GGPS1 ‐associated muscular dystrophy with and without hearing loss

Abstract Ultra‐rare biallelic pathogenic variants in geranylgeranyl diphosphate synthase 1 (GGPS1) have recently been associated with muscular dystrophy/hearing loss/ovarian insufficiency syndrome. Here, we describe 11 affected individuals from four unpublished families with ultra‐rare missense variants in GGPS1 and provide follow‐up details from a previously reported family. Our cohort replicated most of the previously described clinical features of GGPS1 deficiency; however, hearing loss was present in only 46% of the individuals. This report consolidates the disease‐causing role of biallelic variants in GGPS1 and demonstrates that hearing loss and ovarian insufficiency might be a variable feature of the GGPS1‐associated muscular dystrophy.

Bioimaging Cluster of Excellence (MBExC) (BV). L.P. was supported by an MRC strategic award to establish an International Centre for Genomic Medicine in Neuromuscular Diseases (ICGNMD) MR/S005021/1. We acknowledge the support of the MRC Centre for Neuromuscular Disease Biobank and HSS England National Specialist Diagnostic and Advisory Service for Congenital Muscular Dystrophies and Congenital Myopathies. The research leading to these results has received funding from the Muscular Dystrophy Campaign grant "Understanding the function of novel congenital muscular dystrophies and congenital myopathies genes" (Grant reference number: 18GRO-PG24-0271) and from the European Union's Horizon 2020 research and innovation programme grant "SOLVE-RD: solving the unresolved Rare Diseases" (grant agreement No 779257). Sequencing and analysis of Family 3 were provided by the Broad Institute of MIT and Harvard Center for Mendelian Genomics and was funded by the National Human Genome Research Institute, the National Eye Institute, and the National Heart, Lung and Blood Institute grant UM1 HG008900, and in part by National Human Genome Research Institute grants U01 HG0011755 and R01 HG009141. For the purpose of open access, the author has applied a CC BY public copyright license to any author accepted manuscript version arising from this submission.

Introduction
Geranylgeranyl diphosphate synthase 1 (GGPS1) is a member of the prenyltransferase family and encodes a protein with geranylgeranyl diphosphate (GGPP) synthase activity. The enzyme catalyzes the synthesis of GGPP from farnesyl diphosphate and isopentenyl diphosphate. GGPP is an important molecule responsible for the C20prenylation of proteins. Prenylation provides proteins with a hydrophobic C terminus, the consequence of which is a greatly increased capacity to interact with cellular membranes, which have a high concentration of signaling molecules. 1,2 Prenylation is particularly important for small guanosine triphosphatases (GTPases) such as the Rab family, which have crucial roles in vesicular trafficking in mammalian cells. 2 The roles of prenylated proteins in cells are well conserved across species. 1 While monoallelic variants in GGPS1 have been identified as a risk factor for atypical femoral bone fractures in females exposed to bisphosphonates, ultra-rare biallelic pathogenic missense variants in GGPS1 have recently been associated with muscular dystrophy/hearing loss/ovarian insufficiency syndrome. 2,3 Here, we describe 11 affected individuals from four previously unreported British, Iranian, Egyptian, and Brazilian families harboring ultra-rare missense variants in GGPS1 and provide follow-up clinical data from the previously reported two Pakistani cases by Tucker et al. (2020). 3

Subjects and Methods
Two of the novel families reported here were recruited from Iran and Egypt as part of the SYNaPS project at the Institute of Neurology, University College London (UCL). The third and fourth novel families were identified through a collaboration between Great Ormond Street Institute of Child Health (UCL) and the Center for Mendelian Genomics, Broad Institute, USA and School of Medicine of Universidade de Sao Paulo, respectively. The corresponding author was contacted from Tucker et al. (2020) 3 to obtain follow-up details from the reported 2 persons with homozygous GGPS1 variants. To identify the genetic cause of the disease in the affected individuals, exome sequencing on DNA extracted from probands' leucocytes and variant filtering were performed as described (Table 1 for genetic methods). To assess Ggps1 RNA expression in the mouse cochlear epithelium at various stages (embryonic (E) 14, and postnatal (P) 1 and P7), publicly available single cell RNA-seq data were visualized using the gene expression analysis resource (gEAR). 7 The publicly available Mouse Cell Atlas, generated as previously described, 8 includes single-cell RNA sequencing data used to visualize global expression of Ggps1 in the female E14.5 gonad (version 2.0 with >520,000 cells from over 10 mouse tissues grouped into 95 major clusters) and neonatal skeletal muscle excised from the leg (version 1.0 with >400,000 cells from over 10 mouse tissues grouped into 104 major clusters). Clusters showing expression (red) are marked in black numbers representing the cluster code defined on the Mouse Cell Atlas. 8 The study was approved by the ethics Institutional Review Board of UCL and additional local ethics committees of the participating centres. Informed consent was obtained from all families.

Results
The affected individuals are six males and seven females, all of whom were born at full term, with all but two (P10 and P11) conceived from consanguineous unions (Fig. 1A). Eleven affected individuals are alive with a mean current age of 10 AE 6 years (range 4-24), and two individuals died at the ages of 11 months and 30 years old due to respiratory insufficiency and choking, respectively. Antenatal history was unremarkable in all but 3/13 persons who displayed decreased foetal movements. Tables 1 and 2, Table S1, and Videos S1-S6 provide details on the clinical phenotype of the present cohort. The disease onset ranged from prenatal (3/13) to preschooler age (mean age 16 AE 9.4 months, range 4-36). Delayed motor milestones (10/13), weak cry (3/13) and muscle weakness (3/13) were the first presenting symptoms. Mean ages were 9.7 AE 1.7 months for sitting, 1.7 AE 0.3 years for walking, and 9.9 AE 2.9 months for first words. Proximal or generalized muscle weakness (12/ 13) with the mean onset age of 2.7 AE 2.7 years (range 0.6-11), respiratory insufficiency (8/13) requiring noninvasive ventilation (6/13) and tracheostomy (2/13), joint contractures (7/12), congenital or infantile-onset bilateral severe sensorineural hearing loss (6/13) (only present in Families 1, 3, 4, and 5, while in the rest of the persons the absence of sensorineural hearing loss was clinically and objectively confirmed), and gastrointestinal issues (4/ 13) were among the commonly reported symptoms throughout the disease course (Tables 1 and 2). The age of onset for respiratory insufficiency in the affected individuals ranged between 7 months to 20 years old. Only one female person has reached puberty, and at 30 years old, she does not have clinical signs of primary ovarian insufficiency, while follicle-stimulating hormone levels were reported to be high in P11, aged 12 years old. Six affected individuals lost ambulation at the mean age of 8.1 AE 4.7 years. Ninety-two percent (12/13) of individuals displayed slow rates of disease progression. One person (P10) demonstrated improvement in motor abilities and dysphagia symptoms in childhood. Older individuals in our series tended to have a more severe phenotype reflecting the progressive nature of the disease in most instances.
On exome sequencing, an ultra-rare homozygous missense variant in exon 4 of GGPS1 c.269A > G, p.      conserved GGPS1 protein region. The variant is absent in gnomAD and a number of large publicly available databases apart from 2 heterozygous alleles from the Centogene and TOPMed databases, and is predicted to be damaging by multiple in silico tools. (Table S2). The variant segregated with the phenotype within the family (Fig. 1A). Both Family 1 and Family 5 from Tucker et al.
(2020) carried the same recurrent homozygous GGPS1 c.269A > G; p.(Asn90Ser) variant. An ultra-rare homozygous missense variant in exon 4 of GGPS1 c.439A > G, p.(Met147Val) (NM_004837.4) residing within a 9 Mb region of homozygosity was identified in Family 2 (Fig. S1). This variant is located in a highly conserved GGPS1 protein region with predicted-damaging scores on various in silico tools (Table S2). The variant segregated with the phenotype within the family (Fig. 1A). Based on the extremely rare frequency of the identified GGPS1 variants (PS4, PM2), co-segregation in two unrelated families with multiple affected individuals for the GGPS1 c.269A > G, p.(Asn90Ser) variant (PP1 moderate) and segregation in multiple affected family members from different branches for the GGPS1 c.439A > G, p.(Met147-Val) variant (PP1 moderate), as well as predicteddeleterious effect from multiple in silico tools, we classified both variants from Family 1 and 2 as likely pathogenic according to the American College of Medical Genetics and Genomics (ACMG) guidelines for the interpretation of sequence variants. 9 Two ultra-rare missense variants in exon 4 of GGPS1 were identified, in trans, in the proband of Family 3 (P10), c.196A > C p.(Ile66Leu) and c.545T > C p.(Leu182Pro) (Fig. 1). Both variants are absent from publicly available databases and are conserved across species (Table S2) (Fig. S2). The individual's phenotype was felt to be highly specific for GGPS1 related muscular dystrophy (Table 1). Owing to the rarity of the variants, and lack of further segregation and functional study data, these variants are classified as of uncertain significance according to ACMG guidelines. However, given the affected person's phenotype and proximity of the affected residues to key domains within GGPPS we feel they can be considered disease causing in P10.
An ultra-rare missense variant c.770T > G; p.(Phe257Cys) was identified in Family 4. The variant is absent from all publicly available databases, is conserved across species, and is predicted damaging from a variety of in silico tools (Table S2). The variant segregated with the phenotype within the family (Fig. 1A).
Expression patterns of Ggps1 in the mouse inner ear, skeletal muscle, and female gonad at different embryonic and postnatal stages were studied using single-cell RNA sequencing. Single-cell RNA sequencing of the cochlear epithelium in the E14, P1, and P7 mouse shows Ggps1 with a sustained, diffuse expression at these timepoints (Fig. S3). Global view of Ggps1 expression in the E14.5 female gonad showed clusters most densely in the germ cells and embryonic gonad cells (Fig. S4), whereas neonatal skeletal muscle showed several clusters reflecting comparatively lower expression, for example, in clusters representing stromal cells and chondrocytes (Fig. S5).

Discussion
Protein prenylation is an important downstream part of the mevalonate pathway, which is a pivotal process for the synthesis of specialized lipids that play a crucial role in cellular processes at all levels. 1,2,10 The mevalonate pathway has recently been suggested to be a novel pathway for muscular dystrophy, hearing loss, and infertility. 2 Evidence for this comes from the description of 11 individuals from 6 independent families presenting with early   Biallelic pathogenic missense GGPS1 variants identified in Foley et al. (2020) 2 have been shown to moderately impair the enzymatic activity of geranylgeranyl diphosphate synthase (GGPPS) (only for 50%) in vitro, and the possible effects of these missense variants on GGPPS function have been speculated to be subtle, putatively including the impairment of the dynamic subcellular localization of the enzyme for cell-type-specific processes, particularly in muscle, ovary, and the inner ear. It had been suggested that the localization of the missense GGPS1 variants residing outside of the catalytic core of the enzyme could account for the very specific and recognizable clinical phenotype. However, the variants reported in P10 (Family 3) co-locate within the catalytic core of GGPPS and implicate impaired substrate and co-factor binding as potential pathogenic mechanisms, resulting in a phenotype consistent with other reported individuals.
Family 2 in our cohort presents hitherto the largest family (6 affected) with biallelic GGPS1 variants reported. Affected individuals from our report replicated most of the previously described clinical features associated with GGPS1 defects. This includes muscular dystrophy with progressive proximal muscle weakness and episodes of variably elevated CK, scoliosis, and respiratory insufficiency. Muscle histopathology findings were also variable but largely consistent with those previously reported by Foley et al. (2020), 2 including dystrophic changes, regenerating fibers, core-like lesions, internal nucleation, and mitochondrial changes. The lack of dystrophic histopathology and fat infiltration observed in the muscle biopsy and lower limb MRI of Family 3 (P10) may reflect the early age (8 months) at which these investigations were performed, compared to other reported persons. The families described in our report display inter-and intra-familial variability with respect to hearing loss, which was detected only in all affected siblings from Family 1, the probands from Families 3 and 4, and one affected sibling from Family 5. This could suggest that hearing loss might be a variable feature of GGPS1-associated muscular dystrophy and is an interesting observation in light of its diffuse expression in the embryonic and postnatal mouse inner ear. Since most of the subjects with unimpaired hearing in the current report were from Family 2, carrying the GGPS1 c.439A > G, p.(Met147Val) variant, we cannot exclude the possibility of a variant-specific effect. Additionally, our data shows that primary ovarian insufficiency might also be a variable feature of GGPS1 deficiency, despite strong Ggps1 expression in relevant cells in the embryonic mouse female gonads. Although it is difficult to determine the exact molecular cause of the observed clinical variability, we can hypothesize a possible link with localisation of the GGPS1 variants within the gene. One possible limitation of the present study is the absence of in vitro enzyme activity measurements. However, taking into account the findings from the study by Foley et al. (2020), 2 one could expect that the impact of the missense variants from our study on GGPPS function could be subtle as well, with an insignificant reduction of the enzymatic activity of GGPPS. All missense GGPS1 variants c.269A > G, p.(Asn90Ser), c.439A > G, p.(Met147Val), c.196A > C p.(Ile66Leu), c.545 T > C p.(Leu182Pro) and c.770 T > G p(Phe257Cys) reported in the current study are ultra-rare across several large genetic databases including 1,314,000 alleles (Table S2) and uniformly predicted to be damaging in various in silico prediction tools supporting their disease-causing roles.
Collectively, this report consolidates the disease-causing role of biallelic variants in GGPS1, demonstrates that hearing loss and ovarian insufficiency might be variable features of the GGPS1-associated muscular dystrophy and implicates impaired substrate and co-factor binding as potential pathogenic mechanisms in this novel form of congenital muscular dystrophy.

Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article.   Figure S4 Expression of Ggps1 in the mouse embryonic (E14.5) gonad through single-cell RNA-sequencing data. Figure S5 Expression of Ggps1 in mouse neonatal leg muscle through single-cell RNA-sequencing data. Table S1 Extended clinical features of the cases with biallelic GGPS1 variants. Table S2 Population frequencies and in silico pathogenicity predictions for GGPS1 variants reported in this study.
Video S1 Shows P1 with a waddling gait, lower limb weakness, and inverted feet. Video S2 Shows P4 who is unable to stand unassisted and walk due to the lower limb weakness. Video S3 Shows P5 with a suggestion of Gower's sign and waddling gait. Video S4 shows P6 with proximal lower limb weakness. Video S5 shows P7 with positive Gower's sign. Video S6 shows P8 with positive Gower's sign.