Papp–Lantos inclusions and the pathogenesis of multiple system atrophy: an update
- First Online:
- Cite this article as:
- Jellinger, K.A. & Lantos, P.L. Acta Neuropathol (2010) 119: 657. doi:10.1007/s00401-010-0672-3
- 358 Views
Multiple systemic atrophy (MSA) is a progressive, adult-onset neurodegenerative disorder of undetermined aetiology characterized by a distinctive oligodendrogliopathy with argyrophilic glial cytoplasmic inclusions (GCIs) and selective neurodegeneration. GCIs or Papp–Lantos inclusions, described more than 20 years ago, are now accepted as the hallmarks for the definite neuropathological diagnosis of MSA and suggested to play a central role in the pathogenesis of this disorder. GCIs are composed of hyperphosphorylated α-synuclein (αSyn), ubiquitin, LRRK2 (leucin-rich repeat serine/threonine-protein) and many other proteins, suggesting that MSA represents an invariable synucleinopathy of non-neuronal type, a specific form of proteinopathies. The origin of αSyn deposition in GCIs is not yet fully understood, but recent findings of dysregulation in the metabolism of myelin basic protein (MBP) and p25α, a central nervous system-specific protein, also called TPPP (tubulin polymerization promoting protein), strengthened the working model of MSA as a primary glial disorder and may explain frequent alterations of myelin in MSA. However, it is unknown whether these changes represent an early event or myelin dysregulation occurs further downstream in MSA pathogenesis. The association between polymorphisms at the SNCA gene locus and the risk for developing MSA also points to a primary role of αSyn in its pathogenesis, while in a MBP promoter-driven αSyn transgenic mouse model gliosis accompanied the neurodegenerative process originating in oligodendrocytes. Because αSyn represents a major component in both oligodendroglial and neuronal inclusions in MSA, some authors suggested both a primary oligodendrogliopathy and a neuronal synucleinopathy, but current biomolecular data and animal models support a crucial role of the Papp–Lantos inclusions and of aberrant αSyn accumulation as their main constituent, causing oligodendroglial pathology, myelin disruption and, finally, neuronal degeneration in MSA. The relationship between oligodendrocytes involved by Papp–Lantos inclusions and those in degenerating neurons in the course of MSA needs further elucidation.
KeywordsMultiple system atrophyPapp–Lantos inclusionsAberrant α-synucleinGlial cytoplasmic inclusionsOligodendrogliopathyPathogenesis
Since 1969, when the term multiple system atrophy (MSA) was first introduced , there has been remarkable progress in our understanding of this disorder. MSA is now recognized as a progressive, adult-onset neurodegeneration with autonomic failure and motor impairment resulting from poorly l-dopa responsive parkinsonism, cerebellar ataxia, and corticospinal tract dysfunction in any combination [28, 29, 122]. Morphologically, it is characterized as a distinctive α-synuclein (αSyn) oligodendrogliopathy and selective neurodegeneration . Glial cytoplasmic inclusions (GCIs) predominantly composed of αSyn, first reported and comprehensively characterized in 1989 , are now accepted as the hallmarks for the definite neuropathological diagnosis of MSA [59, 111]. They are suggested to represent a primary pathological event in this disorder [58, 60, 82]. Although the underlying molecular mechanisms are still poorly understood, since the first description of the Papp–Lantos inclusions a growing body of evidence has been collected regarding their impact on the pathogenesis of MSA [103, 121]. Recent transgenic mouse models, overexpressing human αSyn under specific oligodendroglial promoters, reproduce the specific GCI pathology and show that oligodendroglial αSyn may trigger neurodegeneration, involving secondary axonal αSyn accumulation, mitochondrial dysfunction, microglial activation, and increased susceptibility to exogenous oxidative stress [101, 102]. These data support the central role of GCI pathology (oligodendrogliopathy) for the progressive neurodegeneration, while others suggested that both the oligodendroglia-myelin-axon mechanism and direct involvement of neurons due to aggregation of αSyn might synergistically accelerate the degenerative process in MSA [118, 124, 125]. Based on currently available data, the present state of cellular pathology of MSA and the impact of the Papp–Lantos inclusions on the pathogenesis of this disorder are summarized. A comparison between GCIs and Lewy bodies (LBs) appears to be out of the scope of this specific review, see [46, 57].
Cellular pathology of MSA
Ultrastructurally, the GCIs represent cytoplasmic aggregates composed of a meshwork of loosely packed filaments or tubules, 15–30 nm in diameter with a periodicity of 70–90 nm and straight filaments, both consisting of polymerized αSyn, and variable types of filaments. These filaments are associated with granulated material related to cytoplasmic organelles such as mitochondria and secretory vesicles [7, 12, 50]. Detailed immunoelectron microscopy of GCIs demonstrated the presence of variable types of filaments, their width being either uniform or showing periodic variation.
Constituents positively identified by routine immunohistochemistry
Candidate proteins that have so far eluded detection by routine immunohistochemistry
14-3-3 protein (in subset of GCIs)
Carbonic anhydrase isoenzyme IIa (MS+)
cdk-5 (cyclin-dependent kinase 5) (MS+)
τ2 (reversible on exposure to detergent) 
Heat shock proteins Hsc70 and Hsp70 (MS+)
Isoform of 4-repeat tau protein (hypo-phosphorylated) (MS+)
Rab5, Rabaptin-5 
Mitogen-activated protein kinase (MAPK)
Other microtubule-associated proteins (MAPs): MAP-1A and -1B; MAP-2 isoform 1, and isoform 4 (all MS+)
Phosphoinositide 3-kinase (P13K) (MS+)
p25α/TPPP (MS+) (tubulin polymerization-promoting protein)
Septin-2, -3, -5, -6 and -9
Ubiquitin (MS+) SUMO-1 (small ubiquitin modifier 1)
p62-co-localization with α-synuclein (inconsistent) 
Metallothionein-III (MT-III) 
Actin, γ-1 and γ-2 propeptides (MS+)
Amyloid-β precursor protein (MS+)
Glial fibrillary acidic protein (GFAP) (MS+)
Myelin basic protein (MBP)-3, -4, -5 (MS+)
Myelin oligodendrocyte glycoprotein (MOG), α- and β-isoforms (MS+)
Myosin (9 distinct isoforms) (MS+)
Neurofilaments (NF-3, NF-HC, NF-LC) (MS+)
Recent quantitative studies of membrane-associated, sodium dodecyl sulfate-soluble αSyn, both the full-length 17 kDa and high molecular weight species by western blotting in MSA brain, confirmed massive accumulation in the putamen  and in the frontal cortex . This distribution could qualitatively correspond to the regional pathological findings including GCIs, although the study suggested that the accumulation of monomeric αSyn may not be merely caused by the formation of GCIs . In the putamen, the area with the most severe αSyn accumulation and abundant GCIs, there was a marked depletion of cytosolic αSyn, supporting the possible role of increased αSyn insolubility in the formation of GCIs [13, 112].
The origin of αSyn deposition and the individual role of each of these polypeptides in GCI formation and MSA pathogenesis are still poorly understood. Although 14-3-3 protein is known to interact with αSyn and to participate in several transduction pathways, αSyn fibrillation without involvement of 14-3-3 or tau proteins appears to be sufficient for GCI formation . However, the accumulation of multifunctional proteins such as 14-3-3 may add to the disruption of cellular homeostasis [27, 51, 55]. GCIs also contain at least two members of the heat shock family, Hsc70 and Hsp70, important molecular chaperones that have also been found in neuronal inclusions of other degenerative disorders . Additional proteins include α- and β-tubulin, several microtubule-associated proteins [1, 6], cyclin-dependent kinase 5 (cdk5) [31, 107], which is important in neuronal signalling, and cyclic adenosine monophosphate-regulated phosphoprotein-32 (DARRP-32) , the latter representing a known substrate of cdk5 . In GCIs, there is also aberrant expression of Rabaptin-5 that is involved in the process of endocytosis in neurons . GCIs are immunopositive for metallothionein (MT)-III, showing co-localization with αSyn, and immunoelectron microscopy revealed MT-III labelling of the amorphous material surrounding αSyn filaments in GCIs, whereas MT-I/II isoforms were not different from those in normal controls, suggesting that MT-III, a low-molecular metal-binding protein, is a specific component of GCIs . In MSA brains, numerous GCIs, NCIs, and dystrophic neurites are also intensively immunoreactive for HtrA2/Omi, a mitochondrial serine protease that is released into the cytosol and promotes apoptotic processes by binding to several members of the inhibitors of apoptosis protein family. Its widespread accumulation in MSA brains as well as in both classic (brainstem-type) and cortical LBs suggests an association with the pathogenesis of α-synucleinopathies .
Recently, LRRK2 (leucin-rich repeat serine/threonine-protein) and Parkin, genetic deficits of which are pathogenic for Parkinson disease (PD) and their protein products have also been localized in LBs [3, 32, 66, 67, 91, 126], and were subsequently detected by immunohistochemistry in GCIs of MSA , while genetic variants of Parkin and PINK1 loci do not play a critical role in the pathogenesis of MSA . LRRK2 was found in most of the enlarged oligodendroglia and co-localized with the majority of αSyn- and p25α-immunopositive GCIs, while only a minority of αSyn-positive and p25α-negative GCIs contained LRKKs. Parkin was observed only in a small proportion of GCIs. These data support previous studies associating these proteins with αSyn inclusion formation [3, 32, 66, 91]. Both LRRK2 and p25α were predominantly found in the earliest stages of GCI formation , while Parkin immunoreactivity was present in more mature GCIs . These data suggest that LRRK2 may contribute in the initial form of GCI formation, similar to that suggested for early brainstem LB formation . p25α, also called tubulin polymerization promoting protein (TPPP) , is an oligodendroglia-specific phosphoprotein [105, 106], present in myelin basic protein (MBP)-immunopositive sheaths and probably linked to myelination . There it may mediate interaction between MBP and tubulin. In vitro, p25α/TPPP is a potent stimulator of αSyn aggregation , while in phosphorylation by Erk2 or cdk5, it shows reduced microtubule-binding properties in vitro . Recent studies have shown that MSA is not a TDP-43 proteinopathy; the minor TDP-43 pathology in a subset of MSA cases could represent an age-related ‘incidental’ phenomenon .
In MSA brain, where p25α strongly co-localizes with oligodendroglial αSyn-positive GCIs [62, 98], it is estimated that up to 50% of oligodendroglia within affected fibre tracts show abnormal accumulation of p25α . However, the physiological co-localization of p25α with MBP is markedly reduced in MSA brain, and p25α appears to relocate within oligodendroglial compartments away from the myelin sheath towards the soma, where it promotes enlargement . Together with the apparent redistribution of p25α, a marked reduction in the total level of MBA and a concurrent increase in detectable MBP degradation products can be seen. LRRK2 immunoreactivity also occurred in degenerating myelin sheaths in MSA, while aggregations of αSyn, p25α and Parkin were occasionally located in dystrophic axons . Relative to GCIs, αSyn-positive NCIs appear in later stages of MSA, when the pons becomes affected [10, 49, 73, 74]. While p25α and αSyn reactivity in GCIs is frequent, this co-localization was seen only in 40% of NCIs in MSA ; virtually all of them being negative for αSyn . Since p25α-positive neurons in both MSA and AD show twisted and straight filaments , its abnormal expression may occur in a variety of neurodegenerative disorders, and may indicate some synergistic reaction with αSyn as has been suggested for αSyn and Tau [25, 120]. Furthermore, αSyn and phosphorylated tau co-occur in certain brain regions in cases of combined MSA and AD, but the proteasomal pathways differ between αSyn- and phosphorylated tau-bearing neurons .
Papp et al.  and Matsuo et al. , using routine immunohistochemistry, have documented alterations of myelin in MSA. They are most frequent in the external capsule, striopallidal fibres, cerebellar white matter, middle cerebellar peduncles and transverse pontine tracts, but they can be identified in otherwise apparently normal areas . GCIs and microglial burden are greatest in mild to moderate white matter lesions and decrease with the severity of gray matter damage in the putamen and SN: this confirms previous findings of decreased GCIs in severely affected areas .
These changes suggest a sequence of events in which early pathogenic signals impede the normal cellular function of p25α in myelin and decrease the stability of MBP. In accordance, subsequent translocation of p25α within oligodendroglial cell bodies may favour the deposition and fibrillation of αSyn, which is suggested to be a relatively late event. With increasing neuronal loss and αSyn aggregation into GCIs, p25α is reduced, while α-β-crystallin immunoreactivity remains consistent, suggesting an early and prolonged cell-stress activation in response to GCI formation . These findings have strengthened the hypothesis of a primary oligodendrogliopathy that precedes neuronal and myelin degeneration in MSA . However, it is presently unknown whether the change in the MBP and p25α interaction represents the incipient event and myelin dysregulation occurs further downstream in MSA pathogenesis .
Data from recent postmortem studies have suggested that NIIs develop early in the disease process in pontine nuclei and inferior olives of MSA and may be an earlier phenomenon than NCI formation [73, 74, 117]. Moreover, neuronal p25α aggregation has been reported in MSA, both independently and associated with αSyn in some NCIs , similar to that seen in LBs in PD and dementia with Lewy bodies (DLB) . This similarity may indicate neuronal dysfunction in α-synucleinopathies through common pathways, which involve cytoskeleton disruption with protein dislocation and aggregation.
Novel evidence for the role of Papp–Lantos bodies in the pathogenesis of MSA
The finding of GCIs as a pathological hallmark of the disease suggests that MSA represents an invariable synucleinopathy of a nonneuronal type [19, 92]. To date, researchers of MSA have not been able to pinpoint the cellar origin(s) of oligodendrocytic αSyn. The assumption that oligodendroglial function, SNCA (synuclein α) metabolism, and GCI formation are pathogenetically intertwined is indirectly supported by the finding of early, but transient expression of endogenous SNCA messenger RNA and αSyn protein in oligodendroglial cell cultures from normal rat brain . While changes in the expression levels of SNCA messenger RNA have not been detected in MSA brain , and in situ hybridization failed to detect evidence for transcripts of the human SNCA gene in oligodendroglia from adult control brain and definite MSA , a genome-wide screening of MSA for candidate single nucleotide polymorphism associations has reported a significant association between these polymorphisms (rs 11931074, rs 3857059, and rs 3822086) at the SNCA locus and the risk for developing MSA [2, 93]. This genetic breakthrough points to a primary role of αSyn processing in the pathogenesis of MSA. Genetic association of SLC1A4, SQSTM1, and EIF4EBP1 with MSA may support the hypothesis that oxidative stress is associated with its pathogenesis .
The density and distribution of GCIs correlate significantly with the duration of illness, the clinical subtype of MSA, the degree of neuronal degeneration in the striatonigral and olivocerebellar systems [37, 44, 77] and of white matter lesions , but does not conform to the pattern of αSyn-positive NCIs and NIIs . The proportion of LRRK2-immunopositive GCIs was negatively associated with an increase in neuronal loss and αSyn-positive dystrophic axons . These latter data indicate that an increase in LRRK2 expression—like that of p25α—occurs early in association with myelin degradation and GCI formation, and that a reduction of LRRK2 expression in oligodendroglia is associated with increased neuronal loss in MSA.
These data suggest that both LRRK2 and p25α expression appear to be increased in the early stages of myelin sheath disruption and degradation prior to GCI formation in MSA. A reduction of their expression in microglia is related to increased neuronal death and possibly suggests an early neuroprotective role of these proteins. The later aggregation of Parkin in GCIs might indicate a later marked abnormality in proteosome function, probably due to the increased cellular oligomeric and fibrillary αSyn observed in MSA oligodendroglia. GCI formation may be linked, either directly or indirectly, to a mechanism that counteracts rather than accelerates nuclear shrinkage of oligodendrocytes lacking GCIs, which also suggests a link between GCI formation and neurodegeneration in MSA . Lastly, there is a reduction in constituent protein expression in degenerating MSA oligodendroglia, which is associated with axonal accumulation of αSyn, parkin and p25α. These relatively late abnormalities may be important for some of the pyramidal tract impairments observed in the absence of significant pyramidal degeneration in MSA . Overall, recent studies provided further insight into the potential sequence of cellular protein abnormalities occurring in MSA. Based on current evidence, two parallel degenerative processes in MSA have been proposed: GCI-linked oligodendrogliopathy with secondary neuronal degeneration, and neuronal α-synucleinopathy associated with aggregate formation (NIIs, NCIs, and dystrophic neurites). These two phenomena might synergistically cause neurodegeneration in MSA .
The role of transgenic mouse models of MSA
The role of αSyn oligodendrogliopathy as a trigger of MSA-like neurodegeneration has been investigated in transgenic (tg) mice with overexpression of human αSyn, reproducing the main features of human MSA: axonal αSyn aggregation and degeneration, mitochondrial dysfunction, microgliosis, and environmental stress. Using a myelin proteolipid protein (PLP) promoter, long-lasting expression of transgenic αSyn in white matter throughout the brain was achieved . The tg αSyn-positive profiles resembled the typical triangular and half moon-shaped GCIs of MSA patients at the light microscopic level . Moreover, the pathological phosphorylation at serine-129 was recapitulated in (PLP)-αSyn mice. Biochemical analysis confirmed the pathological insolubility of hyperphosphorylated tg αSyn in this mouse model. However, in contrast to mice expressing high levels of αSyn in neurons, oligodendroglial αSyn did not readily fibrillize . Likewise, use of a 2′,3′-cyclic nucleotide 3′-phosphodiesterase (CNP) promoter to express human αSyn in tg mouse oligodendrocytes failed to lead to genuine GCIs . In both instances, as well as in a MBP promoter-driven αSyn tg mouse model , gliosis accompanied the neurodegenerative process that apparently originated in oligodendrocytes.
However, complete MSA-like degeneration and MSA-like change in CNS neurotransmitter expression could not be replicated [102, 113]. Biochemical analysis confirmed the pathological insolubility of hyperphosphorylated human αSyn , and the mitochondrial inhibitor 3-nitropropionic acid enhances oxidative modification of αSyn in a tg mouse model of MSA . In another mouse model of MSA, the neuronal accumulation of insoluble αSyn was suppressed by treatment with a microtubule-depolymerizing agent . Despite the failure of the currently available transgenic models to resolve the issue surrounding the origin of αSyn accumulation in human MSA and the exact mechanism through which diseased oligodendrocytes involved by αSyn-positive Papp–Lantos bodies leads to neuronal death, they offer important insights into downstream events. They demonstrate that the accumulation of αSyn within oligodendroglia can promote neurodegeneration either through mitochondrial dysfunction and/or myelin disruption leading to axonal degeneration through microglial activation [95, 102, 123]. These findings strengthen the hypothesis that an oligodendrogliopathy can trigger neurodegeneration when induced by MSA-type synucleinopathy. A recent chip-based microarray study of transcriptome dysregulation in the rostral pons of MSA brains that was shared only in part with changes in PD tissue and other synucleinopathies, but appeared otherwise specific to oligodendrocytic processes seen in MSA, supports the notion that MSA represents a distinct disease entity .
A recent study suggested that in PD, the disease progression is possibly associated with neuron-to-neuron transmission of αSyn via endocytosis, leading to nuclear fraction and caspase 3 activation [15, 16]. In contrast, experiments in the tg (PLP)-Syn mouse model which failed to identify host-to-graft propagation did not confirm oligodendroglia-to-neuron propagation of αSyn, suggesting a different mechanism of neurodegeneration in MSA . Dopaminergic graft reinnervation accompanied by reactive gliosis in a tg MSA mouse model showed migration of oligodendrocytes expressing host-specific αSyn into the graft tissue after 3 months of survival . These data suggest that mechanisms secondary to oligodendroglial αSyn pathology are associated with changes in the microenvironment of degenerating neurons, including oligodendroglial dysfunction [62, 98], microglial activation [38, 102], and mitochondrial dysfunction , that may compromise the connectivity and neuro-restorative outcome of striatal grafts .
Because αSyn represents a major component of both oligodendroglial and neuronal inclusions in MSA, some scientists have postulated two parallel disease processes: one is attributable to degeneration of the oligodendroglia-myelin axis, as in a primary oligodendrogliopathy; and the other is caused by accumulation of αSyn within nerve cells, as in a neuronal synucleinopathy. The involvement of oligodendroglia can be seen early on throughout the central and autonomic nervous system when neuronal loss is still at a relatively early stage [118, 119, 121, 124, 125].
Recently, widespread GCIs without neuronal loss detected in two asymptomatic elderly subjects were considered to represent prodromal MSA or a rare, non-progressive age-related α-synucleinopathy . These findings suggested that accumulation of prefibrillary αSyn species might occur more frequently within neurons of MSA brain than is generally appreciated. In that case, the neurodegenerative process would develop simultaneously with the disruption of the oligo-myelin-axon interface in MSA rather than occur in sequence . In contrast, given the fact that myelin degeneration in MSA is widespread  and that p25α co-localizes with αSyn in GCIs [56, 98], it appears likely that a primary oligodendrogliopathy is the main engine driving the disease process in MSA. Whether the affected oligodendrocytes mediate axonal dysfunction and neurodegeneration in MSA through direct physical contact or whether humoral and cellular effects (or both) promote an inflammatory response that results in cell death remains to be addressed . In a man aged 81 years with isolated REM sleep behavioral disorder, numerous GCIs and NCIs within medulla, pons, cerebellum, dentate nucleus and other brain areas were associated with moderate neuronal loss and gliosis in these areas, suggesting a preclinical phase of MSA, preceding the classical dysautonomic and motor signs .
More than 20 years have passed since both the critical question as to the “nature of the beast” that underlies MSA  was posted and Papp–Lantos inclusions, now accepted as the morphological hallmark of MSA , were reported. At present, the basic mechanisms by which GCI-involved oligodendroglia degenerates and promotes myelin degeneration and neuronal death still remain unknown. Immunohistochemical studies have shown that phosphoinositide 3-kinase (PI3K), which is closely associated with the regulation of apoptosis, is upregulated in many oligodendrocytes and isolated neurons in MSA brain . While apoptosis has been detected in PD neurons  and in autopsy-confirmed MSA, known indices of cell death, such as intranucleosomal DNA fragmentation, expression of apoptosis-related proteins, and activation of caspase-3, are rarely seen within neurons. However, such changes can be seen together with αSyn-positive inclusions within affected microglia and oligodendroglia [41, 42, 86]. Cytoplasmic expression of Bcl-2 and concurrence with αSyn can be detected in around 25% of GCI-positive cells . Since these cells are usually Bcl-2 negative, its expression in MSA tissue may reflect a stress response that is aimed at avoiding cell death through the upregulation of this antiapoptotic protein .
In conclusion, accumulating evidence suggests that aberrant oligodendroglial αSyn accumulation, the main constituent of the Papp–Lantos inclusions, causing oligodendrocyte dysfunction, myelin disruption and, finally, neuronal degeneration, represents the target of and markers for the neurodegenerative process in MSA. It appears that the “nature of the beast”  in MSA now can be slightly better defined, namely, as a unique entity within the spectrum of proteinopathies . Today, we consider MSA to represent a primary oligodendrogliopathy that features early myelin dysfunction, progressive synucleinopathy, and associated axonal damage, thereby resulting in secondary neurodegeneration. Future studies should address perturbations in signalling between oligodendrocytes involved by Papp–Lantos inclusions and degenerating neurons leading to the striking pattern of neurodegeneration observed in MSA.
The authors thank Mr. E. Mitter-Ferstl, PhD, for secretarial and computer work. The work was partially supported by the Society for the Promotion of Research in Experimental Neurology, Vienna, Austria.