Journal of Neurobnmunology, 40 (1992) 197-210

197

,~ 1992 Elsevier Science Publishers B.V. All rights reserved 0165-5728/92/$05.00 JNI 90008

Inflammatory mediators in demyelinating disorders of the CNS and PNS H a n s - P e t e r Hartung ~, Stefan Jung a, Guido Stoll b, Jiirgen Zielasek ", Beate Schmidt ~, Juan J. Archelos a and Klaus V. Toyka a '~Department of Neurology, Julius-Maximilians-Unit,ersitiit, Wiirzburg, German),, and h Department of Neurology, Unicersity of Diisseldo~ Diisseldo~ German)'

Key words: Inflammatory demyelination; Central nervous system: Peripheral nervous system; Multiple sclerosis; Guil[ain-Barr6 syndrome; Experimental allergic encephalomyelitis; Experimental autoimmune neuritis; Effector molecule

Summary Work in both experimental models and human disorders of the central and peripheral nervous system has delineated multiple effector mechanisms that operate to produce inflammatory demyelination. The role of various soluble inflammatory mediators generated and released by both blood-borne and resident cells in this process will be reviewed. Cytokines such as interleukin (IL)-I, interferon (IFN)-y, and tumor necrosis factor (TNF)-a are pivotal in orchestrating immune and inflammatory cell-cell interactions and represent potentially noxious molecules to the myelin sheath, Schwann cells, and/or oligodendrocytes. Arachidonic acid metabolites, synthesized by and liberated from astrocytes, microglial cells and macrophages, are intimately involved in the inflammatory process by enhancing vascular permeability, providing chemotactic signals and modulating inflammatory cell activities. Reactive oxygen species can damage myelin by lipid peroxidation and may be cytotoxic to myelin-producing cells. They are released from macrophages and mieroglial cells in response to inflammatory cytokines. Activation of complement yields a number of inflammatory mediators and results in the assembly of the membrane attack complex that inserts into the meylin sheath-creating pores. Activated complement may contribute both to functional disturbance of neural impulse propagation, and to full-blown demyelination. Proteases, abundantly present at inflammatory loci, can degrade myelin. Vasoactive amines may play an important role in breaching of the blood-brain/blood-nerve barrier. The importance of nitric oxide metabolites in inflammatory demyelination merits investigation. A better understanding of the multiple effector mechanisms operating in inflammatory demyelination may help to devise more efficacious antigen non-specific therapy.

Introduction Correspondence to: H.-P. Hartung, Department of Neurology, Julius-Maximilians-Universit~it, Josef-Schneider-Strasse 11, 8700 Wiirzburg, Germany.

Experiments in cell cultures, immunocytochemical studies of diseased tissue obtained at biopsy or autopsy, and immunological observa-

I gh

tions in experimental animals and in patients have recently delineated common effcctor mechanisms that operate in immunc-mediated demyelinating disorders of both the central and pcripheral nervous system (CNS, PNS). This short review attempts to summarize some of the evidence that implicates inflammatory mediators in the pathogenesis of expemnental autoimmune encephalomyelitis (EAE), experimcntal autoimmunc neuritis (EAN), multiple sclerosis (MS), and the Guillain-Barrd syndrome (GBS). Bloodbornc cells (T cells, m o n o c y t e s / m a c r o p h a g e s , neutrophils) invading the CNS or PNS. and activated resident cells (microglia, astrocytes, pericytes, mast cells, endothelial cells, endoneurial macrophages) can elaborate an array of potentially injurious molecules which usually act lit short range. These include reactive oxygen species, nitric oxide metabolites, eicosanoids, complement components, proteases, vasoactive amines, and cytokines (reviewed in Gallin et al., 1092: Henson and Murphy, 1989). As an example, the multiple actions of macrophages are illustrated in Fig. 1.

oxidative burst i

02 NADPH oxidase

/ OH

superoxide dismutase H202 catalase 02

H20

Fig. 2. Oxidative bur~t. Neutrophib,. macrophage~, and nilcroglial cells are endowed with a metabolic path~ay that. upon membrane perturbation, generates reactive oxygen species from molecular oxygen. The ultimate toxin is sup posed to be OH. produced upon interaction of superoxide anion ( O , ) and h,,drogen peroxide I H ~O • ).

Reactive oxygen species Upon perturbation of their surface membranes by soluble or phagocytic stimuli, neutrophils,

MACROPHAGE

I ICYTOLYSIS

INFLAMMATION INTERLEUKIN 1 [NTERLEUKIN 6 COMPLEMENT FACTORS CLOTTING FACTORS ARACHIDONIC ACID METABOLITES PROTEASES

/,/ ~ IPROTEASES

\' / f

.... " "

l,

''

}

I['40"

i T LYMPHOCYTE ACTIVATION ANTIGEN PROCESSING AND PRESENTATION INTERLEUKIN I

Fig. 1. Central role of macrophages. Macrophage.', and mlcroglial cells arc chief effector celb, of inflammatoo demyclination. They can generate a host of pro-inflammatory molecules. The_,,' can also exert cytotoxic activity by direct physical contact or through release of toxic bioproduct3.

m o n o c y t e s / m a c r o p h a g e s , and microglial cclls undergo a rapid increase in extramitochondrial oxygen consumption and generate toxic oxygen radicals (e.g. superoxide anion, hydrogen peroxide, hydroxyl radicals) (Fig. 2: Colton and Gilbert, 1987; Nathan, 1987: Sonderer et al., 1987). In EAN, elevated release of ox3,'gen radicals from monocytes lind macrophages has been shown (Hartung ct al., 1988a; Stevens et al., 1990) by superoxide anion and hydrogen peroxide production, and chemiluminescence. A pathogenic role was proven by pharmacological experiments in which oxygen radical scavengers (superoxide dismutase and catalase) were administered to animals at various times after induction of EAN. Both scavengers suppressed RAN when given before signs of disease were evident and greatly ameliorated the disease whet] commenced after the onset of clinical signs (Hartung et al., 1988a). In EAE. the iron-chelating agent desferrioxamine that scavenges oxygen radicals was to be found

199

effective in down-regulating disease activity (Willenborg et al., 1988). In canine distemper encephalitis, circulating anti-myelin antibodies from serum or CSF caused a marked secretion of reactive oxygen species in brain macrophages (Griot et al., 1989). In GBS, blood monocytes isolated early in the course of the disease generated much larger amounts of oxygen radicals in response to phorbol diester than controls (Hartung and Toyka, 1990). Reactive oxygen species can damage myelin by lipid peroxidation and have indeed been found to degrade myelin in vitro (Chia et al., 1983: Konat and Offner, 1983; Konat and Wiggins, 1985). Moreover, they can injure endothelial cells and may thereby disturb the blood-brain or bloodnerve barrier. In dog glial cell cultures exposed to xanthine/xanthine oxidase, a system capable of generating superoxide anion, a selective degeneration of oligodendrocytes was noted (Griot et al., 1990).

J nitric oxide pathway] L-arginine FB~I FAD hydroxy-L-arginine N- monomethylarginine

1

(?) citmlline cytotoxicity against tumor cells and parasites: pro - inflammatory? carrier molecules?

nitrite nitrate nitrosamines

guanyl!l cyclase t

Jo. t -cardiovascular actions (vasodilation) -neurotransmitter activity anterograde and retrograde) smooth muscle relaxation)

Nitric o x i d e m e t a b o l i t e s

Nitric oxide and its metabolites are generated in a L-arginine-dependent biosynthetic pathway (Fig. 3). They have recently attracted attention because of their pleiotropic properties which encompass their actions as an endothelium-derived relaxing factor and as neurotransmitters. Furthermore, they represent major defense molecules of immune cells against microbes and malignant cells (Kolb and Kolb-Bachofen, 1992). This biosynthetic pathway utilizing the NO synthase is present in macrophages. We recently investigated whether neonatal rat microglial cells can also generate nitric oxide metabolites. Upon stimulation with lipopolysaccharide (LPS) or interferon (IFN)-7, neonatal rat microglial cells secreted nitrite (Fig. 4) (Zielasek et al., 1992). Nitric oxide production could be inhibited with NgMMA, a structural analog of the precursor amino acid L-arginine and competitive inhibitor of macrophage NO synthase. It is conceivable that microglial cells upon activation by autoreactive T lymphocytes secrete NO in vivo, which might damage myelin or nearby CNS cells. Peripheral blood mononuclear cells and neutrophils from

NO synthase (constitutive or inducible form)

o

N-monomethylarginine

l

Fig. 3. Nitric oxide pathway. Macrophages and microglial cells can generate from L-arginine: nitric oxide, nitrite, nitrate, molecules with pleiotropic effects that may engage not only in microbicidal and cytostatic action but also in inflammation. BH 4, tetrahydro-biopterin: FMN, flavin mononucleotide: FAD, flavin adenine dinucleotide. 5045 4O 35

.~ 30 N 25 ~

20 15

0-

o

o.1

1 G-IFN (U/ml)

lO

lOO

Fig. 4. Nitrite production by microglial cells. Neonatal rat microglial cells were incubated in the presence of indicated concentrations of recombinant rat IFN-~, (G-IFN). Nitrite production was quantitated by the GrJess reaction after incubatkm for 48 h (cf. Zielasek et al., 1992).

2{Itl

rats with hyperacute EAE have recently been found to generate increased amounts of both reactive nitrogen and oxTgcn intermediates. Secretion was furthcr enhanced when these cclls were exposed to antigen-stimulated myelin basic protein-reactive T cell lines (MacMicking et al., 1992). Future studies shoul{-I determine whether inhibition of nitrite/" nitric oxide production in rive may abrogate the development of autoimmune inflammatory and demyelinating nerxous system diseases.

Eicosanoids Arachidonic acid metabolites, the prostaglandins and thromboxanes derived from the cyclooxygenase pathway and the leukotrienes and H E T E s (hydroxyeicosatetraenoic acids) generated in the lipoxygenase pathway are potent inflammatol3,' mediators (Salmon and Higgs, 1987: Lewis et al., 1990). The main source of eicosanoids are macrophages. In the CNS, astrocytes and microglia are also capable of producing products of both pathways (Fontana et al., 1987; Hartung and Toyka, 1987a,b; Hartung et al., 1988b, 1989; Murphy et al., 1988). Stimuli that precipitate cicosanoid release include LPS, interleukin-1/3 and tumour necrosis factor-a (TNF-a). Arachidonic acid derivatives are instrumental in turning on and maintaining inflammaton.,' responscs: they function as chemoattractants for neutrophils and monocytes/" macrophages, cause adherence of platelets and leukocytes to the cndothelium, enhancc vasopermeability, promote edema, thereby allowing influx of proteinaccous fluid and ingress of leukocytes which they, in turn, also stimulate to release additional compounds of pro-inflammatory potential, such as oxygen radicals and hydrolytic enzymes (Salmon and Higgs, 1987: Lewis et al., 1990). Furthermore, eicosanoids modulate functional activities of T and B lymphocytes as well as of macrophages, contributing to immunoregulation. In EAE, increased CNS levels of PGE were found (Bolton et al., 1984a,b). LTC4 and LTD4, in particular, have been demonstrated to breach the b l o o d - b r a i n barrier (cf. Hartung and Toyka, 1987b). Dual cyclooxygenase and lipoox3,'genase blockade by pharmacological corn-

pounds such a,,, BW755(" attcnuatc EAE (Simmons ct al., 1c)92L In MS, increased P(iE plt)duction by peripheral blood monocytcs cx rive ha, been noted. In one report, this was correlated with disease exacerbation, PGE release pcaking shortly betZ~re exaccrbation and dropping precipitously during full discase activity (Dorc-Duffy cl a[., 1985. 1986: Merrill et al.. 1t~89). CSF Icxcls of PGE were recorded as elevatcd in one and lox~ercd in another study, but [cukotricnc ('4 levels were not increased Ireviewcd in Hartung and Heiningcr, 1989: Merrill ct al.. 1983). Enhanced immunoreactivity for both IL-I and P(}E localizcd to astrocytes has been observed in multiple sclerosis plaques (Hofmann ct al., 198(~). Ill EAN, macrophages were noted to rclcasc increased amounts of PGE and LT(/4 (Hartung ct al., 1988c). Pharmacological inhibition of eicosanoid production suppressed or markedly attenuated disease severit$ as shown by both by clectrophysit)[og} 1t111_t histological exanlination. In these intcrventional studio,,,, administration of the pharmacological Jnhibitors of arachidonatc ccm\'cr~ion at an early stagc of L,-XN resulted in a grcatl.'y re{-luccd nunlbcr of ED1 + illacrl)phages inlilti-ating nerve. However. when drug adnlinistration \vas instituted after ttlC appearance of o\crt llCUrologlcal sign,s and electrophysiolc)gical c~idcncc of nerve dysfunction, cellular infiltrates at day 21 p.i. were pi-c,~cnt in the nerve although to a much lesser degree than fotind in .sham-treated con-

trois. The benefical cffccts of thc agents tit this stage could therefore not be solely attributed to blockade of macrophage influx into the lesion. It is likely that those macrophages prcscnt in thc nerve could no longer generatc arachidonic acid derived chemotactic signals, and hence further accumulation of cells with cnstiing amplification of inllamrnatory damage was prevented (Hartung ct al.. 1988c).

Proteases There are indications that in MS brains a number of proteases are increased in concentration with some being localized to astrocytes. Myelin injury results in proteolytic degradation of

201 MBP which can be detected in CSF (reviewed by Beyer and Whitaker, 1985). Neutral proteases can degrade myelin in vitro (Beyer and Whitaker, 1985). In one study, using freshly isolated myelin, a cooperative action of complement, plasminogen and macrophagesecreted plasminogen activator was required to initiate demyelination (Cammer et al., 1989). Since local effector cells in the CNS can synthesize all of these compounds, such a mechanism of myelin injury may be relevant to some diseases. It should be noted that myelin contains endogenous proteases that under appropriate conditions could induce MBP degradation (Beyer and Whitaker, 1985). A pathogenic role of proteases is also suggested by the efficacy of protease inhibitors to suppress EAE. Likewise, in EAN increased protease contents were found in diseased nerves (Sobue et al., 1982). Macrophage-derived neutral proteases and phospholipases have been implicated in myelin damage in vitro (Cammer et al., 1978; Trotter and Smith, 1986). Microinjection of protease K into rat sciatic nerves in situ produced inflammatory demyelination (Westland and Pollard, 1987). Treatment of EAN with protease inhibitors delayed development of the disease to some extent (Schabet et al., 1991).

Complement Upon activation, the complement system produces a number of ligands that bind to surface receptors on inflammatory cells. The terminal complement complex (TCC, also called membrane attack complex) might insert into the myelin sheath and damage cells by pore-formation. The major source of complement at inflammatory loci are macrophages, but astrocytes and microglia have also been shown to synthesize at least some components of complement (Levi-Strauss and Mallat, 1987). C9 neoantigen has been detected in the lesions of EAE, EAN, MS, and GBS (reviewed by Compston, this issue; Compston et al., 1991). Evidence for activation of the complement system has been adduced by detecting C9 consumption in CSF of MS patients, increased concentrations in plasma and CSF of the soluble membrane attack complex in MS and GBS (Fig.

COMPLEMENT ACTIVATION IN GBS s C 5 b - 9 (pg/ml)

HC (n-20)

GBS (n-38)

ON (n-28)

Fig. 5. Complement activationin Guillain-Barr~syndrome.To search for evidenceof complement activationin Guillain-Barrg syndrome, plasma samples were taken within 4 days of disease onset and assayed for the presence of soluble terminal complement complex (sC5b-9) by ELISA. HC. healthy controls; ON, patients with other non-inflammatory,non-demyelinating neurological disorders, n, number of patients.

5; Compston et al., 1986, 1991; Sanders et al., 1986; Koski et al., 1987, Koski, 1990) and of the low molecular mass anaphylatoxic peptides C3a and C5a in the CSF of MS and GBS patients (Hartung et al., 1987). We have found TCC deposited on the surface of Schwann cells and their myelin sheath and to some extent in the extracellular space at sites of impending demyelination before the onset of clinical signs of EAN and for a short period thereafter (Fig. 6; Stoll et al., 1990). Interestingly, no TCC immunoreactivity was seen in the distal stump of transected sciatic nerves after axotomy. This argues against nonspecific activation by macrophages which are abundantly present in Wallerian degeneration. The early transient deposition of TCC on Schwann cells and myelin sheaths before overt demyelination suggested that complement activation played a pathogenic role in the initiation of immune-mediated myelin damage. This notion was supported by the finding that complement depletion by administration of cobra venom factor delayed the onset of EAN and EAE (Feasby et al., 1987; Linington et al., 1989). The terminal complement complex is also required for de-

2~)2

;.~_~..~,,~ _~ ~ ~.-

~. ~ ~ ~ ' ~ ' ~; . '2 ~ ~~~--'~

~,~.~-,, -.:.,,.,

~ ~_

.~:.

~~

:~ ~ - - z - y . --,

... ~ ~,i.~ ~.~

~,~~".,~ • ~"~,~ ~"~_~,,~.~ - ' ~ ' ~ ~ - U"~:" : ~ L ~ ........~ ~ , ~ -

"

~i.~-

E

,~

~ ~

. ~

:

-

I ..~::"

~i~'.; . ~7---

i_~-

Fig. 6. lmmunocytoehemical localization of the terminal complement complex (TCC) in experimental autoimmune neuritis before overt clinical disease and demyelination. Ventral roots from day 11 after immunization were stained with a polyclonal rabbit anti-C5b-9 antibody. The reaction product (arrows) is localized on the surface of myelin sheaths and myelinating Schumann cells (*) (A-E). (A) On a cross-section, the TCC-labelling predominantly appears on nerve fibers near venules (arrows) and other areas of the nerve root are spared. (B-E) Longitudinal sections. Note that single nerve fibers are labelled (arrows in B, C, E), whereas neighboring ones are not (arrowheads). Between these fibers, additionally, TCC-positive cells (open arrows) can be seen. The distribution of these cells is similar to that of W3/13-positive leukocytes as shown in (F) (arrows). Scale bar = 10 # m (from Stoll et al. (1992), with permission of the Editor).

21)3 myelination of organotypic cultures of dorsal root ganglion cells to occur when serum from GBS patients is added (Sawan-Mane et al., 1991). Oligodendrocytes underwent reversible injury upon exposure to complement in vitro. The relevance of these findings is discussed elsewhere (Compston, this issue; Compston et al., 1991). What are the mechanisms involved in complement activation and complement-mediated destruction of myelin? Apart from antibodies, a 56-kDa protein of central myelin and the PO protein of peripheral myelin can directly activate the complement system (Koski, 1990). TCC insertion into the myelin membrane opens transmembrahe pores (Shin and Carney, 1988). This might permit an influx of proteases such as those secreted by macrophages/microglia, which could then break down MBP. Transmembrane flow of calcium could activate calcium-dependent proteases, including integral myelin proteases, again with ensuing cleavage of MBP, and myelin splitting and vesiculation at the interperiod line. Transient opening of the myelin membrane may locally perturb the ionic milieu and thereby impede nerve impulse propagation. This may be one possible mechanism for transient functional disturbances encountered in MS patients. C3a, C5a and TCC stimulate phospholipase A 2 and transmethylation whereas they inhibit acyltransferase (Shin and Carney, 1987). As a consequence, eicosanoids are generated which can further amplify inflammatory responses (see above). Activation of pericellular (macrophage-, microglia-, and astrocyte-derived) complement by accessible myelin components might start a positive feedback loop resulting in potentiated mye,lin injury. These complex interactions could account for the pattern of myelin destruction, i.e. vesicular dissociation of myelin in immediate proximity to macrophages and phagocytic attack of the myelin sheath through coated pits. It should be noted in passing that macrophages could, in an antibodydependent, cell-mediated cytotoxic (ADCC) fashion, be armed via the Fc receptor with antibodies against MBP or degradation products and be guided to their target. In addition, they could interact via their C R 1 / C R 3 receptors with C3b deposited along the myelin sheath.

Vasoactive amines Vasoactive amines, histamine, serotonin, and catecholamines augment vascular permeability and promote edema. Histamine is capable of degrading myelin in vitro. On the other hand, MBP and P2 can elicit mast cell degranulation with attendant release of vasoactive amines and proteases (Johnson et al., 1988). Based on these findings and morphological studies, a role for mast cells in demyelination has been proposed (Bo et al., 1991; Dietsch and Hinrichs, 1989, 1991). The relevance of vasoactive amine-induced disruption of the blood-brain barrier is indicated by the suppression of EAE achieved with inhibitors of their release (Goldmuntz et al., 1986; Stanley et al., 1990).

Cytokines Potent inflammatory agents are interleukin-1 (IL-1), interleukin-6 (IL-6), interferon-y (IFN-y), and tumor necrosis factor-a (TNF-c~) (for review see Arai et al., 1990; Cerami, 1992; Nathan and Sporn, 1991). Interleukin-1 is synthesized by macrophages, microglia and astrocytes (cf. Frei, this issue). Its pleiotropic effects include chemotaxis, induction of increased adherence, enhanced vascular permeability, stimulation of the release of platelet-activating factor, and prostacyclin by endothelial cells, and of PGE and TXB2 by neutrophils, macrophages, astrocytes, and microg[ia (Hartung et al., 1989). It also increases expression of adhesion molecules on immunocompetent cells and astrocytes (Frohman et al., 1989; Loughlin et al., 1992). Injection of IL-1/3 into the rabbit eye produces full-blown inflammation in the retina (Martiney et al., 1990) IL-1 immunoreactivity was demonstrated in active MS plaques and elevated IL-1 concentrations were measured in brains of EAE rats (Hofmann et al., 1986; Fontana et al., 1987). Apparently, it is not detectable in the CSF of MS patients (Maimone et al., 1991; Tsukuda, 1991). IL-1/3 also induces TNF-c~ gene expression and production in astroglial cells (Chung and Benveniste, 1990; Bethea et al., 1992). Interferony, predominantly produced by CD4 + T lympho-

204

cytes of the Thl inflammatots, phenotype, also exerts a multitude of inflammatory effects. It upregulates MHC class II and ICAM expression

on macrophages, astrocytes (e.g. Fierz et al., 1985: Sun, 1991), and microglia, enhances vascular permeability and stimulates the release of oxygen

•.,

.,,.,

. :~ . ~ . - ~

. ~ 4~"7. ,.',,. ~..~_

Fig. 7. Localization of IFN-y in EAN before onset of clinical disease. 1 # m thick cryosections of rat [umbosacrat ventral roots on day 12 after immunization. Paired serial sections were stained for IFN-y with the monoclonal antibody DB-I (A, C) and for rat T cells/granulocytes with mAb W3/13 (B), respectively. Note that most cells exhibiting cytoplasmic IFN-), immunoreactivity in (A) are W 3 / 1 3 positive on their surface in (B) identifying them as T cells/granulocytes (arrowheads denote identical cells in (A) and (B). Schwann cells (arrow) were IFN-y-negative. Many IFN-y-positive cells (C) were identified as granulocytes on morphological grounds by their typical segmented nuclei in a subsequent section (D) which was solely stained with hematoxylin and eosin. All section were counterstained with hematoxylin and eosin. Scale bar = 10 # m (from Schmidt et al. (1992). with permission of the Editor).

205 radicals from macrophages. Its pathogenic role has been investigated in detail in EAN. We localized IFN-y in nerves from rats with EAN and T cell line-mediated EAN. Immunoreactivity was mainly associated with T cells and was detectable transiently prior to onset of clinical disease but not thereafter (Fig. 7; Schmidt et al., 1992). In parallel, serum levels of INF-7 transiently increased. A pathogenic role was suggested by two experiments. In one group of animals with EAN, disease severity was greatly enhanced when recombinant rat INF-7 was injected systemically with a massive increase in inflammatory infiltrates and major histocompatibility complex (MHC) class II expression on macrophages and T ceils. Conversely, neutralization of endogenously produced IFN-7 by monoclonal antibody ameliorated disease (Strigard et al., 1989; Hartung et al., 1990). The timing of these interventions was critical corroborating the observation in situ (Fig. 7; Schmidt et al., 1992). Similarly, pharmacological blockade of IFN-7 synthesis down-regulates EAN (Tsai et al., 1991). Apart from promoting cellular influx into the nerve, IFN-y apparently also acts by inducing oxygen radical production in macrophages. Peritoneal macrophages collected from EAN animals were primed to greatly enhanced superoxide anion and hydrogen peroxide release, whereas macrophages isolated from EAN rats treated with the monoclonal antibody to IFN-y had a greatly diminished capacity to generate toxic oxygen species (Hartung et al., 1990). These results stand in apparent contrast to observations made in EAE in which treatment with anti-IFN-7 monoclonal antibodies actually enhanced the disease in mice (Billiau et al., 1988; Voorthuis et al., 1990; Duong et al., 1992). This is surprising in view of the documented effects of IFN-y both in vitro and in vivo. For example, it was shown that following an intracerebral injection of IFN-7 into rat brain, lymphocytes and m o n o c y t e s / m a c r o p h a g e s were massively recruited into brain parenchyma, and MHC class I antigen expression was enhanced on endothelial and ependymal cells whereas MHC class II antigens were found to be increased on microglial, ependymal and perivascular cells throughout the hemispheres (Steiniger and van der Meide, 1988; Vass and Lassmann, 1990; Sethna and Lampson,

1991). Microinjection of IFN-7 into the lumbosacral spinal cord of rats produced severe inflammatory changes (Simmons and Willenborg, 1990). In EAE, a CD4 + suppressor cell population could be isolated from spleens of rats in the recovery phase of the disease that inhibited IFN-7 but not IL-2 production (Karpus and Swanborg, 1989). Apparently, the disturbance of a cytokine network by selective inhibition of one of multiple cytokines can yield unpredictable net results. In a longitudinal study of MS patients, a significant relation was observed between clinical attacks and changes in mitogen-driven synthesis of IFN-7 and TNF-a by peripheral blood leukocytes (Beck et al., 1988). A most telling piece of evidence implicating IFN-y in inflammatory demyelination comes from a clinical trial in which the systemic administration of IFN-y to MS patients resulted in clinical exacerbations, increased numbers of HLA-DR expressing circulating monocytes, enhanced proliferative responses of peripheral blood T cells and NK cell activity (Panitch et al., 1987). TNF-a, synthesized predominantly by macrophages but also by T cells, astrocytes and microglial cells (cf. Frei, this issue; Lieberman et al., 1989; Sawaa et al., 1989), has been linked to the inflammatory demyelinating process of EAN, EAE and MS. In EAN, we could show that TNF-a was transiently present in inflamed nerves but appeared later than IFN-y (Stoll et al., in preparation). Apart from its stimu[atory activity on mediator release from macrophages it has been shown to be myelinotoxic in organotypic cultures (Brosnan et al., 1988; Selmaj and Raine, 1988). Encephalitogenicity of autoreactive T cells has been attributed to their capacity to secrete TNF-a (Powell et al., 1990; Hershvitz et al., 1992), and direct injection of TNF-a into the spinal cord of healthy rats produced meningitis and perivascular cuffing (Simmons and Willenborg, 1990). Intraperitoneal injection of recombinant human TNF-a prolonged and worsened EAE in Lewis rats (Kuroda and Shimamoto, 1991). Antibodies to TNF-a abrogated demyelination in MBP T cell line-mediated EAE of mice. In one study, the different susceptibility of rat strains to develop EAE has been ascribed to differential TNF-a gene expression in astrocytes (Chung et al., 1991). TNF-a was identified immunocytochemically in

2l)t~

the lesions of MS brains (Hofmann et al., 1989: Selmaj et al., 1991a) and aberrant cytokine production by monocytes from MS patients has been reported (Merrill eta[., 1989; Rudick and Ransohoff, 19921. Increased serum and CSF concentrations of this cytokine have been noted in patients with actively progressive MS (Sharief and Hentges, 199l; Trotter et al., 1991). However, other studies had previously failed to detect significantl.v increased TNF-a' levels in serum or CSF (Gallo et al., 1989; Maimone et al., 19911. This may reflect different disease activities of patients under investigation or be due to different technologies used to measure TNF-a'. Lymphotoxin (TNF-/3) has also been implicated in the effector phase of inflammatotT demyelination. First, lymphotoxin was immunocytochemically identified in active MS plaques (Selmaj et aI., 1991b). Secondly, co-incubation of MBPreactive T cell lines with an antibody to lymphotoxin prevented disease upon transfer (Ruddle et al., 1990). The precise mechanisms underlying lymphotoxin effects are yet unknown but may involve nuclear disintegration with DNA fragmentation, i.e. apoptosis (Selmaj et al., 1991c).

Conclusions Work cited in this short review reveals that multiple effector mechanisms operate to produce functional disturbances and tissue damage in inflammato~' demyelinating processes of the CNS and PNS. Early in the course of these disorders, a number of mediators breach the blood-brain or blood-nerve barrier to permit the influx of pathogenic molecules and the migration of inflammatoo: cells from the peripheral immune compartments to the brain and the PNS. Once focussed into the lesion, these blood-borne as well as resident effector cells, upon appropriate stimulation, release additional mediators that further orchestrate cellular interactions and eventually attack oligodendrocytes, Schwann cells a n d / o r the myelin sheath to produce dysfunction and tissue injury. Unfortunately, the pathogenic role of these mediators and factors cannot always be correctly predicted. Time and sequence of action may be more important than previously

appreciated. Antigen non-specific interventions aimed at preventing an amplification of this process by interfering with the synthesis or activities of these mediators should prove useful in the treatment of the human diseases once their mode of action has been defined.

Acknowledgements Work from the authors" laboratories has been supported by grants from Deutsche Forschungsgemeinschaft (SFB 200, B5; Ha 1563/4-1, Bundesministerium fiir Forschung und Technologic (0l KD 9001). and the Hermann and Lilly Schilling Stiftung TSE 013/65). We are indebted to Dr. J.D. Pollard for helpful discussions and appreciate excellent secretarial assistance by Mrs. B. Goebel.

References Arai, K., Lee, F.. Mjyajima, A.. Miyatake, S.. Arai. N. and Yokata, T. (199111 Cytokines: Coordinators of immune and inflammatory responses. Annu. Rex. Biochem. 59, 7S3 836. Beck, J., Rondot, P.. Catinot, L., Falcoff, E., Kirchner, H. and Wietzerbin. J. (1988) Increased production of interferon -y and tumor necrosis factor precedes clinical manifestation in multiple sclerosis: do cytokines trigger off exacerbation>'? Acta Neurol. Scand. 78. 318-323. Bethea, J.R., Chung, I.Y., Sparacia, S.M., Gillespie, G.Y. and Benveniste, E.N. (19921 lnterleukin-1/3 induction of tumor necrosis factor-alpha gene expression in h u m a n astroglial cells. J. Neuroimmunol. 36, 179-191. Be!,er. C.T. and Whitaker, J.N. (19851 Proteinases in inflammatory demyelinating disease. Springer Semin. Immunopathol. 8, 235 250. Billiau, A., Heremans, H_ Vandekerkeho~e, F., Dijkmans, R., Sobis, [-t., Meulepas, g. and Carton. H. (19881 Enhancement of experimental allergic encephalomyelitis in mice by antibodies against IFN-y. J. Immunol. 140, 1506 1510. Bo, L.. Olsson, T., Nyland, H.. Kriiger, P.C., Taule, A. and Mork, S. (1991) Mast cells in brain during experimental allergic encephalomyelitis in Lewis rats. J. Neurol. Sci. 11)5, 135-142. Bohon, C,, Turner. A.M. and Turk, J.B. (1984a) Prostaglandin levels in cerebrospinal fluid from multiple >clerosis patients in remission and relapse. J. Neuroimmunol. 6, 151 159. Bolton, C., Gordon, D. and Turk, J.B. (1984b) Prostaglandin and thromboxane levels in central nervous system tissues

207 from rats during the induction and development of EAE. Immunopharmacology 7, 101-108. Brosnan, C.F., Selmaj, K. and Raine, C.S. (1988) Hypothesis: A role for tumor necrosis factor in immune-mediated demyelination and its relevance to multiple sclerosis. J. Neuroimmunol. 18, 87-94. Cammer, W., Bloom, B.R., Norton, W.T. and Gordon, S. (1978) Degradation of basic protein in myelin by neutral proteases secreted by stimulated macrophages: A possible mechanism of inflammatory demyelination. Proc. Natl. Acad. Sci. USA 75, 1554-1558. Cammer, W., Brosnan, C.F., Bloom, B.R. and Norton, W.T. (1981) Degradation of the P0, P1, and P2 proteins in peripheral nervous system myelin by plasmin: Implications regarding the role of macrophages in demyelinating diseases. J. Neurochem. 36, 1506-1514. Cerami, A. (1992) Inflammatory cytokines. Clin. Immunol. Immunopathol. 62, $3-S10. Chia, L.S., Thompson, J.E. and Morcarello, M.A. (1983) Disorders in human myelin induced by superoxide radical: an in vitro investigation. Biochem. Biophys. Res. Commun. 117, 141-146. Chung, I.Y. and Benveniste, E.N. (1990) Tumor necrosis factor-alpha production by astrocytes. Induction by lipopolysaccharide, IFN-gamma and IL-I beta. J. Immunol. 144, 2999-3007. Chung, I.Y., Norris, J.G. and Benveniste, E.N. (1991) Differential tumor necrosis factor a expression by astrocytes from experimental allergic encephalomyelitis-susceptible and -resistant rat strain. J. Exp. Med. 173, 801-811. Colton, C.A. and Gilbert, D.L. (1987) Production of superoxide anion by a CNS macrophage, the microglia. FEBS Lett. 223, 284-288. Compston, D.A.S., Morgan, B.P., Oleesky, D., Fifield, R. and Campbell, A.K. (1986) Cerebrospinal fluid C9 in demyelinating disease. Neurology 36, 1503-1506. Compston, D.A.S., Scolding, N., Wren, D. and Noble, M. (1991) The pathogenesis of demyelinating disease: insights from cell biology. Trends Neurosci. 14, 175-182. Dietsch, G.N. and Hinrichs, D.J. (1989) The role of mast cells in the elicitation of experimental allergic encephalomyelitis. J. Immunol. 142, 1476-148l. Dietsch, G.N. and Hinrichs, D.J. (1991) Mast cell proteases liberate stable encephalitogenic fragments from intact myelin. Cell. Immunol. 135, 541-548. Dore-Duffy, P., Ho, S.-Y. and Longo, M. (1985) The role of prostaglandins in altered leukocyte function in multiple sclerosis. Springer Semin. Immunpathol. 8, 305-319. Dore-Duffy, P., Donaldson, J.O., Koff, T., Longo, M. and Perry, W. (1986) Prostaglandin release in multiple sclerosis: correlation with disease activity. Neurology 36, 15871590. Duong. T.T., St. Louis, J., Gilbert, J.J., Finkelman, F.D. and Strejan, G.H. (1992) Effect of anti-interferon-y and antiinterleukin-2 monoclonal antibody treatment on the development of actively and passively induced experimental allergic encephalomyelitis in the SJL/L mouse. J. Neuroimmunol. 36, 105-115.

Feasby, T.E., Gilbert, J.J., Hahn, A.F. and Neilson, M. (1987) Complement depletion suppresses Lewis rat experimental allergic neuritis. Brain Res. 419, 97-110. Fierz, W., Endler, B., Reske, K.. Wekerle, H. and Fontana, A. (1985) Astrocytes as antigen-presenting cells. I. Induction of Ia antigen expression on astrocytes by T ceils via immune interferon and its effect on antigen presentation. J. Immunol. 134, 3785-3793. Fontana, A., Bodmer, S. and Frei, K. (1987) Immunoregulatory factors secreted by astrocytes and glioblastoma cell. Lymphokines 14, 91-121. Franciotta, D.M., Grimaldi, L.M.E., Martino, G.V., Piccolo, G., Bergamschi, R., Citterio, A. and Melzi. G.V. (1989) Tumor necrosis factor in serum and cerebrospinal fluid of patients with multiple sclerosis. Ann. Neurol. 26, 787-789. Frohman, E.M., Frohman, T.C., Dustin, M.L., Vayuvegula, B., Choi. B., Gupta, A., van den Noort, S. and Gupta, S. (1989) The induction of intercellular adhesion molecules l (ICAM-1) expression on human fetal astrocytes by interferon-y, tumor necrosis factor a , lymphotoxin, and interleukin-l: relevance to intracerebral antigen presentation. J. Neuroimmunol. 23, 117-124. Gallin, J.J., Goldstein, I.M. and Snyderman, R. (1992) Inflammation: Basic principles and Clinical Correlates. 2rid edn. Raven Press, New York, NY, in press. Gallo, P., Piccinno, M.G., Krzalic. L. and Tavolato, B. (1989) Tumor necrosis factor alpha (TNFc~) and neurological disease: Failure in detecting TNFa in the cerebrospinal fluid from patients with multiple sclerosis. AIDS dementia complex, and brain tumors. J. Neuroimmunol. 23, 41-44. Goldmuntz, E.A., Brosnan, C.F. and Norton, W.T. (1986) Prazosin treatment suppresses increased vascular permeability in both acute and passively transferred experimental autoimmune encephalomyelitis in the Lewis rat. J. Immunol. 137, 3444-3450. Griot, C., Biirge, T., Vandervelde. M. and Peterhans, E. (1989) Antibody-induced generation of reactive oxygen radicals by brain macrophages in canine distemper encephalitis: a mechanism for bystander demyelination. Acta Neuropathol. 78, 396-403. Griot, C.. Vandervelde, M., Richard, A., Peterhans, E. and Stocker, R. (1990) Selective degeneration of oligodendrocytes mediated by reactive ox3'gen species. Free Rad. Res. Comm. 11, 181-193. Hartung, H.P., Schwenke, C., Bitter-Suermann, D. and Toyka, K.V. (1987) Guillain-Barr~ syndrome: Activated complement components C3a and C5a in CSF. Neurology 37, 1006-1009. Hartung, H.P. and Toyka, K.V. (1987a) Leukotriene production by cultured astroglial cells. Brain Res. 435, 367-370. Hartung, H.P. and Toyka, K.V. (1987b) Phorbol diester TPA elicits prostaglandin E release from cultured rat astrocytes. Brain Res. 417, 347-349. Hartung, H.P., Sch~ifer, B., Heininger, K. and Toyka, K.V. (1988a) Suppression of experimental autoimmune neuritis by the oxygen radical scavengers superoxide dismutase and catalase. Ann. Neurol. 23, 453-460. Hartung, H.P., Heininger, K. and Toyka, K.V. (1988b) Pri-

nla~' rat astroglial cultures can generate leukotrienc B4. J. Neuroimmunol. 19, 237 243. Hartung, H.P., Schiller. B.. Heininger, K.. Stoll, G. and Toyka. K.V. ( 1988c~ The role of macrophage,~ and eicosanoids in th'e pathogenesis of experimental allergic neuritis. Serial clinical, electrophysiological., biochemical and morphological observations. Brain I I 1, 103~;- 11)59. Hartung, H.P., Schiller, B., Heininger, K. and Toyka, K.V. (19891 Recombinant interleukin-I/3 stimulates cicosanoid production in rat p r i m a o culture astrocytes. Brain Res. 489. 113 119. Hartung, H.-P. and Heininger, K. (19891 Non-specific mechanisms of inflammation and tissue damage in MS. Res. lmmunol. 1411, 226-233. Hartung, H.P.. Sch~ifer, B., van der Meide, P.H.. Heininger, K. and Toyka, K.V. (19901 The role of interferon-y in the pathogenesis of experimental a u t o i m m u n e disease of the peripheral nervous system. Ann. Neurol. 27. 247-257. Hartung, H.P. and Toyka. K.V. ( 19901 T-cell and macrophage activation in experimental a u t o i m m u n e neuritis and Guillain-Barr~ syndrome. Ann. Neurol. 27 (Suppl.), $57-$63. Hartung. H.P., Stoll, G. and Toyka, K.V. (19921 I m m u n e reactions in the peripheral nervous system. In: P.J. Dyck et al. (Eds. L Peripheral Neuropathy, 3rd edn., Saunders, Philadelphia, PA, in press. Henson, P.M. and Murphy, R.C. (1989) Mediators of the inflammatory process. Handbook of Inflammation, Vol. 6. Elsevier, Amsterdam. Hershviz, R.. Mor, F., Gilat, D., Cohen. I.R. and Lider, O. (19921 T cells in the spinal cord in experimental autoimmune encephalomyelitis are matrix adherent and secrete tumor necrosis factor alpha. J. Neuroimmuol. 37, 161-166. Hofmann, F.M., van Hanwehr, R.I., Dinarello, C.A.. Mizel, S.B. and Merrill, J.K. (19861 Immunoregulalory molecules and Ik-2 receptors identified in multiple sclerosis brain. J. Immunol. 136, 3239-3245. Hofmann, F.M., Hinton, D.R., Johnson, K. and Merrill, J.E. (19891 T u m o r necrosis factor identified in multiple sclerosis brain. J. Exp. Med. 170, 607 612. Johnson, D., Seeldrayers, P.A. and Weiner, H.L. (1988) The role of mast cells in demyelination. I. Myelin proteins are degraded by mast cell proteases and myelin basic protein and P2 can stimulate mast cell degranu[ation. Brain Res. 444, 195-198. Karpus, W.J. and Swanborg, R.H. (1989) CD4 ~ suppressor cells differentially affect the production of IFN-y by effector cells of experimental a u t o i m m u n e encephalomyelitis. J. Immunol. 143, 3492-3497. Kolb. H. and Kolb-Bachofen, ~. (1992) Nitric oxide: a pathogenic factor in autoimmunity. Immunol. Today 13, 157-160. Konat, G.W. and Offner, H. (19831 Density distribution of myelin fragments isolated from control and multiple sclerosis brains. Neurochem. Int. 4, 241 246. Konat. G.W. and Wiggins, R.C. (19851 Effect of reactive oxygen species on myelin proteins. J. Neurochem. 45, 1113-1118. Koski, C i . ( 19901 Characterization of complement-fixing an-

fibodies to peripheral nerve myelin m (_hdilam-Barr5 ~,~,ndrome. Ann. Neurol. 27 (Suppl. L $44-$47. Koski, C.L.. Sanders, M.K.. Swo~eland. P.T.. Lawley, T.S.. Shin, M i . . Frank, M.M. and Joiner, K A . (1c1871 Acti\alion of terminal components of complement in patients ~ith Guillain-Barr~ syndrome and other demyelinating neuropathies. J. Clin. ln~est. 80, 14,92-1497. Kuroda, Y. and Shimamoto, Y. (19911 H u m a n tumor necrosis factor-alpha augments experimental allergic encephalomyelitis in rats. J. Neuroimmunol. 34, 159 165. Levi-Strauss, M. and Mallat. M. (19871 P r i m a l ' cultures of murine astrocytes produce C3 and factor B t~o components of the alternative pathways of complement activation. J. Immunol. 13c;. 2361 2366. Lewis, R.A., Austen, K.F. and Soberman, R.J. (1991D keukotrienes and other products of the 5-1ipoxygenase path~ay. N. Kngl. J. Med. 323, 645-655. kieberman. A.P., Pitha, P.M., Shin. H.S. and Shin. M i . (1989) Production of tumor necrosis factor and other cytokines by astrocytes stimulated with lipopolysaccharide or a neurotropic visur. Proc. Natl. Acad. Sci. LISA 86, 63486352. Linington, C., Morgan, B.P., Scolding, N.J., Wilkins. P., Piddleston, S. and Compston, D.A.S. (19891 The role of complement in the pathogenesis of experimental allergic encephalomyelitis. Brain [12, 895 911. Loughlin, A.J.. Woodroofe, M.N. and Cuzner, M i . (It1921 Regulation of Fc receptor and major histocompatibilt.v complex antigen expression on isolated rat microglia by tumour necrosis factor, interleukin-I and lipopolysaccharide: effects on interferon-y induced activation. Immunology 75, 170 175. MacMicking, J.D.. Willenborg, D.O,, Weidemann, MJ., Rockett, K.A. and Cow,den, W.B. (19921 Elevated secretion of nitrogen and oxs'gen intermediate,,, by inflammato~,,. leukocytes in hyperacute experimental autoimmune encephalomyelitis: E n h a n c e m e n t by the soluble products of encephalitogenic T cells. J. Exp. Med. 176, 3(/3 307. Maimone, D., Grego~', S., Arnason, B.G.W. and Reder, A.T. ( 1991 ) Cytokine levels in the cerebrospinal fluid and serum of patients with multiple sclerosis. J. Neuroimmunol. 32, 67 74. Martiney, J.A., Litwak, M., Berman, J.W., Arezzo, J.C. and Brosnan, C.F. (i9901 Pathophysiologic effect of interleukin-lg4 in the rabbit retina. Am. J. Pathol. 137, 1411 1423. Merrill, J.E., Gerner, R.H., Myers, L.W. and Ellison, G.W. (19831 Regulation of natural killer cell cytotoxicity by prostaglandin E in the peripheral blood and cerebrospinal fluid of patients with multiple sclerosis and other neurological diseases. J. Neuroimmunol. 4, 223-237. Merrill, J.E., Strom, S.R., Ellison, G.W. and Myers, L W . (19891 In vitro study of mediators of inflammation in multiple sclerosis. J. Clin. Immunol. 9, 84-96. Murphy, S., Pearce, B.. Jeremy, J. and Dandona, P. (19881 Astrocytes as eicosanoid-producing cells. Gila 1, 241-245. Nathan, C.F. (1987) Secretors' products of macrophages. J. Clin. Invest. 79, 319-326.

209 Nathan, C. and Sporn, M. (1991) Cytokines in context. J. Cell Biol. 113, 981-986. Panitch, H.S., Hirsch, R.K., Schindler, J. and Johnson, K.P. (1987) Treatment of multiple sclerosis with gamma interferon: exacerbation associated with activation of the immune system. Neurology 37, 1097-1102. Powell, M.B., Mitchell, D., Lederman, J., Buckmeier, J., Zamvii, S.S., Grahman, M., Ruddle, N.H. and Steinman, L. (1990) Lymphotoxin and tumor necrosis factor-alpha production by myelin basic protein-specific T cell clones correlates with encephalitogenicity. Int. lmmunol. 2, 539-544. Ruddle, N.H., Bergman, C.M., McGrath, M.S., Lingenheld, E.G., Grunnet, M.L. and Padula, S.J. (1990) An antibody to lymphotoxin and tumor necrosis factor prevents transfer of experimental allergic encephalomyelitis. J. Exp. Med. 172. 1193-1200. Rudick, R.A. and Ransohoff, R.M. (1992) Cytokine secretion by multiple sclerosis monocytes. Arch. Neurol. 49, 265-270. Salmon, J.A. and Higgs, G.A. (1987) Prostaglandins and leukotrienes as inflammatory mediators. Br. Med. Bull. 43, 283-296. Sanders, M.E., Koski, C.L., Robbins, D., Shin, M.L., Frank, M.M. and Joiner, K.A. (1986) Activated terminal complement in cerebrospinal fluid in Guillain-Barrt} syndrome and multiple sclerosis. J. Immunol. 136, 4456-4459. Sawaa, M., Kondo, N.. Suzumura, A. and MaronouchL T. (1989) Production of tumor necrosis factor-alpha by microglia and astrocytes in culture. Brain. Res. 491,394-397. Sawant-Mane, S., Clark, M.B. and Koski, C.L. (1991) In vitro demyelination by serum antibody from patients with Guillain-Barr¢} syndrome requires terminal complement complexes. Ann. Neurol. 29, 397-404. Schabet, M., Whitaker, J.N., Schott, K., Stevens, A., Ziirn, A., Biihler, R. and Wieth61ter, H. (1991)The use of protease inhibitors in experimental allergic neuritis. J. Neuroimmunol. 31,265-272. Schmidt, B., Stoll, G., van der Meide, P.H.. Jung, S. and Hartung, H.P. (1992) Transient cellular expression of interferon-gamma in myelin-induced and T cell line-mediated experimental autoimmune neuriits. Brain, in press. Selmaj, K.W. and Raine, C.S. (1988) Tumor necrosis factor mediates myelin and oligodendrocyte damage in vitro. Ann. Neurol. 23, 339-346. Selmaj, K., Raine, C.S. and Cross, A.H. (1991a) Anti-tumor necrosis factor therapy abrogates autoimmune demyelination. Ann. Neurol. 30. 694-700. Selmaj, K., Raine, C.S., Cannella, B. and Brosnan, C.S. (1991b) Identification of lymphotoxin and tumor necrosis factor in multiple sclerosis lesions. J. Clin. Invest. 87, 949-954. Selmaj, K., Raine, C.S., Farooq, M., Norton, W.T. and Brosnan, C.F. (1991c) Cytokine cytotoxicity against oligodendrocytes: apoptosis induced by lymphotoxin. J. Immunol. 147. 1522 1529. Sethna, M. and Lampson, L.A. (1991) Immune modulation within the brain: recruitment of inflammatory cells and increased major histocompatibility antigen expression following intracerebral injection of interferon-',/. J. Neuroimmunol. 34, 121-132.

Sharief, M.K. and Hentges, R. (1991) Association between tumor necrosis facor-a and disease progression in patients with multiple sclerosis. N. Engl. J. Med. 325. 467-472. Shin. N.S. and Carney, D.F. (1988) Cytoxic action and other metabolic consequences of terminal complement proteins. Progr. Allergy 40, 44-81. Simmons, R.D. and Willenborg, D.O. (1990) Direct injection of cytokines into the spinal cord causes autoimmune encephalomyelitis-like inflammation. J. Neurol. Sci. 100, 3742. Simmons, R.D., Hugh, A.R., Willenborg, D.O. and Cowden, W.B. (1992) Suppression of active but not passive autoimmune encephalomyelitis by dual cyclo-oxygenase and 5lipoxygenase inhibition. Acta Neurol. Scand. 85, 197-199. Sobue, G.. Yamato, S., Haryama, M.. Matsuoka, Y., Uematsue, J. and Sobue, I. (1982) The role of macrophages in demyelination in experimental allergic neuritis. J. Neurol. Sci. 56, 75-87. Sonderer, B., Wild, P., Wyler, R., Fontana, A., Peterhans, E. and Schwyer. M. (1987) Murine glia cells in culture can be stimulated to generate reactive oxsgen. J. Leukocyte Biol. 47. 463-473. Stanley, N.C., Jackson, F.L. and Orr, E.L. (1990) Attenuation of experimental autoimmune encephalomyelitis by compound 48./'80 in Lewis rats. J. Neuroimmunol. 29, 223-228. Stevens, A., Lang, R., Schabet, M., Wieth61ter, H. and Hammann, K. (1990) Spontaneous chemiluminescence activity of peripheral blood cells during experimental allergic neuritis (EAN). J. Neuroimmunol. 27, 33-40. Stoll, G., Schmidt. B., Jander, J., Toyka, K.V. and Hartung, H.P. (1991) Presence of the terminal complement complex (C5b-9) precedes myelin degradation in immune-mediated demyelination of the peripheral nervous system. Ann. Neurol. 30, 147-155. Strigard, K.. Holmdahl, R.. van der Meide, P.H., Klareskog, K., and Olsson. T. (1989) In vivo treatment of rats with monoclonal antibodies against gamma interferon: effects on experimental allergic neuritis. Acta Neurol. Stand. 811. 201. Sun, D. (1991) Enhanced interferon-y-induced la-antigen expression by glial cells after previous exposure to this cytokine. J. Neuroimmunol. 34. 205-214. Traugott, U. and Lebon, P. (1988) Multiple sclerosis: involvement of interferons in lesion pathogensis. Ann. Neurol. 24. 243- 251. Trotter, J. and Smith, M.E. (1986)The role of phospholipases from inflammatoD' macrophages in demyelination. Neurochem. Res. II, 349-361. Trotter. J.L., Collins, K.G. and van der Veen. R.C. (1991) Serum cytokine levels in chronic progressive multiple sclerosis: interleukin-2 levels parallel tumor necrosis factor-t~ level. J. Neuroimmunol. 33, 29-36. TsaL C.P., Pollard, J.D. and Armati, P.J. (1991) Interferon-,/ inhibition suppresses experimental allergic neuritis: modulation of major histocompatibility complex expression on Schwann cells in vitro. J. Neuroimmunol. 31, 133-145. Tsukada, N., Miyagi, K., Matsuda, M., Yanagisawa, N. and Yone, N. (1991) Tumor necrosis factor and interleukin-I

21ii in the CSF and sera of patients with multiple sclerosis. ,I. Ncurok Sci. 102, 230-234. Voorthuis~ J.A.C., Uitdehaag, B.M.J., de Groot, C.J.A., Goede, P.H., van der Meide. P.H., and Dijkstra. C.D. 11990) Suppression of experimental allergic encephalomyelitis by intraventricular administration of interferong a m m a in Lewis rats. Clin. Exp. Immunol. 81, 183-188. Westland, K. and Pollard, J.D. (1987) Proteinase induced demyelination: An electrophysiological and histological study. J. Neurol. Sci. 82, 41-53.

Willenborg. D.O., Bowern, N.A., Danta, G. and Dohert~, P.D. (1988) Inhibition of allergic encephalomyelitis by the iron-chelating agent desferrioxaminc. J. Neur~fimmunol, 17, 127-135. Zielasek, J., Tausch, M., Toyka, K.V. and Hartung, H.P, 11992) Production of nitrite by neonatal rat microglial cells/' brain macrophages. Cell. lmmunol. 14 l, 11 I - 120.