Spinocerebellar Ataxia Type 3 (SCA3): Thalamic Neurodegeneration Occurs Independently from Thalamic Ataxin‐3 Immunopositive Neuronal Intranuclear Inclusions

Abstract

In the last years progress has been made regarding the involvement of the thalamus during the course of the currently known polyglutamine diseases. Although recent studies have shown that the thalamus consistently undergoes neurodegeneration in Huntington’s disease (HD) and spinocerebellar ataxia type 2 (SCA2) it is still unclear whether it is also a consistent target of the pathological process of spinocerebellar ataxia type 3 (SCA3). Accordingly we studied the thalamic pathoanatomy and distribution pattern of ataxin‐3 immunopositive neuronal intranuclear inclusions (NI) in nine clinically diagnosed and genetically confirmed SCA3 patients and carried out a detailed statistical analysis of our findings. During our pathoanatomical study we disclosed (i) a consistent degeneration of the ventral anterior, ventral lateral and reticular thalamic nuclei; (ii) a degeneration of the ventral posterior lateral nucleus and inferior and lateral subnuclei of the pulvinar in the majority of these SCA3 patients; and (iii) a degeneration of the ventral posterior medial and lateral posterior thalamic nuclei, the lateral geniculate body and some of the limbic thalamic nuclei in some of them. Upon immunocytochemical analysis we detected NI in all of the thalamic nuclei of all of our SCA3 patients. According to our statistical analysis (i) thalamic neurodegeneration and the occurrence of ataxin‐3 immunopositive thalamic NI was not associated with the individual length of the CAG‐repeats in the mutated SCA3 allele, the patients age at disease onset and the duration of SCA3 and (ii) thalamic neurodegeneration was not correlated with the occurrence of ataxin‐3 immunopositive thalamic NI. This lack of correlation may suggest that ataxin‐3 immunopositive NI are not immediately decisive for the fate of affected nerve cells but rather represent unspecific and pathognomonic morphological markers of SCA3.

INTRODUCTION

Spinocerebellar ataxia type 3 (SCA3 or Machado–Joseph disease, MJD), first described among Portuguese families living in the United States, represents the most frequent autosomal dominant cerebellar ataxia in Europe and the United States 30, 31, 32, 61, 63, 64, 67, 71). Because of its underlying genetic defect SCA3 is assigned to the growing group of CAG‐repeat or polyglutamine diseases 20, 44, 45, 46, 63, 67). These currently untreatable neurological disorders include not only SCA3 and five additional spinocerebellar ataxias (ie, SCA1, SCA2, SCA6, SCA7 and SCA17), but also Huntington’s disease (HD), and spinobulbar muscular and dentatorubro‐pallidoluysian atrophy, all of which are characterized molecular biologically by the presence of expanded and meiotic unstable (CAG_n) trinucleotide repeat sequences at specific gene loci. These pathological expanded CAG‐repeats encode abnormally long polyglutamine tracts in the disease proteins 44, 46, 63, 72).

The SCA3 gene has been mapped on chromosome 14q24.3–q32.2 and contains 12–40 CAG‐repeat sequences in healthy individuals, but is expanded to sequences of approximately 56–84 CAG‐repeats in affected patients and at‐risk carriers 11, 17, 31, 32, 44, 62, 63, 64). The encoded disease protein (ie, ataxin‐3) represents a protein of unknown physiological function and is widely expressed in neural and non‐neural human tissue. In SCA3 patients its mutated form tends to aggregate into neuronal intranuclear inclusion bodies (NI) within surviving nerve cells 12, 17, 32, 44, 45, 46, 49, 61, 67, 72). As they were originally observed in select degenerated brain regions of SCA3 patients 45, 46, 61, 67) and thereafter in some brain regions believed to be spared 20, 40, 67, 69, 70), the relationship between these ataxin‐3 immunopositive NI and neurodegeneration as well as their precise role in the pathogenesis of SCA3 still remain unclear 12, 40, 44, 61, 67, 69, 70). Although progress has been made in the past concerning the central nervous distribution of these NI 12, 40, 44, 45, 57, 58, 61, 67, 69, 70), this is the first SCA3 study that (i) provides detailed comparisons of the anatomical distribution of neurodegeneration with that of NI in a given central nervous system (CNS) region; and (ii) considers potential influential factors (ie, onset of disease, duration of disease, length of the expanded CAG‐repeats) with respect to the distribution pattern of ataxin‐3 containing NI.

The human thalamus is integrated into a variety of functional systems which are clearly affected (ie, somatomotor and somatosensory systems) or seemingly spared in SCA3 (ie, auditory, visual, olfactory, limbic and higher‐order processing sensory systems) 1, 28, 29, 38, 56) and according to our initial SCA3 case study may include severely and markedly degenerated as well as apparently spared nuclei (56). As (i) the thalamus therefore, represents an adequate CNS model for the systematic comparison of the pathoanatomy of SCA3 and the distribution pattern of ataxin‐3 immunopositive NI; and (ii) the advantages of an exhaustive comparison of the entire CNS pathoanatomy of SCA3 with the central nervous distribution pattern of ataxin‐3 immunopositive NI are questionable, we decided to extend our recent case study (56) and performed a systematic study of the thalamic degeneration and distribution pattern of ataxin‐3 immunopositive NI in nine clinically diagnosed and genetically confirmed SCA3 patients and carried out a detailed statistical analysis of our findings.

PATIENTS AND METHODS

Clinical course.  Our nine SCA3 patients (three females and six males; mean age at death: 59.0 ± 17.4 years; Table 1) descended from well‐known SCA3 families in the northern part of the Netherlands (cases 1–4, 6, 8, 9; Table 1) or the Northrhine‐Westfalia area in Germany (cases 5, 7; Table 1). Disease onset was determined as the time when the patient (or a close relative) noticed without a doubt the first neurological signs of SCA3, whereby the mean age at disease onset was 37.6 ± 14.1 years (Table 1).

Table 1.

Synopsis of the spinocerebellar ataxia type 3 patients (SCA3) and control individuals studied. Case number, age, gender (F = female, M = male), clinical diagnosis (SCA3), and number of expanded CAG‐repeats (CAG) in the mutated SCA3 allele.

Case	Age	Gender	Clinical diagnosis	CAG
1	24	M	SCA3	82
2	45	M	SCA3	69
3	52	M	SCA3	69
4	56	M	SCA3	74
5	60	M	SCA3	73
6	66	M	SCA3	70
7	73	F	SCA3	71
8	75	F	SCA3	68
9	80	F	SCA3	65
10	17	M	Control	n.d.
11	40	M	Control	n.d.
12	50	M	Control	n.d.
13	53	M	Control	n.d.
14	54	F	Control	n.d.
15	57	M	Control	n.d.
16	60	M	Control	n.d.
17	61	M	Control	n.d.
18	63	M	Control	n.d.
19	65	M	Control	n.d.
20	69	F	Control	n.d.
21	74	F	Control	n.d.
22	80	F	Control	n.d.

In seven of our nine SCA3 patients, the initial disease symptoms were gait and stance ataxia (cases 2, 4–9; 1, 2), while one patient initially presented with dystonia followed by gait and stance ataxia (case 1; 1, 2) and the other with diplopia followed by ataxic symptoms (case 3; 1, 2). All of the SCA3 patients included in this study suffered from progressive gait, stance, limb and truncal ataxia, progressive incoordination of the upper and lower limbs, dysarthria and dysphagia (mean age at onset of dysphagia: 51.1 ± 19.1 years; Table 2), dysmetrical saccades and saccadic smooth pursuits and were eventually wheel‐chair‐bound. Additional disease symptoms included: abducens paresis (cases 1–4, 6, 8; Table 1), writing problems (cases 1, 2, 6, 8; Table 1), reduced epicritic sensation in the upper and lower limbs (cases 2–9; Table 1), and cognitive decline (cases 5, 8; Table 1). Eight of our SCA3 patients died from aspiration pneumonia (cases 1, 2, 4–9; 1, 2) and one patient committed suicide (case 3; 1, 2).

Table 2.

Select clinical data of the spinocerebellar ataxia type 3 (SCA3) patients examined. Onset of initial symptoms (years), duration of SCA3 (years), onset of dysphagia (years), underwent aspiration pneumonia (yes: +; no: −).

Patient	Initial disease symptom	Onset of initial symptoms	Duration of disease	Onset of dysphagia	Aspiration pneumonia	Cause of death
1	Dystonia	13	11	17	+	Aspiration pneumonia
2	Gait and stance ataxia	25	20	30	+	Aspiration pneumonia
3	Diplopia	30	22	42	−	Suicide
4	Gait and stance ataxia	30	26	45	+	Aspiration pneumonia
5	Gait and stance ataxia	38	22	60	+	Aspiration pneumonia
6	Gait and stance ataxia	50	36	60	+	Aspiration pneumonia
7	Gait and stance ataxia	50	23	67	+	Aspiration pneumonia
8	Gait and stance ataxia	47	27	62	+	Aspiration pneumonia
9	Gait and stance ataxia	55	25	77	+	Aspiration pneumonia

Genetic analysis.  In all SCA3 patients the clinical diagnosis was confirmed by genetic analysis, which was performed by genotyping the DNA extracted from peripheral lymphocytes with polymorphic dinucleotide repeat sequences that flank the MJD1 gene locus on chromosome 14q24.3–q32.2 31, 68). It revealed sequences of 65–82 CAG‐repeats in the diseased alleles (Table 1), thereby identifying all of the clinically diagnosed SCA3 patients as carriers of the mutated SCA3 gene.

Tissue treatment.  In this postmortem study the brains of the nine clinically diagnosed and genetically confirmed SCA3 patients together with those of 13 representative control individuals without medical histories of neurological or psychiatric diseases (four females and nine males; mean age at death 57.5 ± 15.9 years) were examined (Table 1). The examination of the brains was approved by the Ethical Board of the Faculty of Medicine at the J. W. Goethe‐University, Frankfurt/Main. The degenerative findings in the thalamus of the 75‐year‐old female SCA3 patient (case 8; Table 1) have been recently reported in a case study (56).

Autopsies were performed on the SCA3 patients within 18.9 ± 12.5 h (range 8–48 h) and in the control individuals within 20.4 ± 10.1 h (range 12–38 h) postmortem. After fixation of the brains in a 4 % non‐buffered, aqueous formaldehyde solution, the brainstems together with the cerebella of the control individuals and the SCA3 patients were severed perpendicular to the long axis of the brainstem at the level of the inferior colliculus.

For purposes of routine neuropathological examination (performed by R.A.I. de Vos) 24 tissue blocks from the right cerebral hemisphere and midbrain of the SCA3 patients were embedded in paraffin, cut into 6‐µm‐thick frontal sections and stained with gallocyanin (21). The left cerebral hemispheres together with the midbrain of the SCA3 patients and the control individuals were embedded in polyethylene glycol (PEG 1000, Merck, Darmstadt, Germany) 4, 65) and then cut into an uninterrupted series of 100 µm thick frontal sections.

Each serial collection of the first, 11th, 21st etc. of the 100‐µm left hemispheral sections was stained for neuronal lipofuscin pigment (aldehydefuchsin) and Nissl material (Darrow red) 8, 60) and employed for the anatomical delineation of the thalamic nuclei 22, 23, 24, 28, 38, 56), as well as for the pathoanatomical assessment of neurodegeneration in the SCA3 patients. The second, 12th, 22nd, etc. serial left hemispheric sections of the SCA3 patients and control individuals were stained according to a modified Heidenhain procedure to analyze the structural integrity and myelinization of the thalamic peduncles and traversing thalamic fiber tracts 25, 47). Furthermore, reactive astrocytes were highlighted by treating the collections of the third, 13th, 23rd, etc. left hemispheric sections with a rabbit polyclonal antibody against glial fibrillary acidic protein (GFAP) (1:500, Dako, Glostrup, Denmark). Finally, in all of the SCA3 patients and control cases immunocytochemistry was performed to visualize ataxin‐3 immunopositive NI. To this end a rabbit polyclonal anti‐ataxin‐3 antibody (1:500) was applied to the serial collections of the fourth, 14th, 24th, etc. serial hemispheric sections of the SCA3 patients and control cases 45, 46). The specifity of the immunostaining was analyzed by omission of the primary antibodies. For single immunostaining, the hemispheric sections were incubated with the primary antibodies for 12–18 h at room temperature, followed by incubation with biotinylated anti‐rabbit immunoglobulins for 90 minutes at room temperature. Bound antigens were visualized with the AB complex (Vectastain, Vector Laboratories, Burlingame, CA, USA) and 3,3‐diaminobenzidine‐tetra‐HCl/H₂O₂ (DAB D5637, Sigma, Taufkirchen, Germany). Nickel‐intensification was applied to select tissue sections to improve the contrast of the ataxin‐3 immunostainings.

In addition, in each instance representative tissue sections from the neo‐ and allocortex (ie, hippocampus and entorhinal region) were stained with a modified Gallyas silver iodide method to show Alzheimer’s disease (AD)‐related neurofibrillary changes. Neuropathological classification of AD‐related cortical neurofibrillary pathology was performed on these sections according to the acknowledged Braak and Braak staging system (6).

The human thalamus is among the targets of the Parkinson’s disease (PD)‐ and AD‐related intraneuronal inclusion body pathologies 5, 7, 53, 54). Therefore, in all control individuals and SCA3 patients we performed immunostainings to visualize alpha‐synuclein‐positive Lewy bodies and Lewy neurites as well as AD‐related intraneuronal cytoskeletal pathology consisting of a distinct and abnormally phosphorylated form of the cytoskeletal protein tau. This was achieved by pre‐treating select 100‐µm hemispheral sections through the thalamus of all control individuals and SCA3 patients with 100% formic acid for 10 minutes according to a standard protocol designed to inhibit endogenous peroxidase and prevent non‐specific binding, followed by incubation for 12–18 h in the affinity purified alpha‐synuclein antiserum (Afshp) at a dilution of 1:2000 (13). Subsequent to processing with biotinylated secondary antibodies (anti‐sheep immunoglobulins for 90 minutes), reactions were visualized with the AB complex and 3,3‐diaminobenzidine‐tetra‐HCl/H₂O₂ (DAB D7679). Likewise, select hemispheral sections through the thalamus of the control individuals and SCA3 patients were immunotreated with the anti‐tau antibody AT8 (Innogenetics, Ghent, Belgium) to render visible the abnormally phosphorylated cytoskeletal tau protein (3). Incubation of free‐floating sections was carried out for 18 h at room temperature (antibody dilution 1: 2000) and incubation with the second biotinylated antibody (anti‐mouse immunoglobulins) was performed for 2 h. Immunoreactions again were visualized by means of the AB complex and 3,3‐diaminobenzidine‐tetra‐HCl/H₂O₂ (DAB D7679).

Subsequent to tissue treatment we assessed the extent of nerve cell loss and reactive astrogliosis in the thalamic nuclei and the loss of fibers and myelin within the reaches of the thalamus of the SCA3 patients and classified these pathological findings into none discernible (0), obvious (+), or severe (++) (Table 3), while the presence of ataxin‐3 immunopositive NI was categorized as absent (0), an isolated few (+), or numerous (++) (Table 3). All investigations of unconventional thick sections including those immunostained with the polyclonal anti‐ataxin‐3 antibody, the anti‐tau antibody AT8 and the alpha‐synuclein antiserum were performed by U. Rüb.

Table 3.

Extent of thalamic neuronal loss in the spinocerebellar ataxia type 3 (SCA3) patient 5 (N, neuronal loss; none discernible 0, obvious +, severe ++) and presence of ataxin‐3 immunopositive intranuclear nuclear inclusions (NI; absent 0, an isolated few present +, numerous present ++). Abbreviations: AND = anterodorsal nucleus; ANP = anteroprincipal nucleus; LD = laterodorsal nucleus; PT = parataenial nucleus; PV = paraventricular nuclei; MD = mediodorsal nuclei; PU m = pulvinar, medial nucleus; CL = Central lateral nucleus; CEM = central medial nucleus; CU = cucullar nucleus; PC = paracentral nucleus; CM = centromedian nucleus; LI‐SG = limitans‐suprageniculate‐complex; PF = parafascicular nucleus; SPF = subparafascicular nucleus; LGB = lateral geniculate body; LP = lateral posterior nucleus; MGB = medial geniculate body; PU a = pulvinar, anterior nucleus; PU i = pulvinar, inferior nucleus; PU l = pulvinar, lateral nucleus; VA = ventral anterior nucleus; VL = ventral lateral nucleus; VPL = ventral posterior lateral nucleus; VPM = ventral posterior medial nucleus; VPM pc = ventral posterior medial nucleus, parvocellular part; RT = reticular nucleus.

Case	1		2		3		4		5		6		7		8		9
	N	NI	N	NI	N	NI	N	NI	N	NI	N	NI	N	NI	N	NI	N	NI
Anterior thalamic nuclei
AND	−	+	−	+	−	++	−	+	−	++	−	+	−	+	+	+	−	+
ANP	−	+	−	+	−	+	−	+	−	++	−	+	−	+	−	+	−	+
LD	−	+	−	+	−	++	−	+	−	++	−	+	−	+	+	+	−	+
Midline thalamic nuclei
PT	−	+	−	+	−	+	−	+	−	++	−	+	−	+	−	+	−	+
PV	−	++	−	+	−	+	−	+	−	++	−	++	−	++	−	+	−	+
Medial thalamic nuclei
MD	−	+	−	++	−	++	−	+	−	++	−	++	−	++	−	+	−	+
PU m	−	++	−	+	−	++	−	++	−	++	−	++	−	++	+	++	−	+
Rostral intralaminar nuclei
CL	−	+	−	+	−	+	−	+	−	++	−	+	−	++	−	+	−	+
CEM	−	+	−	+	−	+	−	+	−	++	−	++	−	++	−	+	−	+
CU	−	+	−	++	−	++	−	++	−	++	−	++	−	++	−	+	−	++
PC	−	+	−	+	−	+	−	+	−	++	−	+	−	+	−	+	−	+
Caudal intralaminar nuclei
CM	−	+	−	+	−	+	−	+	−	+	−	+	−	+	−	+	−	+
LI‐SG	−	+	−	++	−	++	−	+	−	++	−	++	−	++	−	+	−	++
PF	−	+	−	+	−	+	−	++	−	++	−	++	−	++	−	+	−	+
SPF	−	++	−	+	−	++	−	++	−	++	−	++	−	++	−	+	−	+
Lateral basal nuclei
LGB	+	+	−	+	−	++	−	+	−	++	+	+	−	++	+	+	−	+
LP	+	+	−	+	−	++	−	+	−	+	+	++	−	+	+	+	−	+
MGB	−	+	−	+	−	++	−	+	−	++	−	+	−	+	−	+	−	+
PU a	−	+	−	+	−	++	−	++	−	+	−	++	−	++	−	+	−	+
PU i	++	++	+	++	+	++	+	+	−	++	+	+	−	++	++	+	+	+
PU l	+	++	+	++	+	+	+	+	−	++	+	+	−	++	+	+	+	+
VA	++	++	+	+	+	+	+	+	+	++	+	+	+	+	++	+	+	+
VL	++	+	+	+	+	+	+	+	+	++	+	+	+	++	+	+	+	+
VPL	+	+	+	+	+	+	+	+	−	+	+	+	−	+	+	+	+	+
VPM	+	+	−	+	−	++	−	+	−	+	+	+	−	++	+	+	−	+
VPM pc	−	+	−	++	−	+	−	+	−	+	−	++	−	+	−	+	−	+
RT	++	+	+	+	+	+	+	+	+	+	+	+	+	+	++	+	+	+

Statistical analysis.  The nonparametric Kruskal and Wallis H‐test was applied to determine whether either the extent of the individual overall thalamic degeneration in a given SCA3 patient (as expressed by a score summing up the assessed extent of neurodegeneration of the affected thalamic nuclei of a given SCA3 patient according to Table 3), the extent of the degeneration of a specific thalamic nucleus or the presence of NI in specific thalamic nuclei of our SCA3 patients were dependent upon the individual length of the CAG‐repeats in the mutated SCA3 allele, the age at disease onset or the duration of SCA3. Subsequent to the H‐test a nonparametric trend test for the detection of significant linear, quadratic or cubic trends in association was carried out. In addition, the correlation between the severity of neurodegeneration of specific thalamic nuclei and the presence of NI was described using Kendalls’s rank correlation coefficient tau (τ) (2). This coefficient was also used to describe the correlation between the extent of thalamic neurodegeneration and the distribution and severity of GFAP‐immunopositive astrogliosis in each given SCA3 patient.

RESULTS

Degenerative changes.  Upon macroscopic examination marked atrophy of the basis of the pons, medial cerebellar peduncle and bulge of the inferior olive was observed in all of our SCA3 patients (Figure 1). Routine light microscopical examination of the gallocyanin‐stained thin tissue sections through the right cerebral hemispheres and midbrain of all SCA3 patients disclosed considerable neuronal loss in the internal segment of the pallidum, in the red and subthalamic nuclei, as well as in the substantia nigra, whereas the thin sections through the right thalami revealed no indications for breakdown of nerve cells.

Figure 1.

The brainstem in spinocerebellar ataxia type 3 (SCA3). (A) Anterior aspect of the pons and medulla oblongata of a representative 53‐year‐old male control individual (case 13; Table 1) in comparison with (B) that of a 24‐year‐old male patient who suffered 11 years from SCA3 (case 1; 1, 2) and (C) that of a 56‐year‐old male patient who suffered 26 years from SCA3 (case 4; 1, 2). The basis of the pons (asterisk), the medial cerebellar peduncle (arrows) and the bulge of the inferior olive (arrowheads) of both SCA3 patients were markedly atrophic. Scale bar is valid for A, B and C. Abbreviations: V = trigeminal nerve; VI = abducens nerve; VII = facial nerve; VIII = vestibulocochlear nerve.

The investigation of unconventionally thick pigment‐Nissl‐stained tissue sections through the SCA3 patients’ left hemispheres and midbrains confirmed the findings of the routine neuropathological examination, that is, damage to the internal segment of the pallidum, the red and subthalamic nuclei, and to the substantia nigra. However, in contrast to the routine neuropathological examination, we also found that the thalamus consistently underwent neurodegeneration in our SCA3 cases. The thalamic nuclear groups differed with respect to their vulnerability for the neurodegenerative process of SCA3, whereby the distribution pattern of thalamic neurodegeneration was not associated with the descent of the SCA3 patients (2, 3; Table 3). In general, the lateral basal group with its motor and sensory nuclei and the extraterritorial reticular nucleus were most frequently and severely involved, while the nuclei of the anterior, midline, medial and intralaminar groups, most of which are assigned to the limbic system, were either spared or only occasionally involved (2, 3; Table 3). More specifically, the motor ventral anterior and ventral lateral and the extraterritorial reticular nuclei were consistently degenerated (Figure 2A,B,E,F; Table 3) and the somatosensory ventral posterior lateral nucleus and the visual lateral and inferior subnuclei of the pulvinar affected in seven of the SCA3 patients (2, 3; Table 3). Finally, neurodegeneration occurred in the visual lateral geniculate body of four (Figure 2D,H; Table 3), in the motor lateral posterior nucleus and the somatosensory ventral posterior medial nucleus of three (Figure 2C,G; Table 3), and in the limbic anterodorsal and laterodorsal nuclei (Figure 3A,E; Table 3) and in the limbic medial subnucleus of the pulvinar of one SCA3 patient (Figure 3B,F; Table 3).

Figure 2.

Motor and sensory nuclei of the thalamus in spinocerebellar ataxia type 3 (SCA3). (A) Frontal section through the dorsolateral aspect of the motor thalamic VA of a 54‐year‐old female control case without neurological or psychiatric diseases in his medical history (case 14; Table 1) and (B) the dorsolateral portion of the motor thalamic VL of a typical 63‐year‐old male control case (case 18; Table 1). (C) The somatosensory VPL and VPM of a typical 65‐year‐old male control case (case 19; Table 1) and (D) the visual LGB of a 74‐year‐old female control case (case 21; Table 1). (E) Severe neuronal loss in the VA of a 24‐year‐old SCA3 patient (case 1; 1, 2, 3) and (F) marked nerve cell loss in the VL of a 52‐year‐old female SCA3 patient (case 3; 1, 2, 3). (E,F) Note the additional degeneration of the extraterritorial RT of both SCA3 patients. (G) Obvious neuronal loss in the VPL and VPM of a 66‐year‐old male SCA3 patient (case 6; 1, 2, 3) and (H) pronounced loss of nerve cells in the deep magnocellular layers of the LGB of the same SCA3 patient (case 6; 1, 2, 3). Framed area in (H) is shown in greater detail at top right and along with remaining nerve cells in the deep magnocellular layers I and II of the LGB shows numerous extraneuronal lipofuscin granules (arrowheads). Arrows and arrowheads indicate surviving nerve cells in the VA, VL, VPL and VPM. (A–H: aldehydefuchsin–Darrow red staining, 100‐µm PEG sections). Abbreviations: ATW = triangular area of Wernicke; CM = centromedian nucleus; LGB = lateral geniculate body; RT = reticular nucleus; VA = ventral anterior nucleus; VL = ventral lateral nucleus; VPL = ventral posterior lateral nucleus; VPM = ventral posterior medial nucleus.

Figure 3.

Limbic thalamic nuclei and nuclei of the pulvinar in spinocerebellar ataxia type 3 (SCA3). Frontal sections through the limbic thalamus of a 40‐year‐old male control case without a medical history of neurological or psychiatric diseases (case 11; Table 1) depicting (A) the laterodorsal nucleus (LD) and (B) the medial part of the medial nucleus of the pulvinar (PU m). Frontal sections through the pulvinar of representative 50‐year‐old male control case (case 12; Table 1) with (C) the lateral (PU l) and (D) the inferior nuclei of the pulvinar (PU i). (E) Marked damage to the LD and (F) the PU m of a 75‐year‐old female SCA3 patient (case 8; 1, 2, 3). Framed areas in (E) and (F) are shown in greater detail at top or bottom right and along with surviving LD und PU m nerve cells show extraneuronal lipofuscin granules (arrowheads). (G) Considerable neuronal loss in the PU l and (H) PU i of a 66‐year‐old male SCA3 patient (case 6; 1, 2, 3) [framed areas in (G) and (H) are shown at top right]. Arrowheads point to surviving nerve cells. (A–H: aldehydefuchsin–Darrow red staining, 100 µm PEG sections). Abbreviations: LD = Laterodorsal nucleus; PU i = Inferior subnucleus of the pulvinar; PU I = Lateral subnucleus of the pulvinar; PU m = Medical subnucleus of the pulvinar; RT = Reticular nucleus.

In all of the SCA3 patients the thalamic peduncles and the fiber tracts traversing the thalamus were unremarkable.

Application of the Kruskal and Wallis H‐test revealed that the extent of overall thalamic degeneration in all of the SCA3 patients and the extent of nerve cell loss in all affected thalamic nuclei of the SCA3 patients was independent of the individual length of the CAG‐repeats in the mutated SCA3 allele, the age at disease onset and the duration of SCA3 (all P values > 0.10). In addition, no significant linear, quadratic or cubic trends in association were detected (all P values > 0.10).

Associated tissue changes.  The thalami of all control individuals and SCA3 patients were devoid of PD‐related alpha‐synuclein immunopositive inclusion bodies (Lewy bodies, Lewy neurites). The severity of AD‐related cortical neurofibrillary pathology in the control individuals did not go beyond that encountered in cortical stage II, and in the SCA3 patients it corresponded to stage 0 (case 1; Table 1), stage I (cases 2–4, 6; Table 1), stage II (cases 7–9; Table 1) or stage V (case 5; Table 1). AD‐related AT8‐immunopositive cytoskeletal changes occurred in the thalami of eight of our SCA3 patients (cases 2–9; Table 1). With the exception of one patient, who displayed a few isolated AT8‐immunopositive nerve cells in the degenerated thalamic anterodorsal nucleus (case 8; Table 1), the AD‐related cytoskeletal pathology was confined to limbic thalamic nuclei spared by neurodegeneration (ie, anterodorsal, laterodorsal, paraventricular and all intralaminar nuclei).

In contrast to the control individuals, all of the thalamic nuclei of the SCA3 patients displayed ataxin‐3 immunopositive NI, irrespective of whether they underwent neurodegeneration (Figure 4A–D; Table 3). According the Kruskal and Wallis H‐test the occurrence of NI in all thalamic nuclei was independent of the length of the CAG‐repeats in the mutated SCA3 allele, the age at disease onset and the duration of SCA3 (all P values > 0.10), whereby no significant linear, quadratic or cubic trends in association were detected (all P values > 0.10). In addition, calculation of Kendall’s rank correlation coefficients did not disclose significant correlations between the severity of degeneration of the thalamic nuclei and the presence of NI in them (all P values > 0.10).

Figure 4.

Neuronal intranuclear inclusion bodies (NI) in degenerated and spared thalamic nuclei in spinocerebellar ataxia type 3 (SCA3). Ataxin‐3 immunoreactive NI in (A) a surviving nerve cell (arrow) of the degenerated ventral anterior thalamic nucleus of a 60‐year‐old male SCA3 patient (case 5; 1, 2, 3) and (B) in a remaining nerve cell (arrow) of the affected ventral lateral thalamic nucleus of a 73‐year‐old female SCA3 patient (case 7; 1, 2, 3). Ataxin‐3 immunoreactive NI in (C) a nerve cell (arrow) of the unaffected parafascicular thalamic nucleus of a 52‐year‐old male SCA3 patient (case 3; 1, 2, 3) and (D) in a nerve cell (arrow) of the untouched anterodorsal thalamic nucleus of a 80‐year‐old female SCA3 patient (case 9; 1, 2, 3). (A–D: anti‐ataxin‐3 immunostaining with nickel‐intensification, 100‐µm PEG sections).

In contrast to the control individuals studied, a marked to severe GFAP‐immunopositive astrogliosis was present in all of the thalamic nuclei and the extraterritorial reticular nucleus of all SCA3 patients studied (data not shown). The distribution and severity of the GFAP‐immunopositive astrogliosis in all of the SCA3 patients studied was not correlated with the extent of thalamic neurodegeneration (all P values > 0.10).

DISCUSSION

The present pathoanatomical study for the first time shows that the human thalamus like in other polyglutamine diseases such as HD or SCA2 22, 23, 55, 56) represents a consistent target of the pathological process underlying SCA3. As we applied acknowledged and sensitive staining methods 3, 4, 7, 13, 54), but failed to recognize noteworthy PD‐ or AD‐related inclusion body pathologies beyond of some of their limbic thalamic nuclei, the thalamic involvement observed in our SCA3 patients cannot be explained by the occurrence of concomitant PD‐ or AD‐related inclusion body pathologies but rather is accountable to the degenerative process of SCA3.

The distribution pattern of thalamic neurodegeneration observed in our SCA3 patients shows both similarities with as well as distinguishing features to that observed in SCA2 and HD. Similar to the situation in SCA2, thalamic neurodegeneration in SCA3 predominantly involves the nuclei of the lateral basal group and the extraterritorial reticular nucleus and affects the nuclei assigned to the limbic system only in the minority of patients 55, 56). However, in contrast to the consistent involvement of the visual lateral geniculate body in SCA2 (55) and that of the limbic mediodorsal nucleus and the motor centromedian and parafascicular nuclei in HD intimately connected with the neostriatum 22, 23) these thalamic nuclei, according to our present findings, are only occasionally among the targets of or spared by the pathological process of SCA3.

In the present study we provide a comparison of the findings of routine neuropathological light microscopic examination performed on conventional thin tissue sections through the right cerebral hemispheres and midbrain of our SCA3 patients with the results obtained during detailed pathoanatomical light microscopical analysis carried out on unconventionally thick serial tissue sections through their left cerebral hemisphere and midbrain. This pathoanatomical analysis not only confirmed the findings of routine neuropathological examination of the SCA3 patient’s brains with respect to neurodegeneration, but disclosed also an additional involvement of their thalami. This additional identification of thalamic involvement upon pathoanatomical analysis repeatedly shows that this labor‐intensive and time‐consuming unconventional approach may reveal more precise information with respect to the extent and severity of the pathological process underlying neurodegenerative diseases 21, 50, 60). It therefore represents a useful complementation of the routine neuropathological approach in the field of basic research, which primarily is concerned with diagnostic duties and responsibilities in everyday medical life. Because of its focus even on subtle pathological alterations, application of the pathoanatomical method to unconventionally thick serial brain tissue sections recently contributed to the elaboration of the extent of the degenerative process underlying some spinocerebellar ataxias 15, 50, 51, 55, 56, 57, 58, 59, 60).

We observed consistent marked atrophy of the pons, medulla oblongata, and medial cerebellar peduncles in our SCA3 patients upon macroscopic examination, whose extent was comparable to that recently reported in our cohort of SCA2 patients. Although atrophic changes of the pons, medulla oblongata, and medial cerebellar peduncles may be more severe in some SCA2 patients 9, 33, 63), our and other macroscopic SCA3 findings 9, 41, 42, 62, 66) are in contradiction with the view of SCA3 as a neurodegenerative disease which is not accompanied by gross brainstem atrophy (39) and are in agreement with the opinion that SCA2 and SCA3 cannot be reliably distinguished based on macroscopical criteria 33, 62).

Upon statistical analysis we failed to establish a significant association between thalamic neurodegeneration and the individual length of CAG‐repeats, age at disease onset and duration of SCA3. Although, this may at first glance indicate that thalamic neurodegeneration is independent from these three factors, it cannot be excluded that additional and up to now unknown disease‐modifying factors obscured an association of these three fixed factors with thalamic neurodegeneration. Search for such disease‐modifying factors is required and may lead to rational therapies that slow or prevent neuronal loss in SCA3 patients, in particular if those factors themselves could be proven to be subject to exertion of influence.

In contrast to SCA2, the occurrence of NI represents a widely acknowledged morphological feature of SCA3 12, 26, 34, 40, 43, 44, 45, 46, 69, 70), but their precise role in the pathogenesis of SCA3 still is uncertain. Although, it remains an open question whether NI are pathogenic structures directly linked to the disease process of SCA3, some researchers assume a toxic property of NI for affected nerve cells while others speculate about a potential protective role 12, 20, 44, 70).

Based on the findings of postmortem studies the following conclusions could be drawn concerning the significance of NI in the pathological process of SCA3: (i) NI represent only unspecific markers of SCA3 and are not directly decisive for the destiny of nerve cells. (ii) NI are directly involved in the pathogenesis of SCA3 neurodegeneration, exerting either a protective or a toxic effect on affected nerve cells. The assumption of a protective effect could explain the presence of NI in the well‐preserved thalamic nuclei of our SCA3 patients and would suggest that NI perhaps evolve comparatively late or delayed in affected thalamic nuclei and help to stabilize their nerve cells. On the other hand, the assumption of a toxic effect could explain the presence of NI in all degenerated thalamic nuclei of our SCA3 patients and would suggest that NI perhaps evolve comparatively late in unaffected CNS regions, helping to destabilize and ultimately contribute to break‐up of their nerve cells.

In the present study we provide the first detailed comparison of the pathoanatomy and distribution pattern of ataxin‐3 immunopositive NI in a given CNS region, which shows that NI are present in all thalamic nuclei irrespective of their function or anatomical connections. These findings (i) confirm the suggestions of previous studies that SCA3‐related NI may occur in obviously degenerated CNS regions as well as in CNS regions believed to be spared in SCA3 patients 20, 40, 44, 45, 61, 67, 69, 70) and (ii) are comparable to the results of a recent SCA7 study (51). As was the case with thalamic neurodegeneration, statistical analysis revealed no association between the occurrence of thalamic ataxin‐3 immunoreactive NI and the individual length of CAG‐repeats, age at disease onset and duration of SCA3. Similar to thalamic neurodegeneration, these initial findings regarding the developmental features of ataxin‐3 immunoreactive NI suggest that there might be additional, albeit still unknown factors, which exert influence on the emergence of ataxin‐3 immunoreactive NI. Furthermore, statistical analysis underpinned the lack of an association between thalamic neurodegeneration and the presence of ataxin‐3 immunoreactive NI. Accordingly, the present thalamic data favor the view that ataxin‐3 immunoreactive NI are not directly decisive for the fate of affected nerve cells, exerting either an immediate protective or an immediate toxic effect on affected nerve cells and we suggest to regard NI for the time being only as unspecific but pathognomonic morphological markers of SCA3.

However, postmortem studies at best can only provide correlative associations, yet do not allow statements with respect to causative associations. Therefore, precise answers to the question concerning the exact pathogenetic role of NI in SCA3 currently seem to be subject to studies involving experimental animal models, of which the most useful should reflect the damage to SCA3 brains. Accordingly, development of animal SCA3 models should take the central nervous distribution pattern of neurodegeneration in SCA3 brains into account, which now is almost complete 11, 17, 30, 50, 52, 56, 57, 58, 59, 69, 70).

Upon analysis of the medical records and pathological reports of the SCA3 patients we found that (i) dysphagia occurred in all of them; (ii) dysphagia was associated with aspiration pneumonia in the vast majority of them; and (iii) aspiration pneumonia occurred in nearly 90% of them and constituted the most frequent cause of death in our sample of SCA3 patients (Table 2). A detailed investigation of the ingestion‐related brainstem nuclei in six of these patients (cases 1–5, 9; Table 1) showed their widespread involvement: the motor trigeminal, principal trigeminal, and facial nuclei were consistently degenerated, whereas the hypoglossal, ambiguus, parvocellular reticular and dorsal motor vagal nuclei underwent neurodegeneration in the majority of these SCA3 patients and the mesencephalic trigeminal, spinal trigeminal and solitary nuclei and the intermediate reticular zone were involved in the minority of them (52).

In view of these clinical findings, that aspiration pneumonia apparently may represent an incisive complication and a very frequent cause of death in SCA3 patients suffering from neurogenic dysphagia, preventive research and improvement of management of dysphagia and treatment strategies aimed at preventing aspiration pneumonia in SCA3 is imperative.

The functional consequences resulting from thalamic degeneration offer explanations for a variety of clinical and paraclinical SCA3 findings and have been discussed in detail in previous papers 22, 23, 55, 56). Therefore, we have chosen here to restrict the discussion only to the functional consequences which may result from the mostly affected thalamic nuclei. Because of their integration into the re‐entrant motor cerebellothalamocortical and basal ganglia‐thalamocortical circuits 1, 10, 14, 16, 28, 35) and the current knowledge regarding the functional consequences of their impairment 1, 14, 16, 36) degeneration of the thalamic ventral anterior and ventral lateral nuclei together with that of other components of these circuits (pallidum, subthalamic and pontine nuclei, cerebellar cortex) most likely contributed significantly to gait, stance, limb and truncal ataxia, as well as dysarthria in our SCA3 patients. The extraterritorial reticular nucleus of the thalamus by way of its inhibitory GABAergic projections channels the motor output of the ventral anterior and ventral lateral thalamic nuclei destined for the cerebral cortex 19, 27) and loss of its GABAergic inhibitory nerve cells is advantageous for thalamocortical data flow (19). Accordingly, the degeneration of the reticular thalamic nucleus of our SCA3 cases may have compensated, in part, for the motor consequences resulting from the lesions in the interconnected ventral anterior and ventral lateral thalamic nuclei. In addition, as the inferior and lateral subnuclei of the pulvinar are integrated into the visual attentional networks of the human brain 18, 48), it is conceivable that involvement of these subnuclei of the pulvinar represents the morphological basis of the visual attention deficits of SCA3 patients that have been demonstrated by a recent neuropsychological study (37). Finally, damage to the ventral posterior lateral thalamic nucleus conducting epicritic data from the limbs and the trunk to the primary somatosensory cortex 28, 29) offers an explanation for the epicritic somatosensory deficits seen in our and other SCA3 patients 11, 17, 32).