J Clin Pathol: Mol Pathol 2001;54:419–426 419

Multiple system atrophy: cellular and molecular

pathology
D J Burn, E Jaros

Abstract by Graham and Oppenheimer in 1969 to avoid

Multiple system atrophy is an adult onset “unnecessary confusion caused by inventing
neurodegenerative disease, featuring par- new names”.2 Recent consensus criteria rec-
kinsonism, ataxia, and autonomic failure, ommend that patients with predominantly par-
in any combination. The condition is kinsonian features should be designated
relentlessly progressive and responds MSA-P, whereas those with predominantly
poorly to treatment. Death occurs on cerebellar features are designated MSA-C.3
average six to seven years after the onset These terms are more clinically appropriate
of symptoms. No familial cases of multiple than the pathologically derived terms striatoni-
system atrophy have been reported, and gral degeneration and olivopontocerebellar
no environmental factors have been ro- atrophy types of MSA, respectively.
bustly implicated as aetiological factors. The description of argyrophilic glial cyto-
However, analytical epidemiological stud- plasmic inclusions (GCIs) in the oligodendro-
ies are hampered because the condition is cytes of patients with phenotypically varied
relatively rare. The discovery of the glial MSA provided persuasive evidence that MSA
cytoplasmic inclusion (GCI) in 1989 is a clinicopathological entity.4 5 Subsequently,
helped to define multiple system atrophy similar inclusions were identified in glial nuclei
as a clinicopathological entity, and drew (glial nuclear inclusions (GNIs)), and in
attention to the prominent, if not primary, neuronal cytoplasm and nuclei (neuronal cyto-
role played by the oligodendrocyte in the plasmic inclusions (NCIs), and neuronal nu-
pathogenesis of the condition. Subse- clear inclusions (NNIs), respectively), and also
quently, GCIs were shown to be positive in axons, although all at lower frequencies than
for á-synuclein, with immunostaining for the GCIs.6 However, the mechanisms under-
this protein indicating that white matter lying inclusion body formation, and how it
pathology was more widespread than had relates to glial and neuronal loss in MSA,
previously been recognised. The presence remain to be determined. Understanding these
of á-synuclein in GCIs provides a link with processes will be essential before an eVective
Parkinson’s disease, dementia with Lewy treatment to halt or reverse the disease can be
bodies, and neurodegeneration with brain developed.
iron accumulation, type 1 (or Recently, an abnormally insoluble
Hallervorden-Spatz syndrome), in which filamentous/tubular form of á-synuclein pro-
á-synuclein is also found within Lewy tein was shown to be a major component of
bodies. This has led to the term “synuclei- GCIs, NCIs, and NNIs, but not of GNIs (fig
nopathy” to embrace this group of condi- 1).7–10 The accumulation of this synaptic
tions. The GCIs of multiple system protein in MSA provides an unexpected and as
atrophy contain a range of other cytoskel- yet unexplained link with Parkinson’s disease,
etal proteins. It is unknown how fibrillo- dementia with Lewy bodies, and neurodegen-
genesis occurs, and whether there is eration with brain iron accumulation, type 1
primary oligodendrocytic dysfunction, (NBIA 1 or Hallervorden-Spatz syndrome), in
which then disrupts the neurone/axon as a which á-synuclein is an important component
consequence of the glial pathology, or of the Lewy bodies. MSA, Parkinson’s disease,
Regional whether the oligodendrocytic changes dementia with Lewy bodies, and NBIA 1 are
Neurosciences Centre, merely represent an epiphenomenon. thus sometimes referred to as “synucleinopa-
Newcastle General Further research into this devastating thies”.11 12
Hospital, Westgate condition is urgently needed to improve
Road, Newcastle upon our understanding of the pathogenesis, Presentation and natural history of MSA
Tyne NE4 6BE, UK and also to produce new treatment ap- Although parkinsonism develops in most pa-
D J Burn tients with MSA (84–100%), “pure” cerebellar
proaches.
Institute for the Health (J Clin Pathol: Mol Pathol 2001;54:419–426) forms of the disorder are uncommon (0–
of the Elderly, 16%).13 Autonomic failure was reported in
Keywords: multiple system atrophy; glial cytoplasmic 78% of 188 pathologically confirmed cases of
Department of
inclusion; á-synuclein; oligodendrocyte
Neuropathology, MSA.14 Symptomatic orthostatic hypotension
Newcastle General occurring within one year of onset of parkin-
Hospital Multiple system atrophy (MSA) is a devastat- sonism has recently been shown to predict
E Jaros
ing adult onset, sporadic neurodegenerative MSA in 75% of cases, although this study was
Correspondence to: disease of unknown aetiology, characterised by retrospective, and the numbers involved were
Dr Burn parkinsonism, ataxia, and autonomic failure, in small.15 The parkinsonism of MSA is more
d.j.burn@ncl.ac.uk any combination. The age adjusted prevalence variably responsive to levodopa than Parkin-
Accepted for publication of MSA in the UK is 4.4/100 000.1 The term son’s disease itself. Up to one third of patients
8 March 2001 “multiple system atrophy” was first proposed may show a moderate or good response to this

www.molpath.com
420 Burn, Jaros

drug, but this usually declines over the first one Neuropathology of MSA and
to two years of treatment.14 clinicopathological correlation
A male predominance of 1.4 : 1 was re- Macroscopically, the brain in MSA shows
ported by Quinn in a review of 231 pathologi- varying degrees of atrophy of the cerebellum,
cally confirmed MSA cases.14 If confirmed, this cerebellar peduncles (especially the middle and
observation may have aetiological relevance. inferior peduncles), pons, medulla, and also the
For example, there may be greater environ- posterolateral putamen. There may be loss of
mental exposure to putative toxins in men, or pigment in the substantia nigra and also
endogenous protective factors (hormonal per- discolouration of the striatum (notably the
haps) in women. putamen). Excessive iron accumulation has
The mean age at onset in 203 pathologically been demonstrated within the striatum to
confirmed cases of MSA was 54.3 (range, account for this pigmentary change.19
33–78) years.16 The upper limit of the age range Oligodendrocyte GCIs and GNIs have a so
must be viewed with a degree of caution, how- called “system bound” distribution in the
ever, because clinicopathological series are suprasegmental motor systems (primary
prone to bias, and older patients are less likely motor, and higher motor areas of the cerebral
to undergo postmortem examination.17 A cortex, pyramidal, extrapyramidal, and cortico-
population based study is necessary to confirm cerebellar systems), in the supraspinal auto-
the age range and mean age of disease onset. nomic systems, and in their targets.20–22 Neu-
The mean disease duration was only 6.2 ropathological changes in neurones follow a
(range, 0.5–24) years in a recent meta-analysis similar system bound distribution and include
of 433 pathologically confirmed cases,18 indica- variable neuronal loss, and densities of NCIs
tive of a relentlessly progressive illness. This and NNIs in the striatum, substantia nigra,
review was retrospective, however, and may locus ceruleus, inferior olives, pontine nuclei,
have been biased towards the most severe cerebellar Purkinje cells, dorsal motor nucleus
cases. Cerebellar features were associated with of vagus, nucleus vestibularis, intermedi-
marginally increased survival in this review, but olateral cell column of the spinal cord, and
this did not reach significance. Onuf’s nucleus.16 23 Rare MSA cases showing

Figure 1 á-Synuclein pathology in multiple system atrophy. (A) Glial cytoplasmic inclusions (GCls) (arrows) in
cerebellar white matter with cerebellar granule cells in the lower right corner. (B) Neuronal cytoplasmic inclusions (NCIs)
(double arrows) and an early formation of a neuronal nuclear inclusion (NNI) (triple arrow) in neurones of pontine nuclei
with neurites in the neuropil (arrowheads). (C) A GCI (arrow) and the early formation of an NCI (double arrow) and
NNI (triple arrow) in a neurone of the pontine nuclei with neurites in the neuropil (arrowheads). (D) A GCI (arrow) and
an NNI (triple arrow) in the pontine nuclei. Immunohistochemistry was performed using monoclonal antibodies to
á-synuclein (Novocastra Laboratories, Newcastle upon Tyne, UK) on formalin fixed, paraYn wax embedded sections that
had been pretreated with formic acid; Vectastain Elite ABC peroxidase kit (Vector, Peterborough, UK); DAB;
Haematoxylin counterstain. Scale bars in A to D, 30 µm.

www.molpath.com
Multiple system atrophy 421

additional involvement of frontal or temporal Several studies have looked for polymor-
lobes, including atrophy, and the presence of phisms or mutations in candidate genes, which
GCIs and NCIs have also been reported.24 25 may predispose an individual towards develop-
White matter pathology is also increasingly ing MSA. The apolipoprotein å4 allele is not
recognised in MSA, with the fibre tracts of the over-represented in MSA when compared with
suprasegmental motor and supraspinal auto- controls, and there have been conflicting
nomic systems (see above) bearing the brunt of reports of the association of a cytochrome
demyelination.26 Furthermore, using mono- P-450-2D6 polymorphism with MSA.38–40
clonal and polyclonal antibodies that recognise There is no evidence to suggest that patients
epitope QDENPVV of human myelin basic with MSA have expansions at the SCA1 and
protein, accesible only in areas of myelin SCA3 alleles.41 Furthermore, there is no
degeneraiton, Matsue and colleagues have evidence to support an association between
demonstrated that in MSA myelin degenera- polymorphisms in the H5 pore region of the
tion and abnormal oligodendrocytes were human homologue of the weaver mouse gene,
widespread, even in areas where GCIs were not hiGIRK2, the insulin-like growth factor 1
detectable, and where myelin appeared intact receptor gene (linked with a decreased intra-
with standard myelin stains.27 cellular response to insulin-like growth factor 1
A correlation has been established between in the cerebellar cortex of lurcher mice), or the
akinesia and the degree of nigral and putaminal ciliary neurotrophic factor gene.41
cell loss, although rigidity relates only to this It seems improbable that a mutation in the
last feature.16 Ataxia correlates with the degree á-synuclein gene underlies protein accumula-
of olivopontocerebellar atrophy and pyramidal tion in MSA. Recent studies have not found a
signs with pyramidal tract pallor.16 Recently, a mutation in the entire coding region of the
loss of Betz cells was documented in all of á-synuclein gene in patients with pathologi-
seven patients with MSA studied, six of whom cally confirmed MSA.42 43 However, mutations
had pyramidal signs documented before in the regulatory or intronic regions of the gene
death.28 Some groups have found an associ- have not been ruled out.
ation between postural hypotension and inter- Polymorphisms in the á-synuclein gene
mediolateral cell column degeneration,16 but might increase the risk of developing MSA, by
this finding has not been confirmed by others.29 promoting á-synuclein protein aggregation. To
A severe loss of catecholaminergic neurones in date, polymorphisms have been identified in
the rostral ventrolateral medulla has been the promoter sequence, and in the intron 4
noted in patients with MSA.30 This area is sequence of the á-synuclein gene.44 A combina-
involved in the control of sympathetic cardio- tion of allele 1 of the á-synuclein promoter
vascular outflow. polymorphism and the ApoE4 allele has been
There is limited information available about reported to increase the relative risk for devel-
striatal dopamine receptors in MSA and their oping sporadic Parkinson’s disease 12.8 fold.45
correlation with extrapyramidal features. A Polymorphisms in codons 1 to 39 of the
relative preservation of putaminal cell counts in á-synuclein gene, a domain related to interac-
patients with MSA responding to levodopa has tion with the recently identified protein,
been reported,31 and resistance to levodopa synphilin-1, or polymorphisms in the
might be the result of a loss of putaminal synphilin-1 gene itself, or in the genes of other
dopamine D2 receptors.32 Patients with MSA protein interacting partners of á-synuclein,
who are not responsive to levodopa may have may also need to be considered.46 The number
more severe topographical degeneration of the of á-synuclein protein interacting partners has
putaminal eVerent terminals in the ventrola- expanded to include 14-3-3 protein chaper-
teral portion of the globus pallidus.33 In ones, protein kinase C, extracellular regulated
contrast, a case of MSA has been reported kinase, and BAD, a Bcl-2 homologue that
where there was no significant response to regulates cell death.47
dopaminergic treatment, yet there was no Increased expression of a brain specific pro-
evidence of putaminal cell loss at necropsy.34 tein called ZNF231 in cerebellar neurones has
Furthermore, in vivo positron emission tomog- been reported to occur in patients with MSA.48
raphy (PET) studies using the dopamine D2 The gene is located on chromosome 3p21 and
receptor ligand 11C-raclopride have demon- encodes a neuronal double zinc finger protein
strated only a modest 15% reduction in striatal with a nuclear targeting sequence, suggesting
D2 sites.35 The failure to respond to levodopa that it might function as a transcription regula-
was thought to reflect “loss of other basal gan- tor. The importance of this finding is as yet
glia connections”. uncertain, but it is possible that patients with
MSA diVer from unaVected individuals by
sequence polymorphisms within, and flanking,
Genes, polymorphisms, and MSA the putative functional motifs of the ZNF231
MSA, as reflected in its current definition, is gene.
regarded as a sporadic disease.36 Familial cases
have not been described, although clinical Glial cytoplasmic inclusions:
symptoms of MSA were reported by a characteristics and composition
significantly larger group of patients’ relatives Argyrophilic oligodendroglial inclusions were
than controls in one study.37 However, a self first described in the brains of patients with
administered questionnaire was used to elicit MSA in 1989 and became known as glial cyto-
symptoms from the relatives in this series, plasmic inclusions.4 Subsequent studies have
leading to potential bias. confirmed the sensitivity and specificity of

www.molpath.com
422 Burn, Jaros

Table 1 Immunocytochemical characteristics of glial The small heat shock protein and molecular
cytoplasmic inclusions in multiple system atrophy51 56 chaperone, áB-crystallin, is a normal compo-
Positive immunoreactivity Negative immunoreactivity
nent of the central nervous system, where it is
expressed primarily in oligodendrocytes and to
á-Synuclein Tau* a lesser degree in astrocytes.57 It is also an
Ubiquitin Neurofilaments
áB-crystallin Glial fibrillary acidic protein important protein component of GCIs. áB-
á-Tubulin and â-tubulin Myelin basic protein crystallin binds to 20S proteasome, thereby
Mitogen activated protein 2 Vimentin regulating its proteolytic activity,56 in addition
Mitogen activated protein 5 Actin
Cyclin dependent kinase 5 Desmin to binding to intermediate filaments.57
Mitogen activated protein kinase Myosin Ubiquitin is also involved in the 26S protea-
Midkine Cytokeratin
Rab5
some dependent proteolytic process and may
Rabaptin-5 play a protective role against neurodegenera-
tion.58 Although the inclusions of MSA may be
*Glial cytoplasmic inclusions are generally negative for
phosphate dependent tau antibodies and positive for normal
identified by ubiquitin immunostaining, results
adult tau. on isolated GCI proteins suggest that they are
poorly ubiquitinated.56
In sections of MSA brains, antibodies to
GCIs for MSA.5 21 49 Double staining tech- á-synuclein immunolabel a greater number of
niques using markers for oligodendroglia, such GCIs than do anti-ubiquitin antibodies, indi-
as myelin basic protein, Leu-7, carbonic anhy- cating that á-synuclein is a major component of
drase enzyme II, and transferrin, confirm the GCIs, and that the accumulation of
localisation of the inclusion to this cell type.49 50 á-synuclein precedes its ubiquitination.8 11
In contrast, GCI containing cells stain nega- á-Synuclein, also referred to as the precursor of
tively for glial fibrillary acidic protein (an the non-amyloid component of plaques
astrocytic marker), and for class II major histo- (NACP), is a 140 amino acid protein that is
compatibility antigen (a microglial marker). normally localised in the human brain to
Using routine light microscopy, GCIs are presynaptic nerve terminals.59 It is natively
faint eosinophilic inclusions that eccentrically unfolded and highly soluble,60 but can poly-
displace the nucleus. By virtue of its selective merise into filaments under a variety of in vitro
dark staining of inclusions with a clean conditions, including increased temperature
background, the Gallyas silver technique is the and concentration, acidic pH conditions,
method of choice for demonstrating GCIs. longer time lag, and increased iron
This technique shows that the GCIs vary in concentrations.61–64 Hence, the formation and
morphology from sickle shaped to flame accumulation of á-synuclein filaments in
shaped to ovoid, and occasionally superficially GCIs, and in Lewy bodies,7–9 11 has been
resemble neurofibrillary tangles.51 GCIs are speculated to result from altered intracellular
essentially negative with other commonly used conditions.61 62 It is of interest that in the basal
stains, including phosphotungstic acid haema- ganglia of patients with MSA, total iron
toxylin, periodic acid SchiV, Masson tri- concentrations are raised,65 and GCIs have
chrome, Alcian blue, thioflavine S, Congo red, been found within oligodendroglial cells con-
oil red O, and Sudan black B.23 52 taining iron pigment, although inclusions have
At the ultrastructural level, GCIs are non- also been found in cells with no evidence of
membrane bound cytoplasmic inclusions com- pigment accumulation.19 Oligodendrocytes are
posed of loosely aggregated filaments/tubular the predominant iron regulatory cells in the
structures (20–40 nm in cross sectional diam- brain,66 but it is not known whether oli-
eter) and granular material that may ensnare godendrocytes in patients with MSA show
cytoplasmic organelles, such as mitochondria abnormal concentrations or activities of ferritin
and secretory vesicles.4–6 49 (the iron sequestration protein) or transferrin
Table 1 lists the immunocytochemical char- (the iron transport protein).
acteristics of GCIs. Notably, GCIs are immu- Full length á-synuclein is present in GCIs
noreactive for ubiquitin and are generally and NCIs,11 67 although more vigorous antigen
negative for tau. Conflicting data exist regard- retrieval is required for immunohistochemical
ing whether GCIs stain for tau, but it is now detection of other than C-terminal epitopes.67
believed that if the inclusions contain tau, it is In contrast, very little full length á-synuclein
largely non-phosphorylated, in contrast with appears to be present in immunoisolated GCIs,
the neurofibrillary pathology of Alzheimer’s and the C-terminal truncated form of
disease.53 Thus, phosphate dependent tau anti- á-synuclein may predominate.56 In addition to
bodies will not label GCIs in MSA. GCIs are the formation of GCIs, there is evidence for a
immunoreactive with several other cytoskeletal more widespread modification of á-synuclein
proteins, including á-tubulin and â-tubulin,4 solubility in MSA than is obvious from the
mitogen activated protein 5 (MAP5),54 GCI distribution.26 In particular, in MSA, and
MAP2,55 and also cyclin dependent kinase 5 also in preliminary studies with Lewy body
(cdk5) and MAP kinase (MAPK), both of disease, an increased ratio of sodium docecyl
which are known to phosphorylate MAP2.55 sulphate soluble to buVer soluble á-synuclein
This led to the hypothesis that there is a close has been seen, leading to the proposal that this
association between the GCI and the microtu- property may be a biochemical “fingerprint”
bular cytoskeleton, with aberrant or ectopic for the synucleinopathies, even in the absence
expression of the neuronal kinases cdk5 and of inclusion bodies.26
MAPK causing abnormal phosphorylation of Using immunoelectron microscopy on iso-
microtubular cytoskeletal proteins.55 lated sarcosyl insoluble filaments extracted

www.molpath.com
Multiple system atrophy 423

from MSA brains, and PER4 antiserum to the GCIs, however, neuronal inclusions stain posi-
C-terminus of á-synuclein, some filaments tively both for ubiquitin and á-synuclein.19
appear to be twisted, with a width alternating Antibodies to á-synuclein, but not to ubiquitin,
between 5 and 18 nm, and an apparent period also reveal numerous degenerating neurites in
of 70–90 nm, whereas other filaments appear the white matter of patients with MSA.8 26 This
to be straight, with a uniform width of 10 nm.11 suggests that a hitherto unrecognised degree of
The diVerences in morphology and diameter pathology may be present in the axons of
between the isolated á-synuclein filaments and patients with MSA, although whether
the aggregated filamentous/tubular structures neuronal/axonal á-synuclein pathology pre-
seen in sections of GCIs (see above) are cedes glial á-synuclein pathology or myelin
thought to indicate that although á-synuclein degeneration (see above) has not yet been
may play a key role in fibrillogenesis, other determined.
cytoskeletal proteins (for instance á-tubulin
and â-tubulin) are involved in filament forma- Cell death of oligodendrocytes and
tion.56 62 Furthermore, non-cytoskeletal ele- neurones in MSA
ments also appear to be involved, as demon- It is something of a paradox that although
strated most recently with antibodies to MSA produces clinical symptoms typical of
midkine, a new neurotrophic factor found to grey matter dysfunction, the hallmark patho-
label most GCIs intensely. With immunoelec- logical lesion aVects myelin producing cells.51
tron microscopy, midkine positive, granule How oligodendroglial dysfunction might lead
coated fibrils appear to be essential constitu- to regional neuronal loss remains unexplained.
ents of GCIs.68 Midkine is a heparin binding There is no evidence to suggest that oli-
growth factor, implicated in various biological godendrocytes can be subtyped according to
phenomena such as neuronal survival, diVeren- the neuronal populations they subserve. In-
tiation, and migration, angiogenesis, and car- deed, single oligodendrocytes may myelinate
cinogenesis.69 Midkine is strongly expressed in axons of diVerent anatomical tracts. Further-
the nervous system during the midgestation more, GCIs involve all morphological types of
period, probably by astrocytes. It is absent from oligodendrocytes (perivascular, perifascicular,
normal adult brain, except in ischaemic zones and perineuronal) with varying frequency in
surrounding cerebral infarction.70 diVerent anatomical regions.23 There are there-
The reason for the expression and accumu- fore no clues to point towards a “selective vul-
lation of á-synuclein, a nerve terminal protein, nerability” of a particular subgroup of oli-
in the GCIs of MSA brains is unknown. In cul- godendrocytes. Nevertheless, because GCIs
tured rat oligodendrocytes the expression of and oligodendroglial loss show a pronounced
á-synuclein mRNA and protein is developmen- preponderance over NCIs and neuronal loss, it
tally regulated,71 hence the accumulation is has been suggested that oligodendroglial
more likely to be a consequence of altered pathology may be the primary lesion of MSA,
rather than de novo expression. One possibility rather than an epiphenomenon.21–23 It has also
is that oligodendrocytes in MSA brains may been shown that in MSA pronounced DNA
have an impaired ability to degrade fragmentation, characteristic of apoptotic cell
á-synuclein, which they normally produce at death, occurs almost exclusively in oli-
very low concentrations.9 Alternatively, selec- godendrocytes in a distribution pattern similar
tive upregulation in the expression of to that of GCIs.22 However, it appears that GCI
á-synuclein in glial cells could occur in formation may represent a diVerent stage in
response to certain pathologies.7 The accumu- oligodendroglial pathology because only oli-
lation in GCIs of tau and MAP2 also appears godendrocytes lacking GCIs express Bax, the
to be a consequence of altered rather than de apoptosis promoting protein. In contrast, the
novo synthesis because both of these princi- GCI containing oligodendrocytes express
pally neuronal proteins are expressed in imma- Bcl-2, the apoptosis suppressor protein, indica-
ture oligodendrocytes grown without axonal tive of an activated repair mechanism.22 This
contact in tissue culture.72 73 The aberrant raises the possibility that in MSA regional
expression of the neuronal kinases cdk5 and neuronal/axonal pathology and withdrawal of
MAPK (see above), of the neuronal endocyto- axon derived trophic factors normally prevent-
sis regulatory proteins Rab5 and Rabaptin-5,74 ing apoptosis (see below) precede the dysfunc-
and of the neuronal survival and diVerentiation tion of oligodendrocytes, and demyelination in
factor midkine (see above), normally expressed the related white fibre tracts.
by astrocytes, suggests even more profound The absence of DNA fragmentation in MSA
alterations in the phenotype of the MSA neurones may indicate that they are destroyed
oligodendrocytes. Whether this is indicative of either by necrosis or by a form of programmed
the existence of a repair mechanism68 or of a cell death other than apoptosis.22 Nonetheless,
more profound defect in the oligodendrocyte– increased p53 immunoreactivity in the striatal
axon–neurone communication, involving and midbrain neurones of patients with MSA
oligodendroglial/neuronal trophic factors, re- has been interpreted as evidence of neuronal
mains to be established. apoptosis75 because p53 gene upregulation is
The neuronal inclusions (NCI and NNI) in known to precede apoptosis. More recently,
MSA are less frequent than GCIs, and their immunoreactivity for the calcium binding pro-
ultrastructure reveals a composition of gran- teins calbindin and parvalbumin has been
ules and filaments that tend to be associated shown to be greatly decreased in the cerebellar
with a more diverse range of cellular organelles, Purkinje cells of patients with MSA, whereas
when compared with GCIs. In common with immunoreactivity for both Bax, the apoptosis

www.molpath.com
424 Burn, Jaros

promoting protein, and the Bcl-x, the apoptosis in the cortex or in the minor aVerents originat-
suppressor protein, was increased. The expres- ing in the substantia nigra. The striatal
sion of Bcl-2, another apoptosis suppressor neurones themselves are not BDNF immuno-
protein, was restricted to a subpopulation of reactive.82 So far, there have been no reports of
granule neurones.76 It was suggested that a alterations in glial growth factors in patients
diminished calcium binding capacity of MSA with MSA. Glial growth factors are derived
Purkinje cells could lead to a change in the from both astrocytes and neurones. White
regulation of proteins of the Bcl-2 family, matter astrocytes produce platelet derived
favouring the initiation of apoptosis. growth factor (PDGF), which has been shown
A further clue to the regional selectivity seen to act as a survival factor for newly formed oli-
in MSA and apoptosis may lie in the study of a godendrocytes.84 Neurones/axons generate sev-
family of cysteine proteases called caspases, eral isoforms of neuregulins, which are mem-
which act as the executioners of apoptosis. bers of a large trophic family all originating
Cytoskeletal proteins and enzymes essential for from a single gene. Some of the isoforms
cell repair may be substrates for caspase promote proliferation and survival of oli-
processing in several neurodegenerative dis- godendrocytes through their interaction with
eases.77 Caspase-3 cleaves the antiapoptotic the erb-B family of receptors, expressed by the
protein Bcl-2. Peptide caspase inhibitors have oligodendrocytes.85 Glial growth factor, one of
been shown to protect against 1-methyl-4- the neuregulin isoforms, promotes the prolif-
phenylpyridinium induced apoptosis in cul- eration and survival of oligodendrocyte pro-
tured cerebellar granule neurones.78 Further- genitors but inhibits their further diVerentia-
more, the distribution of caspase-3 may tion.86 ARIA, another neuregulin isoform, acts
contribute towards the regional vulnerability as a morphogen for developing oligodendro-
seen in the substantia nigra in idiopathic cytes by promoting the extension and complex-
Parkinson’s disease, with dopaminergic neu- ity of their processes.87
rones expressing caspase-3 degenerating pref- Most recently, it has been demonstrated that
erentially.79 Similar studies in MSA have not yet the sensitivity of oligodendrocytes to astrocyte
been reported. derived PDGF is greatly enhanced by an extra-
The activation of microglial cells may be the cellular matrix glycoprotein, laminin-2,88 which
final common pathway, contributing both to is expressed by Purkinje cell axons.89
demyelination and neuronal removal, irrespec- Laminin-2 also appears to play an important
tive of the mode of cell death. Microglia express part in the signalling mechanisms that stimu-
proinflammatory peptides, which may be a late oligodendrocytes to elaborate the myelin
result of activation of nuclear factor êB sheath.90 Laminin-2 is a heterotrimer com-
(NF-êB). AVected brain areas of patients with posed of a longer á2 chain and shorter â1 and
MSA show strong immunoreactivity for nu- ã1 chains. The á2 chain is associated with neu-
clear Rel A p65 (a subunit of the NF-êB/Rel ritic processes, synapses, and dendritic spines
family), which is almost exclusively localised in of limbic neuronal populations, whereas the ã1
activated microglia. Nuclear translocation of chain is found in cell bodies of essentially all
Rel A is not detected in striatal tissue of normal neurones in the central nervous system.91 92
controls and patients with Parkinson’s disease. Based on the restricted distribution of the á2
This suggests that NF-êB/Rel A complexes chain to the limbic areas, it has been proposed
may play a role in mediating microglial activa- that isoforms of the laminin á chains, other
tion in MSA.80 Finally, PK 11195 selectively than the already known á1 and á2 chains, are
binds to benzodiazepine sites on activated expressed by diVerent neuronal systems.91
microglia. 11C PK 11195 positron emission Thus, it is conceivable that polymorphisms in
tomography has recently demonstrated acti- the genes encoding laminin á chains, which are
vated microglia in vivo in the putamen, shared by the suprasegmental motor system,
pallidum, substantia nigra, and pontine region the supraspinal autonomic systems, and their
in four patients with MSA.81 targets, could be risk factors for MSA, and may
underlie the system bound distribution of the
Trophic factors and extracellular matrix oligodendroglial and neuronal pathology in this
Trophic factors have been studied in several disease.
neurodegenerative diseases, both for their
potential pathogenic role and also for their Conclusion
possible therapeutic benefit. Brain derived The aetiology of MSA remains elusive. The
neurotrophic factor (BDNF) is abundantly and discovery of á-synuclein within GCIs has pro-
widely expressed in neurones of the adult vided a recent impetus to research, and also an
mammalian brain. It has a neurotrophic eVect elusive link with Parkinson’s disease.
on many neuronal types, including nigral Nevertheless, MSA is distinguished from other
dopaminergic and striatal neurones.82 BDNF neurodegenerative diseases by the prominent,
positive neurites have been shown to be more if not primary, involvement of the glial cell. In
abundant in the striatum of patients with MSA contrast with Parkinson’s disease, dementia
than in the striatum of those with Parkinson’s with Lewy bodies, and a host of other neurode-
disease and normal controls. The upregulation generative tauopathies, familial MSA has never
of BDNF has been interpreted as a protective been documented, and there are no clues as yet
mechanism against progressive degeneration of to a genetic predisposition.
the striatal neurones in MSA.83 However, it is It is likely that future progress in unravelling
not clear whether the upregulation has oc- the cause of MSA will need to include large
curred in the major striatal aVerents originating multicentre clinical studies, applying validated

www.molpath.com
Multiple system atrophy 425

diagnostic criteria to yield a sizeable cohort of 24 Konagaya M, Sakai M, Matsuoka Y, et al. Multiple system
atrophy with remarkable frontal lobe atrophy. Acta
well characterised MSA cases. This cohort will Neuropathol (Berl) 1999;97:423–8.
form the necessary basis for genetic and patho- 25 Shibuya K, Nagatomo H, Iwabuchi K, et al. Asymmetrical
temporal lobe atrophy with massive neuronal inclusions in
logical studies. The potentially central role of multiple system atrophy. J Neurol Sci 2000;179:50–8.
á-synuclein in fibrillogenesis, and its interac- 26 Dickson DW, Liu W, Hardy J, et al. Widespread alterations
of alpha-synuclein in multiple system atrophy. Am J Pathol
tion with other cytoskeletal proteins within 1999;155:1241–51.
glial cells is likely to provide a fruitful avenue 27 Matsuo A, Akiguchi I, Lee GC, et al. Myelin degeneration in
multiple system atrophy detected by unique antibodies. Am
for further research. At the same time, the tem- J Pathol 1998;153:735–44.
poral and spatial involvement of the pathologi- 28 Tsuchiya K, Ozawa E, Haga C, et al. Constant involvement
of the Betz cells and pyramidal tract in multiple system
cal process in MSA needs to be investigated in atrophy: a clinicopathological study of seven autopsy cases.
more detail, and potentially disruptive factors Acta Neuropathol (Berl) 2000;99:628–36.
29 Gray F, Vincent D, Hauw JJ. Quantitative study of lateral
of the oligodendrocyte–neurone–axon func- horn cells in 15 cases of multiple system atrophy. Acta Neu-
tional unit should be explored further. ropathol (Berl) 1988;75:513–18.
30 Benarroch EE, Smithson IL, Low PA, et al. Depletion of
catecholaminergic neurons of the rostral ventrolateral
medulla in multiple systems atrophy with autonomic
1 Schrag A, Ben-Shlomo Y, Quinn NP. Prevalence of progres- failure. Ann Neurol 1998;43:156–63.
sive supranuclear palsy and multiple system atrophy: a 31 Fearnley JM, Lees AJ. Striatonigral degeneration: a clinico-
cross-sectional study. Lancet 1999;354:1771–5. pathological study. Brain 1990;113:1823–42.
2 Graham JG, Oppenheimer DR. Orthostatic hypotension 32 Churchyard A, Donnan GA, Hughes A, et al. Dopa
and nicotine sensitivity in a case of multiple system atrophy. resistance in multiple system atrophy: loss of dopamine D2
J Neurol Neurosurg Psychiatry 1969;32:28–34. receptors. Ann Neurol 1993;34:219–26.
3 Gilman S, Low PA, Quinn N, et al. Consensus statement on 33 Ito H, Kusaka H, Matsumoto S, et al. Striatal eVerent
the diagnosis of multiple system atrophy. J Neurol Sci 1999; involvement and its correlation to levodopa eYcacy in
163:94–8. patients with multiple system atrophy. Neurology 1996;47:
4 Papp MI, Kahn JE, Lantos PL. Glial cytoplasmic inclusions 1291–9.
in the CNS of patients with multiple system atrophy (stria- 34 Wenning GK, Quinn N, Magalhaes M, et al. “Minimal
tonigral degeneration, olivopontocerebellar atrophy and change” multiple system atrophy. Mov Disord 1994;9:161–
Shy Drager syndrome). J Neurol Sci 1989;94:79–100. 6.
5 Kato S, Nakamura H, Hirano A, et al. Argyrophilic ubiqui- 35 Brooks DJ, Ibanez V, Sawle GV, et al. Striatal D2 receptor
tinated cytoplasmic inclusions of Leu-7-positive glial cells status in patients with Parkinson’s disease, striatonigral
in olivopontocerebellar atrophy (multiple system atrophy). degeneration, and progressive supranuclear palsy,
Acta Neuropathol 1991;82:488–93. measured with 11C-raclopride and positron emission tom-
6 Papp MI, Lantos PL. Accumulation of tubular structures in ography. Ann Neurol 1992;31:184–92.
oligodendroglial and neuronal cells as the basic alteration 36 Wenning GK, Wagner S, Daniel S, et al. Multiple system
in multiple system atrophy. J Neurol Sci 1992;107:172–82. atrophy: sporadic or familial? Lancet 1993;342:681.
7 Arima K, Ueda K, Sunohara N, et al. NACP/alpha- 37 Nee LE, Gomez MR, Dambrosia J, et al. Environmental-
synuclein immunoreactivity in fibrillary components of occupational risk factors and familial associations in multi-
neuronal and oligodendroglial cytoplasmic inclusions in ple system atrophy: a preliminary investigation. Clin Auton
the pontine nuclei in multiple system atrophy. Acta Res 1991;1:9–13.
Neuropathol (Berl) 1998;96:439–44. 38 Iwahashi K, Miyatake R, Tsuneoka Y, et al. A novel
8 Gai WP, Power JH, Blumbergs PC, et al. Multiple-system cytochrome P-450IID6 (CYPIID6) mutant gene associ-
atrophy: a new alpha-synuclein disease? Lancet 1998;352: ated with multiple system atrophy. J Neurol Neurosurg Psy-
547–8. chiatry 1995;58:263–4.
9 Tu PH, Galvin JE, Baba M, et al. Glial cytoplasmic 39 Bandmann O, Wenning GK, Quinn NP, et al. Arg296 to
inclusions in white matter oligodendrocytes of multiple Cys296 polymorphism in exon 6 of cytochrome P-450-
system atrophy brains contain insoluble alpha-synuclein. 2D6 (CYP2D6) is not associated with multiple system
Ann Neurol 1998;44:415–22. atrophy. J Neurol Neurosurg Psychiatry 1995;59:557.
10 Wakabayashi K, Hayashi S, Kakita A, et al. Accumulation of 40 Cairns NJ, Atkinson PF, Kovacs T, et al. Apoprotein E e4
alpha-synuclein/NACP is a cytopathological feature com- allele frequency in patients with multiple system atrophy.
mon to Lewy body disease and multiple system atrophy. Neurosci Lett 1997;221:161–4.
Acta Neuropathol (Berl) 1998;96:445–52. 41 Bandmann O, Sweeney MG, Daniel SE, et al. Multiple-
11 Spillantini MG, Crowther RA, Jakes R, et al. Filamentous system atrophy is genetically distinct from identified inher-
alpha-synuclein inclusions link multiple system atrophy ited causes of spinocerebellar degeneration. Neurology
with Parkinson’s disease and dementia with Lewy bodies. 1997;49:1598–604.
Neurosci Lett 1998;251:205–8. 42 Ozawa T, Takano H, Onodera O, et al. No mutation in the
12 Galvin JE, Giasson B, Hurtig HI, et al. Neurodegeneration entire coding region of the alpha-synuclein gene in patho-
with brain iron accumulation, type 1 is characterized by logically confirmed cases of multiple system atrophy. Neu-
alpha-, beta-, and gamma-synuclein neuropathology. Am J rosci Lett 1999;270:110–12.
Pathol 2000;157:361–8. 43 Chen R, Forno LS, Di Monte DA, et al. Mutation screening
13 Wenning GK, Quinn NP. Parkinsonism. Multiple system in the á-synuclein gene in MSA. Parkinsonism and Related
atrophy. Baillieres Clin Neurol 1997;6:187–204. Disorders 1999;5(suppl):S28.
14 Quinn N. Multiple system atrophy. In: Marsden CD, Fahn 44 El-Agnaf OM, Curran MD, Wallace A, et al. Mutation
S, eds. Movement disorders 3. Oxford: Butterworth- screening in exons 3 and 4 of alpha-synuclein in sporadic
Heinemann; 1994:262–81. Parkinson’s and sporadic and familial dementia with Lewy
15 Wenning GK, Scherfler C, Granata R, et al. Time course of bodies cases. Neuroreport 1998;9:3925–7.
symptomatic orthostatic hypotension and urinary inconti- 45 Krüger R, Vieira-Saecker AMM, Kuhn W, et al. Increased
nence in patients with postmortem confirmed parkinsonian susceptibility to sporadic Parkinson’s disease by a certain
syndromes: a clinicopathological study. J Neurol Neurosurg combined a-synuclein/apolipoprotein E genotype. Ann
Psychiatry 1999;67:620–3. Neurol 1999;45:611–17.
16 Wenning GK, Tison F, Ben Shlomo Y, et al. Multiple system 46 Engelender S, Kaminsky Z, Guo X, et al. Synphilin-1 asso-
atrophy: a review of 203 pathologically proven cases. Mov ciates with a-synuclein and promotes the formation of
Disord 1997;12:133–47. cytosolic inclusions. Nat Genet 1999;22:110–14.
17 Maraganore DM, Anderson DW, Bower JH, et al. Potential 47 Ostrerova N, Petrucelli L, Farrer M, et al. á-Synuclein
selection bias in autopsy series of Parkinson’s disease and shares physical and functional homology with 14-3-3
related disorders. Neurology 1998;50(suppl 4):A98–9. proteins. J Neurosci 1999;19:5782–91.
18 Ben-Shlomo Y, Wenning GK, Tison F, et al. Survival of 48 Hashida H, Goto J, Zhao N, et al. Cloning and mapping of
patients with pathologically proven multiple system ZNF231, a novel brain-specific gene encoding neuronal
atrophy: a meta-analysis. Neurology 1997;48:384–93. double zinc finger protein whose expression is enhanced in
19 Dickson DW, Lin W, Liu WK, et al. Multiple system a neurodegenerative disorder, multiple system atrophy
atrophy: a sporadic synucleinopathy. Brain Pathol 1999;9: (MSA). Genomics 1998;54:50–8.
721–32. 49 Arima K, Murayama S, Mukoyama M, et al. Immunocyto-
20 Papp MI, Lantos PL. The distribution of oligodendroglial chemical and ultrastructural studies of neuronal and
inclusions in multiple system atrophy and its relevance to oligodendroglial cytoplasmic inclusions in multiple system
clinical symptomatology. Brain 1994;117:235–43. atrophy. 1. Neuronal cytoplasmic inclusions. Acta Neu-
21 Inoue M, Yagishita S, Ryo M, et al. The distribution and ropathol (Berl) 1992;83:453–60.
dynamic density of oligodendroglial cytoplasmic inclusions 50 Murayama S, Arima K, Nakazato Y, et al. Immunocyto-
(GCIs) in multiple system atrophy: a correlation between chemical and ultrastructural studies of neuronal and
the density of GCIs and the degree of involvement of stria- oligodendroglial cytoplasmic inclusions in multiple system
tonigral and olivopontocerebellar systems. Acta Neu- atrophy. 2. Oligodendroglial cytoplasmic inclusions. Acta
ropathol (Berl) 1997;93:585–91. Neuropathol (Berl) 1992;84:32–8.
22 Probst-Cousin S, Rickert CH, Schmid KW, et al. Cell death 51 Castellani R. Multiple system atrophy: clues from inclu-
mechanisms in multiple system atrophy. J Neuropathol Exp sions. Am J Pathol 1998;153:671–6.
Neurol 1998;57:814–21. 52 Chin SS, Goldman JE. Glial inclusions in CNS degenerative
23 Lantos PL, Papp MI. Cellular pathology of multiple system diseases. J Neuropathol Exp Neurol 1996;55:499–508.
atrophy: a review. J Neurol Neurosurg Psychiatry 1994;57: 53 Cairns NJ, Atkinson PF, Hanger DP, et al. Tau protein in the
129–33. glial cytoplasmic inclusions of multiple system atrophy can

www.molpath.com
426 Burn, Jaros

be distinguished from abnormal tau in Alzheimer’s disease. 73 Vouyiouklis DA, Brophy PJ. Microtubule-associated pro-
Neurosci Lett 1999;230:49–52. teins in developing oligodendrocytes: transient expression
54 Kobayashi K, Miyazu K, Katsukawa K, et al. Cytoskeletal of a MAP2c isoform in oligodendrocyte precursors. J Neu-
abnormalities in patients with olivopontocerebellar rosci Res 1995;42:803–17.
atrophy—an immunocytochemical and Gallyas impregna- 74 Nakamura S, Kawamoto Y, Nakano S, et al. Expression of
tion study. Neuropathol Appl Neurobiol 1992;18:237–49. the endocytosis regulatory proteins Rab5 and Rabaptin-5
55 Nakamura S, Kawamoto Y, Nakano S, et al. Cyclin- in glial cytoplasmic inclusions from brains with multiple
dependent kinase 5 and mitogen-activated protein kinase system atrophy. Clin Neuropathol 2000;19:51–6.
in glial cytoplasmic inclusions in multiple system atrophy. J 75 De la Monte SM, Sohn YK, Ganju N, et al. p53- and
Neuropathol Exp Neurol 1998;57:690–8. CD95-associated apoptosis in neurodegenerative diseases.
56 Gai WP, Power JH, Blumbergs PC, et al. Alpha-synuclein Lab Invest 1998;78:401–11.
immunoisolation of glial inclusions from multiple system 76 Wullner U, Weller M, Kornhuber J, et al. Altered expression
atrophy brain tissue reveals multiprotein components. J of calcium- and apoptosis-regulating proteins in multiple
Neurochem 1999;73:2093–100. system atrophy Purkinje cells. Mov Disord 2000;15:269–75.
57 Head MW, Goldman JE. Small heat shock proteins, the 77 Schulz JB, Weller M, Moskowitz MA. Caspases as treatment
cytoskeleton, and inclusion body formation. Neuropathol targets in stroke and neurodegenerative diseases. Ann Neu-
Appl Neurobiol 2000;26:304–12. rol 1999;45:421–9.
58 Alves-Rodrigues A, Gregori L, Figueiredo-Pereira ME. 78 Du C, Dodel RC, Bales KR, et al. Involvement of caspase-3-
Ubiquitin, cellular inclusions and their role in neurodegen- like cysteine protease in 1-methyl-4-phenylpyridinium-
eration. Trends Neurosci 1998;21:516–20. mediated apoptosis of cultured cerebellar granule neurons.
59 Jakes R, Spillantini MG, Goedert M. Identification of two J Neurochem 1997;69:1382–8.
distinct synucleins from human brain. FEBS Lett 1994; 79 Hartmann A, Hunot S, Faucheux Y, et al. Is caspase-3 a vul-
345:27–32. nerability factor and final eVector in apoptotic death of
60 Weinreb PH, Zhen W, Poon AW, et al. NACP, a protein dopaminergic neurons in Parkinson’s disease? Parkinsonism
implicated in Alzheimer’s disease and learning, is natively and Related Disorders 1999;5(suppl):S18.
unfolded. Biochemistry 1996;35:13709–15. 80 Schwarz SC, SeuVerlein T, Liptay S, et al. Microglial activa-
61 Hashimoto M, Hsu LJ, Sisk A, et al. Human recombinant tion in multiple system atrophy: a potential role for
NACP/á-synuclein is aggregated and fibrillated in vitro. NF-kappaB/rel proteins. Neuroreport 1998;9:3029–32.
Relevance for Lewy body disease. Brain Res 1998;799:301– 81 Gerhard A, Banati R, Cagnin A, et al. In vivo imaging of
6. activated microglia with [11C]PK11195 positron emission
62 Giasson BI, Uryu K, Trojanowski JQ, et al. Mutant and wild tomography in patients with multiple system atrophy. Mov
type human alpha-synucleins assemble into elongated fila- Disord 2000;15(suppl 3):215–16.
ments with distinct morphologies in vitro. J Biol Chem 82 Murer MG, Yan Q, Raisman-Vozari R. Brain-derived
1999;274:7619–22. neurotrophic factor in the control of human brain, and in
63 Hashimoto M, Hsu LJ, Xia Y, et al. Oxidative stress induces Alzheimer’s disease and Parkinson’s disease. Prog Neurobiol
amyloid-like aggregate formation of NACP/alpha- 2000;63:71–124.
synuclein in vitro. Neuroreport 1999;10:717–21. 83 Kawamoto Y, Nakamura S, Akiguchi I, et al. Increased
64 Ostrerova-Golts N, Petrucelle L, Hardy J, et al. The A53T brain-derived neurotrophic factor-containing axons in the
alpha-synuclein mutation increases iron-dependent aggre- basal ganglia of patients with multiple system atrophy. J
gation and toxicity. J Neurosci 2000;20:6048–54. Neuropathol Exp Neurol 1999;58:765–72.
65 Dexter DT, Jenner P, Schapira AH, et al. Alterations in lev-
els of iron, ferritin, and other trace metals in neurodegen- 84 Barres BA, Hart IK, Coles HS, et al. Cell death and control
erative diseases aVecting the basal ganglia. The Royal of cell survival in the oligodendrocyte lineage. Cell
Kings and Queens Parkinson’s Disease Research Group. 1992;70:31–46.
Ann Neurol 1992;32(suppl):S94–100. 85 Burden S, Yarden Y. Neuregulins and their receptors: a ver-
66 Connor JR. Iron acquisition and expression of iron satile signalling molecule in organogenesis and oncogen-
regulatory proteins in the developing brain: manipulation esis. Neuron 1997;18:847–55.
by ethanol exposure, iron deprivation and cellular dysfunc- 86 Cannoll PD, Kraemer R, Teng KK, et al. GGF/neuregulin
tion. Dev Neurosci 1994;16:233–47. induces a phenotypic reversion of oligodendrocytes. Mol
67 Duda JE, Giasson BI, Gur TL, et al. Immunohistochemical Cell Neurosci 1999;13:79–94.
and biochemical studies demonstrate a distinct profile of 87 Vartanian T, Corfas G, Li Y, et al. A role of the acetylcholine
alpha-synuclein permutations in multiple system atrophy. J receptor-inducing protein in oligodendrocyte develop-
Neuropathol Exp Neurol 2000;59:830–41. ment. Proc Natl Acad Sci U S A 1994;91:1126–30.
68 Kato S, Shinozawa T, Takikawa M, et al. Midkine, a new 88 Frost EE, Buttery PC, Milner R, et al. Integrins mediate
neurotrophic factor, is present in glial cytoplasmic neuronal survival signal for oligodendrocytes. Curr Biol
inclusions of multiple system atrophy brains. Acta Neu- 1999;9:1251–4.
ropathol (Berl) 2000;100:481–9. 89 Powell SK, Williams CC, Nomizu M, et al. Laminin-like
69 Zhang N, Deuel TF. Pleiotrophin and midkine, a family of proteins are diVerentially regulated during cerebellar
mitogenic and angiogenic heparin-binding growth and dif- development and stimulate granule cell neurite outgrowth
ferentiation factors. Curr Opin Hematol 1999;6:44–50. in vitro. J Neurosci Res 1998;54:233–47.
70 Wang S, Yoshida Y, Goto M, et al. Midkine exists in 90 Buttery PC, Vrench-Constant C. Laminin-2/integrin inter-
astrocytes in the early stage of cerebral infarction. Dev actions enhance myelin membrane formation by oli-
Brain Res 1998;106:205–9. godendrocytes. Mol Cell Neurosci 1999;14:199–212.
71 Richter-Landsberg C, Gorath M, Trojanowski JQ, et al. 91 Hagg T, Portera-Cailliau C, Jucker M, et al. Laminins of the
Alpha-synuclein is developmentally expressed in cultured adult mammalian CNS; laminin-á2 (Merosin M-) chain
rat brain oligodendrocytes. J Neurosci Res 2000;62:9–14. immunoreactivity is associated with neuronal processes.
72 Muller R, Heinrich M, Heck S, et al. Expression of Brain Res 1997;764:17–27.
microtubule-associated proteins MAP2 and tau in cultured 92 Tian M, Hagg T, Denisova N, et al. Laminin-á2 chain-like
rat brain oligodendrocytes. Cell Tissue Res 1997;288:239– antigens in CNS dendritic spines. Brain Res 1997;764:28–
49. 38.

www.molpath.com