Neuropathological features of genetically confirmed DYT1 dystonia: investigating disease-specific inclusions

Dystonia is a common and heterogeneous neurological disorder. Recently, dystonia has been defined as a movement disorder characterized by sustained or intermittent muscle contractions causing abnormal, often repetitive, movements, postures, or both with genetic heterogeneity reported. The term `dystonia musculorum deformans’ was proposed in 1911 to describe a disorder in children with twisted postures, muscle spasms, gait abnormalities and clonic, rhythmic jerking movements [1]. In childhood onset cases dystonia tends to become generalized, while remaining focal or segmental in adult onset cases [2]. A newly proposed classification system classified dystonia according to two axes: clinical characteristics and etiology [3]. Dystonia linked to a 3 base pair deletion (ΔGAG) mutation in exon 5 of the TOR1A gene on chromosome 9q34.11 is called DYT1 dystonia [4]. The new classification system classified DYT1 dystonia as isolated dystonia in which dystonia is the only clinical sign with the exception of tremor [3]. The prevalence of isolated dystonia was estimated at 330 per million in Rochester, Minnesota [5] and is about 152 per million in the European population [6]. In DYT1 the inheritance is autosomal dominant with the disease onset in childhood [7] and the most common clinical features include muscle contractions affecting the leg or arm, often progressing to generalized involvement with severe disability [8]. The mutation has been identified in many diverse ethnic groups [9],[10] and has a low penetrance of 30-40% [10]-[12].

The TOR1A gene encodes a 332 amino acid protein, torsinA, which has a cytoplasmic localization in cells. TorsinA is widely expressed throughout the central nervous system in humans, but is found at particularly high levels in dopaminergic neurons of the substantia nigra, locus coeruleus, Purkinje cells, cerebellar dentate nucleus, basis pontis, thalamus, hippocampal formation, oculomotor nuclei and frontal cortex [13]-[16]. The exact function of torsinA remains unclear in humans. However, mutant torsinA protein has been shown to have aberrant cellular localization and impaired protein interactions and is associated with defective synaptic vesicle formation and altered development of neuronal pathways [17].

Much of the research in primary dystonia has focused on the role of the basal ganglia in disease [18]-[20]. However, imaging studies of patients carrying a mutation in TOR1A causing primary dystonia revealed an increase in metabolic brain activity not only in the basal ganglia but also in the cerebellum [21],[22]. Neuropathological investigation of clinically diagnosed primary dystonia cases has been generally disappointing with no specific abnormalities observed [23]. In genetically confirmed DYT1 cases, no evidence of neuronal loss, inflammation or altered localization of torsinA could be identified [15],[23]-[26]. A possible reduction in striatal dopamine and the size of pigmented neurons in the substantia nigra have been suggested [15],[23],[26]. The most interesting observation to date has been the finding of perinuclear intraneuronal inclusions immunoreactive for ubiquitin, torsinA and lamin A/C in the periaqueductal grey matter, cuneiform and pedunculopontine nuclei [27]. Similar inclusions were reported in some of the DYT1 mouse models produced by expression of transgenic human torsinA but this has not been a consistent finding [28],[29]. The intriguing observation of brainstem inclusions in DYT1 cases has not, so far, been replicated.

The aims of this current study were: 1) to report the neuropathological features of seven previously unreported genetically proven DYT1 cases, 2) to determine whether any pathological features were consistently observed in all cases and could be regarded as disease-related and 3) to determine whether the perinuclear intracytoplasmic inclusions previously described in DYT1 dystonia are a consistent feature of the disease.

There are only five neuropathological studies of genetically confirmed DYT1 mutation cases in the literature [23]. We aimed to significantly contribute to the field of dystonia research by assessing the largest case series of DYT1 yet investigated to determine whether consistent neuropathological features could be identified as a hallmark of the disease. This study also addressed the question of whether neuronal perinuclear intracytoplasmic inclusions of the type previously associated with early-onset DYT1 dystonia could be confirmed in our cases [27].

We confirmed that all of the cases carry a 3-basepair deletion (ΔGAG) in exon 5 of the TOR1A gene which is a common mutation in DYT1 dystonia. Among the seven cases studied, two cases (cases 3 and 5) were symptomatic and five were asymptomatic carriers for DYT1 dystonia (cases 1, 2, 4, 6, and 7). Using a similar IHC approach to that applied in a previous neuropathological study of DYT1 dystonia [27] we did not replicate the finding of ubiquitin positive inclusions in the midbrain and pons with the exception of inclusions that were also immunoreactive for tau or α-synuclein, representing NFTs and LBs respectively. We also screened for intracellular inclusions in other brain regions and could not demonstrate abnormality other than sparse Alzheimer and LB pathology. The remaining pathological features identified in the cases were those of gliosis, mild-moderate Purkinje cell depletion, CAA, SVD of varying severity and cerebral infarcts. None of these changes can be regarded as disease specific and are most likely to be due to the advanced age at death of the cases (75-90 years).

We examined anatomical regions implicated in dystonia and found no pathological intracellular inclusions or evidence of more than mild neuronal loss in the caudate, putamen, subthalamic nucleus, thalamus, pigmented neurons of the substantia nigra and globus pallidus, other than that associated with small infarcts in the latter in two cases [35].

The reason that we failed to replicate the findings of McNaught et al.[27] is uncertain. With the exception of case 4 in the McNaught series, who was aged 33 years at death, the ages of our 2 case series were similar excluding the possibility that the inclusions observed are age related. One difference between the two studies is that all four cases described by McNaught et al. manifested clinical dystonia compared with only two cases in our series, the remaining five cases that we investigated were non-manifesting carriers of the gene mutation. If the neuronal inclusions are only present in patients with clinical dystonia this may have reduced the likelihood that we would have detected inclusions in our cases. However, similar electrophysiological and functional imaging abnormalities have been reported in cortical motor circuits of manifesting and non-manifesting DYT1 carriers [36],[37]. An alternative explanation could be that different antibodies were used in the two studies and this remains a possibility. Given the limitation that only a small amount of tissue was available for study we chose ubiquitin and p62 as screening antibodies as they target abnormally folded proteins for degradation by the proteasome and autophagy respectively. As both are sensitive markers of inclusions formed from a wide variety of different proteins in neurodegenerative diseases and can recognise proteinaceous inclusions for which the principle protein component has not been identified we argued that if DYT1 dystonia is associated with intraneuronal inclusions these would be highlighted by one or both of these antibodies [38],[39]. Any inclusions identified could then have been further defined using other specific antibodies such as lamin A/C or torsinA and the affected neurons would have been immunophenotyped to determine whether cholinergic neurons were affected as described in the McNaught study. A limitation of our study was that, due to the restricted tissue available for examination, not all regions were present in each case. It therefore remains a possibility that because of these sampling issues we did not identify rare inclusions. Future studies in systematically sampled cases would be required to resolve this issue. Assessment of cell inclusions using a reliable human specific torsinA antibody would be highly desirable. We have previously tested 4 different antibodies: torsinA [13],[27],[40], torsinA (TA913) [41], anti-torsinA (ab34540) and torsinA (Santacruz s-20) using western blots, IHC and immunofluorescence (unpublished data). However, none of these was found to be reliable and therefore could not be included in the study.

Despite the limitations of this study due to its retrospective nature and the restricted tissue availability our observations support the previous findings that in primary dystonia, there is no overt neurodegeneration or cell loss [23]. The pathological change of brainstem inclusions reported in DYT1 dystonia [27] was not replicated in our case series. Hence this study supports the previous findings that overt cell loss or other obvious neuropathological defects such as inclusions do not seem to be consistently observed in DYT1 dystonia. However, subtle morphological defects of the type seen in the DYT1 mouse models (e.g. abnormalities of neuronal size/shape, dendrite structure, spines, etc.) cannot be excluded. The analogy is similar to most human mental retardation syndromes, where cell loss or inclusions are absent, but marked abnormalities of neuritic structure are thought to play an important role. The subtle morphological changes may not be identifiable using current methods. In this regard, generation of cellular disease models using the induction of pluripotent stem (iPS) cells from patient skin fibroblast has opened new horizons [42],[43]. Understanding the cellular pathways influenced by particular mutations and the neuronal networks affected may inform future neuropathological studies.

We report a detailed neuropathological study of 7 DYT1 cases, the largest DYT1 case series yet investigated. The anatomical regions implicated in dystonia showed no consistent disease specific pathological features. Our observations support the previous findings that understanding the biochemical changes may be more relevant than morphological alterations in isolated dystonia.

Reema Paudel 1A, B, C, 2A

Aoife Kiely 1C, 2B

Abi Li 1C

Tammaryn Lashley 1C

Rina Bandopadhyay 1C

John Hardy 1A, 2B

Hyder A Jinnah 2B

Kailash Bhatia 2B

Henry Houlden 1A, 2B

Janice L Holton 1A, B, C, 2B