Introduction

The classification of hereditary ataxias represents a significant challenge due to the large number of neurological and metabolic diseases that present with cerebellar dysfunction and the phenotypic heterogeneity in known genetically defined disorders. Indeed, ataxia is a presenting feature in degenerative disorders that target mainly the cerebellum, but it may be present in hereditary spastic paraplegias, inborn errors of metabolism, and various encephalopathies. Proper classification and phenotypic understanding is of primary importance in this field where the high prevalence of repeat expansion disorders, which are not adequately covered by the next-generation sequencing (NGS) techniques [1, 2], precludes NGS as a first diagnostic step and requires phenotypic evaluation to perform custom gene testing when applicable. Nevertheless, autosomal recessive cerebellar ataxias have remained an ill-defined and disorganized group of disorders for two main reasons. First, unlike the dominant ataxias that have been organized with a numerical naming system, recessive disorders presenting with ataxia have been named in a highly heterogeneous manner according to clinical features, physicians’ surname, or regions of high prevalence. Second, several recessive multisystemic or complex metabolic disorders present with ataxia, such that it is difficult to properly circumscribe this group of disorders and classify it in a meaningful way for both clinicians and researchers. Hence, the Society for Research on the Cerebellum and Ataxias (SRCA) Task Force on the Classification of Recessive Cerebellar Ataxias was created in 2016 to regroup a panel of international ataxia experts in order to propose a classification relevant to clinical practice and researchers. As a first step, we undertook a systematic scoping review of the literature to identify all recessive disorders presenting with ataxia, select those in which cerebellar degeneration was a core feature, and propose a first classification. This systematic scoping review has been previously published [3] and served as the basis for the current work.

Recently, the Movement Disorder Society Task Force on Classification and Nomenclature of Genetic Movement Disorders proposed a revised naming system based on the gene name associated with a phenotypical prefix. They presented a list of 92 gene-defined recessive disorders associated with ataxia for which this naming system would be applied and an exhaustive list of disorders that may occasionally present with ataxia [4]. This represents a useful reference for interpretation of NGS results. However, in a significant number of listed disorders, the cerebellum is only one of the many affected organs in multisystemic and metabolic disorders. For example, maple syrup urine disease, caused by BCKDHB mutations, and congenital disorders of glycosylation 1a, 1c, and 1q have been included. These disorders are inborn errors of metabolism characterized by developmental delay, hypotonia, and metabolic defects, and ataxia is only mild, found in a minority of patients, or present solely during episodes of metabolic decompensation. Hence, there remains a need for a classification system that focuses on disorders affecting primarily the cerebellum and organizes clinical and paraclinical information to promote an understanding of cerebellar disorders useful not only to ataxia experts but also to general neurologists, learners, patients, and researchers.

The objective of this task force was to build a consensus on the classification of autosomal recessive ataxias in order to develop a general approach to a patient presenting with ataxia, organize disorders according to clinical presentation, and define this field of research by identifying common pathophysiological mechanisms in recessive disorders presenting with ataxia. This aims at bringing together clinicians and researchers to promote a common understanding of recessive cerebellar disorders in order to advance research and improve patient care.

Materials and Methods

The first step was to identify all recessive disorders presenting with ataxia. Recessive cerebellar ataxias were defined as disorders with autosomal recessive inheritance characterized by a cerebellar motor syndrome of gait ataxia, dysmetria, adiadochokinesia, nystagmus, and dysarthria associated with cerebellar degeneration as demonstrated by imagery or pathology. A pathogenic mutation had to be identified in at least two independent families for a specific gene to be included. Purely malformative disorders were excluded, and disorders with complex phenotypes where ataxia is a secondary or late feature were also excluded. We conducted a systematic scoping review of the literature to identify relevant reports. The methodology and results of this systematic review have been published previously [3]. In the first publication, this review process had allowed the identification of 2354 records and was current as of September 2016. The literature search was updated and is current as of October 2018.

The second step was to regroup a panel of 12 international ataxia experts to create a logical classification system and build a consensus. Ataxia experts were identified from various geographical regions and areas of expertise within the field of ataxias, ensuring proper representation of regional differences in prevalence and clinical approach to ataxias. Discussions spanned over 2 years, included meetings at two SRCA international conferences, and concerned general orientation, clinical approach, specific disorders, classification issues, and regional specificities. The first author (MB) reviewed identified records for inclusion, extracted clinical, epidemiological, and molecular data to build the classifications and wrote the text integrating all authors’ input and comments. All authors approved the final manuscript and list of included disorders.

Results

The final list of included autosomal recessive cerebellar ataxias is presented in Table 1 and includes 59 primary recessive ataxias, which regroup 15 disorders that are more prevalent and widely distributed and 44 disorders that are less frequent and reported only in certain populations or few families. Because ethnic and regional specificities are an essential element to consider in the appraisal of a patient with a recessive ataxia, areas where the disorder has been reported to date are listed. Metabolic or mitochondrial disorders where ataxia is only a secondary nonspecific finding in a multisystemic phenotype were excluded, as cerebellar pathology is not central in these disorders. However, clinicians must bear in mind that some of these disorders may present with a milder juvenile or adult onset phenotype where cerebellar ataxia may predominate, for example, in Niemann-Pick disease type C, Tay-Sachs disease, sialic acid storage disorders, congenital disorders of glycosylation, and Zellweger spectrum disorders. As some of these metabolic disorders may benefit from early treatment, clinicians must keep a high index of suspicion to test for these disorders, and they should be included in large NGS gene panels for ataxia. These and other complex disorders that may occasionally present with ataxia are presented in Table 2. This second list is not exhaustive and presents only the main or most frequent disorders occasionally associated with ataxia. Disorders in which the cerebellar phenotype is not clearly established have been excluded.

Table 1 Primary autosomal recessive cerebellar ataxias
Table 2 Other metabolic or complex autosomal recessive disorders that have ataxia as an associated feature

Clinical Approach to a Patient Presenting with Ataxia

  1. 1.

    The first step in evaluating a patient with ataxia is to perform a detailed clinical evaluation that includes a clinical history, a family history, a targeted neurological and systemic physical evaluation, and relevant paraclinical tests. The temporal course is a central element in determining the underlying etiology. Indeed, a chronic progressive evolution over months to years, without trauma or toxin exposure, is suggestive of a hereditary disorder, whereas acute or subacute onset points towards an acquired etiology. A clinical history and physical examination are essential to assess the severity of the cerebellar syndrome and the presence of associated neurological features or systemic involvement. Headache, fever, or an associated autoimmune disorder should prompt the consideration of acquired etiologies. A detailed family history should be obtained to search for relatives with similar symptomatology. Laboratory tests may be useful to rule out acquired causes or as biomarkers for certain disorders. Neuroimaging, preferably with magnetic resonance imaging, is an essential tool to evaluate the presence of cerebellar atrophy or signal anomalies, to search for associated pontine atrophy, and to rule out space-occupying lesions. Electromyography and nerve conduction studies can prove the presence of clinically suspected or subclinical neuropathy and provide evidence of associated myopathy.

  1. 2.

    Following the clinical assessment, one should verify that acquired and treatable causes for ataxia have been excluded. These include vascular disease, trauma, infection, primary or metastatic tumor, excess alcohol consumption, vitamin deficiency, Creutzfeldt-Jakob disease, and immune-mediated cerebellar ataxias such as multiple sclerosis, gluten ataxia, anti-GAD (glutamic acid decarboxylase) ataxia, and paraneoplastic cerebellar degenerations. Clinical evaluation should reveal previous exposure to toxins or traumatic injuries, along with specific signs and symptoms suggestive of infectious, vascular, or metastatic disease. Laboratory tests are useful to identify vitamin deficiencies or autoimmune conditions. Specifically, testing for antibodies involved in paraneoplastic or autoimmune cerebellar degeneration may be particularly useful for patients with a subacute progression, older age at onset, and absence of family history. The paraneoplastic antibodies most associated with cerebellar degeneration are anti-Yo, anti-Hu, anti-Tr, and anti-mGluR1 antibodies; the tumors most often involved are breast and gynecological tumors, Hodgkin lymphoma, and small-cell lung carcinoma [218]. Large paraneoplastic autoantibody panels are now available and may reduce the delay associated with serial testing.

  1. 3.

    Once acquired causes have been ruled out, a genetic etiology may be considered, especially in the presence of a positive family history, early onset, chronic progressive course, or with a set of clinical signs and symptoms that is reminiscent of a well-described genetic disorder. One should bear in mind that a negative family history does not rule out a genetic cause, and sporadic cases may be due to recessive or mitochondrial inheritance, de novo mutations, genetic anticipation, incomplete penetrance, variability in disease expression, paternity error, gonadic mosaicism, or incomplete phenotyping of family members. Indeed, recessive disorders may appear as sporadic in small kindred or with incomplete family history. In other cases, a complete family history should allow identification of the mode of transmission.

  1. 4.

    If autosomal recessive inheritance is suspected, the next step in clinical evaluation is to consider age at onset and clinical signs and symptoms to evaluate if the clinical picture is reminiscent of a well-described disorder. Presentation in infancy suggests ataxia telangiectasia or autosomal recessive ataxia of Charlevoix-Saguenay. Childhood or teenage onset should raise the suspicion for Friedreich ataxia, ataxia with oculomotor apraxia 1 and 2, and POLG-related disorders. Finally, recessive ataxia with onset in adulthood is evocative of autosomal recessive cerebellar ataxia 1 and 3 and spastic paraplegia 7. However, there are large variations in the age at onset of most of the presented disorders, and Friedreich ataxia is one of the best examples with some patients presenting with late-onset (> 25 years of age) or very-late-onset Friedreich ataxia (> 40 years of age). Clinical signs and symptoms may provide clues to identify the mutated gene. Indeed, certain discriminating clinical features or combinations of neurological symptoms may be helpful to guide the clinician towards specific genes (Fig. 1 and Table 1). As one may observe in Fig. 1, none of the autosomal recessive ataxias reported up to now presents with a pure cerebellar phenotype. Even SYNE1-related autosomal recessive cerebellar ataxia 1, which used to be the prototype of a pure cerebellar phenotype [21], has recently been reported to be associated with upper and/or lower motor neuron involvement in 58% of cases, with some rare patients presenting with a very severe early-onset neuromuscular phenotype [22]. The presence of motor neuron involvement, polyneuropathy, extrapyramidal movement disorders, eye movement abnormalities such as oculomotor apraxia, intellectual impairment, and associated multisystemic involvement may guide the clinician towards a particular diagnosis. Some clinical syndromes are particularly evocative of specific disorders. Multisystemic involvement with sensory loss, muscle weakness, cardiomyopathy, diabetes, optic atrophy, and sensorineuronal hearing loss is characteristic of Friedreich ataxia, which is the prototype of a disorder associated with mitochondrial dysfunction. Other associated disorders present with similar features and occasionally epilepsy, retinal involvement, or ophthalmoplegia, such as POLG-related disorders, autosomal recessive cerebellar ataxia 2, and Marinesco-Sjogren syndrome. Extrapyramidal involvement with oculomotor apraxia, elevated α-fetoprotein, and occasional polyneuropathy are typical findings of ataxia telangiectasia, ataxia telangiectasia-like disorder, spinocerebellar ataxia recessive 26, and ataxia with oculomotor apraxia types 1, 2, and 4. Nevertheless, autosomal recessive ataxias are characterized by important phenotypic variability and significant clinical overlap between different pathologies, such that predicting the mutated gene according to the clinical phenotype is prone to errors even for ataxia experts [219]. Some laboratory tests may serve as useful biomarkers for recessive ataxias. Altered levels of vitamin E, α-fetoprotein, albumin, coenzyme Q10, cholesterol, cholestanol, lactate, sex hormones, and gonadotropins have been associated with specific disorders (see Table 1). Dosing of immunoglobulins, very long chain fatty acids, and hexosaminidase may be relevant according to clinical suspicion.

    Fig. 1
    figure 1

    Clinical classification of autosomal recessive ataxias. The gene associated with each primary recessive ataxia is classified according to the most frequent clinical syndrome described for this disorder. Note that some disorders have more complex or variable phenotypes and are placed in the overlapping areas between two categories. Genes presented in larger font represent the most prevalent ataxias

  1. 5.

    Once the clinical assessment is complete, genetic testing is indicated to confirm the mutated gene or allow a more specific diagnosis if the clinical picture is nonspecific. Initial testing should include searching for the Friedreich ataxia-associated trinucleotide repeat expansion in the FXN gene considering the high prevalence of this mutation, its incomplete coverage through the next-generation sequencing methods [1], and the heterogeneous clinical phenotype. Searching for a FXN repeat expansion can be done with frataxin protein analysis or gene analysis with Southern blot or PCR. Moreover, clinicians may consider testing for another specific gene through Sanger sequencing or multiplex ligation-dependent probe amplification (MLPA) if the clinical and paraclinical data are highly evocative of a particular disorder, if there is a confirmed mutation in a relative or in isolated populations where selected disorders are highly prevalent. Finally, a panel for the dominantly inherited CAG-repeat expansion spinocerebellar ataxias may also be considered as part of the initial assessment if family history is inconclusive regarding the mode of inheritance and considering the high prevalence of these mutations and their incomplete coverage through the next-generation sequencing methods [1].

  1. 6.

    If single gene testing does not provide a molecular diagnosis, one should consider the high-throughput NGS methods either with a multigene panel, whole exome sequencing, or whole genome sequencing. Several studies have demonstrated the efficacy and cost efficiency of multigene panels [220], targeted exome sequencing [219, 221], or whole exome sequencing [222, 223], with a diagnostic yield varying between 18 and 80%. The highest yield is obtained for patients with early-onset ataxia and positive family history and consanguinity among parents. NGS panels allow for better coverage of included genes and reduce the volume of genetic variants that are unrelated to the clinical phenotype, while exome sequencing may reveal mutations in genes that were not previously known to be associated with ataxia [1]. Whole genome sequencing may be considered in selected cases with appropriate genetic counseling, but its diagnostic yield is uncertain [224]. Once genetic testing is completed and a pathogenic mutation has been identified, it is of primary importance to provide specialized genetic counseling for the patient and his or her relatives along with symptom management and disease treatment when available. Figure 2 presents a graphical summary of the proposed clinical approach.

    Fig. 2
    figure 2

    Graphical summary of the clinical approach to a patient presenting with ataxia

Pathophysiological Mechanisms Underlying Autosomal Recessive Cerebellar Ataxias

The importance of a proper recessive ataxia classification goes beyond the clinical diagnosis perspective. Autosomal recessive ataxias can be regrouped according to the deficient cellular and metabolic pathways involved, which provide a better understanding of cerebellar physiology and of its selective vulnerability to certain metabolic defects. This is also essential from a therapeutic perspective, as disorders that belong to the same metabolic pathway may respond to the same treatment options, indicating potential for drug repurposing. Figure 3 presents a pathophysiological classification of autosomal recessive ataxias. Certain genes are presented more than once since some proteins are involved in several metabolic pathways or may interfere with other cellular processes as they accumulate in neurons or glial cells. Table 3 presents a more detailed listing of the pathogenic pathways involved along with relevant references.

Fig. 3
figure 3

Pathophysiological classification of autosomal recessive ataxias. A Purkinje cell is depicted along with a granule cell and parallel fibers. Subcellular organelles and structures are represented graphically. Each gene is classified at one or more subcellular localizations according to the different metabolic pathways involved

Table 3 Detailed pathogenic mechanisms involved in autosomal recessive cerebellar ataxias

Certain pathways are predominantly involved, notably mitochondrial dysfunction, which may result from abnormal mitochondrial DNA maintenance with progressive mutagenesis, defective mitochondrial protein synthesis and quality control, increased levels of reactive oxygen species and oxidative stress, deficient coenzyme Q10 metabolism, altered mitochondrial dynamics, defective mitochondrial chain assembly, or abnormal mitochondrial RNA maturation and processing (Table 3). Interestingly, many of the disorders caused by mitochondrial dysfunction also present with a mitochondrial clinical syndrome as shown in Fig. 1. Disorders of DNA repair mechanisms are also common, with double-strand break repair pathway or single-strand break repair complexes predominantly involved. Pathogenic mutations in these genes are also associated with a susceptibility to ionizing radiations and predisposition for cancers, but the neurological syndrome is characterized by cerebellar involvement and extrapyramidal movement disorders. It remains debated whether defective DNA repair is the main pathogenic mechanism causing the neurological phenotype [230], but the fact that several interacting genes in this pathway are involved in degenerative cerebellar ataxias suggests that the cerebellum has a peculiar susceptibility to DNA damage for which the underlying mechanism is not understood. Finally, altered synaptic morphology or synaptic dysfunction of Purkinje cells (PC) is frequently involved in recessive ataxias and is associated with aberrant morphology at the PC/parallel fiber synapse, impaired dendritic architecture, or dysregulation of glutamate transmission. Other disorders have been implicated in synaptic dysfunction through indirect evidence, for example, SLC9A1, which localizes in presynaptic terminals and is involved in the modulation of synaptic activity [254, 275]. Of interest, many of these disorders are characterized by significant cognitive impairment that goes beyond what is expected in the cerebellar cognitive-affective syndrome and cause intellectual disability, developmental delay, or dementia, highlighting the importance of synaptogenesis in cognitive development.

Discussion

We present a new clinical classification of autosomal recessive ataxias in parallel with a pathophysiological classification. The objective of this classification is to provide a tool for clinicians and researchers that facilitates the understanding of this complex group of disorders and defines this field of research. This work is based on the results of our systematic scoping review of the literature [3]. We updated this literature review and regrouped a panel of 12 international ataxia experts to build a consensus on the definition and classification of cerebellar ataxias. The task force vision is that a classification goes beyond the listing of disorders and must organize diseases in a way that allows better understanding and clinical mastery of this group of disorders. Hence, we proposed a clinical classification along with a pathophysiological classification, which enabled us to observe that there is significant overlap between these two classifications, highlighting how clinical presentation is in some cases a good projection of the underlying biochemical defect. This has potential applications from bench to bedside since treatments that address a specific pathogenic pathway may have therapeutic potential in all disorders in which this pathway is affected. The clinical classification is presented along with a structured clinical approach to a patient presenting with ataxia, which is intended as a clinical tool for expert and nonexpert clinicians. Despite the increasing accessibility of the NGS techniques, there remains a critical place for clinical judgment in the prescription of genetic tests and interpretation of results, taking into account the technical limitations and risk of finding variants of unknown significance. Recently, Renaud and colleagues published the results of a diagnostic algorithm for recessive ataxias that integrates 124 clinical features to propose three potential diagnoses among a list of 67 recessive disorders that may present with ataxia [285]. This is a very promising tool, but its pragmatic impact on molecular testing strategy, final diagnostic rate, patient management, or time efficiency remains to be validated. In the meantime, it is essential for clinicians to be at ease with a general approach to recessive ataxias with the NGS techniques often permitting molecular diagnosis when the clinical picture is nonspecific.

One of the major strengths of this classification proposal is that it is based on a consensus from a panel of international ataxia experts, thereby ensuring a proper representation of regional differences in the prevalence and clinical approach to ataxias. Moreover, the literature search was based on a systematic scoping review of the literature whose methodology has been published before and which permitted an unbiased appraisal of all potentially relevant articles. Nevertheless, there are some limitations to this classification proposal that are inherent to classifying a group of diseases that evolves very rapidly and that is highly heterogeneous. First, as new evidence emerges regarding the identification of novel ataxia-associated genes and as new phenotypes are described for previously described disorders, this classification will need to be updated. This was highlighted by the significant additions to the list of primary recessive ataxias since the original systematic review was conducted in 2016. Indeed, many new genes and new phenotypes of previously described genes have been reported in only 2 years, which suggests that there is a need for periodic updates to the present classification or an online resource. Moreover, several decisions were made in the elaboration of this classification regarding general orientation, purpose of a classification, inclusion of specific disorders, and classification categories. The lists presented here offer in our opinion the best compromise between synthesis and exhaustiveness for the expert and nonexpert clinician.

Compared with a previously published report by the Movement Disorders Society Task Force [4], we decided to exclude disorders in which cerebellar involvement is a minor or late finding in a complex multisystem phenotype or disorders that are already classified on their own, such as genes associated with Joubert syndrome. The objective was to identify the core disorders that are involved in autosomal recessive ataxias in order to define this field of research and build a classification that would be accessible for all clinicians. Indeed, with the progressive advent of affordable NGS diagnostic testing, we believe that it is most important for clinicians to be at ease with one classification and familiar with the most frequent disorders in their unique ethnical and clinical context. Disorders in which ataxia has been reported as a rare or late finding should be included in large NGS testing strategies, but in our opinion should not be categorized as primary ataxias per se. From this perspective, our classification complements the proposal by the Movement Disorders Society Task Force.

There remain some important challenges to be addressed in the field of autosomal recessive ataxias. First, the issue of a proper nomenclature system has been much debated. Recently, the Movement Disorders Society Task Force proposed a revised naming system based on an ataxia prefix associated with the gene name [4]; this was part of a larger effort to revise the nomenclature of all genetic movement disorders. This system overcomes the limitations of the numbered nomenclature, notably unconfirmed genes, and erroneously attributed phenotypes, but its ease of use by nonexperts and patients remains uncertain. Moreover, some disorders were assigned as many as three phenotypic prefixes while some other disorders that are among the most prevalent causes of recessive ataxia, such as POLG, were not assigned an ataxia prefix. Hence, there remains a debate concerning the attribution of prefixes and the integration of this naming system with other fields in neurology and other specialties as many genes involved in ataxia have very complex multisystem phenotypes. Finally, one of the most important challenges in this field of orphan diseases is to develop targeted treatment strategies that address the pathogenic mechanism underlying symptom progression. To this end, we believe that identifying common pathophysiological pathways may provide an opportunity for drug repurposing or enlarge the number of patients that are admissible for drug trials in order to find treatments for these rare but debilitating diseases.

Conclusion or Summary

We present a clinical and a pathophysiological classification of autosomal recessive cerebellar ataxias along with a clinical approach to a patient presenting with ataxia. This classification is the result of a consensus among a panel of international experts, and it promotes a unified understanding of autosomal recessive cerebellar disorders for clinicians and researchers.