Clinical and Molecular Spectrum of Autosomal Recessive CA8‐Related Cerebellar Ataxia

Based on a limited number of reported families, biallelic CA8 variants have currently been associated with a recessive neurological disorder named, cerebellar ataxia, mental retardation, and dysequilibrium syndrome 3 (CAMRQ‐3).

Carbonic anhydrase VIII, an enzymatically inactive member of the zinc metalloenzyme family that otherwise catalyzes the reversible hydration of carbon dioxide, is encoded by the CA8 gene.CA8 exhibits a distinct expression pattern primarily in the Purkinje cells within the cerebellum, 1 but it is also found in various peripheral tissues.In 2009, Türkmen et al 2 published the first report of biallelic variants in CA8 causing cerebellar ataxia, mental retardation, and dysequilibrium syndrome 3 (CAMRQ-3, Online Mendelian Inheritance in Man [OMIM] 613227).][4][5][6] Despite these descriptions, the current knowledge regarding CA8-related disease (CA8-RD) remains limited.
In this comprehensive study, we investigate CA8-RD by conducting an extensive literature review alongside an in-depth exploration of neurological, neuroradiological, and molecular observations of newly identified patients.Additionally, we provide functional insights into the role of CA8 in neurodevelopment through the generation and analysis of ca8 knockout zebrafish models.

Methods
This study identifies 19 CA8-RD patients from nine families, integrating data from a literature review of 12 cases.Clinical, genetic, and radiological assessments, including Scale for the Assessment and Rating of Ataxia (SARA) evaluations, were conducted.Exome sequencing (ES) and zebrafish experiments were used for genetic characterization.Descriptive statistics and GraphPad Prism 9 were used for analysis, with significance set at P < 0.05.Supplementary Data S1 and S2 provides detailed methodologies.
Najmabadi et al, 4 whose sexes are not known), from 14 families.Consanguinity was reported in 11 families.The median age at the last follow-up was 15.45 AE 8.46 years (media AE standard deviation [SD], range: 4-35 years; available in 19/27 patients).The median time of follow-up of patients was 6.30 AE 1.63 years (media AE SD, range: 5-10 years; available in 13/19 patients).The results in the upcoming sections will be presented with denominators specified for each characteristic based on available data in this cohort (19 patients) or including the total number of patients (27 individuals).Figure 1A contains detailed pedigrees for the newly reported families, as well as the family previously documented, 3 for which new information is available.S1).The mean age at sitting was 2.68 AE 0.15 years (media AE standard error [SE], range: 6 months-10 years; available in 16 patients), at crawling was 2.74 AE 0.28 years (media AE SE, range: 6 months-9 years; available in nine patients) and at walking was 9.57 AE 0.31 years (media AE SE, range; 3-22 years; available in 16 patients) (Fig. 1C).Eight individuals achieved independent walking, whereas six required assistance, such as walkers or handholds.Moreover, two individuals were non-ambulatory, and five of them exhibited an unsteady gait or could only walk short distances.Except for P6, 16 of 27 patients presented with speech delay.The mean age of first words was 6.45 AE 0.38 years (media AE SE, range: 1-14 years; available in 11 patients).Patients who are able to speak produce a limited vocabulary (10 words), exhibit poor syntax and phonology, or demonstrate dysarthria while speaking.After a febrile illness, P2 reported developmental regression.

Varying Degrees of ID and Behavioral
Characteristics Are Observed Seventeen of 27 patients were diagnosed with ID ranging from mild to moderate at various ages (Fig. 1B).Two of 27 patients had a normal intelligence quotient (IQ) (P6 and reported by Richmond et al 6 ), and 16/27 patients had a low IQ, whereas the remaining nine of 27 patients did not undergo IQ testing.Except for 6/27 patients who attended a mainstream school with special education needs for physical needs, the majority of patients attended a specialized school.Four of 10 patients were able to perform activities of daily living with relative ease, whereas 6/10 patients had varying degrees of difficulty.Others exhibited signs of impulsivity, tantrums, and autism spectrum disorders (ASD) in addition to aggression and apathy.

Clinical Variability of Ataxia and Mild Motor
Manifestations in CA8-RD The median age at the onset of the disease was 0.72 AE 0.37 years (media AE SD, range: neonatal period to 1.5 years; available in 12 patients).Except for P1 (febrile seizure), P2 (motor regression after febrile seizure), and P7 (brain malformation detected during pregnancy), 19/24 patients initially displayed motor delay as the initial symptom.

Cerebellar signs
Gait ataxia was observed in 25/27 patients, dysarthria in 21/27, dysmetria in 17/27 patients, dysdiadochokinesia in 16/27 patients, truncal ataxia in 15/27 patients, and intention tremor in 13/27 patients (Videos S1-S9).Additionally, wide-based stance was noted in 15/27 patients, tandem ataxia in 9/27 patients, and head titubation in 2/27 patients (Fig. 1D).Ten of 27 patients demonstrated a quadrupedal gait.In terms of disease progression, ataxia was progressive in 10/15 patients and improved in 5/15 patients.The mean follow-up duration for patients, as reported by the physician documenting the symptom progression, is 6.3 AE 0, 10 years (mean AE SE, range: 5-10 years; available for 15 patients).Two of 12 patients had mild ataxia, 5/12 had moderate ataxia, and 5/12 had severe ataxia.According to the SARA scale (available in eight patients), P1 had the most severe symptoms, whereas P6 had the least.P1 scored highest in all tests: gait, stance, sitting, speech disturbance, finger chase, nose-finger test, fast alternating hand movements, and heel-shin slide (total score of 29 points).P6 had the lowest overall score (16 points).

Motor manifestations and movement disorders
Hypophonia and hypomimia were observed in 6/27 patients.Rigidity was noted in 3/27 patients, postural instability in 5/27 patients, and bradykinesia was evident in 5/27 patients.P1 displayed freezing during turns.Additionally, dyskinesia or dystonia was reported in 6/27 patients, and intentional tremor was present in 12/27 patients.None of the patients displayed rest tremor.Importantly, it should be noted that the patients in this study did not receive treatment for bradykinesia; hence, dyskinesia/dystonia cannot be attributed to medication adverse effects.

Pyramidal signs
Ten of 27 patients exhibited axial hypotonia, whereas 8/27 patients exhibited upper or lower limb spasticity.Mild muscle weakness was observed in 8/27 patients, mainly in the upper limbs, whereas distal muscle Movement Disorders, Vol.39, No. 6, 2024 atrophy was reported in P13.Muscle stretch reflexes were mainly hyperreflexia, and the plantar reflex was upgoing in most patients.

Oculomotor abnormalities
Strabismus was observed in 7/27 patients.Ten of 27 patients displayed horizontal nystagmus, and one patient (P3) presented with vertical nystagmus.Abnormal pursuit and saccades were observed in 11/27 and 9/27 patients, respectively.P4 presented with vertical gaze palsy.

Other neurological symptoms
Four of 27 patients exhibited hypoesthesia for pain and temperature.

Seizures
P1 and P2 had febrile seizures at 5 months and 1.5 years of age, respectively, whereas P6 had a single seizure with a normal video electroencephalography during gastroenteritis at 7 months of age.In addition, a 7-year-old patient described by Paternoster et al 5 exhibited focal seizures.
During the diagnostic process, multiple metabolic tests were performed with normal results.Electromyography/nerve conduction studies were performed for some patients, and the results were normal, except for P1 (severe sensory and moderate motor demyelinating polyneuropathy in the lower limbs) and P2 (demyelinating sensory-motor polyneuropathy in the upper and lower limbs) (Supplementary Data S3).
Table 1 illustrates the phenotypic spectrum of core clinical features in our cohort.A comprehensive summary of all phenotypic characteristics evaluated is provided in the Supplementary Data S1.

Clinical Heterogeneity in Symptom Progression
among Patients with CA8-RD In 8/18 patients, information about the progression of symptoms was available and exhibited considerable heterogeneity.P7 experienced developmental delay and walking difficulty by 1.5 years, with the subsequent onset of gait ataxia and speech abnormalities at 4 years.P8 presented with developmental delay at 9 months, speech abnormalities at 2 years, and behavioral abnormalities at 3 years.P10 and P11 manifested hypotonia, motor regression, head titubation, and tremors.P16 demonstrated an inability to walk without support and later speech problems at 5 years.P17 showed an inability to walk without support and speech problems at 1 year.P18 experienced difficulties in crawling at 9 months and subsequently lost the ability to walk without support, along with speech problems.P30 showed gait and truncal ataxia, tremor, dysmetria, brisk reflexes, distal muscle atrophy, extensor plantar reflex, abnormal saccades, and epilepsy, with dysarthria reported at 16 years old.

Sequential Neuroimaging Studies Reveal Progressive Cerebellar Atrophy, Particularly Progressive Superior Vermis Atrophy, and Prenatal Brain Malformation
Brain magnetic resonance imaging (MRIs) were available for 12/18 patients, with repeated imaging available for 7/18 patients, enabling a comparative analysis of the progression of neuroradiological findings.The mean age at the first MRI was 7.87 AE 0.66 years (mean AE SE, range: 6 monthsÀ23 years; available in 12 patients), and at the last MRI, it was 9.05 AE 0.59 years (mean AE SE, range: 4-17 years; available in seven patients).Neuroimaging findings demonstrate selective involvement of the cerebellum in all patients with "isolated cerebellar atrophy", including moderate to severe atrophy, particularly involving the superior aspect of the vermis and hemispheres, without evidence of discrete signal changes (Fig. 1E).The mean age at the diagnosis of cerebellar atrophy on MRI was 6.51 AE 0.99 years (mean AE SE, range: 6 months to 17 years; available for seven patients).There was no evidence of signal changes or malformative features involving the supratentorial compartment of the brain.High-resolution imaging confirmed the normal appearance of the cranial nerves.Magnetic resonance (MR) spectroscopy and positron emission tomography were not performed, except for one patient where MR spectroscopy showed normal peaks and ratios between major metabolites.The pregnancy of P12 was interrupted based on brain malformations (anencephaly) identified during prenatal ultrasounds.Case-based descriptions of the imaging findings are provided in Table 1.

The Genetic Spectrum of CA8-RD
ES and variant filtering in eight newly identified families revealed seven homozygous CA8 (NM_004056.6)variants segregated with the phenotype within the families (Supplementary Table S2).The observed variants included two missense: c.513G>C, p.(Gln171His) (families 2 and 5) and c.653A>G, p.(Tyr218Cys) (family 6); two nonsense c.811G>T, p.(Gly271Ter) (family 1) and c.751C>T, p.(Arg251Ter) (family 7, aborted fetus), and three splicing c.100 + 1G>A, (p.?) (family 4), c.418-3C>G, (p.?) (family 3), and c.738 + 1G>A, (p.?) (family 9) variants.Five of seven variants are absent in gnomAD V4.0.0 and the rest of the two variants are present at ultra-low allele frequencies.The missense variants have very high Combined Annotation Dependent Depletion (CADD) scores (30 and 34), reside in the evolutionarily constrained regions of the CA8 protein (Genomic Evolutionary Rate Profiling [GERP], scores 4.69 and 5.89), and are predicted to be deleterious.All splicing variants are predicted to lead to aberrant splicing, according to SpliceAI (https://spliceailookup. broadinstitute.org/).Based on the American College of Medical Genetics (ACMG) guidelines, all seven variants are classified as pathogenic.The two nonsense and two missense CA8 variants uncovered in the newly identified patients are predicted to reside in the minimum binding site of the CA8 protein, which is important for its function. 1,12No other pathogenic or likely pathogenic variants were identified in other disease-associated genes that could have potentially explained the phenotypes of the patients in the correct zygosity.No clear genotypephenotype correlation was established.

ca8 Knockout Does Not Trigger Obvious Gross Anatomical Defects in Zebrafish
To study the impact of CA8 bi-allelic knockout in a whole organism context, we deployed a F0 crispantknockout approach in the zebrafish animal model. 13,14he CA8 homolog in the zebrafish (ie, dre-ca8 or ca8) encodes a protein with >80% homology with its human protein counterpart. 15We used Cas9 protein injections complexed with four guide RNAs targeting the first two ca8 exons to ensure proper bi-allelic disruption (Fig. 2A, Supplementary Data S1).Interestingly, ca8-KO in zebrafish did not trigger any obvious gross malformations or anatomical defects in all the conditions tested (Fig. 2B,C).ca8-KO did, however, significantly reduce the total body length of the animal at the early stage of their development, but this difference became non-significant starting at 7 days post fertilization (dpf) (Fig. 2D).Most organs except the brain appeared unchanged, with, for example, normal cardiovascular development and heart function (Figs.2B and 4F).

dre-ca8 KO Impairs Brain Development and Motor Function without Affecting Motor Neurons
Although the gross development of the ca8-crispant zebrafish appeared normal, abnormal brain development was observed in all conditions tested.Observed at 24 hours post fertilization (hpf), the brain of ca8crispant animals present some robust and distinctive defects when compared to the injected controls (Fig. 3A,B).For example, the mid-brain to hind-brain boundary could not be distinguished in the crispant animals, suggesting probable associated cerebellum neurodevelopmental defects (the cerebellum being the most anterior part of the hindbrain in the zebrafish. 16These defects are clearly observed when looking at dorsal views of the injected animals (Fig. 3C), with distinctive abnormal brain ventricle formation (Fig. 3D,E).
Associated with these aforementioned neurodevelopment defects, ca8-KO also affected the animal motor behavior and motor response to external stimulation.At 28 hpf, the mean duration of spontaneous 989 coiling was reduced in ca8-crispant embryos (Fig. 4A).
The response to touch was also impaired, with both the coiling-response at 30 hpf (Fig. 4B) and the swimming-response at 48 hpf significantly reduced in ca8-KO animals (Fig. 4C).Interestingly, those motor defects in the crispants were not associated with any detectable motor neuron phenotype.Analyzed motor neurons appeared normal both in terms of axonal development (Fig. 4D) and anatomical features (Fig. 4E).

Literature Review
Twelve patients have been described previously, 2-6 although the data from two patients reported by Kaya et al 3 are not mentioned.The data extracted from published cases are presented in the Data S1.

Discussion
This study presents a cohort of 27 affected individuals who harbor biallelic CA8 variants associated with CA8-RD.Using a systematic analysis approach, we provide a comprehensive characterization of the clinical, molecular, and imaging features of this rare syndrome, and based on our findings, we may propose a more suitable name for this disorder.Furthermore, our zebrafish experiments revealed that ca8-KO did not induce gross anatomical defects, but it significantly impacted brain development and motor function in the absence of detectable motor neuron phenotypes.
Knowledge about CA8-RD remains limited.Individuals with biallelic CA8 variants exhibit motor and language developmental delay, variable ID, and ataxia, along with isolated cerebellar atrophy, which mainly affects the superior aspect of the vermis and hemispheres.Notably, CA8 patients may also display hypomimia or bradykinesia, developmental regression, and febrile seizures.However, quadrupedal gait is observed in less than half of the patients.Therefore, we endorse the proposal by Richmond et al 6 to use the term "CA8-related cerebellar ataxia" regardless of the presence of ID or quadrupedal gait in these patients.
Significantly, the term CAMRQ likely originated from its initial designation as Uner Tan syndrome.This syndrome was originally described in 2005 in five individuals presenting with "quadrupedal gait, flexed head and body, primitive speech, severe mental retardation, and mild cerebellar signs with a disturbed conscious experience".Consistent with Tan's psychomotor theory suggesting a punctuated human evolution (the transition from habitual quadrupedality in our common ancestors to habitual bipedality in Homo erectus with a bipedal walking pattern and a larger brain may have resulted from sudden mutations in specific genes rather than a gradual evolution), Uner Tan syndrome is considered a devolution of the human being. 17,18enetic defects in the same "special genes" responsible for the acquisition of bipedalism are thought to lead to Uner Tan syndrome. 19Although all the genes currently associated with CAMRQ syndrome invariably exhibit neurodevelopmental abnormalities, CA8 defects appear to present with additional signs of neurodegeneration.Cerebellar atrophy is observed in CA8-RD in contrast to other CARMQ-associated genes exhibiting neurodevelopmental cerebellar abnormalities.Unlike many other degenerative conditions, 20 CA8-RD cases in our cohort did not exhibit significant signal changes in the cerebellar cortex (the "bright-cerebellum"), involvement of the brainstem, or notable changes in the supratentorial brain compartment.Therefore, it is crucial to underscore the robustness of our MRI evidence, which demonstrates selective cerebellar atrophy, supporting the differentiation between CA8-related ataxia and other subtypes of the syndrome.
The association of CA8-RD with neurodegeneration may be underpinned by a molecular mechanism involving the interaction of CA8 with the inositol 1,4,5-triphosphate receptor type 1 (ITPR1)-mediated calcium signaling pathway.ITPR1, an endoplasmic reticulum Ca 2+ channel, controls mitochondrial Ca 2+ signaling.It is abundantly expressed in the Purkinje cells and plays a pivotal role in cerebellar coordination by facilitating motor learning and synaptic plasticity. 1 It has been suggested that many genes involved in genetic cerebellar ataxias act through the ITPR1-dependent signaling pathway, including CA8-RD.Importantly, of all CAMRQ-associated genes, only CA8 operates through the ITPR1 pathway, putting it aside from the rest of the genes associated with CAMRQ.Although CA8 inhibits the ITPR1 protein, CAG repeats in the genes, associated with cerebellar ataxia, appear to activate ITPR1.Loss of CA8 is associated with aberrant excitatory inputs to the Purkinje cells. 21olecular defects in the CA8 and the ITPR1 genes have overlapping clinical features.Monoallelic and biallelic variants within the ITPR1 gene are associated with spinocerebellar ataxia (SCA) 15 (MIM: 606658), 22 Gillespie syndrome (MIM: 206700), 21 and SCA29 (MIM: 117360). 23The latter two syndromes, similar to CA8-RD, present with slowly progressive or non-progressive cerebellar ataxia on the background of neurodevelopmental abnormalities.Furthermore, quadrupedal gait, akin to CA8-RD, has been documented in some patients with biallelic variants in ITPR1 (autosomal-recessive SCA29). 17,18However, despite the overlapping molecular mechanisms and clinical features, ITPR1 and CA8 remain in two distinct phenotyping categories in the OMIM database: SCA and CAMRQ.Hence, based on these findings, we advocate for the inclusion of CA8-RD in the phenotypic realm of recessive cerebellar ataxias (SCARs) and suggest its nomenclature as autosomal-recessive CA8related cerebellar ataxia.Cerebellar atrophy is characterized by a gradual enlargement of the interfoliar space within the cerebellum, 20,24 indicating irreversible tissue loss, and is often associated with CA8-RD.Our cohort also showed predominant atrophy of the superior vermis and cerebellar hemispheres.In the advanced stages of the disease, these regions exhibited atrophy.Notably, similar progressive superior vermis atrophy has been observed in other conditions like ataxia of Charlevoix-Saguenay and ataxia-telangiectasia, 25 despite the distinct initial phenotype of CA8-RD.
Clinically, CA8-RD, ataxia telangiectasia, and Charlevoix-Saguenay ataxia exhibit distinct clinical profiles.Ataxia telangiectasia manifests with delayed motor and language development, immunodeficiencies, telangiectasias, other associated movement disorders, and mild and slowly progressive cerebellar atrophy. 26harlevoix-Saguenay ataxia typically features progressive ataxia, normal cognition or rarely ID, prominent spasticity, polyneuropathy, retinal optic nerve hypermyelination, and cerebellar atrophy. 27Other differences are summarized in Supplementary Data S4.
Parkinsonism-ataxia syndromes are known for their clinical variability and have been associated with several spinocerebellar ataxias (SCA2, SCA3, SCA6, SCA8, and SCA17) and metabolic disorders (neurodegeneration with brain iron accumulation, Fahr's disease, Wilson disease, and mitochondrial disorders). 28ix patients within our cohort exhibited hypomimia, bradykinesia, or freezing during turns, indicating that CA8-RD should be considered for inclusion in parkinsonism-ataxia syndrome classifications.It is noteworthy to mention that these symptoms were detected in patients who were relatively young.Nevertheless, it is crucial to exercise prudence when making assumptions regarding the course of degeneration, given the potential variability in the trajectory of motor symptoms.Additional data are required to validate any forecasts of age-related progression.
Our functional investigation using the zebrafish animal model provides evidence that disruption of ca8 gene function impairs early neurodevelopment.Morphologically, disruption of ca8 delayed the appearance of embryonic brain landmarks such as the midbrainhindbrain boundary and the development of the forebrain, midbrain, and hindbrain ventricles (Fig. 3).Although more in-depth analysis would be required, the apparent midbrain-hindbrain boundary abnormalities suggest a probable cerebellum neurodevelopmental defect.Interestingly, functionally, these defects were associated with impaired motor behavior and motor response to stimulation, despite the absence of any detectable spinal motor neuron phenotype (Fig. 4).These observations align with the manifestation of cerebellar ataxia in patients, suggesting that the zebrafish model could be a valuable platform for delving into the underlying pathological mechanisms.Furthermore, the reduction of spontaneous movement/coiling suggests the presence of bradykinesia/parkinsonism-like defects that would warrant further investigation at a later development stage.Importantly, these findings are consistent with those of Aspatwar et al, 15 and Huang et al 29 who investigated the inhibition of ca8 mRNA in zebrafish embryos using translation-and splice-blocking morpholinos.Both groups reported reduced activity of ca8 morphants despite the normal appearance of Rohon-Beard mechanosensory neurons and spinal motor neurons, suggesting impaired cerebellar function.Aspatwar et al, 15 also observed abnormal cerebellar morphology by electron microscopy accompanied by apoptosis in ca8 morphant embryos.These findings are consistent with the cerebellar atrophy observed in patients with biallelic CA8 variants.
As a limitation of this study, it is crucial to highlight the challenges associated with relying on retrospective data and patient or caregiver recall.As the study is retrospective, the accurate recording of symptom onset age is inherently limited.Many patients, now in adulthood, experienced the initiation of symptoms during their pediatric years, and unfortunately, this specific data was not adequately documented during their transition to adulthood.The reliance on patient or caregiver recall introduces a level of uncertainty, as memories may be subject to recall bias.Furthermore, in disorders with prolonged evolution, symptoms may dynamically evolve, making it challenging for patients to precisely identify the initiation of specific symptoms.We acknowledge these inherent complexities and potential limitations, underscoring the need for caution in the interpretation of the data and emphasizing the importance of prospective studies with meticulous data collection methodologies in future research endeavors.
Additionally, it is important to note that the progression of symptoms was based on the subjective determination of treating physicians.To address this, future studies should administer all ataxia scales longitudinally at different time points to ensure a rigorous assessment of disease progression.Another limitation of this study is that some of the data are derived from medical literature, where data collection has not been conducted uniformly.
In summary, our study has expanded the current knowledge of CA8-RD by providing a more comprehensive understanding of its clinical and molecular spectrum.The ca8-KO zebrafish experiments reveal the impact on brain development and motor function.Our findings also suggest that progressive superior vermis atrophy can serve as a useful clinical marker for the diagnosis of this disease.Last, we recommend using the term "CA8-related cerebellar ataxia" when describing this disorder.

FIG. 1 .
FIG. 1. (A) Presents the pedigrees of the families involved in the study and the segregation of CA8 variants.(B) Provides a visual representation of the distribution of individuals categorized by intellectual disability/developmental disorders (ID/DD), severity of ID/DD, ataxia course, quadrupedal gait, bradykinesia/parkinsonism and dyskinesia/dystonia. (C) Presents a schematic representation of sitting and walking in the new patients of this study.The symbol "*" denotes individuals for whom the age at walking is not available, whereas the symbol "." represents individuals for whom the age at sitting is not available.(D) Displays the prevalence of clinical manifestations among all individuals (n = 27) diagnosed with CA8-related disorders, including both the cohort analyzed in the current study and previously reported cases.(E) Presents the cerebral magnetic resonance imaging of the patients (P1-P4 and P6-P8) in sagittal, axial, and coronal views, which show cerebellar atrophy including the vermis and cerebellar hemispheres.[Color figure can be viewed at wileyonlinelibrary.com]

TABLE 1
Overview of clinical, genetic, and radiological features of the CA8-related disease.

TABLE 1
Note: P12 has been included in this table; however, as can be observed, there is limited clinical data available for this specific case (aborted fetus).Abbreviations: F, Family; P, patient, +, yes; À, no; NA, not available; GDD/ID, Global developmental delay/intellectual disability; NCS, Nerve conduction study.Movement Disorders, Vol.39, No. 6, 2024