intellectual disability, and dysmorphic features. Here, we describe four patients with pathogenic variants in NGLY1. As the clinical features and laboratory results of the patients suggested a multisystem mitochondrial disease, a muscle biopsy had been performed. Biochemical analysis in muscle showed a strongly reduced ATP production rate in all patients, while individual OXPHOS enzyme activities varied from normal to reduced. No causative variants in any mitochondrial disease genes were found using mtDNA analysis and whole exome sequencing. In all four patients, variants in NGLY1 were identified, including two unreported variants (c.849T>G (p.(Cys283Trp)) and c.1067A>G (p.(Glu356Gly)). Western blot analysis of N‐glycanase in muscle and fibroblasts showed a complete absence of N‐glycanase. One patient showed a decreased basal and maximal oxygen consumption rates in fibroblasts. Mitochondrial morphofunction fibroblast analysis showed patient specific differences when compared to control cell lines. In conclusion, variants in NGLY1 affect mitochondrial energy metabolism which in turn might contribute to the clinical disease course.

Western blot analysis of N-glycanase in muscle and fibroblasts showed a complete absence of N-glycanase. One patient showed a decreased basal and maximal oxygen consumption rates in fibroblasts. Mitochondrial morphofunction fibroblast analysis showed patient specific differences when compared to control cell lines. In conclusion, variants in NGLY1 affect mitochondrial energy metabolism which in turn might contribute to the clinical disease course. This overlap is evident in recently published patients with an impaired function of N-glycanase 1 (NGLY1). This genetic defect causes a broad clinical spectrum including hypo-/alacrima, developmental delay, elevated liver enzymes, diminished deep tendon reflexes, and epileptic seizures. [3][4][5][6] NGLY1 is an enzyme involved in the removal of N-glycans from glycoproteins, which precedes the proteolytic degradation of glycosylated proteins via the endoplasmatic reticulum-associated degradation (ERAD) pathway. 7 In two recent articles, it was suggested that NGLY1 is involved in mitochondrial function in Caenorhabditis elegans, mouse, and human and is regulated through nuclear respiratory factor 1 (NRF1). 6,8 Here we report four patients with biallelic variants in NGLY1 who presented with myoclonus epilepsy, peripheral neuropathy, and metabolic markers suggestive for mitochondrial dysfunction. Molecular testing revealed further mitochondrial morphological and functional alterations which have not yet been described for this patient group.
Taken together, these results provide evidence for a possible role for NGLY1 in mitochondrial function.

| Pathology and functional mitochondrial measurements in different tissues
Routine histology and histochemistry were performed in muscle following standard methods. 9 The ATP production from pyruvate oxidation in fresh muscle and the activity of the mitochondrial complexes I to V, citrate synthase, and protein concentration were measured in fresh muscle biopsies, cultured fibroblasts and liver tissue as described previously. 10

| Whole exome sequencing
Whole exome sequencing (WES) and data analysis were performed as described before. 11,12 In short, exome enrichment was performed using the SureSelect Human All Exon 50 Mb Kit V5 (Agilent).
Sequencing was done on a HiSeq4000 (Illumina) with a minimum median coverage of ×80. Read alignment to the human reference genome (GrCH37/hg19) and variant calling was performed at BGI (Copenhagen) using BWA and GATK software, respectively. Variant annotation was performed using a custom designed in-house annotation. Intronic variants (except for splice sites), synonymous changes, and common variants were filtered and excluded from the initial datasets. Patient data were first analyzed using a custom-made virtual gene panel containing mitochondrial disease genes (as described in OMIM) as well all other genes known to encode mitochondrial proteins. As no disease-causing variants were detected, the entire exome was investigated for rare, protein damaging variants. This was done by comparison with the GnomAD dataset, dbSNPv132 and our inhouse variant database with MAF depending on mode of inheritance.

| Western blotting
Western Blot analysis was performed on 600g supernatant from muscle homogenate and fibroblast homogenates (40 μg per lane) from patients and healthy controls. For fibroblasts homogenates, cell pellets containing 5Á10 6 cells were resuspended 1% Triton in 10 mM Tris-HCl (pH 7.6) followed by centrifugation at 4 C 14 000g. Protein concentrations in the supernatant were measured using the U/CSF protein kit (Thermo Fischer Scientific) in the Kone-Lab 20XT semiautomated platform (Thermo Fischer Scientific, Passau, Germany).

| Oxygen consumption measurements
Oxygen consumption rates (OCRs) were measured using the Seahorse XFe96 Extracellular Flux analyzer (Agilent). Control and patient primary skin fibroblasts were seeded at 15 000 per well in cell culture medium (M199 supplemented with 10% v/v FCS (Gibco/Life Technologies) and 1% v/v penicillin/streptomycin (Gibco/LifeTechnologies)) and grown overnight at 37 C with 5% CO 2 . One hour before measurement, culture medium was removed and replaced by Agilent Seahorse XF Base Medium complemented with 10 mM glucose (Sigma), ified for the Seahorse 96 wells format. In short, after completion of OCR measurements, the Seahorse assay medium was replaced by 0.33% Triton X-100, 10 mM Tris-HCl (pH 7.6), after which the plates were stored at −80 C. Before measurements, the plates underwent two thaw-freeze cycles and 3 mM acetyl-CoA, 1 mM 5,5 0 -dithiobis-2-nitrobenzoic acid (DTNB), and 10% Triton X-100 was added. Using a Tecan Spark spectrophotometer (Tecan, Switzerland), background conversion of DTNB was measured at 412 nm and 37 C for 10 minutes at 1 minute intervals. Hereafter, 10 mM of the citrate synthase substrate oxaloacetate was added to start the reaction. Subsequently, the ΔA412 nm was measured again for 10 minutes at 1 minute intervals at 37 C. Citrate synthase activity was calculated from the rate of DTNB conversion in the presence of substrate, from which the background DTNB conversion rate was subtracted, using an extinction coefficient of 0.0136 μmol/cm. Basal respiration was calculated by subtracting the respiration after addition of rotenone and antimycin A from the first four measurements. For each individual cell line, maximal OCR was acquired by titrating FCCP, with a range from 3 to 6 μM. The highest OCR was then used to calculate maximal OCR by subtracting the respiration after addition of rotenone and antimycin A. The average maximal and basal respiration of two control cell lines and individual patient fibroblasts were compared using a Student's t-test; P < .05 was considered to be statistically significant. Respirometry are displayed as mean ± SD. (v/v) FBS and 1% (v/v) pen/strep in a humidified atmosphere (95% air, 5% CO 2 ) at 37 C up to passage number. Cells were grown to 100% confluence before they were seeded at a density of 7000/dish (FluoroDishes; #FD35-100; World Precision Instruments Ltd., Friedberg Germany). Following 5 days of culturing to allow cell spreading, the M199 medium was replaced by a colorless M199 variant without phenol red (#11043-023; Invitrogen). Then, DMSOdissolved tetramethylrhodamine methyl ester (TMRM; #T668; Life Technologies Thermo Fisher Scientific, Waltham, Massachusetts) was added (15 nM; 25 min) to stain the cells (in the dark; humidified atmosphere; 95% air; 5% CO 2 ) at 37 C. Next, dishes were placed on a video-microscope described in detail previously. 13 Fluorescence images were acquired in TMRM non-quenching mode within 15 min after staining 14 using a ×40 objective, 100 ms exposure time, 540 nm excitation, 560DRLP dichroic mirror and 565ALP emission filter.
Microscopy images were background corrected and processed to allow subsequent calculation of mitochondrial morphofunctional parameters. 15 In this analysis, only objects larger than 10 pixels were considered to represent mitochondria. Morphofunctional parameters included Am (a measure of mitochondrial size in pixels), AR (or aspect ratio, a measure of mitochondrial length), F (formfactor, a combined measure of mitochondrial length and degree of branching) and Dm (density mean, reflecting the mitochondrial TMRM fluorescence intensity). For quality control (QC) analysis, a control cell line (CT5120) from a healthy individual was included in all TMRM experiments. If the measured CT5120 parameter values corresponded with those that were previously obtained, the experiment was considered valid and included in the analysis. 15

| RESULTS
Three female patients, two of Dutch and one of Moroccan ancestry, and one male patient of Dutch ancestry were evaluated from an early age onwards because of moderate to severe intellectual disability, developmental delay, muscular hypotonia, and extra pyramidal movements of the extremities (summarized in Table 1).
During MRS, patient 3 received topiramate to treat epilepsy. Brain MRI showed delayed myelination and a diffuse lack of white matter volume in patient 1 ( Figure 1E). Brain MRI of patient 4, performed at age 3 ( Figure 1F) and 8 years ( Figure 1G), showed incomplete myelination and slowly progressive global atrophy. Patient 1 has deceased at the age of 8 years because of respiratory insufficiency.
The family histories of these patients were unremarkable.
As the clinical and metabolic symptoms were suggestive for a mitochondrial disease, all four patients underwent a muscle and skin biopsy in order to investigate mitochondrial function. All F I G U R E 3 Western blot analysis. A, Immunoblot analysis of NGLY1 in muscle extracts from four patients and three controls. The anti-NGLY1 antibody stains a 74-kDa band in the three controls, whereas this band is absent in the four patients. Anti-SDHA (complex II) was used as loading control. B, Immunoblot analysis of NGLY1 in fibroblasts from four patients and three controls. Similar to the muscle sample, a NGLY1 band is present in the control samples, whereas this band is absent in the four tested patients. Anti-SDHA (complex II) was used as loading control patient 2 and patient 4 ( Figure 5C). This suggests that mitochondrial size is normal in patient 1 and 3 but reduced in patient 2 and 4. In general, the patient cells displayed no major alterations in AR, although this parameter was increased in patient 3 ( Figure 5D).
Finally, F appeared reduced in patient 2, patient 3 and patient 4, but was normal in patient 1 ( Figure 5E). The latter result (combined with the results of Am and AR) suggests that mitochondria are smaller and less branched in patient 2 and patient 4, whereas this is not the case for patient 1 and patient 3.

| DISCUSSION
WES is widely and successfully used to identify the causative genetic defects in patients with all kinds of clinical signs and symptoms. This often leads to a broadening of the clinical spectrum of known disorders.
There is a growing number of disorders in which mitochondrial dysfunction is observed but the primary defect cannot be directly linked to known mitochondrial proteins or processes. 22 The four patients described in this study presented with a multi-system disorder including sensorimotor axonal peripheral neuropathy, myoclonic epilepsy which were highly suggestive for a mitochondrial disorder, even more so when taking the metabolic alterations in blood and urine into account (  To date, the proteins that make up the OXPHOS system have never found to be N-glycosylated. This might suggest that NGLY1 influences mitochondrial function in a secondary manner. 26 It has been shown that AMFR, an E3-ubiquitin ligase involved in the ERAD pathway, localizes to mitochondria associated ER membranes and plays a role in the degradation of Mfn1 and Mfn2 mitochondrial fusion proteins. 27 This might indicate that NGLY1 is involved in the degradation of mitochondrial outer membrane proteins. Furthermore, it was shown that mitochondrial function is influenced in NGLY1 patients via glycosylated Nrf1. 28 This article suggests that the mitochondrial network, visualized with an anti-body directed against Hsp60, is fragmented in patient fibroblasts and Ngly1 −/− mouse embryonic fibroblasts. The researchers suggest that the mitochondrial network fragmentation is due to absence of NGLY1 which alters the glycosylation of Nrf1, hampering its function and resulting in a fragmented mitochondrial network. 28 However, we did not note fragmentation of the mitochondrial network when assessing the mitochondrial morphology in patient fibroblasts as observed previously. 8 To date, 13 patients have been investigated with a pathogenic variant in NGLY1 and with clinical signs and symptoms suggestive of a mitochondrial disorder, including intellectual disability, involuntary movements, and muscular hypotonia. [3][4][5][6] However, no conclusive biochemical evidence for a mitochondrial defect has been provided yet.
Here, we describe four patients that show isolated or combined defects in the OXPHOS enzymes combined with a reduction in ATP production in muscle, strongly suggestive of a mitochondrial disorder.
Differences were observed in mitochondrial function in patient fibroblasts, showed that there is a functional relationship between NGLY1 and mitochondria. However, further investigation is needed to further dissect the molecular mechanism connecting NGLY1 to mitochondrial function.