Novel Features and Abnormal Pattern of Cerebral Glucose Metabolism in Spinocerebellar Ataxia 19

Spinocerebellar ataxia type 19 (SCA19), allelic with spinocerebellar ataxia type 22 (SCA22), is a rare syndrome caused by mutations in the KCND3 gene which encodes the potassium channel Kv4.3. Only 18 SCA19/22 families and sporadic cases of different ethnic backgrounds have been previously reported. As in other SCAs, the SCA19/22 phenotype is variable and usually consists of adult-onset slowly progressive ataxia and cognitive impairment; myoclonus and seizures; mild Parkinsonism occurs in some cases. Here we describe a Swedish SCA19/22 family spanning five generations and harboring the T377M mutation in KCND3. For the first time for this disease, 18F-fluorodeoxyglucose PET was assessed revealing widespread brain hypometabolism. In addition, we identified white matter abnormalities and found unusual features for SCA19/22 including early age of onset and fast rate of progression in the late course of disease in the oldest patient of this family. Electronic supplementary material The online version of this article (10.1007/s12311-018-0927-4) contains supplementary material, which is available to authorized users.


Introduction
Spinocerebellar ataxias (SCA) are a heterogeneous and growing group of autosomal dominant diseases enumerated in chronological order from SCA1 to SCA43 according to the current classification in use [1]. Pathological CAG expansions in different genes cause SCA1, SCA2, SCA3, SCA6, and SCA7 which represent up to 60% of all SCA cases; the remaining SCA subtypes are rare and in most cases associated with conventional mutations [2,3]. Spinocerebellar ataxia types 19 (SCA19) and 22 (SCA22) are rare allelic channelopathies; linkage to chromosome region 1p21-q21 was reported first in two large Dutch and Chinese families [4,5]; and the mutations in KCND3, which encodes the Shal-related voltage-gated potassium channel Kv4.3, associated with ataxia were discovered in 2012 by two independent research groups [6][7][8]. Previously, variants and mutations in KCND3 were found to be associated with sudden unexpected death syndrome (SUDS) and later with Brugada syndrome-9 (BRGDA9) [9][10][11]. So far, only one ataxia patient harboring the mutation L450P in KCND3 has been diagnosed with Brugada syndrome A. [10,11]. Only 18 SCA19/22 families and sporadic cases of diverse ethnicities have been described so far [6,7,[10][11][12][13][14][15]. In addition, two variants of unclear significance (VUS) associated with ataxia have been identified in a recent screening study [12]. The SCA19/SCA22 phenotype consists of adultonset and slowly progressive cerebellar ataxia in most cases, frequent cognitive impairment and variable degree of myoclonus, polyneuropathy, and seizures. Mild Parkinsonism has been reported recently in two unrelated Electronic supplementary material The online version of this article (https://doi.org/10.1007/s12311-018-0927-4) contains supplementary material, which is available to authorized users. French SCA19 families [14]. The Thr377Met (T377M) mutation in KCND3 has been described only once in a Japanese patient affected by pure cerebellar ataxia [7]. The aim of this study was to perform a comprehensive characterization of four affected members of a Swedish family spanning five generations; all affected family members harbor the T377M mutation in KCND3. This characterization includes clinical and cognitive evaluations, structural brain imaging and, for the first time for this disease, brain 18

Materials and Methods
Patients Four patients (III:1, III:2, IV:1, and V:1) from the same family of Swedish origin were recruited and underwent standard clinical investigation, psychometric testing, neuroimaging studies, electroneurography (ENeG) and genetic analyses. This study was carried out in accordance with the recommendations of the ethics committee in Stockholm and the radiation protection organization at the Karolinska University Hospital (Etikprövningsnämnden dnr 2016/2538-32) with written informed consent from all patients. All patients gave written informed consent in accordance with the Declaration of Helsinki. Disease status in three affected shown in the pedigree was assigned by history (Fig. 1) [16]; [17], a z score ≤ − 1.5 SD is compatible with a significant cognitive deficit.

Imaging
Standard neuroimaging with 3T MRI scanners were performed in all four patients, an experienced neuroradiologist did a qualitative assessment. Brain metabolism was evaluated with [ 18 F] FDG PET in three patients at the Karolinska University Hospital in Huddinge. Imaging acquisition after intravenous injection of 18 F-FDG (2 MBq/ kg) was performed in a Biograph mCT PET/CT scanner (Siemens). A low-dose CT scan was used for attenuation correction of PET data. All appropriate corrections, including time of flow (TOF), were applied and reconstruction performed with OSEM (5 iterations, 21 subsets, 2.0 mm Gaussian filter). Visual analysis of PET scans was performed as well as semi-quantitatively using the syngo.via program. Comparison with healthy individuals was performed with three-dimensional stereotactic surface projection (3D-SSP) software using the whole brain as a reference region as well as an automated volume of interest  This reduction was mild and non-progressive, found at age 46 years (VOI)-based analysis of FDG uptake in the cortical and subcortical regions. None of the patients were treated with psychotropic medicines that may alter brain glucose metabolism.

Clinical Findings
The clinical features are summarized in Table 1. Age of onset was not possible to determine in the index case (IV:1) who is now 45 years old. She reported clumsiness since early childhood with clear difficulties to perform balance exercises in school. Since age 23 years, the patient has reported a clear   (Table 1). Patients III:1 and III:2 had limited school attendance. Patient III:1 went to elementary school for 7 years while patient III:2 went to a vocational school (2 years) after 8 years of elementary school. Patients III:1 and III:2 declined psychometric testing for the purpose of this study. Patient III:1 had a MoCA score of 15points with major deficits found in visuospatial/executive tasks as well as in attention and delayed recall. Patient III:2 obtained a MoCA score of 24; this examination revealed mainly deficits in attention. Patient IV:1 has a total education of 13 years and has been working as a nursing assistant, her score was 25. Patient V:1 has totally 12 years of education, and in the last 2 years in school, he went through a vocationally oriented program preparing for work with children and youth. A brief cognitive examination with MMT yielded an average result (28 points) while his MoCA score was 22 which is significantly beneath the normal range. The cognitive deficits in patient IV:1 were mild and non-significant. On the other hand, patient V:1 had a significantly low performance in tests assessing visuospatial episodic memory (ROCFT), working memory (Digit span/WAIS), picture naming ability (BNT), word fluency (FAS/COWAT), and executive function/simultaneous capacity (TMT B). In contrast, the result for verbal episodic long term memory was significantly above average (RAVLT/retention). The cognitive features are summarized in Table 2.
None of the four patients had signs of Brugada syndrome or ventricular arrhythmia; there were no cases of SUDS in this family either. Recently, patient III:1 was found to have a bifascicular block, a long-term registration with an ambulatory

Structural Brain Imaging and [ 18 F] FDG PET
The imaging data are summarized in Table 1 and are shown in Figs. 2, 3, 4, 5, 6, 7, and 8. Brain MRI revealed atrophy o f t h e v e r m i s a n d s u p r a t e n t o r i a l w h i t e m a t t e r hyperintensities (WMH) in patients III:1 and III:2 and to a lesser degree in the index case (Figs. 2, 3, and 5). These WMH were particularly widespread in III:1 (Fig. 2). The index case was not affected by any other comorbidity usually associated with WMH, suggesting that these abnormalities may be an underlying feature of SCA19/22. WMH were not evident in the youngest patient (V:1). [ 18 F] FDG PET revealed hypometabolism in the entire cerebellum, thalamus, prefrontal cortex (PFC), and parietal regions of the index case (IV:1) (Fig. 6). Her mother had reduced glucose metabolism not only in the PFC but also in the vermis as wells as in the motor and temporal cortex (Fig. 4). The youngest patient (V:1) had hypometabolism in the PFC, parietal, and inferior temporal regions but surprisingly not in the cerebellum (Fig. 8). In the latter structure, atrophy of the vermis was evident (Fig. 7).

Genetic Analyses
The heterozygous T377M (c.1130C>T) mutation in KCND3 was identified in the index case first and confirmed by Sanger sequencing. The other affected in the family (III:1, III:2, and IV:1) were found to harbor this mutation. T377M is located in a highly conserved amino acid in exon 3 that is predicted to be pathogenic by three different programs (Mutation Taster, PolyPhen-2 and SIFT). In addition, this variant is absent in ExAC and Swedgene databases. The latter contains genome data on 1000 subjects. Taken together, the T377M is now a class 5 variant according to current criteria proposed by the American College of Medical Genetics [18].

Discussion
Our findings cement the T377 variant as a causative mutation for SCA19/SCA22. This is the first time that functional imaging data is reported in SCA19/22; also new is the ethnic background of this family. Ataxia channelopathies are in general characterized by early motor onset, intellectual disability and slow disease progression as demonstrated in a recent genetic screening performed in a large European ataxia cohort [12]. Most of our findings are in line with those conclusions. In the aforementioned screening which included 412 patients, only one clear pathogenic variant and two variants of unclear significance (VUS) in the KCND3 gene were identified. [12]. In a large cohort of 1500 ataxia patients in the UK, only three SCA19/SCA22 patients were identified while a previous screening in a smaller Asian cohort yielded negative results [13,19]. Penetrance of KCND3 mutations associated with ataxia is very high as shown in the Swedish family; reduced penetrance has been described only once for KCND3 mutations (M373I) [6].
The spectrum of KCND3 mutations constitutes a unique cardiocerebral syndrome; KCND3 is the only known ataxia gene associated with cardiac arrhythmia so far. Besides the described associations with SUDS and Brugada syndrome, a recent association with atrial fibrillation has been reported [20]. Accordingly, the KCND3 gene is widely expressed; higher levels of expression are found in the cerebellum and in the heart [21][22][23]. Despite the importance of voltage-gated K + (Kv) channels for action potential repolarization in cardiomyocytes, KCND3 knockout mice do not differ from wild type animals at least in terms of survival as well as for structural and electrophysiological properties of cardiomyocytes [24]. To our knowledge, there is no published Fig. 7 Brain MRI of patient V:1. a Midsagittal and coronal T1weighted image. b Mild cerebellar atrophy data on any potential structural or functional cerebellar abnormalities in this knockout model. Of all the four mutations or variants in KCND3 associated with Brugada syndrome or SUDS, three are located in the C-terminal and one in the S6 transmembrane domain A. [9][10][11]. This scarcity of cases precludes the delineation of clear genotype-phenotype correlations at the moment; however, risk stratification for cardiac arrhythmia is warranted in patients with KCND3 mutations regardless of the mutation type or location within the gene. Other familial potassium channelopathies associated with cerebellar dysfunction include spinocerebellar ataxia type 13 (SCA13), episodic ataxia type 1 (EA1), and sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance (SESAME syndrome) which are caused by mutations in the KCNC3, KCNA1, and KCNJ10 genes, respectively [25][26][27]. Different to SCA13 in which some patients have a non-progressive course [28], SCA19/22 is clearly progressive [6,7]. Interestingly, an association between epilepsy and long QT syndrome (LQTS) has been proposed for mutations in KCNQ1; however, ataxia is, to the best of our knowledge, not part of the disease spectrum associated with KCNQ1 mutations [29]. Ataxia is on the other hand one of several acquired autoimmune disorders in the spectrum associated with antibodies to voltage-gated potassium channel (VGKC) which also include Morvan's syndrome, neuromyotonia, limbic encephalitis, and cardiac arrhythmia [30][31][32].
All the four patients in the Swedish SCA19/22 family display atrophy of the vermis and three of them display varying degrees of white matter hyperintensities (WMH). WMH are common findings among the elderly and of vascular origin; these abnormalities are associated with cognitive decline particularly when they are progressive [33,34]. Different to the . c FDG uptake is also normal in the basal ganglia and thalami on the coronal image (upper case); reference values in VOI templates are displayed in the lower cases of B and C previously described SCA19/22 families, we did not find clear atrophy in the cerebellar hemispheres or the less common cerebral atrophy [4,35]. In contrast to a Dutch SCA19 family, we did not find evidence of myoclonus nor psychiatric manifestations (e.g., impulsivity or depression) [36]. Onset in the index case and patient III:2 in the Swedish family was during early childhood, but the exact timing was not possible to determine. Early onset is in fact very rare for SCA19/22, but slow disease progression is common [15,36]. At least in one of the patients (III:1), there was evidence of global cognitive impairment as evaluated by MoCA. These deficits were predominantly in visuospatial and executive tasks (IV:1 and V:1) and are similar to what have been reported in two unrelated French families [14]. However, the French families display a more complex syndrome that includes Parkinsonism, seizures, as well as psychiatric symptoms. Previously, impairment in the Wisconsin Cart Sorting Test (WCST) suggesting impaired ability for abstract thinking and set-shifting was identified in SCA19 [14,37,36]. Similar deficits have been found in some of the polyglutamine SCAs (SCA1, SCA2, SCA3, and SCA6) [38,39]. The cognitive deficits we have found in this family and their pattern of brain glucose metabolism support the notion that a frontal-executive dysfunction occurs in SCA19 [36]. It is noticeable that the general intellectual ability was not low neither in patient IV:1 nor V:1. However, the contrast between patient IV:1 whose cognitive deficits are rather mild and the pronounced dysfunction in several cognitive domains found in patient V:1 is striking. Mild deficits are hard to reconcile with the widespread brain hypomethabolism in patient IV:1. The reasons for this remarkable intrafamilial variability are unknown at the moment. The main limitations of this study are small sample size and the fact that neuropsychological testing and [ 18 F] FDG PET were performed at different time points.
Reasonably, significant comorbidity contributed to the faster rate of motor progression and low MoCA score in the oldest patient (III:1) in the Swedish SCA19/22 family. Both T2DM and kidney failure are well-known risk factors for cognitive decline and dementia [40,41]. Differently to the described Japanese SCA19/22 patient harboring the T377M mutation in KCND3, we found polyneuropathy in two patients from the Swedish family who were also affected by T2DM. However, this feature has to be interpreted with great caution since damage to periphery nerves is commonly associated with T2DM. Nonetheless, it will be of value to learn more about non-ataxia features in relatives to the Japanese patient harboring the T377M mutation [7].
Few PET studies have been performed in SCA; so far, only two small [ 18 F] FDG PET studies on ataxia channelopathies have been published [42,43]. The pattern of hypometabolism we described in SCA19/22 is reminiscent of the abnormal pattern found mainly in SCA6 and to some degree in SCA2 [43]. In SCA6, a calcium channelopathy, reduced glucose metabolism was found in the cerebellum and frontal and prefrontal cortices [43]. Different to SCA3 we did not find any evidence of abnormal metabolism in the basal ganglia. Our findings regarding variable brain hypometabolism are not specific enough to draw definitive conclusions. Additional studies are needed in order to assess the pattern of FDG metabolism in patients with other KCND3 mutations.
Only one SCA19/22 neuropathological assessment has been reported so far; besides moderate to severe loss of Purkinje cells, reduced levels of KCND3 protein were found in cerebellar homogenates [44]. Also similar to SCA6, some neuronal loss was identified in a few brainstem nuclei [3,45]. Some of the mutant Kv4.3 proteins remain trapped in the endoplasmic reticulum leading to a loss of function [6,7]. Later, evidence of a dominant negative effect in SCA19/22 was provided for some mutations [10,11] which is similar to previous findings in EA1 and SCA13 [26,27,46,47]. Despite these important advances, many aspect of the disease remain to be explored. For instance, long-term follow-up studies are needed to determine the prevalence and risk of cardiac arrhythmia. Additional neuropathological studies of SCA19/22 are needed in order to establish genotype-phenotype and clinic-pathological correlations. Finally, the potential pathogenic role of the two recently identified VUS in KCND3 [12] and the impact of potential gene modulators on disease expressivity remain also to be investigated.
Acknowledgements Our gratitude is to the patients for their kind participation in this study. Also, thanks to Mariann Skoglund for performing one of the psychometric evaluations and to Professor Chantal Tallaksen for providing valuable opinions on the draft.
Author Contributions M. Paucar designed the study and wrote the first draft. I. Savitcheva analyzed the neuroimaging data. Å. Bergendahl wrote part of the first draft by summarizing the cognitive data. J. Laffita reviewed the data and the first draft version. P. Gustavsson and M. Nordenskjöld review of the genetic data and draft. P. Svenningsson contributed to the discussion on study design, interpretation of data, and critical review of the first draft.
Funding Information Martin Paucar was supported by the Stockholm County Council.

Compliance with Ethical Standards
This study was carried out in accordance with the recommendations of the ethics committee in Stockholm and the radiation protection organization at the Karolinska University Hospital (Etikprövningsnämnden dnr 2016/2538-32) with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the ethics committee in Stockholm.

Conflict of Interest
The authors declare that there is no conflict of interest.
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.