Functionally Relevant Maculopathy and Optic Atrophy in Spinocerebellar Ataxia Type 1

Abstract Background Spinocerebellar ataxia type 1 (SCA‐ATXN1) is an inherited progressive ataxia disorder characterized by an adult‐onset cerebellar syndrome combined with nonataxia signs. Retinal or optic nerve affection are not systematically described. Objectives To describe a retinal phenotype and its functional relevance in SCA‐ATXN1. Methods We applied optical coherence tomography (OCT) in 20 index cases with SCA‐ATXN1 and 22 healthy controls (HCs), investigating qualitative changes and quantifying the peripapillary retinal nerve fiber layer (pRNFL) thickness and combined ganglion cell and inner plexiform layer (GCIP) volume as markers of optic atrophy and outer retinal layers as markers of maculopathy. Visual function was assessed by high‐ (HC‐VA) and low‐contrast visual acuity (LC‐VA) and the Hardy‐Rand‐Rittler pseudoisochromatic test for color vision. Results Five patients (25%) showed distinct maculopathies in the ellipsoid zone (EZ). Furthermore, pRNFL (P < 0.001) and GCIP (P = 0.002) were reduced in patients (pRNFL, 80.86 ± 9.49 μm; GCIP, 1.84 ± 0.16 mm3) compared with HCs (pRNFL, 97.02 ± 8.34 μm; GCIP, 1.98 ± 0.12 mm3). Outer macular layers were similar between groups, but reduced in patients with maculopathies. HC‐VA (P = 0.002) and LC‐VA (P < 0.001) were reduced in patients (HC‐VA [logMAR]: 0.01 ± 010; LC‐VA [logMAR]: 0.44 ± 0.16) compared with HCs (HC‐VA [logMAR]: –0.12 ± 0.08; LC‐VA [logMAR]: 0.25 ± 0.05). Color vision was abnormal in 2 patients with maculopathies. Conclusions A distinct maculopathy, termed EZ disruption, as well as optic atrophy add to the known nonataxia features in SCA‐ATXN1. Whereas optic atrophy may be understood as part of a widespread neurodegeneration, EZ disruption may be explained by effects of ataxin‐1 gene or protein on photoreceptors. Our findings extend the spectrum of nonataxia signs in SCA‐ATXN1 with potential relevance for diagnosis and monitoring.

SCA-ATXN1 is typically characterized by an adult-onset cerebellar syndrome, which is typically combined with nonataxia signs. 5,6 These include most prominently pyramidal signs, but also comprise brainstem oculomotor signs, autonomic dysfunction, late cognitive impairment, and a variety of movement disorders, such as chorea, dystonia, myoclonus, and tremor. [7][8][9][10] Retinal or optic nerve affection are well-established clinical characteristics of several spinocerebellar ataxias, such as spinocerebellar ataxia type 7 (SCA-ATXN7), but not yet described in SCA-ATXN1. 11,12 However, a few incidental case series and smaller studies suggest retinal neuroaxonal degeneration or report a potentially distinct maculopathy, including photoreceptor alterations, and a decline of visual function in a subset of SCA-ATXN1 patients. 13,14 This is further sustained by drosophila animal models of SCA-ATXN1 suggesting retinal neurodegeneration. 15 Against this background, we aimed to investigate neurodegeneration of retina and optic nerve, macular anomalies, and their functional relevance in a cross-sectional cohort of 20 index cases with SCA-ATXN1.

Study Sample
Twenty patients with genetically confirmed and manifest SCA1-ATXN1 were prospectively recruited between June 2010 and November 2016 for this cross-sectional observational study from one university ataxia clinic (Charité-Universitätsmedizin Berlin, Berlin, Germany) and examined at the NeuroCure Clinical Research Center (NCRC). The inclusion criterion was genetically confirmed diagnosis of SCA1-ATXN1. Exclusion criteria for this analysis included missing retinal optical coherence tomography (OCT) data. Two patients and three eyes in 2 other patients had OCT of insufficient image quality and were excluded from the statistical (but not from the descriptive) analysis. Nine patients (45%) were previously published in a smaller cohort study by Stricker et al., which reported OCT findings from an earlier time point. 14 Number of CAG-repeats, Scale for the Assessment and Rating of Ataxia (SARA) score 16 within 90 days of the OCT examination and time since onset of permanent gait ataxia were accessible in the clinical database for 16, 15, and 20 patients, respectively (Table 1).  Apart from the ataxia rating, a description of eye movements was available for 20 patients, from which 19 (95%) presented with abnormalities: 16 patients (80%) with saccadic pursuit, 10 patients (50%) with slowing of saccades, 11 patients (55%) with failure of fixation suppression in testing of the vestibular-ocular reflex, 7 patients (35%) with hypo-or hypermetric saccades, and 3 patients (15%) with a gazed-evoked nystagmus. For comparison, we included cross-sectional data of 22 healthy controls (HCs) from the NCRC's normative OCT database matched for sex (χ 2 < 0.001; P > 0.999) and age (W = 200; P = 0.600). Best-corrected high-contrast monocular visual acuity (HC-VA) and 2.5% low-contrast visual acuity (LC-VA), using retroilluminated Early Treatment in Diabetes Retinopathy Study (ETDRS) at 4 m distance and Sloan charts at 2 m distance, respectively, (Precision Vision, Woodstock, IL) as well as monocular color vision assessments applying the Hardy-Rand-Rittler pseudoisochromatic test (HRR) were acquired in 12 patients and 5 HCs. The study was approved by the local ethics committee of Charité-Universitätsmedizin Berlin (EA1/17515) and conducted in accordance with the applicable German laws and the current version of the Declaration of Helsinki. All participants gave written informed consent. All data is reported according to STROBE reporting guidelines.

OCT
The same Spectralis SD-OCT machine (Heidelberg Engineering, Heidelberg, Germany) with automatic real-time (ART) function for image averaging was used in all participants. 17

Statistical Analyses
Group differences in age and sex between SCA-ATXN1 and HCs were tested by χ 2 and Wilcoxon rank-sum tests, respectively. Group differences for OCT parameters and visual acuity assessments as well as correlations with clinical metrics were analyzed by generalized estimating equation models to account for withinpatient intereye associations of monocular measurements. All tests and graphical representations were performed with R software (V3.3.1; R Foundation for Statistical Computing, Vienna, Austria). 23 Statistical significance was established at P < 0.05. Because of the exploratory character of our analyses, we refrained from corrections for multiple comparison.

Data Availability
Data not provided in the article will be shared at the request of other investigators for purposes of replicating procedures and results.

Results
We found pathological disruption of EZ in 5 patients (25%; Table 2; Fig. 1). Two further patients (10%) had unspecific abnormalities (floaters, monocular epiretinal gliosis of the papillomacular bundle). The disruption of EZ was pronounced in the foveal and parafoveal area of 2 patients (Table 2; Fig. 1, Patients 1 and 2), less pronounced in the fovea of 1 patient ( Table 2; Fig. 1, Patient 3), and more temporally located and less pronounced in 2 patients (Table 2; Fig. 1, Patients 4 and 5). HRR indicated an abnormal color vision in 2 patients with EZ disruption showing a mild red-green deficiency binocularly in the first patient and monocularly in the second patient. We then analyzed group differences in neuroretinal content between eyes from patients with SCA-ATXN1, including those with macular anomalies, and HCs (Table 3; Fig. 2). pRNFL as the marker of axonal content in SCA-ATXN1 was reduced compared with HCs. Compared with the device-specific control database, 57% (N = 12) of eyes had pRNFL values within normal limits; for 19% (N = 4) of eyes, pRNFL values were borderline below, and 24% (N = 5) had pRNFL values below normal limits. Also, GCIP as a marker of ganglion cell volume was lower in SCA-ATXN1 compared with HCs. In contrast, INL and outer retinal layers were similar between SCA-ATXN1 and HCs (Table 3; Fig. 2). The 5 patients with disruption of EZ did visually not show signs of more pronounced retinal neuroaxonal loss ( Table 2). We restrained from further statistical analysis because of the small and highly unbalanced subsets. HC-VA as well as LC-VA were reduced in patients with SCA-ATXN1 compared with HCs (Table 3; Fig. 2). Because of the low sample size, we refrained from performing a statistical subgroup analysis and structure-function correlations. In

Discussion
In our study, 25% of patients with SCA-ATXN1 displayed a distinct macular pathology, a disruption of the EZ, suggesting a retinal involvement in this disease. Furthermore, SCA-ATXN1 patients presented with findings (pRNFL and macular GCIP loss) consistent with optic atrophy, which are pointing toward a neurodegenerative process in the optic nerve. Together with the reported visual function impairment, these structural changes should be considered as further nonataxia signs in SCA-ATXN1.
The macular pathologies reported here are in line with three previously published cases. 24-26 Saito et al. described an SCA-ATXN1 patient with RPE reduction in fundus examination, subsequently reduced HC-VA, and loss of color vision and classified the changes as pigmentary macular dystrophy similar to the alterations observed in SCA7. 25 Vaclavik et al. described hypopigmented macular RPE and a loss of photoreceptors with a subsequent central depression of electrical potentials in 2 patients using OCT and multifocal electroretinogram. 26 Lebranchu et al. described altered foveal lamination and abnormal spacing between RPE and external limiting membrane in 4 patients. 24 Despite different terminology and discrepancies in examination methods, provided images in combination with descriptions highly suggest that the case reports and our study all report on the same macular pathology.
It is possible that the macular pathology is caused by a yet unrecognized pleiotropic effect of mutant atxn1 on photoreceptors. Pathophysiological evidence for this was contributed by Fernandez-Funez et al., who induced the expression of polyQexpanded ATXN1 in retinae of drosophila, which subsequently developed a progressive elimination of retinal cell bodies and axonal projections. 15 In this drosophila study, phenotype severity was correlated with the expression level of mutant protein. 15 In our study, EZ disruption was not associated with disease severity or pronounced retinal neurodegeneration, which may simply be a matter of sample size. Alternatively, development of maculopathy may be independent of optic atrophy in SCA-ATXN1, which may be attributed to cell-specific transcriptional effects of expanded ATXN1. Multiple genes associated with macular disease are located in 6p21.1, for example the genes peripherin-2, which is critical for photoreceptor formation and maintenance, and guanyulate cyclase activator 1a, a causal gene for cone dystrophy. 24,27,28 To differentiate a potential causal effect of the CAG-repeat expansion on 6p23 from a nearby accompanying mutation, a gene analyses of the respective loci in affected patients would be necessary in the future.
Next to this first systematic description of the novel ellipsoid zone disruption maculopathy, our results solidify previous reports on optic nerve atrophy as part of the clinical picture of SCA-ATXN1 showing pRNFL and GCIP, but not mRNFL, thinning. Of note, mRNFL is very thin and has a high intertest variability, making a significant group difference in a small sample unlikely. 20,29 Abe et al., in 1997, first reported that SCA-ATXN1 patients showed optic nerve atrophy by fundus examination as well as color vision dysfunction and visual field abnormalities. 30 An earlier study from our group in 2011, 14 a study comprising several SCA genotypes by Pula et al., 13 and a recent case series by Nishiguchi et al. 31 reported temporal RNFL thinning and macular neurodegeneration, respectively. The neuroaxonal damage might point toward concomitant optic neuropathy as part of a central nervous system degeneration in SCA-ATXN1. 32  chronologically interpret retinal neurodegeneration during the disease process. To date, it remains unclear whether retinal changes are detectable in premanifest stages of disease as has been shown for some clinical and brain imaging markers. 33 The reported changes resulted in impaired visual function in our study-with visual acuity decline as well as with color vision deficits in patients with EZ disruption. The magnitude of changes suggest that they are clinically meaningful, and at least the color vision deficits could probably be detected in clinical visual assessments using standard procedures such as Ishihara charts. 34 Whereas the overall limited number of cases with macular pathologies preclude further analyses regarding the origin of dysfunction, the reduced visual acuity could suggest a relevance of optic atrophy for visual acuity loss. Conversely, 2 patients were suggestive of color vision deficiencies, and these 2 patients also presented with macular phenotype. This is further supported by the aforementioned case reports, also reporting, in part, color vision deficits. 24,25,30 Given the rarity of SCA-ATXN1, an important strength of the study is the comparably large sample size. However, the statistical power might have still been insufficient to confirm numerical differences in outer layers where low measurements in eyes with EZ disruption suggest a potential degenerative relevance. Further important limitations of our study include the incomplete clinical and functional (especially visual function) data and the lack of blue autofluorescence and longitudinal data to conclusively confirm a progressive retinal neurodegeneration in SCA-ATXN1. Especially, investigations in nonmanifest carriers would be of high interest in this respect.
Our findings extend the previously established spectrum of nonataxia signs in SCA-ATXN1 with potential relevance for diagnosis and monitoring. Within the scope of the proposed diagnostic algorithm by Rossi et al., 35 the ellipsoid zone disruption maculopathy may be indicative of an SCA-ATXN1 diagnosis. However, additional research is needed to consolidate this notion, especially at disease onset when differential diagnosis is most relevant, and in comparison to other ataxias, both of which were not within the scope of our current study. With 25% of patients presenting with distinct macular changes-a disruption of the EZ-this macroscopic phenotype appears frequently enough to be considered during clinical workup. The reported changes were associated with visual dysfunction, and thus exploration and examination of the visual system should be part of patient counseling in SCA-ATXN1. If confirmed in a longitudinal study, neurodegeneration of the afferent visual system measured by OCT may be used as a sensitive and easily accessible marker of global neurodegeneration in SCA-ATXN1 and-in the light of approaching therapeutic candidates-even serve as outcome candidate for clinical trials. 4,36 Author Roles