Autosomal Recessive Cerebellar Atrophy and Spastic Ataxia in Patients With Pathogenic Biallelic Variants in GEMIN5

The hereditary ataxias are a heterogenous group of disorders with an increasing number of causative genes being described. Due to the clinical and genetic heterogeneity seen in these conditions, the majority of such individuals endure a diagnostic odyssey or remain undiagnosed. Defining the molecular etiology can bring insights into the responsible molecular pathways and eventually the identification of therapeutic targets. Here, we describe the identification of biallelic variants in the GEMIN5 gene among seven unrelated families with nine affected individuals presenting with spastic ataxia and cerebellar atrophy. GEMIN5, an RNA-binding protein, has been shown to regulate transcription and translation machinery. GEMIN5 is a component of small nuclear ribonucleoprotein (snRNP) complexes and helps in the assembly of the spliceosome complexes. We found that biallelic GEMIN5 variants cause structural abnormalities in the encoded protein and reduce expression of snRNP complex proteins in patient cells compared with unaffected controls. Finally, knocking out endogenous Gemin5 in mice caused early embryonic lethality, suggesting that Gemin5 expression is crucial for normal development. Our work further expands on the phenotypic spectrum associated with GEMIN5-related disease and implicates the role of GEMIN5 among patients with spastic ataxia, cerebellar atrophy, and motor predominant developmental delay.

The hereditary ataxias are a heterogenous group of disorders with an increasing number of causative genes being described. Due to the clinical and genetic heterogeneity seen in these conditions, the majority of such individuals endure a diagnostic odyssey or remain undiagnosed. Defining the molecular etiology can bring insights into the responsible molecular pathways and eventually the identification of therapeutic targets. Here, we describe the identification of biallelic variants in the GEMIN5 gene among seven unrelated families with nine affected individuals presenting with spastic ataxia and cerebellar atrophy. GEMIN5, an RNA-binding protein, has been shown to regulate transcription and translation machinery. GEMIN5 is a component of small nuclear ribonucleoprotein (snRNP) complexes and helps in the assembly of the spliceosome complexes. We

INTRODUCTION
Hereditary ataxias are a heterogenous group of neurological disorders affecting individuals of all age groups, and patients present with a disturbance in movement coordination. While cerebellar abnormalities are usually implicated, lesions elsewhere in the neuroaxis can lead to gait abnormalities that mimic the clinical phenotype. In addition, it is well recognized that many of the autosomal recessive cerebellar ataxias can present as spastic ataxias with pyramidal signs (Ruano et al., 2014). While traditionally these were clinically distinct entities, a spectrum of diseases ranging in clinical presentation from pure cerebellar ataxias to spastic paraplegias have been observed more recently (Synofzik and Schüle, 2017). The genetic causes of hereditary cerebellar ataxias and hereditary spastic ataxias are heterogenous, and often next-generation sequencing technologies are used to establish a molecular diagnosis (Németh et al., 2013). To date, more than 150 genes (OMIM categories include: 45 genes as autosomal dominant spinocerebellar ataxias, 29 autosomal recessive spinocerebellar ataxias, 8 spastic ataxias, and 73 spastic paraplegias) have been implicated in the ataxia paraplegia spectrum of diseases (McKusick, 2007). The identification of the molecular diagnosis in these patients is the first step toward understanding the molecular mechanism and is crucial for developing effective targeted therapeutic strategies.
GEMIN5 is an RNA-binding protein involved in regulating multiple aspects of transcriptional and translational processes Martinez-Salas et al., 2020;Moreno-Morcillo et al., 2020). The GEMIN5 protein has 13 tryptophan-aspartic acid dipeptides (WD domains) at the N-terminus, followed by a dimerization domain, and a noncanonical RNA-binding site (RBS1) domain Martinez-Salas et al., 2020;Moreno-Morcillo et al., 2020). GEMIN5 is predominantly a cytoplasmic protein with sparse expression in the nucleus, suggesting that this protein might be involved in regulating functions in both cellular compartments. The WD domain of GEMIN5 assists in the biogenesis of small nuclear ribonucleoproteins (snRNPs), the building blocks of splicing machinery, and aids in the formation of the snRNP complex along with the other complex proteins (SMN, Gemin2-8 and Unrip) (Lozano et al., 2018;Francisco-Velilla et al., 2020). Moreover, the noncanonical RNA-binding sites (RBS), located on the C-terminus, promote GEMIN5 to interact with various mRNAs and regulate their translation. Importantly, mutations affecting the survival motor neuron (SMN) protein have been linked with a neuromuscular disorder, spinal muscular atrophy (SMA), which has been linked with disruption of snRNP complex assembly (Melki et al., 1994;Lefebvre et al., 1995;Lefebvre et al., 1997;Burlet et al., 1998;Pellizzoni et al., 2002). The GEMIN5 protein has numerous functions as it incorporates into cytoplasmic stress granules, binds to ribosomes, and regulates global translation and SMN expression.
Here, we define a novel neurodevelopmental ataxia syndrome in patients with autosomal recessive variants in the GEMIN5 gene. We provide characterization of their neurological phenotype to include a spectrum of the cerebellar ataxia and spastic paraplegia phenotypes among seven unrelated families with nine affected individuals. We employ computational biology, biochemistry, and mouse genetics to examine the functional consequences of perturbing GEMIN5 protein in vitro and in vivo. Our data suggest that loss-of-function of GEMIN5 is detrimental in human patient cells and in mice.

Patient Recruitment
GEMIN5 variants were identified in probands by whole-exome sequencing in clinical diagnostic settings at different sites. We found a subset of our GEMIN5 patients through GeneMatcher (Wang et al., 2007). All patients were evaluated by a neurologist or geneticist at their respective referral centers. We reviewed the clinical information, neurological symptoms, examination, and radiological studies including brain MRI and other clinical evaluations in each patient. All patient information was deidentified. Informed consent was obtained from patients for publication at each site per local institution requirements by the authors.

Genetic Studies and Variant Assessment
Whole-exome sequencing (WES) was performed at different genetic centers using next-generation sequencing techniques, and all variants were confirmed via Sanger sequencing with standard methods.
Family 1: The trio was sequenced at the Yale Center for Genome Analysis (YCGA). Genomic DNA was captured using an IDT xGen exome kit followed by Illumina DNA sequencing. WES data were processed using two independent pipelines at the Family 5: Clinical exome sequencing was performed as described previously (Cousin et al., 2019). The proband and his parents provided written informed consent to a study approved by the Mayo Clinic Institutional Review Board. Family 5 sequence reads were mapped to the reference genome (GRCh37) with BWA-MEM and further processed using GATK Best Practice workflows, which include duplication marking, indel realignment, and base quality recalibration. Single-nucleotide variants and small indels were called with GATK HaplotypeCaller. The compound heterozygous variants originally identified and Sanger confirmed by the clinical laboratory for Family 5 were called with high confidence using the workflow implemented for Family 1. Family 6: Exome sequencing was done as per previously described protocol (Guillen Sacoto et al., 2020;Kour et al., 2021). All the variants are annotated by using the GEMIN5 NP_056280.2 reference transcript in GnomAD and the other databases to estimate the allelic frequency.
Family 7: Whole-exome sequencing (WES) was performed in the two affected siblings, using DNA from peripheral blood leukocytes. The exome was enriched using the SureSelectXT V5 exome kit (Agilent, Böblingen, Germany) and sequenced on a HiSeq2500 sequencer (Illumina, San Diego, CA, United States). High-quality reads were mapped against the human reference genome (hg19), and variants were called following the Genome Analysis Tool Kit version 3.3.0 best practice recommendations (McKenna et al., 2010;Ma et al., 2013). After annotation with Annovar, filtration and downstream analysis was done with FILTUS (Vigeland et al., 2016). The identified GEMIN5 variants were validated by Sanger sequencing.
The damaging index of GEMIN5 variants was determined by using various in silico prediction tools, such as Polyphen2, Provean, SNAP2, MUpro, PhD SNP, and SIFT (Cheng et al., 2006;Ng and Henikoff, 2006;Bromberg and Rost, 2007;Mi et al., 2010;Choi et al., 2012;Sim et al., 2012). Candidate variants were validated by Sanger sequencing and tested for cosegregation in all family members whenever samples were available (Supplementary Figure S1).
All the variants were annotated by using the GEMIN5 NP_056280.2 reference transcript to estimate the allelic frequency in GnomAD and the other databases. The effect of GEMIN5 variants on its structure and folding was predicted by using the molecular graphics tool PyMOL.

Immunoblotting
PBMCs were isolated from whole blood of the proband and the unaffected parents (family 5) by following the Mayo Clinic's Biospecimen Accessioning and Processing (BAP) core's standard protocol with a target count of 10 million cells. The isolated PBMCs were lysed in the buffer containing 150 mM NaCl, 50 mM NaF, 0.1% SDS, 1% NP40, 2 mM EDTA, 1% sodium deoxycholate, 1 mM DTT, 0.2 mM sodium orthovandate, and protease inhibitor (Roche) for 10 min on ice. The protein lysates were sonicated and centrifuged at 16,000 × g for 10 min to remove the debris. The protein concentration of each sample was determined by using Pierce BCA protein assay kit (Thermo-Scientific). Of the total protein, 40 µg was separated in 4%-12% NuPAGE gel (Novex/Life technologies) and transferred to a nitrocellulose membrane (Invitrogen). The membranes were blocked in 2.5% milk (BLOT-QuickBlocker ™ EMD Millipore) in TBST followed by overnight incubation with primary antibody at 4°C. Membranes were washed, incubated with secondary antibody for 1 h at room temperature, and imaged on Odyssey ® CLx (LI-COR Biosciences). The band intensities were measured using Image Studio ™ (LI-COR Biosciences).
All the primary and secondary antibodies were prepared in 2.5% blocking buffer.

Identification of Bi-Allelic GEMIN5 Variants
Here, we report nine patients with spastic ataxia and cerebellar atrophy ( Figure 1). All patients had extensive metabolic and genetic testing, which was unrevealing before next-generation sequencing revealed biallelic variants in the GEMIN5 gene ( Table 1). All variants were rare with allele frequencies ranging from 0 to 4.48e−3 as a heterozygote and showed autosomal recessive inheritance pattern ( Table 3). The   Table 1). The variants identified were missense, frame shift, termination, and a predicted splice site variant. All missense variants affected conserved amino acid residues and were predicted to be pathogenic by various computational prediction tools (Table 4).
Patients, 8/9, presented in the first 2 years of life with concerns for delayed motor development. No childhood motor or cognitive regression was observed in any of the patients. While all patients had signs of motor dysfunction, the neurological exam ranged from hypotonia and tremulousness to severe spastic ataxia. While central hypotonia was observed in some patients, appendicular tone ranged from normal to spastic. Interestingly, all patients presented with normal to brisk reflexes. All ambulatory patients were assessed by neurologists and were noted to be ataxic on clinical examination. The disease progression was variable among our patients, as some of the patients had a static course, while others showed a possibly slow, progressive ataxia. Magnetic resonance imaging (MRI) of the brain showed cerebellar atrophy in all patients (Figure 2).
Stratifying patients with GEMIN5-related disease from this study and our previously published cohort (Kour et al., 2021) by age of onset and severity revealed three main groups based on neurological involvement: 1. Infantile-onset severe global developmental delay with cerebellar atrophy on neuroimaging.
These patients presented in infancy with severe hypotonia and global developmental delay. Neuroimaging confirmed cerebellar atrophy.
2. Late infantile-onset developmental delay and ataxia syndrome with cerebellar atrophy on neuroimaging. 3. Juvenile/adult-onset spastic ataxia syndrome with cerebellar atrophy on neuroimaging.
Other clinical features seen in our patients include cataracts, strabismus, and nystagmus.
Overall, all variants of GEMIN5 appear to perturb GEMIN5 structure and function(s) and result in deleterious neurological symptoms.

Neuroradiological Features
The MRI images of all patients are shown in Figure 2. All patients showed diffuse cerebellar atrophy including vermian and hemispheric involvement of the cerebellum. Patient 5 had minimal brainstem and spinal atrophy as well. Most patients with repeat neuroimaging demonstrated a stable atrophy. At this time, it remains to be clarified if the loss of cerebellar volume is more of a hypoplasia rather than an atrophy and a natural history study with temporal follow-up of patients and imaging will be needed to better understand this.

GEMIN5 Variants Reduce the Levels of snRNP Complex Proteins in Patient Cells
GEMIN5 is a multidomain protein, which acts as a crucial anchor during the assembly of small nuclear ribonucleoproteins (snRNPs), the essential components of the spliceosome (Pellizzoni et al., 2002;Wan et al., 2005;Battle et al., 2006a;Battle et al., 2006b;Kolb et al., 2007;Yong et al., 2010). We asked if the patients carrying compound heterozygous variants of GEMIN5 perturb the levels of GEMIN5 and other SMN complex proteins. We isolated peripheral blood mononuclear cells (PBMCs) from patient 5 carrying p. Trp373* and p. Arg1016Cys variants of GEMIN5, as well as from the unaffected heterozygous parent (father) and performed Western blot (WB) with antibodies to GEMIN5, SMN, GEMIN4, and GEMIN2 to assess protein levels. We found a significant reduction in GEMIN5, SMN, GEMIN4, and GEMIN2 protein levels in the patient sample compared with the unaffected  (Figure 3 and Supplementary Figure S2). These findings suggested that biallelic variants in GEMIN5 result in deleterious changes in the levels of snRNP complex proteins, which is likely contributing to various neurological defects observed in our patients.

GEMIN5 Variants are Predicted to Cause Structural and Conformational Changes and Disrupt its Interaction With Other RNA/ Proteins
GEMIN5 is a highly conserved protein across vertebrates, and mutations in GEMIN5 have been shown to cause a loss of function and loss of protein stability (Kour et al., 2021) ( Table 4). Various in silico tools allow us to predict the impact of an amino acid substitution on the structure and function of a human protein. Using in silico tools, we found that the GEMIN5 variants are predicted to be deleterious in nature and lead to loss of protein function and stability ( Table 4). Amino acid substitution at key sites within a protein due to a single-nucleotide variant (SNV) may result in various conformation changes, including remodeling or alteration of interaction network, salt bridges, and hydrogen bonds. These changes may perturb protein folding kinetics and can cause the destabilization of protein, impairing its subsequent function or interaction with other molecules (Dill et al., 1993;Teng et al., 2010). To determine the effect of GEMIN5 missense variants on the protein conformation and interaction with surrounding amino acids, we used PyMOL to predict the structural changes caused by five possibly deleterious substitutions-three GEMIN5 variants reported in our current study (p. Ser543Gly, p. Pro594Arg, and p. Arg1016Cys) and two GEMIN5 variants from our previous study (p. Gly683Asp and p. Asp704Glu) (Kour et al., 2021). All of the following variants are located within the WD40 repeat domains (PDB ID: 5H1J), except p. Arg1016Cys, which is located in the tetratricopeptide (TPR)like dimerization domain (PDB ID: 6RNS) of the GEMIN5 protein ( Figure 4). By limiting the affected area of mutagenesis to 5 Å, we measured changes in the orientation and conformation of surrounding amino acid by calculating their distance from the substituted GEMIN5 variant. We found that the variant p. Ser543Gly in GEMIN5 may lead to changes in its angular parameters and distance from Ile584, Ser585, and Lys545, respectively, which could result in structural changes in anaphase-promoting complex subunit 4, the WD40 domain (540-588 aa) of GEMIN5 ( Figure 4A). Similarly, we found that the conversion of Pro594, a nonpolar amino acid, to the positively charged arginine might affect its interactions with Glu595, Gln593, and Tyr598 ( Figure 4B). Proline acts as a structural disruptor in alpha and beta helixes, which is a prerequisite for protein folding and structure. Therefore, its conversion to arginine, an amino acid found at the active centers of the protein and required for interactions with phosphorylated substrates, might result in the disruption of GEMIN5 structure and its interaction with other proteins. PyMOL structure prediction suggested a possible change in the interaction and angular distance of 704Glu variant with His681, His708, Asp703, Gly683, and Arg682 ( Figure 4D). Likewise, the p. Gly683Asp variant of GEMIN5, which is in a grove between two WD40 domains, induces an ionic charge that may result in the formation of additional bonds with Arg682 and Phe705 as well as change in the possible conformation due to its interaction with Arg33 ( Figure 4E). The p. Arg1016Cys variant of GEMIN5 lies in the TPR-like dimerization domain, which modulates the interaction of GEMIN5 with other proteins. Thus, the substitution of the highly polar arginine, which is typically found in the interface of the two proteins, to a small and neutrally charged cysteine, may affect its conformation and dimerization function. PyMOL predicted the possible changes in the distance and bonding with Arg1014, Pro1017, and Asp1019 ( Figure 4C). Furthermore, the GEMIN5 variant p. Asp1019Cys has been reported to disrupt the dimerization properties of GEMIN5, suggesting the possible role and significance of Arg1016 and Asp1019 amino acids in protein-protein interactions. Overall, these findings suggest the possible deleterious effect of GEMIN5 variants on its structure conformation, interaction, and function.

Gemin5 is Essential for Development and Knocking out Gemin5 Causes Embryonic Lethality in Mice
To further understand the consequences of loss of function in Gemin5 in vivo, Gemin5 knockout (GEMIN5-DEL558) mice were created through the International Mouse Phenotyping Consortium (Dickinson et al., 2016;Mianné et al., 2017). To successfully generate the Gemin5 knockout mice, CRISPR/Cas9 was utilized to delete 558 nucleotides including exon 7 of the Mus musculus Gemin5 gene. This deletion of 558 nucleotides resulted in a frameshift and formation of a premature stop codon in the Gemin5 gene ( Figure 5A). Heterozygous mice were validated for harboring one copy of Gemin5 WT and one copy of Gemin5 knockout by SR-PCR genotyping ( Figures 5B, C). Heterozygous knockout mice were crossed together to assess the percentage of homozygous pups. No postnatal homozygous pups were observed at the P0 stage. Of the P0 pups, 63% were heterozygous for the Gemin5 knockout allele, while 37% of the pups harbored WT Gemin5 (n = 57).
To assess if pups were viable during embryonic development, mice were genotyped at embryonic day E9.5-10.5 across five separate litters ( Figure 5D). No homozygous embryos were observed at E9.5-10.5, suggesting that Gemin5 is crucial for embryonic development in mice (n = 30). In addition, the body weight of heterozygous Gemin5 knockout mice was assessed over the course of 20 weeks. There was no significant change in the weight of heterozygous mice compared with wildtype control mice (n = 15 heterozygote mice, 1,013 control mice) ( Figure 5E). Since we did not get any Gemin5 null animals at E9.5, we were unable to determine the impact of abolishing Gemin5 on brain development. These findings suggest that the homozygous loss of Gemin5 results in early embryonic lethality, while harboring one copy of defective Gemin5 does not appear to be detrimental and cause any obvious motor defects (Supplementary Figure S3).

DISCUSSION
Here we expand on the neurological phenotypes observed among patients with biallelic pathogenic variants in the GEMIN5 gene.
The clinical spectrum ranges from global developmental delay of early infantile onset with cerebellar atrophy to adult-onset spastic ataxia syndrome with cerebellar atrophy. Hereditary ataxias are a genetically heterogenous group of disorders with more than 150 genes being associated with this phenotype to date. With our cohort, we define an autosomal recessive hereditary ataxia syndrome associated with biallelic variants in GEMIN5. All but our early infantile patients had normal to brisk reflexes suggesting pyramidal tract involvement. The adult patients in our cohort were primarily followed in a spastic paraplegia clinic. Compiling the neurological phenotype from patients we previously reported (Kour et al., 2021) together with our current cohort, 21 of the currently reported 39 patients showed brisk reflexes, 4 had absent reflexes, 9 were normal, and 5 were unknown. Central hypotonia was seen in 30 of the 39 patients, 7 had normal central tone, and 2 were unknown. Of the 39 patients, 18 had appendicular hypertonia. Cerebellar atrophy was characteristically seen in all patients on neuroimaging, suggesting that this is a key feature associated with GEMIN5 variants. We, hence, categorize the neurological phenotype associated with autosomal recessive variants in GEMIN5 as predominantly in the hereditary cerebellar ataxia-spastic paraplegia spectrum of disease. While the hereditary cerebellar ataxias and hereditary spastic paraplegias have been thought to be clinically distinct, epidemiological studies and recent advances in next-generation sequencing technologies have demonstrated that multiple genes are associated with the disease phenotypes and overlaps in clinical presentation as well as molecular pathways.
Furthermore, this GEMIN5 neurodevelopmental ataxia spectrum is characteristically distinct from the predominantly neuromuscular presentation of spinal muscular atrophy (SMA) caused by another member of the SMN protein complex. Of some overlap, pontocerebellar hypoplasia type 1 (PCH1) is a condition with pontocerebellar atrophy along with anterior horn cell involvement of the spinal cord associated with many autosomal recessive genes (Namavar et al., 2011;Wan et al., 2012;Rudnik-Schoneborn et al., 2013;Sánchez-Albisua et al., 2014;Lardelli et al., 2017;Ivanov et al., 2018;Nuovo et al., 2021). These diseases can present with a varying blend of neurological phenotypes including hypotonia to varying degrees of spasticity.
The GEMIN5 variants identified, so far, are missense, deletion, frameshift, nonsense, and splicing defects with autosomal recessive inheritance pattern. The variants identified in our current study, with two exceptions (p. Arg1016Cys and p. Pro594Arg), were never reported in any publicly available databases. The p. Arg1016Cys variant was identified with a frameshift or nonsense variant in three patients from three unrelated families (Table 1). This variant has been reported with allelic frequency of 4.48e−3 (heterozygote), and there are five individuals reported as homozygous. It is believed that the p. Arg1016Cys variant is a hypomorphic variant, and it is not sufficient to manifest any clinical symptoms unless the other allele is fully disrupted. Alternatively, the homozygous individuals in gnomAD may be at a presymptomatic stage.
FIGURE 5 | Knockout of Gemin5 leads to early embryonic lethality in mice. (A) Schematic representation of the wild-type allele of the mouse Gemin5 transcript. The Gemin5 KO mouse (GEMIN5-DEL558) was created by utilizing the EUCOMM/KOMP-CSD allele structure to delete 558 nucleotides from exon 7 of the Gemin5 gene, encompassing a critical exon to induce a premature stop codon and a null Gemin5 allele (Mianné et al., 2017). (B) SR-PCR genotyping detects the presence of the Gemin5 wild-type allele by short-range PCR in Gemin5 KO and control mice (copy number of 2). Samples are genotyped with a WT loss of allele (WT-LOA) assay. This is a FAM-labeled assay that is designed to detect the critical exon that has been floxed. If the animal contains 1 Gemin5 KO allele, the copy number of this assay should drop by 1 (copy number of 1). For autosomal genes that have been targeted, this means the following: WT = 2 copies of the LOA assay, HET = 1 copy of the LOA assay, HOM = 0 copy of the LOA assay. (C) SR-PCR genotyping detects the presence of the cassette lacZ, the Gemin5 wild-type allele, and the Gemin5 KO allele confirming heterozygotes by short-range PCR (copy number of 1 GEMIN5 is expressed predominantly in the cytoplasm as well as in the nucleus, nucleoplasm, and gem bodies, suggesting functions in multiple cellular compartments (Pacheco et al., 2009;Piñeiro et al., 2015;Francisco-Velilla et al., 2016;Lozano et al., 2018;Martinez-Salas et al., 2020;Moreno-Morcillo et al., 2020). Animal model studies suggest that complete or partial reduction of the snRNP complex proteins are detrimental (Saida et al., 2021). Knocking out Smn, Gemin4, and Gemin5 is lethal in mice and Drosophila, suggesting that these proteins are essential for survival (Gates et al., 2004;Gavrilina et al., 2008;Meier et al., 2018). We have previously shown that genetic reduction of endogenous rigor mortis (fly homolog of human GEMIN5) leads to developmental delay, motor dysfunction, and reduced lifespan, similar to human patients with autosomal recessive variants in GEMIN5 (Kour et al., 2021). Our data in human patient cells (Figure 3) suggest that pathogenic variants in GEMIN5 significantly reduce the expression of snRNP complex assembly. Furthermore, knocking out Gemin5 in mice is embryonically lethal ( Figure 5). Although these findings are interesting, we are unable to establish a link between cerebellar atrophy and Gemin5 loss of function due to early embryonic lethality in our knockout Gemin5 mice. We performed in silico analysis to determine the impact of GEMIN5 variants on the protein structure. We found that GEMIN5 missense variants cause significant alterations on protein conformation, such as remodeling the interaction network, peptide chain bending, and the ability to interact with other amino acids. Since the GEMIN5 protein has been shown to dimerize, it is likely that frameshift and termination variants might perturb the kinetics of protein folding and can cause the destabilization of proteins. These alterations might disrupt the ability of GEMIN5 to form a complex with snRNP proteins and exert physiological functions, which in turn lead to neurological symptoms observed in patients with GEMIN5-related disease. We recently showed that reduction of the levels of GEMIN5 (by mutations or shRNA knockdown) is sufficient to disrupt the snRNP assembly in mammalian cells and iPSC neurons (Kour et al., 2021). These findings support the idea that each component of the snRNP complex has an important role to play in maintaining cellular functions. Alternatively, it is possible that loss of normal GEMIN5 function might affect translation of mRNAs either in a tissue-specific manner or ubiquitously leading to defective protein synthesis and their regulatory pathways.
Given the distinct neurological presentation, we propose the term GEMIN5-related neurodevelopmental ataxia with cerebellar atrophy to describe the variability in clinical presentation associated with this gene.

DATA AVAILABILITY STATEMENT
The datasets presented in this article are not readily available due to patient-related privacy regulations and institutional policies.
Requests to access the datasets should be directed to udai.pandey@chp.edu

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The University of Pittsburgh Institutional Review Board (IRB). Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. The animal study was reviewed and approved by The University of Pittsburgh Institutional Animal Care and Use Committee (IACUC) (Protocol #: 21069518).

AUTHOR CONTRIBUTIONS
DR, SK, TF, and UP planned, performed and analyzed the data. All other authors were involved in collecting patient data, analysis, and interpreting the data. All authors read the manuscript and gave their consent to publish.

FUNDING
The C57BL/6NTac-Gemin5 em1(IMPC)H /H mice and phenotypes were generated at the MRC Harwell Institute. The knockout mouse generation protocol has been approved as per MRC Harwell Institute guidelines. The MRC Harwell Institute generated the mice as part of the International Mouse Phenotyping Consortium (IMPC) and has received funding from the National Institutes for Health for generating and/or phenotyping (5UM1HG006348-08) the C57BL/6NTac-Gemin5 em1(IMPC)H /H mice. The research reported in this publication is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Funding and associated primary phenotypic information may be found at www. mousephenotype.org.

ACKNOWLEDGMENTS
UP and DR are thankful to the Children's Neuroscience Institute, Children's Hospital of Pittsburgh, for providing a pilot grant funding.