Increased Meningeal T and Plasma Cell Infiltration is Associated with Early Subpial Cortical Demyelination in Common Marmosets with Experimental Autoimmune Encephalomyelitis

Abstract

Subpial cortical demyelination (SCD) accounts for the greatest proportion of demyelinated cortex in multiple sclerosis (MS). SCD is already found in biopsy cases with early MS and in marmosets with experimental autoimmune encephalomyelitis (EAE), but the pathogenesis of SCD is not well understood. The objective of this study was to investigate whether and, if so, which meningeal inflammatory cells were associated with early SCD in marmosets with EAE. Immunohistochemistry was performed to analyze brain samples from eight control animals and eight marmosets immunized with myelin oligodendrocyte glycoprotein. Meningeal T, B and plasma cells were quantified adjacent to SCD, normal‐appearing EAE cortex (NAC) and control marmoset cortex. SCD areas appeared mostly hypocellular with low‐grade microglial activation. In marmosets with EAE, meninges adjacent to SCD showed significantly increased T cells paralleled by elevated plasma cells, but unaltered B cell numbers compared with NAC. The elevation of meningeal T and plasma cells was a specific finding topographically associated with SCD, as the meninges overlying NAC displayed similarly low T, B and plasma cell numbers as control cortex. These findings suggest that local meningeal T and plasma cell infiltration contributes to the pathogenesis of SCD in marmosets with EAE.

Introduction

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system (CNS) affecting white and grey matter 23. Even though grey matter involvement is underestimated by conventional imaging techniques, neuropathologic studies show that cortical demyelination is common in MS and present in 90% of chronic MS cases 22, 44. Recent histopathologic analysis even demonstrated that different types of cortical lesions were already detectable in patients with early MS 26. Early cortical MS lesions exhibit infiltration by inflammatory cells, whereas most chronic lesions appear hypocellular 3, 4, 35. However, inflammatory cells including macrophages were also detected in progressive MS in active cortical lesions within the parenchyma and as perivascular accumulations 11, 29. Recent studies using advanced imaging techniques have provided evidence that cortical lesions are associated with epilepsy 8 and cognitive deficits 7, 38 in MS.

Pathologic investigations identified three different types of cortical lesions in MS: leucocortical, intracortical and subpial lesions 4, 35. Subpial cortical demyelination (SCD) accounts for the greatest proportion of demyelinated cortex in chronic MS 4 and seems to be a specific feature of MS 11. Meningeal inflammation is thought to play a role in the pathogenesis of SCD as meningeal inflammatory cells were found in topographic association with SCD in patients with early 26 and progressive MS 18, 27, 28. However, the cellular subpopulations and mechanisms contributing to SCD are not well understood. A role of ectopic B cell follicles associated with SCD in chronic MS was postulated in several studies 27, 39, but these findings were not confirmed by other investigations 10, 21. Of note, pronounced meningeal inflammation per se—as found in other infectious or neoplastic diseases—does not lead to SCD 11. These findings highlight that disease‐specific mechanisms are crucial for the pathogenesis of SCD in MS.

Experimental autoimmune encephalomyelitis (EAE) is the most commonly used animal model for MS. EAE in the common marmoset (Callithrix jacchus) bears a closer resemblance to MS than other EAE models because of a high degree of homology with human genes coding for myelin and immune system components 14, 33, 41. The lesion distribution and composition in this model share many of the major features of MS pathology including cortical involvement 36. In marmoset EAE, SCD also accounts for the greatest proportion of demyelinated cortex and shows the lowest density of inflammatory cells 36. White matter (WM) lesions as well as leuco‐ and intracortical lesions display characteristics that closely resemble the T cell plus antibody/complement‐mediated immunopattern II of early actively demyelinating MS lesions 25, 33. Early SCD also shows subpial antibody deposits 33, 40, whereby factors such as meningeal inflammation and pro‐inflammatory cerebral spinal fluid (CSF) changes are likely to contribute to SCD. To elucidate the pathogenesis of early SCD, we investigated whether and, if so, which meningeal inflammatory cells were associated with early SCD in marmosets with EAE. Our study found that the meninges overlying SCD displayed significantly increased inflammatory cell numbers mainly consisting of T cells followed by plasma cells compared with adjacent normal‐appearing EAE cortex (NAC). In contrast, NAC displayed similarly low meningeal T, B and plasma cell numbers as control cortex. These findings indicate that local meningeal T and plasma cell infiltration might play a role for the pathogenesis of SCD in marmosets with EAE. Further studies are required to unravel the direct and indirect mechanisms contributing to SCD in early and chronic MS.

Material and Methods

Animals

To assess meningeal inflammation in common marmosets (Callithrix jacchus) with and without EAE induction, we analyzed brain tissue samples from eight animals with induced EAE and eight age‐ and sex‐matched healthy control animals. The animals were obtained from the marmoset colony at the German Primate Center (Göttingen, Germany). All animal experimentation was carried out in accordance with the European Council Directive of September 2010 (2010/63/EU) and Communities Council Directive 86/609/EEC was approved by the Lower Saxony Federal State Office for Consumer Protection and Food Safety, Germany.

Induction and evaluation of EAE

For the induction of EAE, recombinant myelin oligodendrocyte glycoprotein (rMOG) corresponding to the N‐terminal sequence of rat MOG (amino acids 1–125) was expressed in Escherichia coli and purified to homogeneity as previously described 1. The purified protein was dissolved in 6 M urea, dialyzed against 20 mM sodium acetate buffer (pH 3.0) to obtain a soluble preparation and stored at −20°C. Immunization procedure was performed as previously described 5, 33. In brief marmosets were anesthetized and received subcutaneous injection of rMOG emulsified in incomplete Freund's adjuvant. Clinical symptoms were monitored daily by experienced observers. Animals and analyzed brain regions of control animals and marmosets with EAE were summarized in Table 1. Detailed clinical data on the individual disease course over time were only available for a subset of five of eight EAE animals (Figure 1); these data were obtained using a disability gradient scale specifically developed for marmosets, which was first described by Villoslada et al 42. Our study used a modified CNS disability score, further developed by Boretius et al 5, to monitor the animals everyday after immunization (Figure 1). This CNS disability score measured disease disability by assessing five different central functions (eyelid reflex, nystagmus, automutilation, paresis and grip strength). Experiments were terminated when a CNS score of 7.5 or higher had been reached.

Table 1.

Animals and brain regions investigated

Animals	Age (month)	Sex (M/F)	Evaluated brain regions
Control animals
1	123	M	Parietal
2	N/A	N/A	Frontal, temporal, parietal
3	119	M	Frontal, temporal
4	58	M	Frontal, temporal
5	48	M	Frontal, temporal
6	33	F	Frontal, temporal
7	25	M	Frontal, temporal
8	32	F	Frontal, temporal
EAE animals
1*	46	F	Frontal, temporal, occipital
2*	38	M	Frontal, temporal occipital
3*	49	F	Frontal, parietal, occipital
4*	47	M	Frontal, temporal, occipital
5*	29	M	Frontal, parietal
6	47	M	Frontal, temporal
7	39	M	Frontal, temporal
8	30	F	Frontal, temporal, parietal, occipital

*Disability scores were obtained from EAE animals 1–5 and shown in Figure 1. EAE = experimental autoimmune encephalomyelitis; F = female; M = male; N/A = no data available.

Figure 1.

Central nervous system (CNS) disability scores of marmosets with myelin oligodendrocyte glycoprotein (MOG)‐induced experimental autoimmune encephalomyelitis (EAE). CNS disability scores were obtained from EAE animals 1–5 over the duration of the experiment. CNS disability scores take into account different central functions (eye lid reflex, nystagmus, automutilation, paresis and grip strength). Experiments were terminated when reaching a CNS disability score of 7.5 or higher.

Histology and immunohistochemistry

The animals received a lethal dose of xylazine/ketamine and were perfused transcardially with 200 mL of saline followed by 400 mL of fixative containing 4% 0.1 M sodium‐phosphate buffered paraformaldehyde (pH 7.4). Heads were post‐fixed in buffered paraformaldehyde; after 24 h they were carefully removed from the skull, dissected and embedded in paraffin. Histologic evaluation was performed on 3‐μm thick sections from 16 tissue blocks (EAE animals) and 10 tissue blocks (controls). Immunohistochemistry was performed using the following primary antibodies: rabbit anti‐myelin basic protein (MBP, 1:500, DAKO, Hamburg, Germany) as well as mouse anti‐proteolipid protein for myelin (PLP, 1:250, mAb plpc1, Biozol, Eching, Germany), rat anti‐CD3 for T cells (1:50, mAb CD3‐12, AbD Serotec, Puchheim, Germany), mouse anti‐CD20 for B cells (1:50, mAb L26, DAKO, Hamburg, Germany), mouse anti‐immunoglobulin G (IgG) for plasma cells (1:100, mAb IgG88, BioGenex, Hamburg, Germany) and mouse anti‐myeloid‐related protein 14 (MRP14, clone MAC387, 1:100, GeneTex, Irvine, California, USA) to detect recently invaded macrophages and microglia expressing MRP14 (also referred to as S100A9) under inflammatory conditions 15, 16, 17. To identify meningeal T, B and plasma cells, respectively, adjacent to SCD, NAC and control cortex, we performed double‐immunolabelling combining either mouse anti‐PLP with mouse anti‐IgG or rabbit anti‐MBP with rat anti‐CD3, mouse anti‐CD20 and mouse anti‐MAC387. Bound antibody was either visualized using an avidin‐biotin technique (ABC) with 3,3′‐diaminobenzidine as chromogen (CD3, CD20, IgG) or alkaline phosphatase‐anti‐alkaline phosphatase for visualization of myelin (MBP, PLP) with Fast Blue as chromogen. Control sections were incubated in the absence of primary antibody, with isotype control antibodies or with non‐immune sera. Slides were counterstained using Mayer's hemalaun solution.

Morphometry and data acquisition

All the marmosets with EAE that were included in this study showed SCD as well as adjacent normal‐appearing cortex (NAC) on MBP‐ and PLP‐stained sections. NAC was defined as EAE cortex with regular myelin density. SCD was characterized by loss of myelin in the superficial cortical layers. The analyses included all areas of NAC, SCD and control cortex that were covered by meninges in all the regions (frontal, temporal, parietal and occipital) available for each animal, as indicated in Table 1. The analyses were based on 16 tissue blocks from eight EAE animals and 10 tissue blocks from eight controls, whereby each block showed whole bi‐hemispheric coronal brain sections. To calculate the meningeal cell numbers per cm meninges, the meningeal length was measured above all the analyzable (ie, covered by meninges) parts of control cortex (Figure 2A), NAC (Figure 2B) and SCD (Figure 2B) for each animal using the software CellA (Soft Imaging System, Olympus Soft Imaging Solutions, Münster, Germany) in the photographs taken by a digital camera (Color View II, Olympus, Hamburg, Germany).

Figure 2.

Subpial cortical demyelination in marmosets with experimental autoimmune encephalomyelitis (EAE). A–F. Representative photographs of proteolipid protein (PLP)‐immunostained sections of healthy controls (A,C) and EAE animals (B,D–F). A–B. Overview photographs of the frontal lobe show the cingulate gyrus adjacent to the corpus callosum as well as frontal cortex and subcortical white matter (WM). A. Control animals display regular myelin density in cortex and WM. B. EAE animals show demyelinated lesions in WM and cortex. To assess the meningeal cell density per cm meninges overlying control cortex, normal‐appearing cortex (NAC) and subpial cortical demyelination (SCD), the length of the meningeal surface is measured above each part of control cortex (A, purple line), NAC (B, black line) and SCD (B, black dashed line) in the photographs. C. High‐power photographs of the subpial cortex show intact cortical myelin in controls, D. whereas EAE animals display NAC (left) and SCD (right). E. Intracortical lesions and F, leucocortical lesions involving cortex (GM) and underlying subcortical WM are also present in marmosets with EAE. The border between cortex and WM is indicated by a dashed black line. Arrows mark the borders of demyelinated lesions. Scale bars A,B: 400 μm; C,D: 100 μm; E: 50 μm; F: 75 μm.

To assess the extent of meningeal inflammation in marmosets with EAE compared with healthy controls, CD3‐positive T cells, CD20‐positive B cells and IgG‐positive plasma cells were quantified in the meninges adjacent to SCD and NAC in EAE animals and control cortex in healthy animals as indicated in Figure 2. As meningeal MRP14‐positive macrophages were only found focally adjacent to restricted SCD areas with phagocytic activity in individual EAE cases, no overall representative quantitative data could be obtained. The total number of T, B and plasma cells were then counted along a distance of at least 5.5 mm mean meningeal length per animal. To obtain meningeal cell numbers for each area of interest per animal, the cell numbers were then divided by the length of meningeal surface and expressed as mean meningeal cell numbers per animal for control cortex, NAC and SCD.

Quantification of cortical and WM demyelination

Digital photographs were taken of all the coronal bi‐hemispheric brain sections immunostained with PLP or MBP of each EAE animal using the Keyence Bio REVO BZ9000 microscope (Keyence, Neu‐Isenburg, Germany). To obtain compound images of the whole coronal brain section per tissue block, single images were combined using the BZ‐II Analyzer software (Keyence, Osaka, Japan). The Image J software (National Institutes of Health, public domain) was used to measure the areas of demyelination within cortex and supratentorial WM as well as the total area of cortex and supratentorial WM on each of these digital compound images. Demyelinated lesions were defined as sharply demarcated areas characterized by myelin loss. Cortical lesions were classified as leucocortical, intracortical and subpial lesions as described previously 35. The proportion of demyelinated WM, cortex and cortical lesion types were then calculated and expressed as a percentage.

Statistical analysis

Statistical analyses were carried out using the software package GraphPad Prism 5.01 (GraphPad Prism Software, Ink, San Diego, California, USA). Paired t‐tests were used to test for differences in meningeal cell densities in SCD vs. NAC within EAE animals. To test for altered densities of meningeal inflammatory cells in EAE SCD vs. control cortex, Mann–Whitney U‐tests were used for nonparametric data and independent t‐tests for parametric data. Pearson's correlation coefficients were used to assess associations between the extent of subpial demyelination and the numbers of meningeal immune cells overlying SCD. A probability value of less than 0.05 was considered significant. All data are given as mean ± standard error of the mean.

Results

Clinical course in marmosets with EAE

Marmosets developed the first clinical symptoms between five and 28 days post immunization. The survival time of the animals was between 4 and 8 weeks after immunization. Animals and analyzed brain regions of all control and EAE animals were summarized in Table 1. The CNS disability scores obtained during the disease course from EAE animals 1–5 revealed a progressive involvement of different CNS systems for the duration of the experiment (Figure 1).

Extensive SCD in marmosets with EAE

The histopathologic evaluation revealed the presence of inflammatory demyelinating lesions within the cortex and in cerebral WM in all eight EAE animals, whereas healthy control animals displayed intact cortical and WM myelin (Figure 2A,C). All three main cortical lesion types were observed in the diseased marmosets: leucocortical lesions involving both WM and underlying deep neocortex (type I) (Figure 2F), small entirely intracortical lesions (type II) (Figure 2B,E) and subpial lesions reaching from the pial surface into the cortex (type III) (Figure 2B,D). All EAE animals showed sharply delineated and well‐defined subpial demyelinated cortical lesions as well as adjacent NAC with intact myelin (Figure 2B,D).

Subpial lesions were most frequent in the cingulate cortex, typically extending in a ribbon‐like pattern along the adjacent sulcus (Figure 2B), but were also observed in other frontal, temporal, parietal and occipital areas. Gyral areas displayed NAC as well as SCD (Figure 2B), whereby lesions often appeared wedge‐shaped with the basis at the cortical surface. The quantification of cortical and WM lesion load showed that cerebral WM demyelination was more extensive (37.3% ± 4.4% of WM) than cortical demyelination (20.9% ± 3.5% of cortex) in all EAE animals (Figure 3A, Table 2). The subanalysis of cortical lesions revealed that SCD was the predominant cortical lesion type present in all EAE animals. SCD affected 14.8% ± 2.1% of the total cortical area, whereas intracortical demyelination was found in 4.3% ± 1.1% of the cortex (Figure 3B, Table 2). Only 1.9% ± 0.5% of the cortex showed leucocortical demyelination (Figure 3B, Table 2).

Figure 3.

Percentage of demyelinated area in the cerebral white matter (WM) and cortex in marmosets with experimental autoimmune encephalomyelitis (EAE). A. The quantification reveals extensive demyelination of WM and cortex. B. The subanalysis of cortical lesions identifies subpial cortical demyelination as the predominant cortical lesion type. Data are presented as mean + standard error of the mean (SEM).

Table 2.

Mean percentage of demyelinated area in the cerebral white matter (WM) and cortex in marmosets with EAE

EAE animal	% Demyelinated area		% Cortex
	WM	Cortex	Leucocortical	Intracortical	Subpial
1*	35.4	11.8	0.6	2.5	8.7
2*	50.7	38.3	2.9	9.6	25.8
3*	19.5	21.2	2.7	2.8	15.7
4	22.5	13.0	0.0	0.3	12.7
5	50.2	27.2	3.6	7.3	16.3
6	58.0	26.7	1.9	6.8	18.0
7	33.9	3.2	0.0	0.0	3.2
8	28.3	26.0	3.6	4.8	17.5
Mean	37.3	20.9	1.9	4.3	14.8

*Animals 1–3 displayed focally restricted subpial lesions with phagocytic activation.

EAE = experimental autoimmune encephalomyelitis; WM = white matter.

All subpial lesions were classified as inactive demyelinated lesions based on the absence of myelin‐laden phagocytes. To assess whether the inflammatory phagocytic activation was increased in SCD in marmosets with EAE, we investigated the expression of the macrophage early activation marker MRP14 6. In EAE animals, most SCD areas appeared as hypocellular lesions (Figure 4A) as described previously 29. These lesions lacked MRP14‐positive cells in the lesion centre (Figure 4A) and at the margin (Figure 4B) and displayed no CD3‐positive T cells (Figure 4C). In addition to this predominant hypocellular pattern of SCD, we also found focally restricted SCD areas with MRP14‐positive cells (Figure 4D) and scattered T cell infiltration (Figure 4F) in three of eight EAE animals. These focally restricted SCD areas with phagocytic MRP14 expression typically displayed an increased density of MRP14‐positive cells at the lesion edge, but no signs of active demyelination as evidenced by the lack of macrophages containing MBP‐positive degradation products (Figure 4E). Furthermore, CD20‐positive B cells and IgG‐positive plasma cells were not observed in SCD areas either (data not shown).

Figure 4.

Low‐grade parenchymal inflammation in most subpial cortical demyelination (SCD) areas. A–C. Low‐grade parenchymal inflammation in SCD in five of eight marmosets with experimental autoimmune encephalomyelitis (EAE). D–F. Only a subset of three marmosets with EAE displays focally restricted SCD areas with phagocytic cells and scarce T cell infiltration. Immunohistochemistry was performed for myelin basic protein (MBP) (blue) in combination with either myeloid‐related protein 14 (MRP14) (brown) to detect MRP14‐positive phagocytes (A,B,D,E) or CD3 (brown) to identify T cells (C,F). Most SCD areas display neither phagocytic MRP14 expression in the lesion centre (A) and margin (B) nor T cell (C) infiltration in the parenchyma, accompanied by scarce meningeal macrophage infiltration (A). A subset of EAE animals shows focally restricted SCD areas with MRP14‐positive cells in the lesion centre (D) and margin (E) as well as scattered parenchymal T cell infiltration (F). Scattered meningeal macrophages (D) are observed overlying these focal SCD areas with inflammatory activation, but no signs of active ongoing demyelination evidenced by the lack of MBP‐positive macrophages (D,E). Exemplary MRP14+ cells and CD3+ T cells are marked by arrows and shown in greater detail (insets). Black arrowheads mark the borders of SCD. Scale bars 50 μm.

Increased meningeal T cell density adjacent to SCD

Only individual scattered T cells were present in the meninges of healthy control animals (Figure 5A) and in the meninges overlying NAC in marmosets with EAE (Figure 5B). In contrast, several T cells were detected in the meninges adjacent to SCD (Figure 5C), particularly along the sulci and the cingulate gyrus (Figure 5D). To test whether SCD was associated with increased meningeal T cell numbers, we quantified the density of CD3‐positive T cells in the meninges of EAE animals and healthy controls. Within EAE cases, meningeal T cell densities were significantly increased adjacent to SCD in EAE animals (37.0 ± 5.0 cells/cm) compared with EAE NAC (13.1 ± 2.2 cells/cm; P < 0.001) (Figure 5E). Control cortex (12.1 ± 2.5 cells/cm) and EAE NAC showed similarly low meningeal T cell densities (Figure 5E).

Figure 5.

Increased T cell infiltration in the meninges overlying subpial cortical demyelination (SCD) in experimental autoimmune encephalomyelitis (EAE) animals. A–D. Immunohistochemistry for myelin basic protein (MBP) (blue) and CD3+ T cells (brown) reveals only single meningeal T cells in healthy control animals (A) and in the meninges adjacent to EAE normal‐appearing cortex (NAC) (B). C,D. Increased numbers of infiltrating T cells are observed in the meninges overlying SCD in EAE‐induced animals. D. Particularly high numbers of T cells are observed in the meninges along the sulcus adjacent to the cingulate gyrus. Exemplary CD3+ T cells are marked by arrows and shown in greater detail (insets). E. The quantitative analysis of T cell density reveals significantly increased T cell infiltration in the meninges adjacent to SCD compared with NAC and control cortex. Scale bars 50 μm. ***P < 0.001.

Unaltered meningeal B cell numbers adjacent to SCD

We observed similarly low meningeal B cell numbers in healthy control animals (Figure 6A), EAE NAC (Figure 6B) and EAE SCD (Figure 6C). Similarly, low numbers of B cells were found within the meninges along the sulcus adjacent to the cingulate cortex (Figure 6D). To test whether SCD was associated with meningeal B cell infiltration, we quantified the density of CD20‐positive B cells in the meninges of EAE animals and healthy controls. There were similarly low numbers of B cells in the meninges adjacent to SCD in EAE animals (7.0 ± 2.4 cells/cm), NAC (5.1 ± 1.4) and control cortex (5.0 ± 2.1) (Figure 6E). Lymph node tissue from a healthy marmoset served as a positive control for the CD20 immunostaining (Figure 6F).

Figure 6.

Unaltered meningeal B cell infiltration adjacent to subpial cortical demyelination (SCD). A–D. Immunohistochemistry for myelin basic protein (MBP) (blue) and CD20+ B cells (brown) reveals only single meningeal B cells in healthy control animals (A), in normal‐appearing cortex (NAC) (B) and SCD (C,D) of experimental autoimmune encephalomyelitis (EAE) animals. B. An exemplary CD20+ B cell is marked by an arrow and shown in greater detail (inset). D. Unaltered B cell numbers are also observed along the sulcus adjacent to the cingulate gyrus. E. The quantitative analysis of meningeal B cell density reveals similar numbers of B cells for all analyzed groups. F. Lymph node tissue from a healthy marmoset served as a positive control for the CD20 immunostaining Scale bars A–D: 50 μm; F: 100 μm.

In addition, we analyzed macrophages expressing MRP14 in the meninges, but rarely found MRP14‐positive macrophages infiltrating the meninges adjacent to most SCD areas (Figure 4A), NAC and control cortex. Individually scattered meningeal MRP14‐positive macrophages were only observed adjacent to the focally restricted SCD areas with MRP14‐positive phagocytes in the subset of three of eight EAE animals (Figure 4D).

Elevated meningeal plasma cell density adjacent to SCD

Only a few scattered plasma cells were present in the meninges of healthy control animals (Figure 7A) and in the meninges adjacent to NAC in marmosets with EAE (Figure 7B). In contrast, several plasma cells were detected in the meninges overlying SCD (Figure 7C)—in particular, along the sulci and the cingulate gyrus (Figure 7D). To assess whether SCD was associated with increased meningeal plasma cell numbers, we quantified the density of IgG‐positive plasma cells in the meninges of EAE animals and healthy controls. Within EAE cases, there was a significant increase of meningeal plasma cell densities adjacent to SCD in EAE animals (19.7 ± 5.0 cells/cm) compared with EAE NAC (6.8 ± 0.8 cells/cm; P < 0.05) (Figure 7E). Control cortex (5.3 ± 1.6 cells/cm) and EAE NAC showed similarly low meningeal plasma cell densities (Figure 7E).

Figure 7.

Increased meningeal plasma cell infiltration adjacent to subpial cortical demyelination (SCD). A–D. Immunohistochemistry for proteolipid protein for myelin (PLP) (blue) and immunoglobulin G (IgG)+ plasma cells (brown) reveals only scattered meningeal plasma cell infiltration in healthy control animals (A) and in the meninges adjacent to experimental autoimmune encephalomyelitis (EAE) normal‐appearing cortex (NAC) (B). C,D. Increased numbers of infiltrating plasma cells are observed in the meninges overlying SCD particularly above the cingulate gyrus (D) in EAE‐induced animals. Exemplary IgG+ plasma cells are marked by arrows and shown in greater detail (insets) E. The quantitative analysis of plasma cell density reveals significantly increased plasma cell infiltration in the meninges adjacent to SCD compared with NAC and control cortex. Scale bars 50 μm. **P < 0.01, *P < 0.05.

Predominant T cells and plasma cells in the meninges overlying SCD

Markedly elevated inflammatory cell numbers were only observed in the meninges adjacent to SCD in EAE animals. To assess the relative contribution of T, B and plasma cells in the meninges adjacent to SCD, we directly compared the densities of these meningeal cell populations adjacent to EAE cortex with SCD. Meningeal macrophages were not included in this comparison as they were too scarce to be quantified in most animals. Taken together, our quantitative data identified T cells as the most prominent inflammatory cells (37.0 ± 5.0 cells/cm) (Figure 8). Plasma cells represented the second most common cell population (19.7 ± 5.0 cells/cm), whereas B cells showed only scattered infiltration of the meninges adjacent to SCD (7.0 ± 2.4/cm) (Figure 8). There were no significant correlations between the extent of SCD and the overlying meningeal T (r = 0.177, P > 0.05), B (r = 0.433, P > 0.05) or plasma (r = 0.359, P > 0.05) cell numbers.

Figure 8.

Predominant T and plasma cells in the meninges overlying subpial cortical demyelination (SCD). The quantitative analysis of inflammatory cell density in the meninges adjacent to SCD in experimental autoimmune encephalomyelitis (EAE)‐induced animals identifies CD3+ T cells as prominent inflammatory cells. Immunoglobulin G (IgG)+ plasma cells represent the second most common inflammatory cells in the meninges overlying SCD, while CD20+ B cells show only scattered meningeal infiltration. Data are presented as mean + standard error of the mean (SEM).

Discussion

In the present study, we investigated whether there is a relationship between meningeal inflammatory cell infiltration and SCD in marmosets with early rMOG_1‐125‐induced EAE. This is the first study to provide quantitative evidence that SCD is associated with significantly increased meningeal T cell and plasma cell infiltration in common marmosets with rMOG_1‐125‐induced EAE. In contrast, similarly low densities of B cells were observed in the meninges overlying SCD, NAC and control cortex. Our quantitative evaluation identified meningeal T cells as the most prominent inflammatory cell population followed by plasma cells, whereas B cell numbers resembled control levels.

SCD was the predominant cortical lesion type (71% of demyelinated cortex) in marmosets with EAE, whereas intra‐ and leucocortical lesions accounted for only 20% and 9% of demyelinated cortex, respectively. Subpial cortical lesions are also the most frequent cortical lesion type in MS, which underscores their contribution to disease pathology. Previous studies noted that meningeal inflammatory infiltrates can be observed adjacent to SCD in biopsies from patients with early MS 26 and in marmosets with EAE 36. However, our study is the first to systematically address the association of different meningeal inflammatory subpopulations for adjacent SCD at such an early disease stage. Our analysis of meningeal inflammatory cells in EAE animals demonstrates that both meningeal T and plasma cell numbers showed a significant, approximately threefold elevation adjacent to SCD compared with surrounding NAC. The fact that these cells were only increased within the meninges overlying SCD, but not above NAC or control cortex, highlights their potential role in the pathogenesis of SCD.

Our work and previous studies demonstrate that SCD is an early feature in MS 26 and marmosets with EAE 31, 36. We observed SCD in all investigated EAE animals, whereby most SCD areas showed a lack of T cells and MRP14‐positive phagocytes. Only a subset of EAE animals (three out of eight) displayed focally restricted SCD areas with intracortical phagocytes and scarce T cell infiltration, but without any signs of active ongoing demyelination. These focally restricted SCD areas with increased inflammatory activation are likely to represent an earlier lesion stage of SCD. The low density of T cells and the lack of MRP14‐positive phagocytes in most SCD areas is in line with previously reported findings in marmosets with EAE 36 and chronic MS cases 4, 35. Recent experimental studies in rats with locally induced cortical demyelination confirm that there is only a brief and transient increase of intracortical T cells as well as microglia and macrophage infiltration in SCD during the first week after lesion induction 13, 32, 37. Chronic SCD then displays only low levels of inflammatory cells, but accounts for the greatest proportion of demyelinated cortex in MS 3, 4. Post‐mortem studies in MS indicate that demyelinated cortical lesions including SCD show a high propensity for remyelination 2, 9. Therefore, the myelin in cortical MS lesions is not only subject to local demyelinating factors, but also to repair mechanisms. This suggests a more dynamic cycle of de‐ and remyelination in MS.

In our study, meningeal T cells were the predominant inflammatory cell population associated with SCD in marmosets with EAE. It may not be surprising that T cells predominate among meningeal infiltrating cells in this animal model mainly driven by autoreactive proinflammatory T cells 20. However, anti‐MOG antibodies have been reported to amplify demyelination in marmosets with EAE 30. Therefore, it is likely that plasma cells, which were the second most common inflammatory cell type in the meninges adjacent to SCD, may also contribute to the pathogenesis of SCD. T cells were not only the predominant cell population among meningeal infiltrating cells in our work, but also in previous studies in chronic MS 18. Even though this human post‐mortem study focused on a potential role of meningeal B cells, follicle‐like structures for cortical pathology, T cells were also found to be the predominant meningeal inflammatory cell population in patients with primary and secondary progressive MS. These and our findings indicate that meningeal T cell inflammation and SCD appear to be related not only in marmosets with EAE, but also in patients with MS.

Our study did not find any evidence for increased meningeal B cell numbers or B cell follicles associated with SCD. Given the ongoing debate about the role of these follicle‐like structures in MS 10, 21, the lack of B cell follicles in our study of early SCD may not be surprising. However, our finding of unaltered meningeal B cell levels may appear to be in contrast to observations in early 26 and chronic MS 18, but it is important to note that our animal model rather reflects acute inflammatory changes compared with a more chronic inflammatory environment in MS. Our results of low meningeal B cell levels in marmosets with a disease duration of 10 days or longer are compatible with recent data from rats with focal SCD 13. These experimental findings in rats indicate that the increase of meningeal B cells was a very quick and transient phenomenon peaking already 1 day after lesion induction 13, whereas elevated meningeal T cells appeared to persist over time. Even though we could not detect increased meningeal B cell infiltration in our EAE animals after 4–8 weeks after immunization, recent work in marmosets with EAE underlines that B cells play a crucial role in the pathogenesis of cortical demyelination 19. B cell depletion completely prevented cerebral grey and WM demyelination in marmosets with EAE. However, it is important to note that B cell depletion does not only affect B cells, but also reduces immunoglobulin (Ig) levels as well as T cell proliferation and cytokine production in marmosets with EAE. Our finding of significantly increased meningeal IgG‐positive plasma cells adjacent to SCD is in line with the previously reported subpial Ig deposition on myelin sheaths and microglial infiltration present during the early stage of subpial lesion formation in MOG‐immunized rats 40. Later stages of SCD then showed subpial demyelination with microglial activation and IgG deposits at the lesion margin in this rat model. Their and our findings suggest a relevant role for plasma cells in the pathogenesis of SCD, even though Ig deposits were not evaluated in marmoset EAE in our study. Increased Ig levels were described in the CSF of MS patients and remain a hallmark finding in the diagnosis of MS 34, 43. Primates with induced EAE also showed an increased IgG concentration in CSF within 4 weeks after immunization 12.

All the investigated EAE animals displayed adjacent NAC beside SCD, which is in line with observations in many autopsy cases with MS. The co‐occurrence of NAC and SCD in a given case underlines that local factors seem to be crucial for the pathogenesis of focal subpial demyelination. These findings raise the question whether meningeal inflammatory cells play a causative role in the pathogenesis of SCD or whether they follow SCD. Our study identified meningeal T and plasma cell infiltration as significant factors associated with underlying SCD. A topographic association of meningeal inflammation and SCD was also observed in patients with early 26 and chronic 18 MS. Howell and colleagues demonstrated that meningeal inflammation correlated with the underlying cortical pathology even in progressive MS, whereby T cells were also the predominant inflammatory cell population 18. It is important to note that our animal model allows much earlier investigations of pathologic changes than studies in human post‐mortem tissue. Recent experimental work by Gardner et al indicates that meningeal inflammation peaks before subpial intracortical inflammation and demyelination after subarachnoid cytokine injection in MOG‐pre‐immunized rats 13. Therefore, both their and our findings strongly suggest a causative role of meningeal inflammation in the pathogenesis of SCD. We hypothesize that meningeal inflammatory cells might contribute to SCD either by spreading via meningeal vessels into the subpial space or by releasing pro‐inflammatory factors. However, further animal studies with detailed earlier time course experiments are required to unravel the precise mechanisms leading to SCD.

We identified the highest density of T cells and plasma cells in the cerebral meninges—particularly in the deep sulci—which may represent a microenvironment with low CSF flow. This could support the retention and proliferation of inflammatory cells giving rise to a highly inflammatory milieu with elevated cytokine levels. Our findings are in accordance with results from a recent study demonstrating focal SCD after subarachnoid injection of tumor necrosis factor alpha (TNFα) and interferon gamma (IFNγ) into the sagittal sulcus in MOG‐pre‐immunized rats 13. These rats displayed meningeal infiltration with T cells, B cells and macrophages within the sagittal sulcus accompanied by SCD in the cingulate cortex. Similarly, SCD can also be induced by intracortical injection of TNFα and IFNγ in MOG‐pre‐immunized rats 32. In marmoset EAE and MS, we hypothesize that both meningeal T and plasma cells contribute to subpial lesion formation. On the one hand, pro‐inflammatory T cells secreting IFNγ and other factors may trigger TNFα production in meningeal macrophages and subpial microglia leading to band‐like subpial microglial infiltration. On the other hand, IgG produced by plasma cells is likely to bind to subpial myelin sheaths and thereby initiate subsequent myelin degeneration. The extent of subpial lesion formation appears to depend on specific combinations of major histocompatibility complex I and II isotypes and alleles as evidenced by experimental studies in different rat strains 40. SCD and transient inflammation might then cause damage to underlying subpial cortical components 24, 28.

In conclusion, our study demonstrates that meningeal T cell and plasma cell infiltration is associated with SCD. Adjacent NAC displayed similarly low meningeal T, B and plasma cell numbers as control cortex. These findings suggest that local meningeal T and plasma cell infiltration plays a role in the pathogenesis of SCD in marmosets with EAE. Further studies are required to understand the direct and indirect mechanisms contributing to SCD in early and chronic MS.

Conflict of Interest

The authors have nothing to disclose and no conflict of interest to report.