The Immunopathology of Multiple Sclerosis: An Overview

Abstract

Multiple sclerosis (MS) is traditionally seen as an inflammatory demyelinating disease, characterized by the formation of focal demyelinated plaques in the white matter of the central nervous system. In this review we describe recent evidence that the spectrum of MS pathology is much broader. This includes demyelination in the cortex and deep gray matter nuclei, as well as diffuse injury of the normal‐appearing white matter. The mechanisms responsible for the formation of focal lesions in different patients and in different stages of the disease as well as those involved in the induction of diffuse brain damage are complex and heterogeneous. This heterogeneity is reflected by different clinical manifestations of the disease, such as relapsing or progressive MS, and also explains at least in part the relation of MS to other inflammatory demyelinating diseases.

INTRODUCTION

Multiple sclerosis (MS) belongs to a larger group of inflammatory demyelinating diseases of the central nervous system (CNS), which include—besides the different manifestations of MS—acute disseminated (or hemorrhagic) leukoencephalitis, Devic’s neuromyelitis optica and Balo’s concentric sclerosis (2). Although these diseases differ in clinical course, imaging, pathology and immunopathogenesis, they share some essential structural features of their lesions. They all occur on a background of inflammatory reaction, composed of lymphocytes and activated macrophages or microglia, and show demyelination, in which axons are at least partly preserved. It is widely believed that an inflammatory process of putative autoimmune nature is the driving force of tissue injury in MS (43).

Most studies on MS pathology and pathogenesis have so far concentrated on focal demyelinated lesions in the white matter mainly at the chronic disease stage. This plaque‐centered view has recently been challenged by magnetic resonance imaging (MRI) studies, which revealed a much more widespread and global damage of the brain and spinal cord, in particular in patients at late stages of the disease (65). Furthermore, current anti‐inflammatory, immunomodulatory or immunosuppressive treatments are at least partly effective in the early stages of the disease but are of limited benefit when patients have entered the progressive phase (68). Moreover, many therapeutic strategies, which had proven effective in paradigmatic experimental models of T cell‐mediated inflammatory demyelinating diseases, had no effect or even worsened the disease, when introduced for the treatment of MS patients (34, 37, 43). These discrepancies suggest that MS is a much more complex disease than previously thought. Therefore, a careful and systematic reassessment of the pathology and immunology is mandatory. This approach has to address the nature of the inflammatory response and the mechanisms of tissue injury and repair at different stages of the disease, as well as provide answers on the relationship between MS and other inflammatory demyelinating diseases. During the last years enormous progress has been achieved in this area, which will be summarized in this brief review.

STAGE DEPENDENT PATHOLOGICAL HALLMARKS OF MS

In the majority of MS patients, the disease begins with a relapsing course (relapsing/remitting MS; RRMS) followed after several years by a progressive phase (secondary progressive MS, SPMS; 21, 22). In some patients, the relapsing phase is missed and the disease is progressive from the onset (primary progressive MS, PPMS). Clinical and MRI data suggest that inflammation and the formation of new white matter lesions are the substrate for RRMS (20), while in the progressive phase new inflammatory demyelinating lesions are rare but diffuse atrophy of the gray and white matter and changes in the so‐called normal‐appearing white matter (NAWM) become prominent (65). The number and severity of new relapses during the early stage of MS in part determine the time at which the progressive stage is reached, but it has no influence on the rate of progression once the progressive phase is reached. Furthermore, anti‐inflammatory therapies are of limited efficacy in the progressive phase of the disease. Based on these observations, it has been suggested that in the early phase of the disease inflammation is the driving force, whereas the progressive phase may be underlined by a neurodegenerative process, which develops at least in part independent from inflammation (94). Is this concept supported by neuropathological findings?

The pathology of MS was originally defined as an inflammatory process, associated with focal plaques of primary demyelination in the white matter of the brain and spinal cord (18). Inflammation is dominated by T cells and activated macrophages or microgia. In active lesions this inflammatory process is accompanied by a profound disturbance of the blood brain barrier (41, 48), the local expression of proinflammatory cytokines and chemokines as well as of their cognate receptors (17, 44). Complete demyelination is accompanied by a variably degree of acute axonal injury and axonal loss (31, 90), which in part is counteracted by remyelination (49). Inflammatory demyelinating focal white matter lesions dominate the pathology in acute MS and RRMS.

In the progressive stage of MS, both in patients with SPMS and PPMS, the pathological picture is different (Figure 1; 51). Although focal demyelinated white matter lesions are still present, classical active demyelinating plaques are rare. However, a substantial proportion of preexisting plaques show evidence for a slow and gradual expansion of the lesions at their margins (78). This is characterized by moderate inflammatory infiltrates, mainly composed of T cells, and profound microglia activation. Only few of these activated microglia cells contain myelin degradation products, suggesting a very slow rate of ongoing demyelination. In addition, the NAWM outside of plaques is highly abnormal (3, 4). It is affected by a diffuse inflammatory process and generalized activation of microglia, which is associated with diffuse axonal injury and destruction, followed by secondary demyelination (51). Besides these white matter alterations, the cortex is also severely damaged (12, 71). Extensive cortical demyelination is seen in both the forebrain and the cerebellum, which may affect, in extreme cases, more than 60% of the total cortical area (Figure 1; 51). Active cortical demyelination is associated with inflammatory infiltrates in the leptomeninges.

Figure 1.

Multiple sclerosis pathology: from focal plaques to diffuse brain damage. A. Acute multiple sclerosis (MS) (Pattern II); multiple perivenous inflammatory demyelinating lesions, which form in some places confluent demyelinated plaques. B. Acute MS (Pattern III); focal large demyelinated plaques (green lesions), the right periventricular lesions shows concentric layering. C,D. Benign MS; inactive demyelinated lesions in the spinal cord (D); very few and tiny lesions in the brain white matter and cortex. E. Relapsing/remitting MS; multiple white matter lesions; most of the lesions are remyelinated shadow plaques; the few demyelinated lesions are active; very few and small lesions in the cerebral cortex. F. Secondary progressive MS; multiple large white matter plaques; most plaques are demyelinated, only few are remyelinated; extensive cortical demyelination and multiple lesions in the deep gray matter. G. Primary progressive MS; multiple small focal white matter lesions, most of them remyelinated; extreme cortical demyelination; in addition there is extensive diffuse injury in the normal‐appearing white matter (see figures I–M). H. Secondary progressive MS, large focal demyelinated white matter lesions are associated with extreme cortical demyelination and diffuse white matter atrophy; several lesions are also present in the deep gray matter. I–M. Same patient as shown in figure G; massive injury of the normal‐appearing white matter, consisting of inflammation (CD8 positive lymphocytes are shown in figure I); massive microglia activation with the formation of microglia nodules (shown by staining for HLA‐D in figure J,K) and expression of inducible nitric oxide synthase (shown in figure L); this inflammatory process is associated with extensive axonal injury (staining for neurofilament in figure M). N–O. Secondary progressive MS with massive inflammation in the meninges, composed of T cells (CD8⁺; figure N) and plasma cells (staining for immunoglobulin; figure O). Green: demyelinated lesions in the white matter; blue: remyelinated lesions in the white matter; red: demyelinated lesions in the cortex; pink: demyelinated lesions in the deep gray matter.

These pathological findings are in agreement with alterations predicted from extensive MRI investigations. There is, however, one discrepancy. Whereas pathologically, all forms of active tissue damage in the progressive stage of the disease is invariably associated with inflammation (51), evidence for inflammation, including blood brain barrier damage, is rare or absent on MRI (88, 89). In order to understand the basis for this discrepancy, the nature of the inflammatory response during different stages of MS needs to be analyzed in detail.

THE NATURE OF THE INFLAMMATORY RESPONSE IN MS

If inflammation is the driving force of tissue injury in MS, understanding its immunological mechanisms is essential for the development of effective therapeutic strategies. Intense research over the last decades revealed that an experimental disease shared many essential features with MS, and was mediated by autoimmunity against antigens within the CNS (43). Using the respective models of experimental autoimmune encephalomyelitis (EAE), basic principles of immune surveillance of the brain by T cells, as well as of brain inflammation and immune‐mediated tissue injury, were elucidated. However, extrapolating results obtained in EAE models to MS clinical trials were only partially successful (37, 43). A possible explanation may be that in MS the inflammatory process is driven at least in part by other immune cells than those operating in EAE (34). The other possibility is that the disease is not autoimmune in nature.

Composition of inflammatory cells in MS lesions.  Inflammation in EAE is mediated by Major Histocompatibility Complex (MHC) Class II‐restricted autoreactive T cells. This is reflected by a dominance of CD4⁺ T lymphocytes at least in the early stages of lesion formation (32). Furthermore, immunotherapies, which specifically target these Class II‐restricted cells, are highly effective. In MS, however, the composition of inflammatory infiltrates in the lesions is different. MHC Class I‐restricted CD8⁺ T cells dominate the lesions, regardless of the stage of activity or disease, while the component of CD4⁺ T cells is relatively small (15, 36, 40; reviewed in more detail in the accompanying paper by M. Rodriguez). Clonal expansion of T cells as a potential footprint of antigen‐specific activation is mainly found in the CD8⁺ T cell population (6). MHC Class I antigens are highly expressed within MS lesions, not only on inflammatory cells, but also on neurons and glia (42). In some cases of acute MS CD8⁺ T cells, which express granzyme B as a marker of cytotoxic activation, can be seen in close proximity or attachment to oligodendrocytes or demyelinated axons (67). Thus, in analogy to other inflammatory brain diseases such as Rasmussen’s encephalitis or paraneoplastic encephalitis, which are mediated by MHC Class I‐restricted cytotoxic T cells, CD8⁺ T cells are a major component of the inflammatory response in MS lesions (10). These differences in the inflammatory response between MS and EAE may in part explain the divergent results observed regarding therapeutic response (34).

Another difference in the inflammatory response between EAE and MS appears to be the contribution of B cells and plasma cells. Although plasma cells are rare in the infiltrates in classical active plaques of acute and relapsing MS, they become increasingly prominent with chronicity of the disease (69). The pathogenic role of B cells and plasma cells in MS is not entirely clear, but recent ongoing studies indicate that therapies targeting B cells in MS may be partially effective (24).

Compartmentalization of inflammation in the CNS may drive chronic progressive disease.   As mentioned above, active tissue injury in the progressive stage of MS occurs on the background of an inflammatory reaction (51). Yet, on MRI, leakage of gadolinium‐diethylenetriamine pentaacetic acid, which is regarded as a marker for (inflammation induced) blood brain barrier damage in MS, is rare or absent (88, 89). These data suggest that inflammation in progressive MS may become trapped behind a closed or repaired blood brain barrier. Using a new marker for leaky endothelial cells, it was recently demonstrated that the relationship between blood brain barrier leakage and inflammation in MS is more complicated than expected (41). In agreement with previous neuropathological studies, increased blood brain barrier permeability was not only shown in active lesions, but was also present in inactive plaques as well as in the NAWM (48, 53). Thus, gadolinium enhancement is a rather insensitive marker for blood brain barrier disturbance. Furthermore, many blood vessels with perivascular inflammation became apparent, which did not express the marker for leaky endothelial cells or showed perivascular fibrin leakage. This was particularly prominent in patients with progressive disease and suggests that at this stage, inflammation persists in the CNS compartment behind a closed or repaired blood brain barrier (41).

A possible reason for such a persistence of inflammation may be the formation of lymphatic‐like tissue within the connective tissue compartments of the brain (5). In patients with progressive MS, focal areas of inflammation can be found in the meninges, which closely reflect the structural features of B cell follicles (82). They contain dense clusters of B cells and plasma cells, which surround areas that resemble germinal centers and contain dendritic cells. These structures seem to be abundant in patients with severe and rapid progression of the disease and are topographically related to areas of cortical demyelination. Similar follicle‐like structures are also found in other chronic inflammatory diseases, such as Hashimoto thyroiditis (5). Certain cytokines and chemokines involved in the formation of lymphatic tissue, such as lymphotoxin, BAFF or CXCL 13, are ectopically expressed in chronic inflammatory MS lesions and may be involved in the homing of B cells and the formation of these structures (63).

Taken together, these data suggest that with increasing chronicity of the inflammatory process, the inflammatory reaction in MS becomes compartmentalized within the CNS behind an intact blood brain barrier. Thus inflammatory tissue injury in the progressive stage of the disease seems to be driven by an inflammatory response, which no longer is under the control of the peripheral immune system. This may in part explain the ineffectiveness of current immunotherapies during this stage of the disease.

MECHANISMS OF TISSUE INJURY IN MS ARE HETEROGENEOUS

The broad spectrum of structural and immunological alterations seen in MS lesions derived from different patients and from different stages of the disease suggests that the mechanisms of tissue injury are heterogenous between patients and stage dependent within the same patient. This implies that therapeutic strategies need to take into account not only the differences in pathogenesis between the relapsing versus the progressive stage of the disease, but also potential interindividual differences between patients.

Classical actively demyelinating lesions.   Experimental neuroimmunology shows that inflammatory demyelinating lesions can be induced by a variety of different immune mechanisms, involving cytotoxic T cells (86), specific autoantibodies (56), mechanisms of innate immunity (30) as well as the genetic susceptibility of the target tissue (57). Not surprisingly, elements of these different mechanisms can be seen within active MS lesions (9, 36, 75, 84). Moreover, the analysis of a large sample of brain biopsies and autopsies from patients during the early stage of the disease revealed patterns of demyelination which were homogenous in multiple lesions of the same patient, but different between patients (59). In some patients the lesions were dominated only by T cells and macrophages, whereas in others, profound accumulation of immunoglobulins and complement was observed suggesting the involvement of pathogenic antibodies (84). In other lesions, a pattern of tissue injury which closely mimics that found in acute white matter stroke lesions (1) was observed. Such lesions appear to be driven by an exaggerated production of oxygen and nitric oxide radicals, which may induce a disturbance of mitochondrial function with subsequent histotoxic hypoxia. Finally, in some patients a mild inflammation was associated with profound oligodendrocyte degeneration in the periplaque white matter, suggesting an increased susceptibility of the target tissue for immune‐mediated injury. These different interindividual patterns of tissue injury clearly segregate in patients with fulminate exacerbations of the disease. In such patients this heterogeneity may have therapeutic consequences. Plasma exchange is highly effective in patients with antibody and complement‐associated tissue destruction, while it fails in patients following other patterns of tissue injury (46).

Having said this, a note of caution must be added. The concept of interindividual heterogeneity of lesions was recently challenged in a study describing the coexistence of different patterns of tissue injury within the same patient (9). However, in this study the criterion for defining complement activation, as well as the immunopathological classification of the lesions, was in part different from that described in the original publication (59). Whether the divergent results can solely be explained by methodological differences between the studies, or whether in some patients different mechanisms of tissue injury can occur side by side is not yet clear. In our studies based on a large sample of cases, also including serial biopsies and autopsies, as well as detailed follow‐up studies with MRI, such an overlap was not observed (61).

Slowly expanding lesions of progressive MS.  Slowly expanding lesions in progressive MS differ in several important features from classical active plaques. While in the latter, macrophages with all stages of myelin degradation products are present either throughout the lesion or in a broad rim at the lesion edge, such macrophages are rare or absent in slowly expanding chronic lesions (78). Instead, such lesions are characterized by the presence of a rim of activated microglia at the lesion edge. Only few of these microglia cells contain myelin degradation products. T cell infiltrates are present but sparse and mainly located perivascularly. Active demyelination occurs in close contact to activated microglia, which may even form microglia nodules (78). Although some deposits of complement components were noted, reactivity for the terminal lytic complement complex on myelin was absent (16, 78).

Cortical demyelination.   Cortical lesions in progressive MS differ from white matter lesions in several fundamental aspects (12, 13, 47, 71). They may appear as small intracortical perivascular lesions, or in continuity with subcortical white matter plaques. The most abundant form of cortical demyelination, however, is subpial demyelination, which appears as large band‐like lesions extending from the outer surface of the cortex into its deeper layers (Figure 1; 71). Such lesions are mainly located in cortical sulci and are particularly abundant in deep indentations of the brain surface, such as the insular, the cingulate, the frontobasal and temporobasal cortices and the cerebellum (50, 52). Within the intra‐ and subpial cortical lesions, essentially no T cells or B cells were found; however, there was profound microglia activation (13). This observation, however, is only based on lesions formed in the late progressive stage of the disease. Whether cortical lesions, which can be seen sometimes also in the acute and relapsing stage, show more profound inflammation is currently not known. Inflammation, composed of T cells, B cells and plasma cells, is present in the meninges, which cover the surface of cortical lesions (51). These data suggest that subpial demyelination in the cerebral cortex in progressive MS may be driven by a soluble factor, which is produced by inflammatory cells within the meninges and induces demyelination either directly or indirectly through microglia activation. In autoimmune encephalomyelitis, closely similar cortical lesions can be seen in selected models (64, 73). In EAE, subpial cortical demyelination only develops when the disease is driven by T cells and demyelinating antibodies, and myelin sheaths at sites of demyelination are dressed by precipitates of immunoglobulin and activated complement (64). Whether antibodies are also implicated in the pathogenesis of MS cortical lesions is currently unresolved. The absence of complement reactivity was described in one study (16), although there are some remaining doubts regarding the activity of the respective lesions.

Diffuse white matter injury.  As described above the NAWM is highly abnormal in MS patients (3), in particular in patients in the progressive stage of the disease. Changes in the NAWM consist of a diffuse inflammatory process (51). Inflammatory infiltrates are present in the perivascular space and are also dispersed throughout the tissue. They mainly consist of T lymphocytes, the majority are CD8⁺ MHC Class I‐restricted cells. Inflammation is associated with profound microglia activation. Microglia cells, frequently forming clusters or nodules (Figure 1), express a variety of activation antigens, including MHC antigens, markers for phagocytic activity and in particular footprints of radical production. This profound microglia activation is associated with diffuse axonal injury and loss throughout the NAWM. Importantly, diffuse white matter pallor, a characteristic feature of the pathology of progressive MS, is secondary to axonal destruction and not to primary demyelination.

For a long time, the diffuse changes in the NAWM of MS patients, previously defined as MS encephalopathy (45), was regarded as a secondary consequence of axonal destruction in focal white matter lesions. There is no doubt that secondary Wallerian degeneration occurs in the MS brain (27, 29), and likely contributes to secondary tract degeneration in the NAWM. However, the extent of global injury in the NAWM does not correlate with the number, size, location and destructiveness of focal white matter lesions, both in the brain (51) and in the spinal cord (29). Pronounced inflammation and microglia activation within the NAWM markedly exceeds that seen in classical Wallerian degeneration, and suggests that diffuse axonal injury in the NAWM in MS occurs at least in part independent from focal white matter plaques.

The mechanisms, leading to diffuse injury of the NAWM, are so far poorly understood. It seems to be related to chronic microglial activation, possibly driven by the inflammatory process. The pronounced expression of i‐NOS and myeloperoxidase within activated microglia suggests that oxygen and nitric oxide radicals are involved in this process. A recent study used microarrays to study gene expression in the NAWM of MS patients and compared it with the gene expression profile in the white matter of controls devoid of neurological disease (38). One upregulated gene cluster was involved in hypoxic preconditioning. Another microarray investigation revealed a decreased expression of genes involved in mitochondrial energy metabolism (26). Furthermore, the selective vulnerability of small diameter axons in MS lesions (28) suggests that energy deficiency may represent an important pathogenic element in axonal destruction (26, 85). One mode of action in neurotoxicity, mediated by nitric oxide and oxygen radicals, is the induction of mitochondrial dysfunction by inhibition of enzymes of the respiratory chain (14). Taken together, these findings indicate that radical mediated mitochondrial injury may be an important factor driving progressive axonal dysfunction and loss in MS.

REMYELINATION

The hallmark of MS pathology is the demyelinated plaque with partial axonal preservation. Remyelination, however, occurs in MS lesions either at the peripheral margins of the plaques or even within the whole white matter lesion (62, 74, 77). Complete remyelination gives rise to so‐called “shadow” plaques (81), which are sharply demarcated areas with reduced myelin density and disproportionately thin myelin sheaths. Within active plaques arising during the early stages of MS, extensive remyelination is frequently encountered, which appears to be accomplished via oligodendrocyte progenitor cells recruited to the site of demyelination (58, 76, 79). However, at later stages in the disease, and in particular in patients with progressive MS, remyelination is believed to be sparse or absent. Thus, it was suggested that remyelination, occurring at the early stages of MS, is unstable (35). As chronic demyelination is associated with slowly progressive axonal loss (49), stimulation of remyelination by growth factors or through (stem) cell transplantation is considered an attractive therapeutic option in MS patients.

However, recent evidence challenges this view. Correlating pathology with MRI in MS lesions revealed that more than 40% of MS lesions showed signs of remyelination (8). In addition, we found that in approximately 20% of MS patients, remyelination is so extensive that almost all plaques within the CNS are shadow plaques (70). Unexpectedly, remyelination was more extensive in patients dying at an old age, and extensive remyelination was not restricted to patients with relapsing disease, but was also found in those with SPMS and PPMS.

Little is known why remyelination is extensive in some patients, while it fails in others. Lesion location is likely an important factor, as lesions arising in the subcortical or deep white matter have a higher remyelination potential compared with those present in the periventricular white matter (70). Furthermore, remyelination also depends upon the availability of axons within the plaques and their ability to become remyelinated. In some lesions, axons express PSA‐NCAM, a molecule, which inhibits myelination during embryonic development (19). However, in the series of Patrikios et al (70), patients segregated into two distinctive groups, one with extensive and the other with sparse remyelination. Whether the potential for remyelination is, at least in part, dependent upon genetic factors needs to be determined. Unfortunately, there are no magnetic resonance parameters, which allow for the distinction between remyelinated and demyelinated lesions within a given patient. Such paraclinical tools are urgently needed to define patients, who are best suited for remyelination‐promoting therapies.

OTHER INFLAMMATORY DEMYELINATING DISEASES AND THE SPECTRUM OF MS

Pathologically, MS is an entity within a larger group of inflammatory demyelinating diseases (2). Those affecting the CNS include acute disseminated leukoencephalomyelitis (ADLE), Devic’s neuromyelitis optica and Balo’s concentric sclerosis. Despite some pathological similarities, there are specific differences which suggest that the pathogenic mechanisms of nervous system injury may be different.

ADLE.  The differentiation between ADLE and MS is often difficult based on clinical and imaging criteria, particularly in childhood. ADLE is generally considered a monophasic self‐limiting disease, although relapsing variants have been described. In its extreme form, ADLE can lead to fatal acute hemorrhagic leukoencephalomyelitis. Pathologically, ADLE is defined as an inflammatory disease. Parenchymal lesions are mainly reflected by edema and demyelination, which is generally limited to thin perivenous sleeves. Thus, the pathological difference between ADLE and acute MS resides in the extent of demyelination, with large confluent demyelinated plaques more typical of MS. Yet, transitional forms between ADLE and acute MS have been described (2, 33, 62). Furthermore, an acute disease originally fulfilling all criteria of ADLE in a brain biopsy was several months later, at autopsy, found to be classical MS with confluent demyelinating plaques (11). The key question remains as to why in some patients an inflammatory demyelinating disease may cease after a monophasic attack, whereas others progress into classical chronic MS. Pathology has yet to provide any answers to explain this dichotomy.

Devic’s type of neuromyelitis optica.  Devic’s neuromyelitis optica was originally described as an acute monophasic disease selectively involving the spinal cord and optic nerves (25). However, more recently it became clear that relapsing variants of Devic’s neuromyelitis optica are in fact more common, and lesions may extend beyond the spinal cord and the optic systems (72, 92, 93). A typical hallmark of Devic’s neuromyelitis optica is the presence of longitudinally extensive spinal cord lesions, which traverse more than three intravertebral segments. Such longitudinally extensive spinal cord lesions may occur as an isolated or recurrent form of transverse myelitis, or may represent the initial presentation of Devic’s neuromyelitis optica with subsequent relapses of optic neuritis (91).

The pathology of Devic’s disease differs in several key aspects from that of classical MS (60). Inflammatory infiltrates in active lesions contain not only T cells and macrophages, but also granulocytes and eosinophils. Although some primary demyelination is regularly seen, the lesions are highly destructive, showing a profound loss of both axons and astrocytes. Massive complement deposition is seen in the lesion in a characteristic rim and rosette vasculocentric pattern, which colocalizes with astrocytic processes. The intense perivascular inflammatory reaction gives rise to profound vascular alterations and perivascular hyalinization. Autoantibodies against Aquaporin 4 have recently been identified as a specific serological marker in patients with either Devic’s disease or patients with isolated longitudinally extensive myelitis (54, 55, 91). Furthermore, the astrocytic dressing with complement seen in Devic’s neuromyelitis optica is associated with the selective loss of Aquaporin 4 in the respective lesions, independent of stage of demyelinating activity (80). Although experimental proof is still missing, the immunopathology of the lesions strongly suggests that tissue destruction in Devic’s disease is largely mediated by autoantibodies against Aquaporin 4 in conjunction with complement. Whether the disease is mediated by antibodies alone or requires additional T cell‐mediated inflammation is currently unresolved.

Balo’s concentric sclerosis.  Balo’s concentric sclerosis was originally defined as an acute variant of inflammatory demyelinating diseases, giving rise to large concentric lesions in the brain (7). Concentric lesions are characterized by alternating rims of demyelinated and myelinated tissue within the lesions. Although originally described in central Europe this variant has later been seen most frequently in the oriental population (87). Concentric lesions, like in MS, occur on a background of T cell inflammation with macrophage and microglia activation. Large concentric lesions, typical of Balo’s disease, are very rare, but small rims of concentric layering of myelinated and demyelinated tissue is seen in many active lesions of acute and relapsing MS, when they follow a demyelination pattern of hypoxia‐like tissue injury as described above (9, 83). Several mechanisms have been suggested to be responsible for the formation of concentric lesions, including ischemia (23), remyelination (66) or the diffusion of soluble myelinotoxic factors in the gel matrix of the extracellular space (39). In a recent study, the formation of concentric lesions was thought to reflect tissue preconditioning in a radially expanding lesion, induced by hypoxia‐like tissue injury (83). In the areas of demyelination, massive expression of i‐NOS, suggesting radical mediated mitochondrial injury, was found, while in close vicinity of the lesions and within the strips of preserved myelin, proteins involved in (hypoxic) tissue preconditioning were up‐regulated. Thus, in concentric plaques the layers of preserved myelin appear to remain protected against further damage in expanding lesions by preconditioning. All these results indicate that concentric sclerosis is part of the spectrum of MS, which may occur in unusually severe forms of the disease.

CONCLUSIONS

Recent neuropathological studies have provided fundamental new insights into the pathogenesis of MS and have provided new explanations for the diverse spectrum of lesions seen in this disease. It provides clear evidence that whenever active tissue destruction is seen in this disease, it occurs on the background of inflammation. Thus, neuropathology does not support the concept, that there is a neurodegenerative component in the pathogenesis of MS, which develops independently from inflammation. However, the nature of the inflammatory response is different between the acute or relapsing stage and the progressive stage of the disease. While in the former, new waves of inflammation derived from the peripheral immune system occur in the brain and spinal cord, and give rise to focal plaques of demyelination mainly located within the white matter, inflammation ultimately becomes trapped behind a closed or repaired blood brain barrier in the progressive stage of the disease. This compartmentalized inflammation drives the slow expansion of preexisting demyelinated lesions, as well as diffuse (mainly axonal) injury in the NAWM. When trapped in the meningeal compartment it may induce widespread subpial demyelination in the cerebral and cerebellar cortices. In addition, the dominant mechanisms of inflammatory tissue injury are diverse between patients, resulting in profoundly different phenotypes of focal demyelinating plaques. This destructive process can be counterbalanced by remyelination, which is extensive in a subset of MS patients. It is unlikely that in such a complex pathogenetic scenario a therapy, targeting a single immune mechanism, will be of major benefit for all MS patients at all stages of the disease.

ACKNOWLEDGMENTS

This study was in part funded by the US National Multiple Sclerosis Society (Grant RG 305).