The present study shows for the first time a molecular signature of Perry syndrome; namely, abnormal phosphorylated and truncated forms of TDP-43 associated with NCI, NII and dystrophic neurites. TDP-43 is a DNA binding protein found in the nucleus in physiological conditions, therefore its cytoplasmic localization strongly supports a pathologic role. Furthermore, the pathologic hyperphosphorylated and 25 kDa truncated forms of TDP-43 were detected in brain regions that also had immunohistochemical evidence of pathology (globus pallidus and SN), but not in histologically unaffected regions (temporal cortex). The electrophoretic and solubility profiles of TDP-43 were similar to those found in other TDP-43 proteinopathies [12
]. Given that mutations in GRN
] and TARDBP
] are associated with TDP-43 pathology, these genes were excluded as causes of Perry syndrome.
TDP-43-positive NCI in Perry syndrome were highly pleomorphic, and some resembled the skein-like inclusions detected in motor neurons of ALS [23
]. On the other hand, the distribution of NCI was different. In ALS motor neurons are affected, while in Perry syndrome NCI were most common in the dopaminergic neurons of the pars compacta of the SN. In addition to NCI, TDP-43 was also detected in lentiform shaped NII; however NII were detected in only four of the eight cases, and they were very sparse. NII are found in some cases of FTLD-U and are more frequent in familial disease [26
]. For example, all cases with mutations in GRN
] and the Valosin Containing Protein
gene-associated FTLD have NII [28
]. It should be emphasized, however, that not all cases with NII have a positive family history [27
] and not all cases with a family history have NII. Perry syndrome falls into the latter category.
The morphology and distribution of TDP-43 immunoreactive NCI, NII and dystrophic neurites in cortex and hippocampus has served to delineate four subtypes of FTLD-U [30
]. The TDP-43 pathology in Perry syndrome defies subclassification due to absence of lesions in the cortex and hippocampus and the consistent involvement of the SN and globus pallidus, which are brain regions not even considered in FTLD-U subclassification schemes. The presence of TDP-43 in axonal spheroids, particularly in the substantia nigra and globus pallidus, has not been emphasized in previous immunohistochemical studies using antibodies to TDP-43. While SN TDP-43-immunoreactive axonal spheroids were prominent in Perry syndrome, they are not specific to this condition. Out of 21 cases of pathologically confirmed FTLD-U with subtype 3 TDP-43 [31
] pathology (12 men; mean age [±SD], 75 [±10] years), 16 (76%) had SN TDP-43-positive axonal spheroids (unpublished data). Glial inclusions similar to those noted in Perry syndrome are also found in FTLD-U, ALS and Guam Parkinson dementia complex [25
]. Extramotor pathology is increasingly recognized in ALS, particularly in ALS with dementia, where inclusions have been described in the striatum, globus pallidus, neocortex, hippocampus, SN and inferior olivary nucleus [35
]. The inferior olivary nucleus, which was frequently affected in Perry syndrome, is also commonly affected in FTLD-U, FTLD-MND and ALS [15
In Perry syndrome the most distinctive pathologic feature is the distribution of the inclusions, which primarily involves the extrapyramidal system, but spares the cortex, hippocampus and upper and lower motor neurons. While hippocampal pathology is common in FTLD-U, it was almost absent in Perry syndrome. Sparse neuritic pathology in the pyramidal layer of the hippocampus and rare NCI in the dentate fascia were detected in only a few cases. None of the cases have involvement of the neocortex.
In Perry syndrome the distribution of the pathology only imperfectly explains the clinical features. The extrapyramidal system predominance, especially the severe involvement of the SN, likely accounts for parkinsonism in Perry syndrome patients. In addition, the basal ganglia involvement may explain the poor or transient response to levodopa. Recently, a neuropathologic study on one of the Japanese patients reported herein found selective loss of putative respiratory neurons in the ventrolateral medulla and in the dorsal raphe nucleus, which most probably represents the pathological substrate of hypoventilation [11
]. Depression in Perry syndrome may stem from the loss of aminergic neurons in the locus ceruleus and ventral tegmental area. Frontal-type apathy, which better applies to some patients than true depression, likely results from reduced ventral tegmental area-frontal cortex dopaminergic projections [1
]. No convincing pathological lesion has been found that would explain the severe weight loss displayed by many patients. We found mild pathology in the hypothalamus which may play a role; however, NCI were sparse and not associated with overt neuronal loss and gliosis.
While TDP-43 pathology was initially thought to be a specific marker for FTLD-U and ALS, subsequent studies have shown it in a variety of other disorders, including Guam Parkinson dementia complex [34
], Pick’s disease [13
], Alzheimer’s disease [18
], hippocampal sclerosis [18
], and Lewy body disease [37
]. It remains unclear if TDP-43 plays a direct and pathologically relevant role in these diseases, or if it only represents a secondary phenomenon. In contrast to these conditions, no other pathological hallmark such as α-synuclein (Lewy bodies) or tau-positive (Pick bodies, tangles) pathology is present in Perry syndrome. Although further studies may challenge TDP-43 pathogenicity in Perry syndrome, the absence of other identifiable pathology supports its designation as a TDP-43 proteinopathy.
Given the lack of mutations in either GRN
, the present results point to a missing link between the genetic defect and the deposition of TDP-43-positive inclusions in Perry syndrome. This missing link may include gene-gene, gene-protein or protein-protein interaction, whereby the mutated gene/protein influences the expression or translation of TARDBP
or acts on the processing or posttranslational modification of TDP-43. A biochemical link between progranulin and TDP-43 was recently identified in a cell biology study, in which caspase-dependent cleavage of TDP-43 was produced by suppression of PGRN
expression by small interfering RNA. Caspase-dependent cleavage of TDP-43 was associated with redistribution of TDP-43 from the nucleus to the cytoplasm, reminiscent of the pattern seen in FTLD-U and in Perry syndrome [38
]. This observation supports a pathogenic role of TDP-43 which is likely to be established in a subset of neurodegenerative conditions with TDP-43-positive pathology [39
Our study shows that Perry syndrome is a unique entity, with a distinct TDP-43-positive pathology and a characteristic clinical presentation. Further studies are needed to establish the role of TDP-43 in Perry syndrome and ultimately to identify its genetic cause.