The formation of inflammatory demyelinated lesions in cerebral white matter

Pietro Maggi, Sheila M Cummings Macri, María I Gaitán, Emily Leibovitch, Jillian E Wholer, Heather L Knight, Mary Ellis, Tianxia Wu, Afonso C Silva, Luca Massacesi, Steven Jacobson, Susan Westmoreland, Daniel S Reich, Pietro Maggi, Sheila M Cummings Macri, María I Gaitán, Emily Leibovitch, Jillian E Wholer, Heather L Knight, Mary Ellis, Tianxia Wu, Afonso C Silva, Luca Massacesi, Steven Jacobson, Susan Westmoreland, Daniel S Reich

Abstract

Objective: Vascular permeability and inflammatory demyelination are intimately linked in the brain, but what is their temporal relationship? We aimed to determine the radiological correlates of the earliest tissue changes accompanying demyelination in a primate model of multiple sclerosis (MS), experimental autoimmune encephalomyelitis (EAE) in the common marmoset.

Methods: By 7T magnetic resonance imaging (MRI), T1 maps, proton density, and T2-weighted images were acquired before and after EAE induction in 5 marmosets (every other week before lesions appeared, weekly thereafter). From scans before and after intravenous injection of contrast material, we measured the evolution of lesional blood-brain barrier (BBB) permeability, comparing in vivo MRI to postmortem tissue examination.

Results: On average, BBB permeability increased 3.5-fold (p < 0.0001) over the 4 weeks prior to lesion appearance. Permeability gradually decreased after lesion appearance, with attendant changes in the distribution of inflammatory cells (predominantly macrophages and microglia) and demyelination. On tissue analysis, we also identified small perivascular foci of microglia and T cells without blood-derived macrophages or demyelination. These foci had no visible MRI correlates, although permeability within the foci, but not outside, increased in the weeks before the animals died (p < 0.0001).

Interpretation: This study provides compelling evidence that in marmoset EAE, which forms lesions strongly resembling those of MS, early changes in vascular permeability are associated with perivascular inflammatory cuffing and parenchymal microglial activation but precede the arrival of blood-derived monocytes that accompany demyelination. Prospective detection of transient permeability changes could afford an opportunity for early intervention to forestall tissue damage in newly forming lesions.

Conflict of interest statement

Competing interests

The authors report no relevant conflicts of interest.

© 2014 American Neurological Association.

Figures

FIGURE 1
FIGURE 1
Matched hematoxylin and eosin (H&E) section (left) and ex-vivo T2* magnetic resonance imaging (MRI) slice (right) from WMH3.
FIGURE 2
FIGURE 2
Cumulative volume of brain white matter lesions over time in marmoset experimental autoimmune encephalomyelitis induced with human white matter homogenate (WMH).
FIGURE 3
FIGURE 3
Focal blood-brain barrier permeability, estimated as the change in longitudinal relaxation rate following intravenous administration of gadolinium-based contrast dye (ΔR1), increases before the appearance of brain white matter lesions on magnetic resonance imaging. (A) Normal appearing (extralesional) white mater (blue) and lesion (red) permeability changes over time across all marmosets with experimental autoimmune encephalomyelitis. tlesion=0 denotes the time of lesion appearance on imaging, measured in weeks. Lesion ΔR1 remains stable prior to 4 weeks before lesion appearance (tlesion=−4). After tlesion=−4, ΔR1 increases in lesion areas, peaking at the time of lesion appearance (tlesion=0), and decreases over the 4 weeks thereafter. Permeability in the normal appearing white matter is stable over time (p=0.77). (B) Average changes to within-lesion-area permeability over time (*, p<0.001; **, p<0.0001). Error bars represent standard errors of the mean.
FIGURE 4
FIGURE 4
Non-demyelinated inflammatory nodules are associated with increasing blood-brain-barrier permeability over time. (A) Matched T2PD terminal magnetic resonance imaging (MRI) scan and Iba1 coronal section of the brain (WMH4). The representative inflammatory nodule appears as a small cluster of Iba1+ cells (black arrow) that is not visible, or barely visible, on the matched in vivo MRI (white arrow). Magnified views, red boxes. Scale bars: 300 μm. (B) Histological characterization of the inflammatory nodule shown in A. Only 1 MRP14+ blood-derived early activated macrophage (with MRP14 reaction measured as moderate) is present within the perivascular space, but not within the parenchyma. Iba1+ activated microglia and macrophages are present in the perivascular space and parenchyma. GFAP+ astrocytes are present within the parenchyma and in contact with the central vessel. CD3+ lymphocytes (T cells) are located at the edge of the vessel wall and in the perivascular space but do not extend into the parenchyma. CD20+ cells were detected in the perivascular space (not shown). CNPase (an oligodendrocyte marker) is normally expressed, and luxol fast blue (LFB), which stains myelin in blue, shows no demyelination. Scale bars: 50 μm. (C) Non-demyelinated inflammatory nodules (red), lesions (purple), and normal appearing white matter (blue) permeability, estimated as ΔR1. For lesions, ΔR1 is shown over the 4 weeks prior to lesion appearance on T2PD MRI. For the inflammatory nodules, we assumed that the week the experiment was terminated coincided with the week preceding lesion appearance. There was no difference between slopes of the permeability curves for inflammatory nodules and lesions (p=0.39); however, unlike normal appearing white matter, permeability within inflammatory nodules and lesions was significantly >0 (p<0.001). Error bars represent standard error of the mean.
FIGURE 5
FIGURE 5
Representative magnetic resonance imaging (MRI) and histopathological features of acute lesions. (A, C) Acute lesions, visible on the hematoxylin and eosin section (blue arrow), were first seen on the terminal MRI (red arrow), performed on the day the experiment ended. The time of lesion appearance on MRI is denoted tlesion=0 and is measured in weeks, and these two lesions therefore appeared at most 1 week prior to the end of the experiment (lesion age <1 week). Magnified views (red boxes). Scale bars, 300 μm. (B, D) Histopathological features of the acute lesions shown in A and C, respectively. Blood-derived MRP14+ early activated macrophages are present in the parenchyma but rarely within the perivascular space of the central blood vessel. Iba1+ activated microglia and macrophages are present in both parenchyma and perivascular space. GFAP+ astrocytes are found in contact with the central vessel B and throughout the lesion D. CD3+ lymphocytes (T cells) are located in the perivascular space of the central vessel and sparingly in the adjacent parenchyma. CD20+ lymphocytes (B cells) are located in the perivascular space of the central vessel and also in the parenchyma. CNPase+ oligodendrocytes are diminished in the lesion parenchyma relative to surrounding white matter. LFB stains myelin in blue and demonstrates that the lesions are demyelinated. Bielschowski’s stain (Biels) shows axons that are disrupted and separated by inflammatory cells (red arrows) unlike the tightly apposed axons in the adjacent parenchyma (white arrows). Magnified views (white boxes). Scale bars, 100 μm. Animals: WMH4 (A, B); WMH3 (C, D).
FIGURE 6
FIGURE 6
Representative magnetic resonance imaging (MRI) and histopathological features of subacute lesions. (A, C) The lesion (red arrows) is not visible on the pre-immunization baseline scan (black arrows). The time of lesion appearance on MRI is denoted tlesion=0 and is measured in weeks. At the end of the experiment (“Terminal T2PD MRI”), the lesion in A was 1–3 weeks old and the lesion in C, 3–5 weeks old. Both lesions enlarged in the weeks after they appeared. A blood vessel within the center of the lesion in A appears dark on MRI. The lesions are denoted by blue arrows on the hematoxylin and eosin (H&E) sections. Magnified views (red boxes). Scale bars, 300 μm. (B, D) Histopathological features of the subacute lesions shown in A and C, respectively. Blood-derived MRP14+ early activated macrophages are present at the lesion edge. Iba1+ activated microglia and macrophages are found throughout the lesion, without a clear distinction between center and periphery. Activated GFAP+ astrocytes are found within the lesion parenchyma and particularly in contact with blood vessels. CD3+ lymphocytes (T cells) are present in perivascular spaces and scattered throughout the lesion area. CD20+ lymphocytes (B cells) are only rarely present in parenchyma and perivascular spaces. CNPase+ oligodendrocytes are reduced in the lesion parenchyma. Luxol fast blue (LFB), which stains myelin blue, shows complete myelin loss at the center and milder demyelination at the lesion margins. Bielschowski’s stain (Biels) shows that normal axonal orientation and distribution is disrupted by infiltrating inflammatory cells with regions of complete axonal loss and fragmentation. Magnified views (white boxes). Dark lines and arcs in various histopathological sections are folding artifacts. Scale bars, 200 μm B and 300 μm D. WMH4 (A, B) ; WMH3 (C, D).
FIGURE 7
FIGURE 7
Representative magnetic resonance imaging (MRI) and histopathological features of late-subacute lesions (animal WMH2). (A, C) Lesions (red arrows) are not visible on the pre-immunization baseline scans (black arrows) but appeared more than 5 weeks before the end of the experiment. The lesion in A was between 6 and 8 weeks old, and the lesion in B between 5 and 6 weeks old. Lesion size peaked in the weeks following lesion appearance but gradually diminished thereafter. Magnified views (red boxes). Scale bars, 300 μm. (B) Histopathological features of the late-subacute lesion shown in A. Blood-derived MRP14+ early activated macrophages are no longer seen. Iba1+ activated microglia and macrophages remain perivascularly and in the parenchyma close to the central vessel. Astrocytes show moderate GFAP expression. CD3+ (T cells) and CD20+ (B cells) lymphocytes are still present, although rare, in the parenchyma and perivascular spaces. CNPase+ oligodendrocytes are almost normal, except in the area immediately surrounding the central blood vessel. Luxol fast blue (LFB), which stains myelin blue, shows pale myelin in the residual lesion area close to the central blood vessel, suggesting remyelination. Axons are mildly disrupted by inflammatory cell infiltrate with evidence of dissolution in comparison with surrounding, closely packed axons. Magnified view (white box). Scale bars 100 μm. (D) Histopathological features of the late-subacute lesion shown in C. Luxol fast blue (LFB) shows absent myelin in the center of the lesion and pale myelin at the lesion margins. CNPase expression is reduced at the lesion center but still partially present at the periphery. Iba1+ activated microglia and macrophages are rarely present in the lesion area, but MRP14+ macrophages are absent. CD3+ lymphocytes (T cells) line the vessel wall and persist throughout the lesion and in the perivascular spaces. Scale bars, 100 μm.
FIGURE 8
FIGURE 8
Model of lesion formation in marmoset experimental autoimmune encephalomyelitis. Peripherally primed myelin-reactive T cells (CD3+ T cells) pass into the perivascular space and recognize myelin antigens on the surface of antigen-presenting cells (e.g., Iba1+ perivascular macrophages). Activation of these T cells following antigen presentation then activates both the vascular endothelium (leading to subtle blood-brain-barrier opening) and resident microglia (Iba1+ microglial activation) by secreting soluble factors (cytokines and chemokines). These inflammatory factors enter the parenchyma, initiating tissue damage (including myelin damage) and activating resident microglia. Within a few weeks, the parenchymal side of the blood-brain barrier (the glia limitans) is breached, and early activated macrophages (blood-derived MRP14+ cells) infiltrate the parenchyma to remove myelin.

Source: PubMed

3
Abonneren