Corpora Amylacea from Multiple Sclerosis Brain Tissue Consists of Aggregated Neuronal Cells

In this report, we describe proteomic analysis of corpora amylacea collected by postmortem laser microdissection from multiple sclerosis (MS) brain lesions. Using low level protein loads (about 30 µg), a combination of two-dimensional electrophoresis with matrix-assisted laser desorption/ ionization-time of flight mass spectrometry and database interrogations we identified 24 proteins of suspected neuronal origin. In addition to major cytoskeletal proteins like actin, tubulin, and vimentin, we identified a variety of proteins implicated specifically in cellular motility and plasticity (F-actin capping protein), regulation of apoptosis and senescence (tumor rejection antigen-1, heat shock proteins, valosin-containing protein, and ubiquitin-activating enzyme E1), and enzy-matic pathways (glyceraldehyde-3-dehydrogenase, protein disulfide isomerase, protein disulfide isomerase related protein 5, lactate dehydrogenase). Samples taken from regions in the vicinity of corpora amylacea showed only traces of cellular proteins suggesting that these bodies may represent remnants of neuronal aggregates with highly polymerized cytoskeletal material. Our data provide evidence supporting the concept that biogenesis of corpora amylacea involves degenera-tion and aggregation of cells of neuronal origin.

Interestingly, recent data show that corpora amylacea have an affinity to nucleic acids as shown by in situ hybridization.The binding affinity varied with the nucleotide sequence, the most intense signal being produced by adenosine-2A receptor antisense probe.Since proteinase K pretreatment caused a complete loss of staining, the binding moiety appeares to be partially proteinaceous (Balea et al., 2006).
Although there is no link between corpora amylacea and multiple sclerosis (MS), corpora amylacea have been frequently described in MS brain tissue (Raine, 1994), however, their occurrence was not attributed any pathologic significance.In a recent analysis of multiple sclerosis brain tissue material we detected many old lesions containing abundant of corpora amylacea, while the adjacent tissue was entirely unaffected.These observations suggested that the generation of corpora amylacea was probably a secondary phenomenon in the pathology of MS.In an attempt to analyze the mechanism underlying the generation of corpora amylacea, we assessed their protein composition.For this purpose, corpora amylacea were laser microdissected from MS lesions, solubilized and their major protein components identified according to a generally accepted proteomic protocol.This technique identified a number of cellular proteins that indicated that corpora amylacea are derived from degenerated and aggregated cells of neuronal origin.

Immunohistochemistry.
Frozen sections were stained with the avidin-biotin-peroxidase complex (ABC) technique.Tissue sections were air dried, fixed in acetone, quenched with 0.03% hydrogen peroxide and blocked with normal serum.They were then incubated with primary antibodies followed by biotinylated anti-mouse IgG and the ABC reagent.The same lesions (serial sections) were stained with antibodies detecting demyelinated axons, SMI 32, interacting with abnormally phosphorylated neurofilament H, as well as normal axons, SMI 31, interacting with phosphorylated neurofilament H (both from Sternberger Monoclonals).Serial sections were also stained with anti-galactocerebroside (GC) and antimyelin basic protein (MBP) antibodies to detect oli-godendrocytes, and anti-glial fibrillary acid protein (GFAP) antibodies, to detect astrocytes.
Collection of corpora amylacea.Early autopsy freshly frozen brain tissue from cases of multiple sclerosis was used for this study.Tissue blocks containing chronic active lesions with corpora amylacea as defined by H&E staining were cut into semithin sections.Adjacent unstained sections were also used for laser-capture microdissection by Leica Microsystems (Wetzlar GmbH, Wetzlar, Germany).Corpora amylacea were visible as dense bodies scattered throughout the lesions, often in the gliotic margins of blood vessels.One thousand and seven hundred corpora amylacea were dissected from several lesions and catapulted into collecting tubes.The same volume of tissue from the lesions left after corpora amylacea dissection was used as a control.In addition, we dissected and catapulted fragments of nonaffected tissue from regions beyond the lesions.
Sample preparation.Lysates of corpora amylacea were prepared using 150 µl of lysis buffer (9 M urea, 4% Chaps, 1% DTT, 2% Pharmalyte, and Complete Inhibitors (Boeringer)).A buffer volume approximately equal to the packed tissue volume was used.To improve resolution and recovery of proteins in 2D electrophoresis, corpora amylacea and tissue extracts were treated with PlusOne 2-D Clean-Up Kit (Amersham-Biosciences).Protein concentration in the supernatant fraction was determined by PlusOne 2-D Quant Kit (Amersham Biosciences), using bovine serum albumin as a standard and tissue lysates were stored at -80 o C until analyzed.
Two-dimensional polyacrylamide gel electrophoresis.Proteins were separated by two-dimensional electrophoresis, using ready-made gels with immobilized pH gradients (Amersham-Biosciences).For the first dimension, samples containing approx.30 µg of soluble protein in lysis buffer were mixed with IPG Reswelling Solution (8 M urea, 1% Chaps, 0.4% DTT, 0.5% Pharmalyte) to obtain a final volume of 450 µl.They were then loaded onto 24 cm immobilized pH linear gradient strip gels (IPG, pH 3-10).IEF strips were allowed to rehydrate for 5 h, and isoelectric focusing was performed according to the manufacturer's protocol by gradual increase of voltage (30 V for 5 h, 500 V for 1 h, 1 000 V for 1 h, followed by 70 kVh at 8 000 V), using an IPGphor system (Amersham-Biosciences).SDS electrophoresis was performed on 12.5% polyacrylamide gels using the Ettan Dalt vertical system.Protein spots were visualized by silver-staining according to the method compatible with analysis of proteins by mass spectrometry (Shevchenko et al., 1996).Silver-stained gels were digitized using a laser Image Scanner (Amersham-Biosciences).Computerized 2D gel analysis was performed with the Image Master 2D software package Version 3.0.Protein spots were excised from the gel and subjected to in-gel digestion with trypsin.
Mass spectrometry, data base search and data processing.Proteins in each gel slice were subjected to reduction with 10 mM dithiothreitol, alkylation with 55 mM iodoacetamide, and tryptic digestion with 12.5 µg/ml modified trypsin (Roche Molecular Biochemicals) at 37 o C for 14 h.After in-gel digestion, the product peptides were extracted with 5% formic acid and acetonitrile, vacuum-dried, and dissolved in 5% formic acid.Proteins were identified by peptide mass fingerprinting using MALDI-TOF mass spectrometry and by partial peptide sequencing with ion trap electrospray mass spectrometry.Multiple peptides from each protein were detected, adding confidence to the protein identification.Trypsin autodigestion peaks were excluded from the database searching.Multiple digested peptides were obtained from a single run of each gel slice sample by microcapillary C18 reverse phase chromatography (200 µm × 5 cm capillary; Michrom BioResources, Inc.) and directly applied to the spectrometer.The ion trap was programmed to carry out two successive scans consisting of the first full scan MS over the range 300-2000 m/z and the second data-dependent scan of the most abundant ion in the first scan.Automatic MS/MS spectra were obtained from the highest peak in each full scan by setting a relative collision energy of 30% and an exclusion time of 5 min for molecules of the same m/z value range.We searched the NCBI nonredundant protein data base with the Mascot program (Matrix Science) for high fits with the ion spectrum data generated by LC/MS/ MS.The raw results identified the top 20 candidates for each spectrum.We examined their automatic ordering manually in terms of their reliability scores and MS spectrum profiles to pick up only highly reliable peptide data (sorted data).

Corpora amylacea presence in MS lesions
In this report we attempted to assess the protein composition of corpora amylacea to test the concept that corpora amylacea may arise from injured axons or glial cells.For this purpose, we used samples from chronic MS lesions in which numerous corpora amylacea were present (Fig. 1A).Such profiles are frequently encountered in MS pathology.Corpora amylacea were spread throughout the lesion area with an average density of 6 per mm 2 .Sometimes they were more numerous in close vicinity of blood vessels.Away from lesions, corpora amylacea were absent.The morphology of corpora amylacea was characterized as dense, rounded bodies sometimes with a ring-like structure (starchgrain).Corpora amylacea were removed from the tissue by laser microdissection.In parallel, samples of control white matter (Fig. 1B), normal and degenerated, were taken in a similar fashion and analyzed.Samples of degenerated white matter were collected from areas in close vicinity of corpora amylacea.

2D gel electrophoresis of corpora amylacea lysates
In order to identify the major protein components, the same volumes of isolated corpora amylacea and control white matter were sonicated in lysis buffer (Amersham), the extracted proteins further purified by Clean-Up Kit (Amersham), and separated by two-dimensional gel electrophoresis (Fig. 2A,  B, C).After staining with silver, 34 proteins of interest were excised from the gels, digested with trypsin, and their identities established by matrix-assisted laser desorption/ionization mass spectrometry.The resulting spectra were used to identify the proteins using the MS-FIT search program.Of the 34 spots excised, 24 contained a single protein while 10 spots consisted of more than one protein but they could be identified by the presence of two or more tryptic peptide fragments.Since some proteins were present in several different spots (n = 34) showing similar or different molecular masses, in fact, there were only 24 different proteins.

MALDI analysis
The major proteins identified were members of the cytoskeletal system (Fig. 2C, D).Slices taken from different areas of major spots A and B showed the presence of α and β subunits of tubulin, particularly their α2, β1 and β5 variants.Interestingly, despite its much higher molecular mass, vimentin comigrated with both tubulin subunits.Both major groups of spots, A and B, showed a wide pI in the range of 4.5-5.0 and 4.3-4.9,respectively.Similarly the C group of spots, containing β actin and its β' mutant, had pI which ranged from 4.6 to 5.3.Such wide pI ranges indicate that these proteins exist in several forms differing due to posttranslational modifications.Table 1 shows that in addition to cytoskeletal proteins, corpora amylacea also contained proteins specific for different cellular locations, including: (a) the cell membrane (LDL receptor-related protein 1, LRP-1), (b) the endoplasmatic reticulum (thiol oxidoreductase ER60, heat shock protein gp96), (c) mitochondria (ATP synthase), (d) cytosol (valosin-containing protein p97, alanyl-t-RNA synthetase, protein disulfide isomerase, 60S acidic ribosomal protein), and (e) the nucleus (RuvB-like 2 protein).Obviously, due to the limited number of lasermicrodissected corpora amylacea, low amounts of extracted proteins were loaded.Therefore, only the most abundant intracellular proteins could be detected after separation by electrophoresis.Comparison of proteins found in corpora amylacea with those of white matter extracts showed strong and significant similarities (not shown).However, there were some proteins missing from corpora bodies suggest- Corpora amylacea in multiple sclerosis brain tissue ing that only those directly associated with the actin, cytoskeleton or tubular systems were protected from proteolysis in the neuronal cell aggregates.

Occurrence of corpora amylacea correlates with lack of axonal staining
Using antibodies interacting with nonphosphorylated and phosphorylated neurofilament H, we immuno-stained the lesions with and without corpora amylacea for the presence of demyelinat-ing and non-demyelinating axons and found that in the lesions in which large numbers of corpora were present axonal staining was minimal (Fig. 3A).On the other hand, lesions with preserved axons showed no corpora amylacea (Fig. 3B).

DISCUSSION
In this report we provide for the first time a direct evidence that corpora amylacea represent The corpora amylacea were solubilized and separated by 2D electrophoresis.Slices from CBB-stained gels illustrated in Fig. 2 were subjected to sequencing.Proteins were identified by peptide mass fingerprinting using MALDI-TOF MS and by partial peptide sequencing with ion trap electrospray MS. *Indicates the presence of more than one protein in the analyzed gel slice.remnants of degenerated and aggregated neuronal cells.This concept is supported by proteomic analysis of precisely collected corpora amylacea that were laser microdissected from multiple sclerosis brains.Sequencing of the major protein spots revealed that they were mostly composed of cytoskelatal proteins that constituted the three major filament systems in neurons, i.e. microtubules, intermediate filaments, and actin microfilaments, forming interwoven scaffolding throughout the cytoplasmic space of the cell.

Gel
Microtubules support the assembly of cytoskeletal structures of diverse morphology and function.The basic building blocks of all microtubules are heterodimers composed of α-and β-tubulins encoded by multiple gene families.The bulk of cellular tubulin is cytoplasmic, but a significant fraction is embedded in, or firmly associated with, the plasma membrane and other membranes (Kelly et al., 1983).
Microtubules are often seen in contact with mitochondria (Smith et al., 1977;Heggeness et al., 1978;Bernier-Valentin et al., 1983), synaptic vesicles (Gray, 1975), endoplasmic reticulum (Vogl, 1996), lysosomes (Mithieux, 1988), cilia (Dentler, 1981), and plasma membranes (Hardham & Gunning, 1978).These features are highly ordered and have a uniformly po-larized organization providing great stability.Therefore, in such a state, cytoskeletal components and cellular proteins associated with the cytoskeleton are resistant to proteolytic cleavage, which may explain why several of the proteins found in corpora amylacea were not degraded.Our results are consistent with an earlier concept suggesting that the function of corpora amylacea could be to prevent the recognition of immunogenic proteins (such as classical pathway-specific components, the activation products C3d and the terminal complement complex, the C3 convertase regulator membrane cofactor protein (MCP) and the fluid phase regulators S-protein and clusterin) by lymphocytes and microglia and thus to protect the CNS from further injury (Singhrao et al., 1995).
There was a surprisingly low level of protein within the degenerated white matter surrounding corpora amylacea indicating perhaps that most cellular proteins had accumulated within corpora amylacea leaving only some outside.Furthermore, due to different amounts of α and β tubulin in the degenerated white matter, one may conclude that there was extensive proteolytic degradation around corpora amylacea.
Since in multiple sclerosis tissue both oligodendrocytes and axons are lost (Lassmann et al., 1998), the striking result of the present study is a lack in corpora amylacea of myelin/oligodendrocytespecific proteins such as myelin basic protein, proteolipid protein, and myelin/oligodendrocyte glycoprotein.This may indicate either that oligodendrocyte proteins were not incorporated into corpora amylacea, or that oligodendrocyte proteins are highly sensitive to early proteolysis.The latter explanation seems plausible since in the tissue surrounding corpora amylacea, myelin/oligodendrocyte proteins were also not detected.The prominent presence of microtubule-related proteins might rather suggest that dying axons provided a major component of corpora amylacea.Loss of axons, although known for many years, was only recently reemphasized in studies on multiple sclerosis (Trapp et al., 1998).If corpora amylacea were to be the tombstones of damaged axons, an important question arises, why do they only accumulate in some MS lesions?This might be related to the degree of axonal loss and/or the mechanisms leading to the axonal injury.Interestingly, when lesions containing different amounts of corpora amylacea were stained for the presence of demyelinated and non-demyelinated axons, there was decreased axonal staining with increased prevalence of corpora amylacea.This observation might support the relationship between corpora amylacea and degenerated axons.The mechanism of axonal transection in MS is not known but several immune pathways could lead to axonal injury.For example, Corpora amylacea in multiple sclerosis brain tissue cytotoxic T cells have been identified in close vicinity to transected axons (Giuliani et al., 2003) but other immune mechanisms and glutamate excitotoxicity have also been implicated in axonal loss in multiple sclerosis (Neumann, 2003).Future studies might answer whether corpora amylacea occur in association with any particular pattern of axonal damage.

Figure 2 .
Figure 2. Typical silver-stained 2D gel electrophoregram of corpora amylacea laser microdissected from multiple sclerosis lesions.Proteins present in extracts of the same volume of unaffected tissue (panel A), tissue from the lesions left after corpora amylacea dissection (panel B) and 1700 corpora amylacea bodies (panel C) obtained by laser-capture microdissection, were separated by 2D electrophoresis, stained with silver and identified by sequencing according to a proteomic protocol.Panel D shows a magnified region of the gel with the highest number of cytoskeletal proteins.