OUP user menu

Devic’s disease: bridging the gap between laboratory and clinic

Ralf Gold, Christopher Linington
DOI: http://dx.doi.org/10.1093/brain/awf147 1425-1427 First published online: 1 July 2002

Our first insights into the pathogenesis of multiple sclerosis were obtained following developments in histopathology in the late 19th Century. More than a century later, despite the advanced molecular tools that are now available, significant progress in unravelling the pathogenesis of multiple sclerosis is still very much dependent on similar approaches. The inaccessibility of the CNS for direct study, and the clinical and genetic heterogeneity of multiple sclerosis combine to make the study of multiple sclerosis difficult to say the least (Compston et al., 1998). Recently Hans Lassmann and his colleagues Lucchinetti and Brück proposed a classification for multiple sclerosis lesions based on molecular histopathological findings from diagnostic biopsies and autopsies (Lassmann et al., 2001), an approach similar to that established by Prineas and Raine in the 1970s and ’80s (Raine et al., 1997). Lassmann’s patterns I–IV of demyelination and inflammation are not closely linked to specific relapsing–remitting or progressive disease courses, but may help us to understand the underlying immunological and/or neurodegenerative mechanisms involved in lesion formation. There is a pressing need to identify specific histopathological or paraclinical markers for the different pathogenic mechanisms that may be involved in multiple sclerosis. Only then will we be able to identify the most appropriate ‘customized’ therapeutic regimen for any individual patient using the growing repertoire of immunomodulatory and cytostatic drugs that are now available (Noseworthy et al., 2000).

In this issue of Brain Lucchinetti and colleagues (Lucchinetti et al., 2002) provide detailed molecular studies on more than 80 lesions from nine cases of neuromyelitis optica (NMO; Devic’s disease) using essentially the same criteria as in their paper on the classification of multiple sclerosis lesions (Lassmann et al., 2001). Devic’s disease, originally described as monophasic attack of optic neuritis and myelitis, may pursue a relapsing–remitting course causing severe neurological disability (Wingerchuk et al., 1999). We now know that MRI scans of the head are mostly normal in NMO. Many of us have also seen patients in which this is the case but where the onset of optic symptoms and myelitis are separated by weeks or even months, suggesting that many of these patients may also have NMO. These unique clinical parameters of NMO are now supplemented by histopathological findings. Strikingly, all early demyelinating lesions in NMO are associated with perivascular deposition of immunoglobulins, in particular IgM, local activation of the complement cascade and a marked eosinophilic infiltrate. This combination of features is relatively specific for early NMO lesions, but is also accompanied by immunopathological changes in the CNS such as macrophage/microglia activation and axonal damage that are ubiquitous in all forms of multiple sclerosis. The immunopathology of NMO is highly suggestive of an antibody‐dependent, complement‐mediated pathogenesis extenuated by the recruitment and degranulation of eosinophils. The presence of eosinophils and small numbers of CCR3 positive lymphocytes may indicate the active involvement of a T helper (Th)‐2 T cell subset response in NMO, but this requires confirmation. In particular future studies have to demonstrate the presence of eotaxin and IL‐4 within these early lesions.

The animal model for multiple sclerosis, experimental autoimmune encephalomyelitis (EAE), may well help us to understand how these immunological effector mechanisms, in particular autoantibodies, contribute to the pathogenesis of NMO. Augmentation of demyelination in EAE in the Lewis rat by circulating antibodies directed towards myelin oligodendrocyte glycoprotein (MOG) was described in the late 1980s (Linington et al., 1988). Since then a wide array of experimental models have focused on MOG as autoantigen and generated variants of EAE in rats and mice which more closely mimic the complex of pathologies seen in multiple sclerosis (Gold et al., 2000).

All the salient immunopathological features of NMO are reproduced in Brown Norway (BN) rats immunized with MOG (Stefferl et al., 1999). MOG‐induced EAE in this strain follows a fulminant clinical course associated with widespread demyelination and axonal injury. The lesional distribution also exhibits a preference for the optic tract (Fig. 1) and spinal cord similar to that seen in NMO. More importantly, the inflammatory infiltrate contains large numbers of eosinophils and demyelination is associated with complement deposition. Recruitment of eosinophils into the CNS was not seen in demyelinating lesions induced by the co‐transfer of MOG‐specific antisera and encephalitogenic Th‐1 MOG‐specific T cells in BN suggesting that eosinophil recruitment was dependent on a MOG‐specific Th‐2 T cell response. However, it was not possible to identify a clear highly polarized MOG‐specific Th‐2 T cell response. In particular, no enhanced secretion of the classical Th‐2 associated‐cytokine IL‐4 was observed and the MOG‐specific antibody response included both Th‐1‐associated (IgG2b) and Th‐2‐associated (IgG1 and IgE) isotypes.

Fig. 1 Coronal section of basal forebrain from rat with MOG‐induced EAE. Note almost complete demyelination of the right optic nerve, whereas the left optic nerve only exhibits small demyelinated foci (arrows). Figure courtesy of Dr Maria Storch, Graz.

These findings demonstrate that even in inbred strains with a high susceptibility for Th‐2 driven autoimmune responses, it is difficult to disentangle the effector pathways involved in lesion formation and ultimately tissue destruction. As yet all that is known about the molecular basis for the induction of this NMO disease model is that while the eosinophilic component is controlled by non‐MHC genes, the intensity and pathogenicity of the antigen‐specific autoimmune response is determined by multiple genes within the MHC. Does this mean that susceptibility to NMO in man is determined by a similar interplay between ‘background’ and MHC loci? Just as unclear is why the immune system in NMO selectively attacks the optic tract and spinal cord while sparing the brain. As these regions of the CNS are not thought to possess a different repertoire of glial cells or myelin proteins, is it due to differences in the structure/stability of the blood–brain barrier as discussed by the authors, or alternatively due to subtle regional differences in the ability of the CNS to process and present antigen to T cells?

Nevertheless, despite our limitations in understanding the genetics, immunology and cellular biology of NMO, the findings of Lucchinetti and colleagues are of substantial therapeutic importance. NMO is associated with clearly defined MRI and CSF abnormalities that should allow us to rapidly identify these patients. The current study clearly demonstrates that humoral effector mechanisms have a central role in NMO, and acute therapeutic interventions may focus on interrupting antibody/complement‐dependent effector mechanisms. This could include plasmapheresis in cases where glucocorticosteroids are ineffective (Keegan et al., 2002), intravenous immunoglobulins or complement inhibitors such as soluble CR‐1. Long‐term immunotherapy of NMO may be more problematic as drugs such as glatiramer acetate have the potential to augment Th‐2‐type responses (Duda et al., 2000). In this case it may be worthwhile to consider treatment with highly active immunosuppressants such as mitoxantrone (Noseworthy et al., 2000) early in disease, reducing the immunotherapeutic regimen as soon as one has achieved stabilization or remission.

Currently it is not possible to identify any of the specific immunopathological features associated with variants of multiple sclerosis such as NMO using non‐invasive imaging techniques, although developments in MRI techniques may soon bridge this gap. Once this goal is achieved it will be possible to identify the pathomechanisms driving disease activity and nourishes the hope that ‘personalized’ therapeutic approaches can be developed for defined multiple sclerosis subtypes.

References