Cortical demyelination can be extensive in chronic multiple sclerosis (MS) patients. Cortical lesions are not associated with lymphocyte infiltration, blood-brain barrier disruption, or complement deposition; therefore, their pathogenesis is unclear. We analyzed the extent and cellular composition of leptomeningeal inflammatory infiltrates and their relationship with cortical demyelinated lesions in brain autopsy samples from 28 chronic MS patients; samples from 6 nonneurological disease control patients were alsostudied. Immunohistochemistry was used to detect meningeal T cells, B cells, macrophages, mature and immature dendritic cells, T-helper cells, (activated) cytotoxic T cells, and plasma cells. Quantitative analysis revealed significant meningeal inflammation in chronic MS patients; T cells were the predominant inflammatory cells. Morphometric analysis was performed on coronal hemisphere sections of the MS cases to assess subpial demyelination; no correlation between the extent of subpial demyelination and extent of meningeal inflammation was identified. Moreover, no differences were observed in the degree or cellular composition of meningeal infiltrates in areas directly adjacent to subpial lesions compared with areas adjacent to normal-appearing gray matter in the MS cases. In addition, no follicle-like structures were found in the MS samples. Our data suggest that the occurrence of cortical lesions is not related to the presence of meningeal inflammation in a large number of chronic MS patients.
Multiple sclerosis pathology
The distinct neuropathologic hallmarks of multiple sclerosis (MS) include multifocal areas of demyelination and axonal loss throughout the brain and spinal cord (1, 2). Although MS has long been regarded as predominantly a white matter (WM) disease, gray matter (GM) involvement has also been extensively described and has recently received increased attention (3-17).
Based on their topological distribution, Bø et al (4) designated 4 distinct types of cortical MS lesions; Lesion types III and IV are referred to as subpial lesions. The type III cortical lesion is the most common cortical lesion type and often affects multiple gyri (4). Studies on its pathogenesis suggest that its underlying disease mechanisms differ between GM and WM lesions (3, 5, 6, 8, 12, 18-20). For example, in contrast to WM lesions, cortical GM lesions are not associated with extensive lymphocyte infiltration, they do not show evidence of blood-brain barrier disruption or astrogliosis, and they are usually not associated with complement deposition (3, 18, 20).
Based on a large postmortem study, Kutzelnigg et al (12) suggested that focal active WM lesions are mainly characteristic of acute or relapsing-remitting phases of MS and that cortical demyelination and diffuse injury of the so-called normal-appearing white matter (NAWM) are more prominent in progressive MS. They additionally found that cortical demyelination and diffuse inflammation throughout the NAWM in these progressive cases coincided with perivascular, parenchymal, and meningeal inflammation (12).
Several other studies have also demonstrated meningeal inflammation in MS patients (21-23). Recently, ectopic B-cell follicles with germinal center formation in the meninges of a significant proportion of secondary progressive MS patients were demonstrated (19, 24). The presence of these meningeal follicles correlated with early disease onset, short disease duration, high expanded disability status scale scores, and extensive cortical subpial demyelination (19). These observations suggest that leptomeningeal inflammation may play a pathogenetic role in the development of subpial demyelination, perhaps by the elaboration of soluble (i.e. myelinotoxic) substances that diffuse into the cortex and mediate damage (19).
Thus, there is considerable recent interest in a possible relationship between meningeal inflammation and the development of cortical lesions in chronic MS (25, 26), but a systematic investigation of the type and composition of meningeal inflammation and their relationship to subpial cortical demyelination has not been undertaken. Our aim was to characterize meningeal inflammation in a large and unselected sample of chronic MS patient autopsy samples and to investigate possible regional and global correlations between meningeal inflammation and subpial demyelination.
Materials and Methods
Human Postmortem Brain Tissue
Brain tissue samples were obtained from the Netherlands Brain Bank (Amsterdam, The Netherlands). A total of 93 paraffin-embedded tissue blocks from 28 MS patients and 14 blocks from 6 donors without neurological disease were selected based on the presence of leptomeninges in the sections. In addition, for a subset of 21 MS patients, 47 coronal full hemispheric sections were selected for morphometric analysis of the extent of subpial demyelination. For 7 MS patients, no coronal slices were available, and 21 standard-size tissue blocks were used for assessment of subpial demyelination. Clinical data of the MS and nonneurological disease control patients are summarized in Table 1.
The age of MS patients at the time of death ranged from 41 to 84 years (mean ± SD, 63 ± 13 years) with a mean postmortem interval of 8 hours 42 minutes (SD, ±3 hours 39 minutes). The ages of the controls ranged from 45 to 99 years (mean ± SD, 73 ± 21 years) with a mean postmortem interval of 29 hours 0 minutes (SD, ±16 hours 17 minutes). The study was approved by the local institutional ethics review board, and all donors or their next of kin provided written informed consent for brain autopsy and use of material and clinical information for research purposes.
Paraffin sections (5 μm thick) were collected on Superfrost Plus glass slides (VWR International, Leuven, Belgium) and dried overnight at 37°C. Sections were deparaffinized in a series of xylene (3 × 5 minutes), 100% ethanol, 96% ethanol, 70% ethanol, and water. Endogenous peroxidase activity was blocked by incubating the sections in methanol with 0.3% hydrogen peroxide. Sources and antigen retrieval protocols used for the antibodies are listed in Table 2.
Tissue sections were stained with anti-myelin proteolipid protein (PLP) or anti-myelin basic protein (MBP) for the detection of demyelination. Serial sections of these blocks were also stained for T cells (CD3), B cells (CD20), macrophages (CD68), mature and immature dendritic cells (dendritic cell-specific ICAM3-grabbing nonintegrin [DC-SIGN]) and activated cytotoxic T cells (granzyme B). All primary antibodies were diluted in 0.01 mol/L phosphate buffered saline, pH 7.4 (PBS) containing 1% bovine serum albumin (Roche Diagnostics, Mannheim, DE). For primary staining with anti-PLP, anti-MBP, anti-CD3, anti-CD20, anti-CD68, and anti-granzyme B, the sections were rinsed in PBS (3 × 10 minutes), incubated with EnVision horseradish peroxidase complex (DAKO, Glostrup, Denmark); 3,3′diaminobenzidine-tetrahydrochloride dihydrate (DAKO) was used as a chromogen. For staining with anti-DC-SIGN, the sections were rinsed in PBS (3 × 10 minutes) and incubated with biotin-labeled swine anti-rabbit immunoglobulins F(ab′)2 (1:500) (DAKO), diluted in 10% normal swine serum (DAKO), 10% normal human serum, and PBS with 1% bovine serum albumin for 30 minutes. The sections were then rinsed for 3 × 10 minutes and incubated for 1 hour at room temperature with streptavidin-biotin-peroxidase complexes (StreptABComplex; DAKO). Sections were rinsed for 3 × 10 minutes and incubated with 3,3′diaminobenzidine-tetrahydrochloride dihydrate. After a short rinse in tap water, all sections were counterstained with hematoxylin for 1 minute and intensely washed with tap water for 5 minutes. Tonsil tissue was used as a positive control for the leukocyte markers; negative controls included omission of either primary or secondary antibodies. Sections of a subset of the material were also stained for T-helper cells (CD4), cytotoxic T cells (CD8), and plasma cells (CD138). Immunostaining for CD4, CD8, and CD138 was performed using the Bond Max (Vision Biosystems, Mount Waverley, Victoria, Australia).
Morphometry and Quantification
The coronal hemisphere and standard sections immunostained with anti-MBP or anti-PLP antibodies were scanned using an Agfa Duoscan T2000XL scanner for preparation of digital images and prints. The cortical demyelinated areas in the plane of the section were analyzed using light microscopy, and the same orientations were applied for the tissue on prints. The areas of subpial demyelination were measured on the digital images using ImageJ software (freely downloadable from U.S. National Institutes of Health, Bethesda, MD; http://rsb.info.nih.gov/ij/index.html).
In MS and control tissues, random areas in the meninges were chosen and photographed at 50× original magnification. Cells were counted in corresponding areas on serial sections that were stained for the leukocyte markers. Cell counts were done in a total of 50 meningeal areas in the control tissue samples. For MS tissue samples, meningeal areas that were adjacent to subpial lesions were counted separately from those that were adjacent to normal-appearing GM (NAGM). Cell counts were made in a total of 432 meningeal areas, of which 297 areas were adjacent to NAGM and 135 areas were adjacent to subpial lessions. On a subset of the meningeal areas investigated, semiquantitative analysis was performed in order to discriminate between the amount of T helper cells and cytotoxic T cells in the MS meninges on the basis of anti-CD4 and anti-CD8 immunostaining. The presence of CD4+ cells and CD8+ cells was determined as follows: 0, no immunopositive cells; +, 1 to 10 immunopositive cells; ++, 10 to 20 immunopositive cells; +++, greater than 20 immunopositive cells. The meningeal lengths of each area investigated were measured using ImageJ software. Calibration of the software program was performed using a standard millimeter scale.
Data were analyzed using GraphPad Prism software (GraphPad Software, San Diego, CA); data distribution was tested for normality. Because all variables were not normally distributed, the nonparametric Mann Whitney U test was used for comparing MS meningeal leukocyte infiltrates with those in the controls. Within the MS group, meningeal infiltrates adjacent to subpial lesions and adjacent to NAGM were also compared using Mann Whitney U statistics. Possible correlations between the degree of meningeal inflammation and the extent of subpial demyelination were investigated with the nonparametric Spearman rank correlation. Differences were considered significant at p < 0.05.
Extent and Composition of Meningeal Inflammatory Infiltrates
Meninges from chronic MS patients contained significantly more leukocyte infiltrates than meninges of controls (p < 0.05; Table 3). Most of the meningeal infiltrates in the MS patients consisted of CD3+ T cells, CD68+ macrophages, and DC-SIGN+ dendritic cells Table 3. By contrast, CD20+ B cells and granzyme B+ activated cytotoxic T cells were far less common (Table 3). CD138+ plasma cells were only occasionally observed in the meninges of the chronic MS patients (Fig. 1H). Semiquantitative analysis revealed that most T cells in the meninges of MS patients were CD8+ cytotoxic T cells and to a lesser extent CD4+ T helper cells (Figs. 1I, J).
Meningeal inflammation does not correlate with cortical demyelination in chronic multiple sclerosis (MS) brains. (A) An anti-myelin proteolipid protein-immunostained hemisphere section shows extensive subpial demyelination (outlined in green line) in a representative chronic MS case (Case 12). Morphometric analysis indicated that 43% of the cerebral cortex (CTX) in this plane of section was demyelinated. Only subpial lesions (Types III and IV) lesions were measured. (B) Lack of correlation between meningeal inflammation (mean total cells per millimeter) and extent of subpial demyelination (percentage demyelinated CTX) in chronic MS patients (Spearman ρ = −0.01; p = 0.73). (C-G) Despite extensive subpial demyelination in Case 12, there is little meningeal inflammation as indicated by anti-CD3 (C), anti-CD20 (D), anti-CD68, (E), anti-granzyme B (F), and anti-dendritic cell-specific ICAM3-grabbing nonintegrin (G) immunostaining. (H) There are occasional CD138-immunopositive plasma cells in the meninges (M). (I, J) A minority of T cells consists of CD4-immunopositive T-helper cells (I); the majority consists of CD8-immunopositive cytotoxic T cells (J) (arrows [I, J]). Original magnification: (C-J) 200×; (insets C-E, G, H) 400×; inset (H) 1,000×.
Relationship Between Meningeal Infiltrates and Cortical Demyelination: Global Associations
To determine whether meningeal inflammation was globally associated with subpial demyelination, the percentages of types III and IV cortical lesions of each MS patient (Fig. 1A) were compared with the extent of meningeal inflammation. The percentage areas of subpial demyelination did not correlate with the extent of meningeal inflammation (mean total cells per millimeter) (Fig. 1B; p = 0.73). Not every chronic MS patient had leukocyte infiltrates in their meninges; therefore, the results were heterogeneous. Hence, extensive subpial demyelination (Figs. 1A) could be observed in the absence of meningeal infiltrates (Figs. 1C-G). There was also no correlation between the extent of meningeal inflammation and disease duration (p = 0.54) (data not shown); finally, no differences in the extent of meningeal inflammation were found between primary and secondary progressive MS cases (p = 0.60).
Relationship Between Meningeal Infiltrates and Cortical Demyelination: Regional Associations
Because meningeal inflammation did not correlate with overall subpial demyelination areas, further comparison was made between leptomeningeal inflammatory cell counts adjacent to subpial lesions and counts in areas adjacent to NAGM. No differences were identified between infiltrates of T cells, macrophages, activated cytotoxic T cells, or dendritic cells in the meninges adjacent to subpial lesions (Fig. 2A) and those adjacent to NAGM (Fig. 2B) (Table 4; Figs. 2C-L). Moreover, no follicle-like structures were observed in these chronic MS tissue samples.
Focal meningeal infiltrates are not associated with cortical demyelination in chronic multiple sclerosis (MS) brain tissue. (A, B) Immunohistochemical staining with anti-myelin proteolipid protein in a representative case. Subpial gray matterdemyelinated lesion (GML) (arrowheads) (A) normal-appearing gray matter (NAGM) (B). (C-L) Immunostaining on MS brain tissue containing meninges (M) directly adjacent to a GML (C, E, G, I, K) revealed no differences in the extent of meningeal leukocyte infiltration from that in areas directly adjacent to NAGM (D, F, H, J, L). (C, D) anti-CD3, Case 1; (E, F) anti-CD20, Case 27; (G) anti-CD68, Case 27; (H) anti-CD68, Case 2; (I) anti-granzyme B, Case 25; (J) anti-granzyme B, Case 24; (K, L) anti-dendritic cell-specific ICAM3-grabbing nonintegrin, Case 20. Original magnification: (A, B) 50×; (C-H, K, L) 100×; (I, J) 200×; (insets C-H, K, L) 400×; (insets I, J) 1,000×.
We demonstrate that meningeal inflammation and subpial demyelination can be extensive in a large unselected subset of chronic MS patients. We also show that leukocyte infiltrates in these cases are predominantly composed of CD8+ T cells, macrophages, and dendritic cells. Additionally, in agreement with previous findings (4, 12), morphometric analysis demonstrated extensive subpial demyelination in a substantial proportion of chronic MS patients. We also provided a detailed characterization of meningeal inflammation in a large, unselected cohort of chronic MS autopsy material and showed that meningeal leukocyte infiltrates are predominantly composed of (CD8+) T cells, macrophages and dendritic cells. Moreover, we found no correlation between the extent of meningeal inflammation and the extent of overall subpial demyelination; regional analysis demonstrated that subpial demyelination was also not associated with adjacent meningeal inflammation. MS cortical lesions were previously found not to be associated with lymphocyte infiltration (3, 14). Our study extends these findings in that subpial demyelination was also not associated with T cell infiltration in the MS meninges.
It has recently been shown that T cells accumulate early in the disease course in mice with experimental autoimmune encephalomyelitis in the meninges (27); by analogy, the T cell mediated meningeal inflammation observed in MS might reflect overall white matter disease activity. This could explain T cell infiltrates in the meninges of the MS patients, but further analysis should confirm this. Although the numbers of B cells in meningeal areas directly adjacent to subpial lesions were slightly increased compared with areas directly adjacent to NAGM, B cells in the MS meninges were relatively sparse overall. Therefore, their direct pathogenetic roles are unclear.
The formation of leptomeningeal ectopic B-cell follicles has been associated with severe cortical pathology and more aggressive clinical disease in a subset of secondary progressive MS patients (19, 28). In these studies, it was further shown that type III cortical lesions extending over large cortical areas were predominantly found in SP-MS patients exhibiting ectopic B-cell follicles (19), whereas secondary progressive MS patients lacking ectopic B-cell follicles showed a more heterogeneous cortical pathology (19). In relation to this, it was suggested that myelinotoxic substances diffusing from the subarachnoid and perivascular spaces might cause subpial demyelination. Secondary progressive MS patients harbouring ectopic B-cell follicles in the meninges had several clinico-pathological characteristics, including a shorter disease duration, earlier onset of the disease and more extensive subpial demyelination (19). In the present study, however, we found extensive subpial demyelination in the absence of substantial B-cell infiltration or ectopic B-cell follicles. It is possible that subpial cortical lesions might have been induced by antecedent inflammatory infiltrates earlier in the disease course, but the lack of correlation between the degree of meningeal inflammation and disease duration argues against this possibility. We cannot, however, rule out the possibility that differences between the previous and the current findings might be caused by differences in disease onset or other parameters in the MS cases studied.
Interest in B cells in MS has received impetus following new treatment options of relapsing-remitting MS patients. Rituximab, an anti-CD20 monoclonal antibody, was shown to deplete CD20+ B cells that correlated with a significant and sustained reduction of gadolinium-enhancing lesions and of relapses in relapsing-remitting MS patients (29). Because plasma cells do not express CD20, treatment of relapsing-remitting MS patients with rituximab did not result in the reduction of total antibody levels (29); antimyelin antibodies might not relate to MS disease progression (30). It is possible, however, that a negative effect on the antigen-presenting capacities of B cells (31-33) might have played a role in effects of rituximab on white matter lesion development (34). Whether rituximab treatment exerts an effect on the progression of relapsing-remitting MS patients and on the development of subpial cortical GM lesions remains elusive and warrants further investigation.
In mice with experimental autoimmune encephalomyelitis, T-cell recognition of their cognate antigen presented by dendritic cells and macrophages in the meninges and perivascular spaces is sufficient for inducing inflammation in the brain and resulting neurological dysfunction (27, 35). We recently reported the presence of large amounts of extracellular myelin in the meninges of MS patients and proposed that this might lead to the induction or continuation of neuroinflammation (36). Alternatively, the presence of brain-derived antigens inside and outside the central nervous system could suppress immune responses leading to tolerogenic effects or even neuroprotection (37-45).
Because immune cells produce neurotrophic factors (46-48), meningeal inflammation may also exert beneficial effects on the underlying cortex. Indeed, extensive cortical remyelination has also recently been described in the cerebral cortex of chronic MS patients (49). Because in the present study no discrimination was made between NAGM and remyelinated cortex, the areas designated as NAGM might have included remyelinated areas. Since no electron microscopy was performed, we felt unable to reproducibly assess remyelination in the cortex on the basis of PLP immunostainings. Future studies are warranted to investigate whether the extent of cortical remyelination is correlated with meningeal inflammation in chronic MS.
Although GM pathology is present in the earliest stages of MS (50-54), it only becomes prominent in the progressive phase of the disease (55-57). Interestingly, GM atrophy accelerates after secondary disease progression to a greater extent than WM atrophy (58). In addition to the elaboration of myelinotoxins from meningeal inflammatory infiltrates, several pathogenetic mechanisms have been proposed for GM damage; these include mitochondrial dysfunction (59, 60), the exhaustion of compensatory mechanisms for adaptive cortical reorganization (61, 62), remyelination (63), redistribution of sodium channels on demyelinated axons (64), and the expression of neurotrophic factors by immune and central nervous system resident cells (46, 65). Whether and to what extent each of these different pathogenetic mechanisms plays a role in the development of subpial lesion formation remains elusive, but the present findings indicate that at least in a subset of chronic MS patients, the development of subpial lesions is not related to the presence of meningeal inflammation.
The authors thank the Netherlands Brain Bank (director I. Huitinga, PhD) for providing the material for this study. Furthermore, the authors thank W. H. Gerritsen, E. M. Montagne, and J. Meijer for technical assistance.
Evert-Jan Kooi is supported by Grant No. 02-298MS and Grant No. 06-587MS from the Dutch MS Research Foundation. Jeroen Geurts and Jack van Horssen are supported by Grant No. 05-358c from the Dutch MS Research Foundation.
Online-only color figures are available at http://www.jneuropath.com.
. Activated human T cells, B cells, and monocytes produce brain-derived neurotrophic factor in vitro and in inflammatory brain lesions: A neuroprotective role of inflammation? J Exp Med 1999;189:865–70
. Gene expression analysis of normal appearing brain tissue in an animal model for multiple sclerosis revealed grey matter alterations, but only minor white matter changes. J Neuroimmunol 2008;205:10–19