Oculomotor Phenotypes in Autosomal Dominant Ataxias

ORIGINAL CONTRIBUTION Oculomotor Phenotypes in Autosomal Dominant Ataxias

Ned Buttner, MD, PhD; Daniel Geschwind, MD, PhD; Joanna C. Jen, MD, PhD; Susan Perlman, MD; Stefan M. Pulst, MD; Robert W. Baloh, MD

Objective: To quantify the oculomotor features of the SCA2, being present in 100% of patients with SCA2. Im- common spinocerebellar ataxia (SCA) syndromes. paired vestibulo-ocular reflex gain occurred with SCA3 only. Patients with SCA6 had prominent deficits in smooth Setting: University ataxia clinic. tracking but normal saccade velocities and vestibulo- ocular reflex gain. Patients: Twenty probands with documented SCA mutations. Conclusions: The oculomotor findings are consistent with pure cerebellar involvement in SCA6, pontine in- Methods: Electro-oculographic recordings of sac- volvement in SCA1 and SCA2, and vestibular nerve or cadic, smooth pursuit, optokinetic, vestibular, and visual- nuclei involvement in SCA3. These phenotypes can be vestibular eye movements. useful for clinical diagnosis and for investigating the mechanism of system specificity with the SCA syn- Results: Distinct phenotype and genotype patterns were dromes. identified with modest overlap between patterns. Slow- ing of saccade peak velocities occurred only in SCA1 and Arch Neurol. 1998;55:1353-1357

LASSIFICATION OF theauto- of the repeat expansion for each is roughly somal dominant cerebellar similar, with less than about 30 repeats be- ataxias (ADCAs) has long ing asymptomatic and more than about 40 been a source of confusion being symptomatic. The size of the repeat and controversy. Harding1 correlates with disease severity and age at Cseparated this collection of hereditary, late- onset.3 Repeat expansion constitutes the mo- onset, cerebellodegenerative disorders into lecular basis of anticipation, which typi- types I through III. The most common type, cally occurs with paternal transmission. ADCA I, presents with a range of findings in- SCA6 is the lone exception to these rules, cluding ataxia, pyramidal and extrapyrami- with a smaller, stable repeat expansion dal signs, and ophthalmoplegia. ADCA II is thought possibly to cause a loss of func- similarbutalsoincludesretinaldegeneration, tion or dominant negative effect.5,6 while ADCA III involves relatively pure cer- Each SCA mutation can produce a ebellar signs. The advent of molecular genet- wide range of phenotypes, but also a single ics has shown this classification to be geneti- phenotype may arise from several differ- cally heterogeneous, composed of a variety ent genotypes.6-9 Thus, although molecu- of distinct spinocerebellar ataxias (SCAs).2 lar genetic advances have raised the pros- SCAs1through4areformsofADCAI,SCA5 pect of distinct genotype-phenotype From the Department of and SCA6 are forms of ADCA III, and SCA7 correlations, progress here has been slow Neurology (Drs Buttner, so far is the only form of ADCA II. The genes to date. We now report a dissociation be- Geschwind, Jen, Perlman, Pulst, forSCA1,SCA2,SCA3/Machado-Josephdis- tween several SCAs with respect to a single and Baloh), Neurogenetics ease (MJD), SCA6, and SCA7 have been phenotype: oculomotor function. Program (Drs Geschwind, Jen, cloned and found to contain expanded CAG and Baloh), and Division of triplet repeats.3,4 The genes for SCA4 and Head and Neck Surgery RESULTS (Dr Baloh), UCLA School of SCA5 have been linked to chromosomes 16 Medicine, Los Angeles, Calif; and 11, respectively, but have not been VISUAL FIXATION 3 and Department of Neurology, cloned. Cedars-Sinai Medical Center, Most SCAs share properties typical of Gaze-evoked nystagmus was a common fea- Los Angeles (Dr Pulst). the CAG-repeat disorders.2,3 The size range ture of this population, found in every pa-

ARCH NEUROL / VOL 55, OCT 1998 1353

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 PATIENTS AND METHODS (3°-36°) and interstep intervals (0.5-2.5 seconds). Horizon- tal smooth pursuit was induced by moving the laser dot sinusoidally at frequencies of 0.1, 0.2, and 0.4 Hz (peak ve- PATIENTS locities of 11.3/s, 22.6/s, and 45.2/s, respectively). Spontaneous nystagmus occurred in the primary position, while gaze- We performed electro-oculography in 20 probands with di- evoked nystagmus occurred on 30° horizontal or vertical gaze agnoses of SCA1, SCA2, SCA3/MJD, or SCA6. Most have deviation. Rebound nystagmus was induced by saccades back been followed up on a regular basis in the ataxia clinic at to the primary position after a sustained gaze deviation, and the University of California, Los Angeles, Neurological Ser- always beat in a direction opposite to that of the preceding vices for many years. All had clinically obvious disease of gaze deviation.13 Fixation instability, such as square-wave jerks, duration ranging from 2 to 30 years (Table 1). was examined with eyes in the primary position for 30- second epochs. MOLECULAR GENETIC ANALYSIS Horizontal optokinetic nystagmus (OKN) was ob- tained with subjects seated in the center of a drum (diam- DNA was isolated from peripheral leukocytes or lympho- eter, 1.3 m) made of heavy black cloth with 30 wide ver- blastoid cell lines as previously described.7 The SCA1 and tical white stripes placed at 15.6° intervals. Patients were SCA3/MJD alleles were amplified and analyzed on agarose asked to watch the stripes passing by without following them and acrylamide gels by standard methods. For SCA3/MJD, to the periphery. Sinusoidal rotation of the drum at 0.05 the following primer pairs were used: MJD52 and MJDB Hz (peak velocity, 60°/s) stimulated the entire visual field. 5Ј-GTAACCTTGCTCCTTAATCC-3Ј. For SCA2 allele ana- The vestibulo-ocular reflex (VOR) was tested with passive lysis, primers SCA2-A (5Ј-GGGCCCCTCACCATGTCG-3Ј) sinusoidal rotation of the subject at 0.05 and 0.4 Hz (peak and SCA2-B (5Ј-CGGGCTTGCGGACATTGG-3Ј) were velocity, 60°/s and 30°/s, respectively). Both VOR tests were used to amplify the SCA2 repeat.7,11 performed in darkness with the subject’s eyes open. Vi- For SCA6, primers F1 and R15,6 were added to 20 to sual vestibular interaction was tested in 2 ways. First, the 40 ng of human genomic DNA with standard buffer and visual VOR was tested by rotating the subject in the light nucleotide concentrations, in a final volume of 20 µL. Af- with the optokinetic drum stationary. Second, ability to ter an initial 5-minute denaturation at 95°C, 35 cycles of suppress the VOR with fixation (VOR-fix) was assessed 95°C denaturation (90 seconds), 62°C annealing (30 sec- by having the patient fixate a light-emitting diode that onds), and 72°C extension (60 seconds), followed by a fi- was attached to the sinusoidally rotating chair (0.05 Hz, nal extension of 72°C for 5 minutes were performed. Ex- 60°/s).10,12 Informed consent was given by all patients. panded alleles were reamplified with the use of phosphorus 32 end-labeled primer R1 and separated by electrophore- EYE MOVEMENT DATA ANALYSIS sis through 6% polyacrylamide sequencing gels and analyzed.11 The methods for our online computer data analysis have been reported previously.10,12 Briefly, eye position signals EYE MOVEMENT RECORDINGS were differentiated, and saccades were identified on the basis of their characteristic velocity profile. Peak velocity was Direct-current electro-oculogram recordings were per- averaged for saccade amplitudes of 10°, 20°, and 30° (bin formed in accordance with established protocols.10,12 width, ±2°). Saccade latency (time from target displace- Briefly, surface electrodes were placed at the inner and ment to eye movement) and saccade accuracy ([saccade am- outer canthi of each eye for recording horizontal eye plitude/target amplitude] ϫ 100) were determined for each movements. Vertical nystagmus and blink artifact were identified saccade. We then calculated the average sac- monitored with electrodes attached above the eyebrow cade latency and the number of hypermetric saccades and below the lower lid. Calibration was performed (Ͼ100% accuracy) for the entire test. before each subtest by having the patient look at targets Gain measurements of VOR, OKN, and smooth pur- in primary position and at 15° horizontally. The record- suit were computed as follows. The fast components were ing signal was amplified and filtered (cutoff frequency at removed, and the resulting gaps in the slow eye velocity 42 Hz, −3 dB), then digitized for online analysis (sam- record were filled by connecting the points at each end of pling rate, 200 Hz). a missing segment with a quadratic regression line. A fast To monitor saccades, smooth pursuit, spontaneous and Fourier analysis was then performed, giving the ampli- gaze-evoked nystagmus, and fixation instability, we asked the tude of the fundamental and first 5 harmonics. The ampli- subject to fixate a red laser dot projected on a white screen. tude of the fundamental was then compared with stimu- Horizontal saccades were induced by moving the laser lus velocity for computation of gain. Normative data have dot in a stepwise pattern with pseudorandom amplitudes been previously reported.10

tient with SCA1, SCA3, and SCA6 examined. Three patients tients with SCA2. Spontaneous downbeat nystagmus was with SCA2 did not have gaze-evoked nystagmus, probably even more specific, occurring in 60% (3/5) of the patients because they could not generate corrective fast components. with SCA6 but in none of the other groups. Rebound and downbeat nystagmus were more specific. Re- Square-wave jerks were the only saccadic intru- bound nystagmus was a prominent feature of SCA3, occur- sions identified. They occurred in 43% (3/7) of patients ring in 100% (7/7) of patients. In comparison, rebound nys- with SCA3, 33% (1/3) of patients with SCA1, and 20% tagmus was observed in 40% (2/5) of the patients with SCA6, (1/5) of patients with SCA2. Ocular flutter was not 33% (1/3) of the patients with SCA1, and none of the pa- seen.

ARCH NEUROL / VOL 55, OCT 1998 1354

Saccades VOR Diagnosis Pursuit, OKN, VOR-Fix, and Patient Age, Duration, PV, Hypermetric, Latency, 0.4 Hz, 0.05 Hz, 0.05 Hz, 0.4 Hz, 0.05 Hz, No. y y °/s % s 45°/s 60°/s 60°/s 30°/s 60°/s SCA1 (33)† (67) (67) (33) (67) (0) (0) (67) 1.1 50 10 280 0 308 0.27 0.12 0.33 0.40 0.15 1.2 29 3 530 31 218 0.73 0.63 0.54 0.60 0.06 1.3 58 8 600 24 274 0.63 0.43 0.67 0.81 0.35 SCA2 (100) (40) (60) (40) (0) (0) (0) (50) 2.1 49 6 295 30 192 0.90 0.75 0.45 0.61 0.06 2.2 43 11 238 0 245 0.58 ...... 0.68 . . . 2.3 42 2 325 19 209 0.78 0.67 0.67 0.64 0.25 2.4 19 10 80 0 241 0.86 ...... 0.75 . . . 2.5 39 9 185 0 258 0.43 ...... 0.78 . . . SCA3 (0) (86) (14) (43) (86) (57) (43) (43) 3.1 27 3 490 30 210 0.72 0.46 0.20 0.14 0.08 3.2 49 21 475 2 254 0.18 0.14 0.11 0.18 0.03 3.3 47 9 550 31 189 0.60 . . . 0.25 0.20 0.20 3.4 50 8 440 24 210 0.41 0.12 0.17 0.40 0.25 3.5 44 7 586 43 205 0.71 0.24 0.54 0.47 0.12 3.6 41 3 476 38 203 0.15 0.19 0.23 0.56 0.25 3.7 40 7 445 30 219 0.80 0.59 0.28 0.38 0.06 SCA6 (0) (40) (0) (100) (100) (0) (0) (100) 6.1 58 9 454 10 183 0.44 0.22 0.64 0.41 0.44 6.2 62 15 575 8 176 0.20 0.33 0.53 0.60 0.67 6.3 56 5 550 5 201 0.23 0.08 0.66 0.82 0.40 6.4 52 7 490 41 205 0.34 0.15 0.64 0.90 0.39 6.5 65 30 480 21 203 0.44 0.14 0.89 0.65 0.61 Reference NA NA Ͼ400 Ͼ14 Ͻ234 Ͼ0.58 Ͼ0.50 0.25-0.96 0.35-1.0 Ͻ0.13 values‡

*PV indicates peak velocity average for 30° saccades; OKN, optokinetic nystagmus; VOR, vestibulo-ocular reflex; VOR-Fix, VOR with fixation; NA, not applicable; and ellipses, absent fast phases. †Percentage abnormal. ‡From Moschner et al.10

SACCADES their characteristic velocity profile with a threshold for computer detection. When saccade velocity drops below this Peak saccade velocity was depressed in 100% (5/5) of the threshold,contaminationofthepursuitrecordbyslowcatch- patients with SCA2, but in none of those with SCA3 or upsaccadesaberrantlyelevatesthesmooth-pursuitgain.Nev- SCA6 (Table 1). One of 3 patients with SCA1 had slow ertheless, patients with SCA6 had a more severe pursuit defi- saccades. Figure 1 illustrates the dissociation between cit than any other group. Of the patients with SCA6, 100% involvement of the fast eye-movement system in SCA2 (5/5) displayed decreased smooth-pursuit gain, while a num- (patient 2.4 from Table 1) vs SCA3 (patient 3.2 from Table ber of patients from other groups demonstrated normal pur- 1). The eye-movement recording on the left shows pro- suitwithoutslowsaccades(Table1).Inconcertwithdecreased found impairment of saccade velocity in SCA2; here the smooth-pursuit gain, OKN gain and VOR-fix gain were each patient is unable to generate saccade velocities greater noted to be impaired in all patients with SCA6. than 100°/s at a wide range of target amplitudes. ProlongedsaccadelatencieswerecommoninSCA1and VESTIBULO-OCULAR REFLEX SCA2 but did not occur in SCA6 (Table 1). Saccade hypermetria was most common in SCA3, occurring in 86% (6/7) The VOR gain was normal in patients with SCA1, SCA2, and of the patients. However, saccade hypermetria occurred in SCA6 but decreased in patients with SCA3. Of the SCA3 all groups: 67% (2/3) of patients with SCA1, 40% (2/5) of group, 57% (4/7) were outside the normal range, and an- those with SCA2, and 40% (2/5) of those with SCA6. other 29% (2/7) were in the low normal range. Three of 5 patients with SCA2 could not produce corrective fast com- SMOOTH PURSUIT, OKN, AND FIXATION ponents so that their eyes became pinned during large- SUPPRESSION OF VOR amplitude (191°), low-frequency (0.05 Hz) rotation (Table 1). However, each of these had normal VOR gain when tested Smooth pursuit was depressed in all 4 groups, but SCA6 was with small-amplitude (12°), higher-frequency (0.4 Hz) ro- affected most severely. Pursuit data in SCA1 and SCA2 may tation where they did not require corrective fast components. be artificially high, because the decreased saccade velocities seen in these disorders can lead to contamination of the COMBINED RESULTS smooth-pursuit record by slow fast phases. As described in the “Patients and Methods” section above, saccades are iden- The combination of saccade, pursuit, and VOR data dis- tified and removed from the pursuit record on the basis of sociated the SCAs into relatively distinct phenotypic groups

ARCH NEUROL / VOL 55, OCT 1998 1355

300 300 Peak Velocity, Peak Velocity, 100 100

–40–30–20–101 10203040–40–30–20–101 10203040 Amplitude, Degrees Amplitude, Degrees

15° ° Eye 15

2 s 2 s

Target 15° 15°

Figure 1. Saccadic eye movements in patient 2.4 with spinocerebellar ataxia 2 (left) and patient 3.5 with spinocerebellar ataxia 3 (right). Horizontal monocular electro-oculogram recordings. Plots of peak velocity vs amplitude for each saccade in the random sequence (top). Dotted lines show normal range.

normal saccades and mild to moderately decreased pur- Table 2. Type and Degree of Horizontal Eye Movement suit. SCA6 had profound abnormalities of smooth- Abnormalities With Different SCA Syndromes* pursuit, OKN, and VOR-fix gains in conjunction with normal saccades and normal VOR gain. Saccade Pursuit/OKN VOR VOR-Fix Our sample size was too small to reliably correlate the Diagnosis Velocity Gain Gain Gain severity of eye movement abnormalities with disease du- SCA1 Moderate Moderate Normal Moderate ration or repeat length. For SCA1, patient 1.1 had the long- SCA2 Severe Mild Normal Mild est disease duration and also demonstrated the slowest sac- SCA3 Mild? Moderate Moderate Mild SCA6 Normal Severe Normal Severe cades. For SCA2, the 3 patients (2.2, 2.4, and 2.5) with the slowest saccades had the longest repeat lengths. Patient 2.1 *SCA indicates spinocerebellar ataxia; OKN, optokinetic nystagmus; had the smallest repeat length and the longest disease du- VOR, vestibulo-ocular reflex; and VOR-Fix, VOR with fixation. ration with intermediate slowing. For SCA3, the 2 patients with the longest disease duration had the most severely depressed VOR gain. For SCA6, there was no relationship between disease duration or repeat size and the degree of impairment of smooth pursuit and OKN gain.

1.0 1.0 0.9 COMMENT 0.9 6 0.8 0.8 2 0.7 Our findings are consistent with previous studies with a few

0.7 2 0.6 VOR Gain exceptions. Numerous clinical studies have identified su- 6 6 2 7-9 0.6 6 0.5 pranuclear ophthalmoplegia in patients with SCA3/MJD. 0.5 6 We found normal horizontal saccade velocities in all 7 of 1 1 2 0.4 ourpatientswithSCA3.Burketal14 reportedmildlydecreased 0.4 1 2 0.3 3 horizontal saccade velocities in 30% of their 32 patients with 0.3 3 0.2 SCA3. However, saccade velocity measurements in their pa- 0.2 3 3 0.1 3 3 tients with SCA3 were not significantly different from those 0.1 3 0 100 in controls. How do we explain this apparent conflict be- 0.0 0.1 200 tween clinical observations and quantitative eye-movement 0.2 0.3 300 0.4 Smooth Pursuit0.5 Gain 400 data? Few details regarding the nature of the supranuclear 0.6 0.7 500 0.8 Saccade Velocity ophthalmoplegia seen with SCA3 have been reported, but, 0.9 600 1.0 on the basis of our clinical experience, vertical-gaze disor- Figure 2. Combined smooth pursuit, saccade, and vestibulo-ocular reflex ders are much more common than horizontal-gaze disor- (VOR) data for each patient with spinocerebellar ataxia 1, 2, 3, and 6, ders in SCA3/MJD. This vertical-horizontal dissociation in showing separation into distinct phenotypes. eye movement involvement is common in diseases that af- fect the basal ganglia. Since we measured only horizontal (Table 2, Figure 2). SCA1 and SCA2 had involvement saccades,abnormalitiesofverticalsaccadeswouldnotbeiden- of the fast eye movement system with decreased saccade tified. Also, since supranuclear gaze disorders are reported velocity and/or saccade latency in conjunction with nor- to increase with disease duration,9 we probably would have mal VOR gain and mild to moderately decreased pursuit. identified a few cases of SCA3 with slow horizontal saccades SCA3 showed a decrease in VOR gain in conjunction with if we had studied more patients with longer disease dura-

ARCH NEUROL / VOL 55, OCT 1998 1356

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 tion. Only one of our patients with SCA3 had had symptoms can in fact be detected with quantitative eye movement formorethan10years.Klostermannetal15 reportedimpaired testing. VOR responses in their patients with SCA1, yet all of our patients with SCA1 had normal VOR gain. On review of their Accepted for publication February 4, 1998. eye movement recordings, however, the problem was im- This work was supported by grants AG9063, PO1 paired fast components rather than VOR slow components, DC02952, NS011849, and DC00008-21 from the Na- similar to the problem we had with the low-frequency data tional Eye Institute, Bethesda, Md. inpatientswithSCA2.Testingwithsmaller-amplitude,higher- Wegratefullyacknowledgethepatients,theirfamilies,and frequency stimuli would probably have shown normal VOR thestaffintheUniversityofCalifornia,LosAngeles,AtaxiaClinic. responses in their patients with SCA1. KathleenJacobson,BA,conductedmostofthetestingandhelped It is of interest to compare SCA phenotypes with those preparethemanuscript.HudaZohgbi,MD,generouslyprovided of older diagnostic categories, such as olivopontocerebel- DNA from a known SCA6 heterozygote. lar atrophy and cerebello-olivary atrophy.1 Patients previ- Reprints: Robert W. Baloh, MD, Department of Neu- ouslydiagnosedashavingolivopontocerebellaratrophyshare rology, UCLA, Box 951769, Los Angeles, CA 90095-1769 similar oculomotor abnormalities with patients now known (e-mail: [email protected]). to carry mutations for SCA1, SCA2, and SCA3. The olivo- pontocerebellar atrophy phenotype included impaired sac- REFERENCES cade velocity and latency, now seen to be characteristic of SCA1 and SCA2, and included depressed VOR now seen 1. Harding AE. Clinical features and classification of inherited ataxias. Adv Neurol. to be characteristic of SCA3.10 Cerebello-olivary atrophy and 1993;61:1-14. 2. Rosenberg RN. Autosomal dominant cerebellar phenotypes: the genotype has SCA6 share the common findings of normal saccade veloc- settled the issue. Neurology. 1995;45:1-5. ityandVORgainbutseverelyimpairedsmoothpursuit,OKN, 3. Nance MA. Clinical aspects of CAG repeat diseases. Brain Pathol. 1997;7:881-900. and VOR-fix gains with both downbeat and rebound nys- 4. David G, Abbas N, Stevanin G, et al. Cloning of the SCA7 gene reveals a highly 10 unstable CAG repeat expansion. Nat Genet. 1997;17:65-70. tagmus. Manypatientsdiagnosedashavingcerebello-olivary 5. Zhuchenko O, Bailey J, Bonnen P, et al. Autosomal dominant cerebellar ataxia atrophy likely carried the expanded repeat in the ␣1A- (SCA6) associated with small polyglutamine expansions in the ␣-1A voltage- 5 dependent calcium channel. Nat Genet. 1997;15:62-69. voltage–dependentcalciumchannelnowtermedSCA6. Also 6. Geschwind DH, Perlman S, Figueroa CP, Baloh RW, Pulst SM. Spinocerebellar of note is the similarity in oculomotor profiles of SCA3 and ataxia type 6 (SCA6). Neurology. 1997;49:1247-1251. Friedreich ataxia. Impaired VOR, saccade dysmetria, and 7. Geschwind DH, Perlman S, Figueroa CP, Treiman LJ, Pulst SM. The prevalence 10 and wide clinical spectrum of the spinocerebellar ataxia type 2 trinucleotide re- slow-wave jerks are common in both conditions. peat in patients with autosomal dominant cerebellar ataxia. Am J Hum Genet. Evidencefromexperimentalanimalsindicatesthatmany 1997;60:842-850. of the SCA oculomotor abnormalities described above lo- 8. Matilla T, McCall A, Subramony SH, Zoghbi HY. Molecular and clinical correlations in spinocerebellar ataxia type 3 and Machado-Joseph disease. Ann Neurol. calize to the cerebellum. Virtually all patients with SCA dis- 1995;38:68-72. played gaze-evoked nystagmus and varying degrees of in- 9. Durr A, Stevanin G, Cancel G, et al. Spinocerebellar ataxia 3 and Machado- Joseph disease. Ann Neurol. 1996;39:490-499. volvement of the slow visual tracking system, ie, decreased 10. Moschner C, Perlman S, Baloh RW. Comparison of oculomotor findings in the smooth-pursuit gain, decreased OKN gain, and increased progressive ataxia syndromes. Brain. 1994;117:15-25. VOR-fix gain. These abnormalities are associated with le- 11. Pulst SM, Nechiporuk A, Nechiporuk T, et al. Moderate expansion of a normally bial- lelic trinucleotide repeat in spinocerebellar ataxia type 2.Nat Genet. 1996;14:269-276. sions of the posterior vermis, flocculus, and parafloccu- 12. Baloh RW, Langhofer L, Honrubia V, Yee RD. On-line analysis of eye move- lus.16-19 Rebound and downbeat nystagmus localize to the ments using a digital computer. Aviat Space Environ Med. 1980;51:563-567. flocculus and paraflocculus as well.13,16 Saccade dysmetria 13. Hood JD. Further observations on the phenomenon of rebound nystagmus. Ann 19,20 N Y Acad Sci. 1981;374:532-539. localizestothedorsalcerebellarvermisandfastigialnuclei. 14. Burk K, Abele M, Fetter M, et al. Autosomal dominant cerebellar ataxia type I clinical Abnormalities of the fast eye-movement system imply ex- features and MRI in families with SCA1, SCA2 and SCA3. Brain. 1996;119:1497-1505. 15. Klostermann W, Zuhlke C, Heide W, Kompf D, Wessel K. Slow saccades and other tracerebellar disorders, specifically the paramedian pontine eye movement disorders in spinocerebellar atrophy type 1. J Neurol. 1997;244: reticular formation, to explain the decreased saccade veloci- 105-111. ties with SCA1 and SCA2.21 The decreased VOR gain with 16. Zee DS, Yamazaki A, Butler PH, Gucer G. Effects of ablation of flocculus and paraflocculus on eye movements in primate. J Neurophysiol. 1981;46:878-899. SCA3 suggests bilateral involvement of primary vestibular 17. Waespe W, Cohen B, Raphan T. Role of the flocculus and paraflocculus in op- neurons (as has been reported with FA22 or brainstem tokinetic nystagmus and visual-vestibular interactions: effects of lesions. Exp Brain vestibular pathways. Magnetic resonance imaging studies Res. 1983;50:9-33. 18. Suzuki DA, Noda H, Kase M. Visual and pursuit eye movement–related activity in in patients with SCA are consistent with these localizations, posterior vermis of monkey cerebellum. J Neurophysiol. 1981;46:1120-1139. showing pure cerebellar atrophy in patients with SCA6 and 19. Versino M, Hurko O, Zee DS. Disorders of binocular control of eye movements 15 in patients with cerebellar dysfunction. Brain. 1996;119:1933-1950. pontine atrophy in patients with SCA1 and SCA2. 20. Sato H, Noda H. Saccadic dysmetria induced by transient functional decortica- How CAG repeat expansion in different genes can tion of the cerebellar vermis. Exp Brain Res. 1992;88:455-458. produce system-specific abnormality is not known. The 21. Henn V, Lang W, Hepp K, Reisine H. Experimental gaze palsies in monkeys and their relation to human pathology. Brain. 1984;107:619. repeat expansion lies within the coding region of each 22. Spoendlin H. Optic and cochleovestibular degenerations in the hereditary atax- SCA gene and is predicted to give rise to a polygluta- ias, II: temporal bone pathology in two cases of Friedreich’s ataxia with vestibulo- mine tract.5,23,24 Recent reports suggest that the polyglu- cochlear disorders. Brain. 1974;97:41. 23. Orr HT, Chung MY, Banfi S, et al. Expansion of an unstable trinucleotide CAG tamine tract itself acts as a common effector mechanism repeat in spinocerebellar ataxia type 1. Nat Genet. 1993;4:221-226. leading to toxic effects.25,26 If indeed this is the case, in- 24. Kawaguchi Y, Okamoto T, Taniwaki M, et al. CAG expansions in a novel gene for Machado-Joseph disease at chromosome 14q32.1. Nat Genet. 1994;8:221-228. dividual proteins may guide cell-specific expression of 25. Davies SW, Turmaine M, Cozens BA, et al. Formation of neuronal intranuclear the repeat expansion. A first step to answering the above inclusions underlies the neurological dysfunction in mice transgenic for the HD question for the SCAs is to define the extent of system mutation. Cell. 199;90:537-548. 26. Sherzinger E, Lurz R, Turmaine M, et al. Huntington-encoded polyglutamine ex- specificity and the identity of the systems affected. Here pansions form amyloid-like protein aggregates in vitro and in vivo. Cell. 1997; we show that relatively specific phenotypes for the SCAs 90:549-558.

ARCH NEUROL / VOL 55, OCT 1998 1357