Expanding the spectrum of neuronal pathology in multiple system atrophy

Abstract

Multiple system atrophy is a sporadic alpha-synucleinopathy that typically affects patients in their sixth decade of life and beyond. The defining clinical features of the disease include progressive autonomic failure, parkinsonism, and cerebellar ataxia leading to significant disability. Pathologically, multiple system atrophy is characterized by glial cytoplasmic inclusions containing filamentous alpha-synuclein. Neuronal inclusions also have been reported but remain less well defined. This study aimed to further define the spectrum of neuronal pathology in 35 patients with multiple system atrophy (20 male, 15 female; mean age at death 64.7 years; median disease duration 6.5 years, range 2.2 to 15.6 years). The morphologic type, topography, and frequencies of neuronal inclusions, including globular cytoplasmic (Lewy body-like) neuronal inclusions, were determined across a wide spectrum of brain regions. A correlation matrix of pathologic severity also was calculated between distinct anatomic regions of involvement (striatum, substantia nigra, olivary and pontine nuclei, hippocampus, forebrain and thalamus, anterior cingulate and neocortex, and white matter of cerebrum, cerebellum, and corpus callosum). The major finding was the identification of widespread neuronal inclusions in the majority of patients, not only in typical disease-associated regions (striatum, substantia nigra), but also within anterior cingulate cortex, amygdala, entorhinal cortex, basal forebrain and hypothalamus. Neuronal inclusion pathology appeared to follow a hierarchy of region-specific susceptibility, independent of the clinical phenotype, and the severity of pathology was duration-dependent. Neuronal inclusions also were identified in regions not previously implicated in the disease, such as within cerebellar roof nuclei. Lewy body-like inclusions in multiple system atrophy followed the stepwise anatomic progression of Lewy body-spectrum disease inclusion pathology in 25.7% of patients with multiple system atrophy, including a patient with visual hallucinations. Further, the presence of Lewy body-like inclusions in neocortex, but not hippocampal alpha-synuclein pathology, was associated with cognitive impairment (P = 0.002). However, several cases had the presence of isolated Lewy body-like inclusions at atypical sites (e.g. thalamus, deep cerebellar nuclei) that are not typical for Lewy body-spectrum disease. Finally, interregional correlations (rho ≥ 0.6) in pathologic glial and neuronal lesion burden suggest shared mechanisms of disease progression between both discrete anatomic regions (e.g. basal forebrain and hippocampus) and cell types (neuronal and glial inclusions in frontal cortex and white matter, respectively). These findings suggest that in addition to glial inclusions, neuronal pathology plays an important role in the developmental and progression of multiple system atrophy.

Keywords: Lewy body-like; alpha-synuclein; inclusion bodies; multiple system atrophy; neuropathology.

See Halliday (doi:10.1093/brain/awv151) for a scientific commentary on this article. Multiple system atrophy (MSA) is characterized clinically by parkinsonism, cerebellar dysfunction, and dysautonomia, and pathologically by glial cytoplasmic inclusions. Using pathological and clinical data from 35 patients, Cykowski et al. expand the spectrum of neuronal pathology in MSA and argue that neuronal inclusions have a key role in disease pathogenesis.

Figure 1

Spectrum of pleomorphic neuronal inclusions in 14 different MSA patients. Alpha-synuclein-immunopositive inclusions are demonstrated in intranuclear (top row), perinuclear (second row), cytoplasmic (third row), and combined intranuclear/cytoplasmic distributions (fourth row). The last row shows globular cytoplasmic/Lewy body (LB)-like neuronal inclusions; the haematoxylin and eosin correlates, if present, were pale body-like structures (left panel of bottom row). Globular cytoplasmic/Lewy body-like inclusions are shown in the medullary reticular formation (middle panel of bottom row) and roof nuclei of cerebellum (right panel of bottom row) (see text for details). All images are at ×400. Ctx = cortex; nuc = nucleus; V nuc = motor nucleus of the trigeminal nerve.

Figure 2

Spectrum of pathologic burden for neuronal inclusions in five regions. Absent/mild to more advanced pathology is shown from left to right for the inferior olive (top row), anterior cingulate cortex (second row), CA2 sector of hippocampus (third row), dentate gyrus of hippocampus (fourth row), and thalamic nuclei (bottom row). Mild olivary pathology was often associated with plaque-like, immunoreactive processes (left panel of top row, black arrows). With advanced olivary pathology, foci of extensive cell loss with immunonegative astrogliosis and absent GCIs and neuronal inclusions were seen (right panel of top row, white arrow). For mild CA2 neuronal inclusion pathology, there are rare neuronal inclusions (black arrow, left panel of third row) and faint preinclusions (white arrow). A globular cytoplasmic/Lewy body-like neuronal inclusion is seen within a thalamic neuron (black arrow, right panel of bottom row) (see text for details). All images are at ×200.

Figure 3

Interregional correlations in alpha-synuclein burden. This is a graphical demonstration of a correlation matrix. Each point in this matrix represents the strength of correlation between the semiquantitative pathologic grading of any two anatomical regions across all patients studied. The colour scale key indicates the strength of the correlation. Dark purple indicates perfect correlation of pathologic severity in two structures (i.e. a Spearman rho correlation coefficient of 1.0 as seen across the diagonal, from top left to bottom right). For instance, pathologic grading in basis pontis (PN) and the inferior olive (OLV) (row 3, column 4) is very strongly correlated as indicated by the purple colour (actual Spearman rho value is +0.88). Conversely, pathological involvement of cerebellar white matter (CRBLL WM) and basal forebrain/hypothalamus (BF/HY) was not correlated at all, as indicated by the yellow colour at row 5, column 11. Other structures evaluated are as follows: putamen (PT), substantia nigra (SN), corpus callosum (CC), anterior cingulate cortex (CING CTX), isocortex of superior frontal gyrus (FRTL CTX), frontal lobe hemispheric white matter (FRTL WM), hippocampus, and thalamus (THAL). Values of Spearman’s rho ≥ 0.6 or ≤ −0.6 were of interest in this exploratory analysis (P-values were not assigned).

Figure 4

Hierarchy of regional involvement in MSA for 35 patients. The frequency of inclusion pathology is indicated with the most frequent sites of neuronal inclusion pathology in the top tier, including both canonical regions (e.g. pontine nuclei) and non-canonical regions (e.g. anterior cingulate cortex). The downward arrow represents the global increase in alpha-synuclein burden (both glial and neuronal inclusions). In these 35 patients studied, with increasing alpha-synuclein burden (middle and bottom tiers of the hierarchy), additional areas of neuronal pathology could be identified. Some structures, indicated in the two bottom-most tiers, appear to be resistant to neuronal inclusion pathology although glial inclusions were seen. ctx = cortex; RF = medullary reticular formation; PAG = periaqueductal grey of midbrain; nuc = nucleus. 1 = Structures with characteristic neuronal and glial pathology in MSA; 2 = GCIs could be identified but convincing neuronal inclusions were not seen; 3 = GCIs could be identified.