Mechanisms of Immune Regulation in the Peripheral Nervous System

Ralf Gold, MD 1, Juan J. Archelos, MD 2, Hans-PeterHartung, MD 2

1 Department of Neurology, Clinical Research Unit for MultipleSclerosis, Julius-Maximilians-Universität Würzburg, Germany

2 Department of Neurology, Multiple Sclerosis andNeuroimmunology Research Group, Karl-Franzens-Universität Graz, Austria

The peripheral nervous system (PNS) is a targetfor heterogenous immune attacks mediated by dif-ferent components of the systemic immune com-partment. T cells, B cells, and macrophages caninteract with endogenous, partially immune-compe-tent glial cells and contribute to local inflammation.Cellular and humoral immune functions of Schwanncells have been well characterized in vitro . In addi-tion, the interaction of the humoral and cellularimmune system with the cellular and extracellularcomponents in the PNS may determine the extent oftissue inflammation and repair processes such asremyelination and neuronal outgrowth. The animalmodel experimental autoimmune neuritis (EAN)allows direct monitoring of these immune responsesin vivo . In EAN contributions to regulate autoimmu-nity in the PNS are made by adhesion molecules andby cytokines that orchestrate cellular interactions.The PNS has a significant potential to eliminate T cellinflammation via apoptosis, which is almost lackingin other tissues such as muscle and skin. In vitroexperiments suggest different scenarios how specif-ic cellular and humoral elements in the PNS maysensitize autoreactive T cells for apoptosis in vivo .Interestingly several conventional and novelimmunotherapeutic approaches like glucocorticos-teroids and high-dose antigen therapy induce T cellapoptosis in situ in EAN. A better understanding ofimmune regulation and its failure in the PNS mayhelp to develop improved, more specificimmunotherapies.

Introduction The peripheral nervous system (PNS) was counted

among the special immuno-privileged tissues like retina,cornea, anterior chamber of the eye, testis, and liver(105). This concept of immune privilege, originally for-mulated by Medawar (79) defined the protection of tis-sue grafted to certain sites. Later it was extended todescribe seclusion of particular areas of the body fromthe systemic immune compartment. Some ten years agoit was shown that indeed immune surveillance is opera-tive in such tissues (113). Even vulnerable privilegedsites like the nervous system are constantly patrolled byactivated T lymphocytes. Hence specialized anatomicbarriers like the blood-brain or blood-nerve barrier (31)or the absence of lymphatic drainage do not necessarilyguarantee the integrity of these sites. Although the con-cept of immune privilege then had to be abandoned, it isbeyond doubt that these tissues possess specializedmechanisms of immune protection, which ensure rapidand gentle elimination of inflammation once this hasencroached on there. Here we summarize pathways ofinflammation which are operative in the PNS anddescribe intrinsic counterregulatory mechanisms thatcontribute to termination of the assault. Both in vitrodata obtained on glial cell cultures and in vivo experi-ments in experimental autoimmune neuritis (EAN) arediscussed.

Experimental autoimmune neuritis (EAN) - a modelto study immune regulation in vivo

Experimental autoimmune neuritis (EAN) is an ani-mal model mediated by autoantigen-specific T-cells forhuman Guillain-Barré syndrome (GBS), an acutedemyelinating inflammatory disease of the peripheralnervous system (PNS), (reviewed in (46). EAN can beactively induced in Lewis rats by immunization withbovine P2-protein or recombinant human P2-protein,with a peptide (amino acid 53-78) spanning the neurito-genic epitope, or by adoptive transfer of neuritogenic T-cells (AT-EAN) (47, 74). Further autoantigens that have

Brain Pathology 9: 343-360(1999)

SYMPOSIUM: Peripheral Neuropathies

Mechanisms of Immune Regulation in thePeripheral Nervous System

Corresponding author:Hans-Peter Hartung, M.D., Universitätsklinik für Neurologie,Karl-Franzens-Universität, Auenbruggerplatz 22, A-8036 Graz,Austria; Fax: +43-316-325520; E-mail: [email protected]

been identified in EAN models in rats and mice are P0protein (75), myelin basic protein (MBP) (1), possiblyPMP22 (32) and galactocerebroside in rabbits. EANreplicates many clinical, electrophysiological andimmunological aspects of the Guillain-Barré syndromeand hence has been widely used as a model to investi-gate disease mechanisms (50).

The immune response in EAN can be dissected intoan induction and an effector phase. Adhesion molecules(AMs) are critically involved into these different phases

(3, 21). In the induction phase the injected autoantigenis presented to “naive” T cells by professional antigen-presenting cells (APC) such as macrophages or dendrit-ic cells resulting in T cell activation. Two external sig-nals are crucially required for an effective T cell activa-tion by antigen presentation: the antigen-specific signalprovided by the immunogenic peptide and presented inthe context of major histocompatibility complex mole-cules on APC, and the antigen-independent signal calledcostimulation and mediated by AMs of the integrin fam-ily such as aLb2 (LFA-1, CD11a/CD18),a4b1 (VLA-4, CD49d/CD29), or avb3 (CD51/CD61) and AMs ofthe immunoglobulin-superfamily such as ICAM-1(CD54), VCAM-1 (CD106), CD2, CD58, and, of spe-cial functional relevance, CTLA-4 (CD152), B7-1(CD80) and B7-2 (CD86) expressed both on T cells andAPC.

Activated T cells then circulate in the blood, attach tothe venular endothelium in the PNS and penetrate theblood-nerve barrier (BNB). This transendothelial migra-tion gives rise to the effector phase of the immuneresponse in EAN. In the PNS the autoantigen is present-ed by macrophages to T cells. The reactivated CD4+ Tcells amplify the immune response by recruiting furtherT cells and macrophages via chemokines and cytokines(51). The resulting breakdown of the BNB allows thepassage of circulating autoantibodies which are thoughtto synergize with T cells to produce demyelination (52,89).

Adhesion molecules in the induction phase of EAN

Antigen presentation is pivotal in T cell activation.The first important step in the pathogenesis of EAN isthe presentation of autoantigen in the induction phase.This results in the physiological activation of disease-inducing T cells (51). AMs are essential for effectiveantigen presentation in vitro in rodents and humans. Invivo, inhibition of some of these coaccessory moleculessuch as CD2 (60), ICAM-1 (6) or LFA-1 (5) by mono-clonal antibodies (mAbs) prevents adequate T cell acti-vation and attenuates EAN in the rat.After activation, T cells circulate and finally enter thetarget organ at the blood-nerve barrier (BNB)(Figure 1).In the acute phase of EAN, upregulation of AMs such asICAM-1 and VCAM-1 at the endothelial tight junctionson lesion-associated blood vessels parallels clinical dis-ease and parenchymal infiltration (30, 53, 103).

Central role of transendothelial migration in lesionformation. The transendothelial migration of immune

344 R. Gold et al: Immune Regulation in PNS

Figure 1. The intact blood-nerve barrier (BNB) and the extra-cellular compartments in the PNS. In GBS activated T cells andmacrophages enter the PNS at the interface between the cap-illary compartment and the neural environment. Three anatom-ical compartments with a different cellular and ECM composi-tion can be distinguished in the PNS: epineurium, perineuriumand endoneurium. The epineurium surrounds the perineuralensheathment of the fascicles and contains blood vessels, lym-phatics, fibroblasts, collagen type I bundels and fat. The peri-neurium consists of a lamellar arrangement of flattened peri-neurial cells, covered by BM and interspersed collagen fibrillsand tenascin-C. Contiguous perineurial cells are linked togeth-er by tight junctions. They constitute the morphological basis ofthe perineurial diffusion barrier. The endoneurium is composedof collagen type I and III fibrills, and embeds blood vessels,axon-Schwann cell units, fibroblasts, and occasionallymacrophages. The basal lamina surrounding Schwann cells iscomposed of classical ECM proteins such as collagen type IV,fibronectin, laminins, vitronectin, entactin, and heparan sulfate.These are secreted by the Schwann cell. The BNB consists ofan endothelial cell lining with tight junctions, a basement mem-brane (BM), the pericytes and is less tight than the blood-brainbarrier. Depending on the etiology, T cells and macrophageshave a different distribution in these peripheral nerve compart-ments in human immune-mediated neuropathies.

cells through the BNB is a crucial step in the pathogen-esis of immune-mediated disease of the PNS. The pres-ence of a variety of AMs on mononuclear cells and theupregulation of their ligands on endothelial cells duringactive disease point to a common pathogenetic role ofAM in the initiation of tissue inflammation in EAN.This notion was reinforced by therapeutic manipulationwith mAbs in vivo. Archelos and colleagues were thefirst to delineate the involvement of ICAM-1, LFA-1and recently, L-selectin (4) in transendothelial migrationin EAN. Enders et al. extended their findings on VLA-4

and its ligand VCAM-1 (30) (Figure 2). Blockade ofVLA-4 (a4b1) by mab was effective in EAN (Figure 3)and this integrin seems to be the most important AM intransendothelial migration of T cells in rodent EAN andhence the most promising target candidate for therapeu-tic intervention in GBS. Surprisingly VLA-4 andVCAM-1 blockade are equally effective in enhancing Tcell apoptosis in the inflammatory lesion (Gold, Lobb inpreparation).

345R. Gold et al: Immune Regulation in PNS

Figure 2. Transendothelial migration of activated T cells across the BNB in EAN is mediated by integrins. Through the action ofselectins - a family of cell adhesion molecules - T cells establish a loose reversible contact with endothelial cells. This initial rolling isfollowed by a firm irreversible adhesion mediated by integrins a4b1 (VLA-4) and aLb2 (LFA-1) on T cells and their immunoglobulin-like receptors on endothelium (VCAM-1, ICAM-1). Locally released chemokines (black circles) bound to endothelial glycoproteins(glycocalyx) increase integrin adhesiveness, and induce directed movement of T cells and monocytes.The subsequent transendothe-lial migration of T cells through the BNB is thought to be mediated mainly by a4b1/VCAM-1. Engagement of a4b1 with VCAM-1 acti-vates matrix metalloproteinases (MMP) which degrade certain ECM proteins.

The extracellular domain of VCAM-1, ICAM-1 and selectins is shed from the cell surface after T-cell-endothelial-cell interaction,and soluble forms circulate in the blood and cerebrospinal fluid (cVCAM-1, cICAM-1, cL-selectin, cE-selectin). Physiological levels ofcirculating forms of integrin ligands increase with pathology. T cells that have migrated into the PNS interact with local cellular andECM components. The nature of this complex interaction determines the outcome of the immune response in the target tissue.

Triggering of autoreactive T and B cell responsesThe mechanisms underlying generation of autoreac-

tive T cell and B cell responses are clear in the EANmodel where the autoimmune cascade is deliberatelyand artificially set off. In both acute and chronicimmune neuropathies in man, the conditions allowingemergence of peripheral nerve directed autoreactivityremain obscure.

It is conceivable that myelin-reactive T lymphocytesare contained in the immune repertoire of healthy indi-viduals. Under normal conditions, they are inactive orsilenced by various modes of peripheral tolerance.Tolerance may break down when the individual is con-fronted by an infective organism that happens to shareepitopes with endogenous peripheral myelin proteinssuch as P0 or P2. Such mimicry has been hypothesizedto underlie GBS associated with cytomegalovirus infec-tion (2, 50b, 50c). Given the degeneracy of epitope

recognition by T lymphocytes this notion may be dis-puted. Another possibility is antigen nonspecific T cellactivation mediated by cytokines (13).

This hypothesis may be viewed as a revival of thetime-honored concept of bystander demyelination.Evidence to support a variation on this theme is avail-able.

The possibility of nerve injury mediated by activated,non-neural specific T cells was studied by systemictransfer of ovalbumin-specific T cells, followed byintraneural injection of ovalbumin (52, 89). Rapidendoneurial infiltration of T cells and macrophagesoccured on the side injected with ovalbumin and wasassociated with a marked increase in blood-nerve barri-er permeability. In the casein-injected control nerve onlydegeneration and macrophage infiltration was observedthat declined after 3 days. Histological and clinical fea-tures in this model were similar to those observed in P2-induced EAN. When given in combination with anti-myelin antibodies primary demyelination or axonaldegeneration was demonstrable by electrophysiologicalstudies, thus replicating typical features of GBS.

Molecular mimicry based on structural similaritiesbetween microorganisms and peripheral myelin may bemore important for the generation of autoreactive B cellresponses. There is a large body of evidence indicatingthat infection with certain strains of the gramnegativeenteropathogen Campylobacter jejunielicits antibodyresponses directed to gangliosides and related glycol-ipids on the myelin membrane and axolemma since car-bohydrate sequences of these glycoconjugates are alsocontained in the lipopolysaccharide fraction of themicroorganism (Figure 4) (47, 50c,117).

Adhesion molecules in the effector phase of EANOnce T cells and monocytes have passed the blood-

nerve-barrier they adopt to a completely differentmicroenvironment in the PNS. There neural cells suchas Schwann cells and resident macrophages and theextracellular matrix harbor relevant ligands for receptorson infiltrating immune cells. In this review we will focuson the role of AM with special emphasis on integrinsexpressed in the PNS and with the immune-regulatingand inflammation terminating capacities of Schwanncells and macrophages.

In EAN, infiltrating macrophages express ICAM-1(103) whereas T cells display on their surface VLA-4(30) and LFA-1 (5) but the in vivo adhesion moleculephenotype has not yet been studied in detail. Matrixmetalloproteinases (MMPs) which degrade extracellularmatrix (ECM) molecules are upregulated on T cells


Figure 3. Morphological analysis of adoptive-transfer (AT)-EANtreated with the VLA-4 specific mAb TA-2. Detection of axonaldamage and demyelination in semithin sections from sciaticnerves of the anti-VLA-4 group (A) and from control animals (B)receiving isotype control only. Arrowheads indicate Walleriandegeneration in B. Bar = 10 mm for A-B

(65). In addition, there is growing evidence that AMs ofthe integrin family and their ECM receptors in the PNSare important in the effector phase and may contribute totissue repair (7, 9).

Demyelination and axonal loss modulate integrinsin EAN. In myelin-induced EAN, an EAN variant char-acterized by severe inflammation, prominent demyeli-nation and axonal loss, the expression of integrins fol-

lows a distinct spatiotemporal pattern (90).Interestingly, integrin expression on Schwann cells isassociated with distinct histopathological alterations inEAN and GBS (Figures 5, 6). Additional in vitro stud-ies and observations from human non-inflammatoryneuropathies (see below) suggest that integrin expres-sion is modulated by several factors in the PNS.Demyelination, disruption of the basal lamina, andaxonal loss appear to override some of the cytokine


Figure 4. Molecular mimicry: Cross-reactive antibody responses to C.jejuni and peripheral nerve glycoconjugates. Approximately30% of all Western patients with Guillain-Barré Syndrome develop the neuropathy some 1 - 3 weeks after infection with the gram-negative enteropathogen C.jejuni. The lipopolysaccharide fraction of serovars of C.jejuni associated with GBS contains epitopesshared with glycolipids present in the myelin sheath and axolemma of peripheral nerve. It is conceivable that cross-reactive antibodyresponses ("molecular mimicry") underlie nerve damage in C.jejuni-related Guillain- Barré Syndrome. Modified from (50c, 91b).

effects on the expression of integrins by Schwann cellsfound in vitro (90).

AMs in immune-mediated neuropathiesThe pathogenesis of GBS and other inflammatory neu-ropathies is only partially known. The nature and char-acteristics of the induction phase remain elusive and it isstill obscure if the immune cells relevant for the diseaseexhibit a defined phenotype and specificity. Therefore,descriptive studies on AM in these diseases had to focuson T cells in general, the blood-nerve barrier and therepair phase (50,90,97).

Expression of AMs in GBS and human neu-ropathies. In GBS, we confirmed the association of thedownregulation of a6b4 integrin with inflammatorydemyelination and the neo-expression of integrin sub-unit a5 with additional axonal loss, as was previously

shown in EAN (90). Similarly, a neo-expression of a5on Schwann cells and a transient downregulation of b4have been observed in non-inflammatory chronic axon-al neuropathies in humans (91, 93). These studies inhuman neuropathies indicate that in the PNS structuralfeatures such as the presence or absence of a basal lam-ina, demyelination or axonal loss are important co-determinants of integrin regulation and that an alteredintegrin expression is involved in the repair process assuggested by studies in animal models of regenerationand inflammation (71, 90).

It is well known that some immune molecules arecleaved from the cell surface after ligand engagement atthe BNB. Such shed molecules circulate in blood andare associated with disease severity in some cases.Circulating fragments of integrin a6b4-like immunore-activity have been detected in GBS and suggested to bea putative novel marker of myelin damage (102).Indeed, circulating integrin subunits would be helpfulsensors of disease activity, but molecular identificationof the a6b4-like immunoreactivity in serum is stillneeded to exclude a casual crossreactivity with otherproteins.

In neuropathies associated with vasculitis - aninflammatory reaction at the arterioles supplying thenerves - increased levels of integrin receptors ICAM-1and VCAM-1 were noted on endothelial cells in thePNS and infiltrating cells were positive for aLb2 andaMb2 (17, 85). Further, raised levels of E-selectin, anAM expressed on activated endothelium, were measuredin the blood of these patients (48).

Integrins and the effector phase in GBSThe complexity of the interactions between neural

and immune cells, ECM and humoral inflammatorymediators such as antibodies, cytokines, chemokines,matrix metalloproteinases and trophic factors in thedevelopment of the demyelinating lesion in GBS makeit difficult to study the role of AM such as integrins inthe repair phase in vivo. However, abundant in vitro datacombined with the described expression pattern of inte-grins in vivo suggest a possible central role of integrinsin the genesis of the inflammatory lesion (7).

Integrins, as ECM receptors, may be involved in thisrepair process. Preliminary experimental evidence indi-cates that the complex processes of repair can be target-ed therapeutically. Treatment of rats with a monoclonalantibody to the a1 subunit attenuates EAN and reducesimmune cell infiltration and demyelination (own unpub-lished observation).

The obvious involvement of adhesion molecules in


Figure 5. Expression of integrins (A) on endothelium andfibroblats and (C) on Schwann cell/axon units in experimentalautoimmune neuritis (EAN) and Guillain-Barré syndrome(GBS). (A) Endothelial cells at the BNB express a variety ofintegrins in normal rats. In acute EAN, there is a up- and down-regulation of distinct integrin subunits on endothelial cells,Schwann cell/axon units and fibroblasts. Inflammatory demyeli-nation and concomitant disruption of the basal lamina is asso-ciated with a transient downregulation of a2b1 and a6b4 inSchwann cells. If additional axonal loss is present due to severeinflammation a neo-expression of a5b1 is observed. Integrinexpression normalizes with progressive tissue repair in EAN. Asimilar pattern is observed in sural nerve biopsies obtainedfrom patients with GBS.

the pathogenesis of EAN and GBS has relevant clinicalimplications (81). Antibody targeting of transendothelialmigration of T cells governed by a4b1 provides a real-istic novel therapeutic option in GBS (72).

Local immune activation - the role of Schwann cells

Cellular immune functions. The contribution ofSchwann cells to the initiation and termination of animmune response in the PNS is still a matter of debate.In principle it has been shown that they possess immunemolecules as a basic prerequisite to interact with invad-ing T cells. Schwann cells in vitro constitutively expressat low levels MHC class I but no significant numbers ofMHC class II molecules. Upregulation of MHC class Ior expression of MHC class II can be induced by stimu-lation with interferon-g (IFN-g ) or upon coculture withactivated T cells (8, 66). Interestingly, tumor necrosisfactor-alpha (TNF-a) synergizes with IFN-g and furtherincreases MHC expression which also has functionalimportance (40). Moreover adhesion molecules such asICAM-1 (CD54), which are constitutively expressed oncultured Schwann cells are upregulated by thesecytokines and may exert costimulatory functions. Thuspretreated, Schwann cells have been shown to processexogenous P2 protein or its neuritogenic peptide forantigen presentation (40). Under certain conditions theycan even present endogenous myelin proteins likemyelin basic protein (MBP) to autoaggressive lympho-cytes in a MHC class II restricted manner (114). In thatcase Schwann cells were prepared from older animalswhere myelination was already ongoing. These findingsunderscore that Schwann cells are able to phagocytoseexogenous antigens and degrade them to antigenic pep-tides in endosomal compartments (8). Although the rel-evance of MHC class II expression on Schwann cellshas been questioned in EAN (96) (see below), there isprinciple evidence that Schwann cells are capable of thisimmune function in vivo (14, 88). A possible explana-tion for this discrepancy between the in-vitro and in vivosituation may be provided by the findings of Tontsch etal. (108) and Neumann et al. (83) who have demon-strated that neurons indeed exert a regulatory functionon glial cells.

It is of note that Schwann cells are also the source ofa number of proinflammatory cytokines such as IL-1,IFNa and TNFa. Given the autocrine and paracrine func-tions of these molecules both feedback actions onSchwann cells and modulation of the pericellularimmune circuitry are conceivable (76).

Schwann cells also express molecules that can termi-

nate T cell inflammation and downregulate immunefunctions. Fas and its ligand are central molecules of afamily of death factors that regulate T cell survival in theimmune system (see review in (82). Recent findingsfrom our group show that Schwann cells do not expressFas (CD95) or Fas ligand (FasL) constitutively.However, proinflammatory Th1 cytokines can upregu-late FasL or Fas on the surface of Schwann cells within48 hrs (Wohlleben,et al., manuscript in preparation -Figure 7). Thus several potential scenarios are conceiv-able. First, crosslinking of Fas molecules on invading Tcells by membrane-bound or secreted FasL could elimi-nate the autoaggressive immune effectors. This couldexplain T cell apoptosis observed during the natural dis-ease course of EAN (118); see below). Second, expres-sion of Fas on Schwann cell membranes could renderthem susceptible to T cell attack. Apoptotic eliminationof Schwann cells is observed during EAN (26b, Brück,Gold; unpublished) and may further augment demyeli-nation in the PNS. The ultimate functional aspect of theexpression of Fas/FasL on Schwann cells is not yetunderstood and currently under investigation.


Figure 6. Immunofluorescence detection of S100 (A) and a2(B) on Schwann cell cultures. Schwann cells were purified fromnewborn rat sciatic nerves. Magnification: A, B: x 340.

Humoral immune functions.Apart from these func-tions which are mediated by and dependent on cell-cellcontact, Schwann cells may modulate local immunereactions in the PNS through the elaboration and releaseof humoral factors. Schwann cells produce IL-1, apotent cytokine which promotes T cell activation andproliferation (15), and also IL-6 which may bias thelocal cytokine milieu to a T helper 2 (Th2) type of reac-tion (16). Furthermore, Constable et al. (26) found thatSchwann cells can be induced in vitro to secreteprostaglandin E2 and thromboxane A2. Theseimmunomodulators can inhibit or stimulate T cells,depending on their level of production. Schwann cellsare also endowed with a cytokine-inducible nitric oxidesynthase (iNOS) (41). Simultaneous treatment withIFN-g and TNF-a upregulates iNOS-specific mRNA inSchwann cells within 12 hrs (Figure 8). Nitrite secretionas a measure of NO production was detectable after 24hrs and reached its plateau on day 3. Nitrite release wasinhibited by N-monomethyl L-arginine, a competitiveinhibitor of iNOS and not by N-nitro L-arginine, whichpreferentially blocks the noninducible NO-synthase.This finding ruled out unspecific release of NO bySchwann cell cultures and underscored the presence ofa macrophage-like, inducible form of NO synthase.

Secretion of NO by Schwann cells exerts a strong sup-pressive effect on T cell activation in a coculture model(41). This mechanism has been proven in vitro wherereactive nitric oxide induced apoptosis of autoreactive Tcell lines (119). Thus, Schwann cell-derived reactiveoxygen intermediates have the potency to limit inflam-matory demyelination. Importantly exogenous IL-2could rescue T cells from apoptotic cell death in vitro(see below). Survival of T cells exposed to IL-2 couldoccur via upregulation of antiapoptotic molecules of thebcl-2 family (23) and provide an important counterplay-er to local proapoptotic elements.

It is interesting to consider that in the local tissuemilieu Schwann cells are in close contact withendoneurial fibroblasts, which have been shown torelease interferon-beta (IFN-b). This could in turndownregulate NO production of Schwann cells, verysimilar to recent experiments with glial cells from thecentral nervous system (CNS) (54).

At present it is not clear whether Schwann cells mayalso act by production of endogenous steroid hormones.These neurosteroids are synthesized in CNS (116), butalso in sciatic nerve (80) and may act by their mediatorssuch as lipocortin-1 (annexin-1) to downregulateinflammatory reactions. Lipocortin-1, but notlipocortin-2 or -5 is a potent inhibitor of the antigen-spe-cific activation of pathogenic T cell lines (37).Surprisingly truncation studies indicated that the activi-ty of lipocortin-1 resides within the core region and notin the N-terminal fragment. In a side by side comparisonwith prednisolone lipocortin-1 suppressed T cell activa-tion at a 30-fold lower dosage by weight. Importantly,lipocortin-1 expression in sciatic nerve is also increasedduring recovery from EAN and may contribute to termi-nation of the inflammatory reaction (36).

Schwann cells are endowed on their membrane witha number of regulatory complement proteins such asCR1 (CD35), decay accelerating factor CD55, mem-brane cofactor protein CD46 and CD59 (70, 110, 111).This ensemble of proteins serves to attenuate the proin-flammatory and demyelinating properties of activatedcomplement factors that are involved in the pathogene-sis of autoimmune demyelination in peripheral nerve(69, 95).

On the other hand, while macrophages constitute themajor source of complement at inflammatory foci,Schwann cells can also be induced, e.g. by IFNg, TNFaand IL-1b to generate the central complement compo-nent C3 (28). In Guillain-Barré Syndrome, nerve anti-body titres are paralleled by levels of complement acti-vation products in blood and CSF (C3a, C5a, soluble


Figure 7. FACS analysis of Fas ligand on rat Schwann cells.Neonatal Schwann cells were treated for 2 days with 300 IU/mlIFNg and TNFa. FACS analysis was performed using a biotiny-lated anti-rat FasL IgM antibody (red line) and an isotype con-trol (green area). A weak but significant upregulation of FasLcan be observed.

membrane attack complex). These activated comple-ment factors are also detected in situ (44, 45, 68).

Antigen-specific interaction of Schwann cells and Tcells. At present we do not know which of these differ-ent cellular and humoral properties prevails in vivo.Finally it is clear that Schwann cells are only incompletefacultative antigen-presenting cells (APC). Antigenpresentation by Schwann cells has two effects on T cellactivation (41): (i) there is incomplete T cell activationafter day 2 as reflected by low thymidine incorporationdespite elevated IL-2 receptor, transferrin-receptor andcostimulatory molecules such as CD28 on T cells. (ii) itincreases the susceptibility of T cells to undergo apop-totic cell death when steroid hormones were added tothe culture. The first effect implies that maximal prolif-eration of antigen-specific T cells occurs after 24 hrs,one day earlier than observed with thymic APC.Thereafter T cell activation is markedly diminished,irrespective of the pretreatment of the Schwann cells(41). Similar findings were obtained for astrocytes (39).Possible mediators of this effect have not yet been iden-tified. Secondly, Schwann cells but not professionalAPC exert a priming proapoptotic effect on T cells. Tcells are then rendered susceptible to steroid-mediatedapoptosis. This may have consequences in vivo wheninvading T cells are also exposed to elevated systemicsteroid levels that are measurable in experimentalautoimmune models (78).

The role of macrophagesMacrophages are the predominant cell population in

the affected nerves of animals with experimental neuri-tis (see below) and their selective elimination suppress-es the disease (Figure 9) (49, 58). They also featureprominently in the lesions of GBS (53, 64). Unlike inthe CNS, macrophages in the PNS lying near vessels orscattered in the parenchyma are phenotypically similar.They do not exhibit the sharp distinction that character-izes the perivascular macrophages and microglia in theCNS (reviewed in (42)). Experimental studies withchimeric rats suggest a rapid and significant turnover ofmacrophages within the PNS. Within three months afterestablished chimerism up to 60% of macrophages werereplaced by bone-marrow derived cells (109).

The PNS macrophages resemble their perivascularcounterpart in the CNS in their constitutive expressionof MHC class II molecules (43, 109). Also, in experi-mental neuritis the primary MHC class II positive cellsare macrophages (96). The crucial role of macrophagesin immune-mediated nerve damage centres around anti-


Figure 8. Upregulation of inducible nitric oxide synthase in ratSchwann cells. iNOS mRNA is specifically expressed in ratSchwann cells following stimulation with 300 U/ml g-IFN andTNF-a. Total RNA from 500.000 Schwann cells per sample wasextracted and subjected to Northern blot analysis as describedin material and methods. A specific mRNA species was detect-ed as early as 12 hours after cytokine treatment, furtherincreased by day 1 and remained elevated until day 5, whereasno specific mRNA was found in controls (Co, top panel). TotalRNA was visualized by methylene-blue staining on the sameblot (bottom panel). (reproduced from (41) with kind permissionfrom Academic Press)

Figure 9. Macrophages in experimental autoimmune neuritis.Seventeen days after active immunisation of Lewis rats withperipheral nerve myelin and complete Freund’s adjuvant, i.e. atthe height of clinical disease, macrophages (bold arrows) arethe chief cellular elements in the peripheral nerve lesion. Herethey are shown as they have phagocytosed the myelin sheathin direct contact with the denuded axon (arrows). Semithincryosection.

gen presentation, direct phagocytic attack on myelin,and the release of proinflammatory cytokines includingTNF-a, IL-1 and IL-6 and other noxious molecules(Figure 10). In coculture models release of proinflam-matory mediators by macrophages can induce T cellapoptosis very efficiently (79b). This raises the possibil-ity that macrophages may contribute to termination ofinflammation by apoptosis (see below). Furthermore,strong expression of lipocortin-1 is seen in macrophagesin experimental neuritis (36). In view of the immuno-suppressive effect of lipocortin-1 on autoimmune T cells(see above) this may further help to downregulateautoimmune inflammation.

Investigation of proinflammatory enzymes involvedin the pathogenesis of EAN may help to developimproved therapeutic approaches. Recent studies havefocussed on the expression of matrix metalloproteinases(MMP). Upregulation and expression of MMP-9 (92-kdgelatinase) and MMP-7 (matrilysin), but not of otherMMPs were found by immunocytochemistry, RT-PCR,and zymography (Figures 11, 12) (55, 65). Also, utiliza-tion of a comparable MMP pattern underlined patho-physiological similarities between EAN and GBS(Figure 13) (65). The selective expression of certainMMPs in inflammatory autoimmune disorders of thePNS and their varied cellular localization suggest thatthese enzymes may play a crucial , but multifactorialrole in the process leading to disruption of the blood-nerve barrier. MMPs may therefore serve as potentialtargets for newer therapies that make use of specific

inhibitors capable of preventing activation of MMPs.Such an approach has already been shown to have a ben-eficial effect on the disease course of EAN (92).

Since EAN is a monophasic disease, endogenouscounterregulatory mechanisms must operate to limit theextent of peripheral nerve inflammation. Potential medi-ators are immunoregulatory Th2 cytokines like trans-forming growth factor beta (TGF-b) or IL-10.Regulation of the isoform TGF-b1 and its cellular local-ization have been characterized during disease evolutionin the AT-EAN model (62, 63). By quantitative Northernblot analysis and in situ hybridization immunohisto-chemistry TGF-b1 mRNA was upregulated in sciaticnerves and spinal roots, with peak levels just precedingthe first clinical signs of recovery (Figure 14). In doublelabelling studies, macrophages and a subpopulation of Tcells were identified as the major cellular source (Figure15). Thus the physiological function of TGF-b1 in the


Figure 10. TNFa in experimental autoimmune neuritis.Seventeen days after active immunisation of Lewis rats withperipheral nerve myelin and complete Freund’s adjuvant teasednerve fibre preparations were stained for TNFa. Massiveupregulation is associated with macrophages. Large post-phagocytic macrophages at later stages of demyelination wereTNFa-negative. Intraperitoneal application of an anti-TNFa anti-body significantly reduced the degree of inflammatory demyeli-nation and clinical expression of the disease (cf (104)).

Figure 11. Immunohistochemistry for Gelatinase B in sciaticnerves from Lewis rats with AT-EAN at peak of clinical severity.Positive immunoreactivity was primarily found in the endo- andperineurium (A), and could be localized to infiltrating mononu-clear cells and around blood vessels (B). A x 200; B x 600.

inflamed nerve not only pertains to cellular growth andmatrix formation, but also implies a number of immuno-logical mechanisms. This is underscored by in vivodatathat confirm a beneficial effect of another isoform, TGF-b2 on the disease course of EAN induced by activeimmunization (59). Yet in rapidly developing AT-EANintraperitoneal injection of TGF-b2 failed to abrogatethe disease of diminish its consequences, suggesting thatits therapeutic efficacy is limited. Since peak levels ofTGF-b1 were reached before the height of the disease,one might speculate that other immunoregulatorycytokines such as IL-10 take over that role at laterstages. This was indeed shown in studies comparing theexpression of IL-10 in EAN as a model for immunemediated demyelination and Wallerian degeneration as alesion model (34). Possible cellular sources of IL-10 areSchwann cells, macrophages, and T cells (57). Hereagain the pathophysiological relevance of these findingswas highlighted by therapeutic studies with IL-10 (10).IL-10 effectively suppressed clinical EAN induced byactive immunization with peripheral nerve myelin. Evenin ongoing EAN administration of IL-10 amelioratedthe disease. ELISPOT and isotype assays confirmed thatthis suppression was associated with downregulation ofTh1 responses and upregulation of a Th2 type response.

T cell apoptosis in EAN - a physiological defensemechanism to terminate inflammation

In 1972 Kerr and colleagues used morphological cri-teria to define apoptosis as a specialized form of celldeath (61). Traditionally apoptotic cells were identifiedby electron microscopy (115), but in the meantimenewer techniques such as in situ tailing assays havebecome available (33, 38) which detect associated bio-chemical events of oligonucleosomal DNA fragmenta-tion occuring in most apoptotic cells. Although apopto-sis is a purely morphological term, it has an array ofpathophysiological and functional implications for theimmune system (25). Since the plasma membrane itselfis not a primary target of apoptosis but rather providesspecific signals to phagocytic cells (94), release ofproinflammatory cytoplasmic constituents is avoided. Invulnerable organs such as the nervous system apoptosiswould provide an ideal, noninflammatory mechanism toterminate the autoimmune T cell attack at vulnerablesites. It is assumed that the process of apoptosis is com-pleted with in 4 to 5 hours (20).

Pender and colleagues drew attention to apoptosis asa possible mechanism of cell destruction in inflammato-ry brain lesions of Lewis rat autoimmuneencephalomyelitis (EAE) (86). In the PNS it was not

clear whether T cells become activated on encounteringtheir target antigen in the peripheral nervous system orwhether this occurs in lymphoid organs. Also, little wasknown about the fate of activated T cells once they haveinvaded the PNS. Conceivably, they could either leave


Figure 12. Proteolytic activity in Lewis rat. Detection of gelati-nase activity in sciatic nerve from Lewis rat with AT-EAN bysodium dodecyl sulfate-substrate gel analysis during the clinicalcourse of the disease. The graph demonstrates the diseasecourse of AT-EAN after injection of neuritogenic T cells. Eacharrow marks a time point proteolytic activity has been deter-mined. Gel analysis is shown below: Each zone of clearing(black) represents proteinase activity. Only in animals with max-imal disease activity - at day 6 - proteolytic activity is detectableat the 92 kDa level. In contrast, clearing remains unchanged atthe 72kDa level throughout all time points investigated.

Figure 13. Immunohistochemistry for Gelatinase B in suralnerve biopsy from a patient with GBS. Positive immunoreactivi-ty was primarily localized around blood vessels and to infiltrat-ing mononuclear cells. X 600.

their target tissue and end up in peripheral immuneorgans or may be eliminated in situ. Sites and mode ofelimination and activation were studied in EAN inducedby autoreactive T cell transfer or by active immunisationwith purified bovine PNS myelin (118). In both models,T cell activation as reflected by incorporation of bro-

modeoxyuridine (BrdU) occured in lymphoid organsbefore disease onset and was only rarely observed in situin the inflamed sciatic nerve. There, however, T cellsexhibiting morphological signs of apoptosis weredetected immediately after disease onset. T cell apopto-sis reached its maximum at day 7 after cell transfer inAT-EAN and around day 17 in active EAN and led toelimination of about 10% of the inflammatory T cells.Bearing in mind that programmed cell death is a dynam-ic process, this could render a substantial number of Tcells apoptotic, amounting up to 50% of the original Tcells within 24 hours. Apoptosis of macrophages orSchwann cells was only very rarely seen.

These findings allow some important conclusions tobe drawn regarding immune protection in the centraland peripheral nervous system. Both tissues obviouslytake advantage of specialized, physiological defensemechanisms to limit inflammatory processes. The mag-nitude is much higher in the CNS (99) than in the PNSand almost zero in various human inflammatorymyopathies (101). Even in the immunological setting ofHIV infection, which is characterized by an increasedfrequency of apoptosis in blood and lymphoid organs, Tcell apoptosis is not augmented in nerve or muscle (100)speaking for similar pathomechanisms of HIV-associat-ed neuromuscular disorders and their idiopathic coun-terparts. T cell activation in the PNS appears to followmechanisms quite similar to those operating in the CNS(84) and may therefore not account for this disparity.These data underline that T cell apoptosis in autoim-mune disorders occurs apparently only in immune-pro-tected sites (reviewed in (35).

At present the molecular mediators are still in largepart unknown. To better understand the mechanisms ofapoptosis in vivo it is important to know whether anti-gen-specific cells alone or also recruited T cells aredestroyed within the nervous system. So far this hasonly been studied in EAE. Pender and colleaguesinduced EAE by a T cell clone using Vb8.2 T cell recep-tor, the predominant T cell receptor element in MBP-induced EAE of the Lewis rat (24), and characterizedapoptosis and T cell receptor usage in lymphocytes iso-lated and enriched from spinal cord (106, 107). Theyfound that the frequency of Vb8.2+ T cells was seven-fold higher in the apoptotic population than in non-apoptotic T cells. Furthermore they could not detectrecirculation of Vb8.2+ T cells in peripheral lymphoidorgans. This work suggests, although by indirect evi-dence, that antigen-specific T cells are the prime targetfor destruction by apoptosis. Similar experiments inEAN are very difficult to perform since quantitative


TGF-b1 mRNA

Standard

Figure 14. Transforming growth factor beta mRNA expressionin EAN. Northern blot demonstrating upregulation of TGF-b1mRNA in Lewis rats with P2-T cell transfer EAN at day 7 (C)compared to day 0 (B) and controls (A). Courtesy of Dr. R.Kiefer.

Figure 15. TGF-b in EAN. In situ hybridisation (upper panels)and in situ hybridisation combined with immunocytochemistryfor macrophages (lower left panel) and T-lymphocytes (lowerright panel) 18 days after immunisation of Lewis rats withperipheral nerve myelin. Upper left panel: control. Massiveupregulation of signal for TGF b-1 mRNA, mostly associatedwith T-cells but also with macrophages. Courtesy of Dr. R.Kiefer.

recovery of T cells from inflamed nerve is not that easi-ly achieved as from spinal cord (R. Gold, unpublished).Experiments comparing T cell apoptosis in EANinduced by different autoantigens may also help todelineate the relative contribution of the specific antigenand subsequent T cell receptor activation. Again, only inthe CNS counterpart EAE data have been generatedshowing that the extent of T cell apoptosis in theparenchyma is comparable in MBP, S-100 or MOGinduced EAE (11). In the CNS a specific histopatholog-ical pattern of lesion distribution associated with a dis-tinct autoantigen (67) may account for differences in Tcell apoptosis. In the future studies in mouse mutantsmay yield important insights with regard to the involvedpartners. Recently detailed investigations have providedevidence that the contribution of the Fas/FasL system inEAE is obviously not as important as in other autoim-mune disorders (77).

Therapeutic induction of T cell apoptosis in autoim-mune diseases: studies in experimental autoimmuneneuritis

Treatment regimens in patients with immune-mediat-ed disorders of the nervous system aim to reduceinflammation, demyelination, to prevent primary andsecondary axonal damage, and to accelerate recovery.Therapeutic efficacy of steroids in inflammatory disor-ders of the nervous system has been well established forimmune neuropathies (29, 87), multiple sclerosis andoptic neuritis (12). Corticosteroids exert their beneficialantiinflammatory effect at multiple levels. Besides theirantiedematous properties, they modulate cell activation,cytokine expression, secretion of inflammatory media-tors, and leucocyte migration (reviewed in (18, 19).These effects are all mediated by the cytosolic gluco-corticosteroid receptor. At high doses, glucocorticos-teroids can also directly promote cell death, and this isthought to reflect a nongenomic, physicochemicalaction of glucocorticosteroids (reviewed in (22).Possible candidates for this molecular effect are mem-brane-bound, ligand-gated purinergic P2X receptors(56), which upon activation cause a perturbation inintracellular ion homeostasis.

Based on the therapeutic use of high-dose glucocor-ticosteroids we first investigated whether steroid ‘pulse-therapy’ could induce T cell apoptosis in situ (111).Since a predictable and synchronized disease coursewas highly desirable for these investigations, wefocussed on AT-EAN of the Lewis rat which was pro-duced by transfer of activated, P2-specific T lympho-cytes. To delineate whether the effect of steroid hor-

mones is stage-specific, two pulses of glucocorticos-teroids (10 mg/kg body weight) were administered afterdisease onset or at its presumed maximum. Due to therapid elimination of apoptotic cells (13) rats had to besacrificed 6 hours after the second steroid pulse, beforea clinical effect became apparent. At both time pointsthere was a massive reduction of T cell inflammationand a four- to fivefold increase of T cell apoptosis in theinflamed sciatic nerve. In addition, reduced cellular pro-


Figure 16. Immunocytochemical characterization of apoptoticcells and BrdU incorporation in sciatic nerve from Lewis ratEAN. A-D) T-cell infiltration in AT-EAN on day 7 in sciatic nerverats receiving control serum only (A), rhP2 recipients (B), andrats treated with rhP2 followed by 10 mg/kg methylprednisolone(C,D); double-labelling of apoptotic T-cells. Nuclei with frag-mented DNA are labeled black by in-situ tailing followed by anti-T-cell immunochemistry visualized with alkaline phosphataseand fast red salt. Microphotographs show only some apoptoticT cells in control rats (arrows in A), whilst their number increas-es after P2 injection (B) and is further augmented by glucocor-ticosteroids (C). (D) shows a higher magnification from the areamarked by squares in (C). In some apoptotic T cells the mem-brane signal is still preserved (arrowheads) bar = 20 µm for A-C, 10 mm for D. E) BrdU positive fragments (brown, indicated byarrows) are incorporated in ED-1 positive, red-labelledmacrophages. F) BrdU positive cells in sciatic nerve after anti-gen therapy (arrows) bar = 10 µm for E-F

liferation in lymphoid organs and loss of thymus weightwere observed in steroid recipients reflecting typicalgenomic actions of steroids. These results support theuse of pulsed high dose corticosteroid therapy inautoimmune neurological disorders. In EAE, dose-response relations have been delineated more clearly(98). Here doses of at least 10 mg/kg body weightmethylprednisolone were needed to enhance T cellapoptosis.

Another approach that has been developed for thetreatment of experimental autoimmune disorders is anti-gen-specific therapy. This is based on the observationthat T cell receptor reengagement at an appropriate stageof the cell cycle eventually leads to T cell apoptosis(reviewed in (73). Its efficacy has been demonstrated byCritchfield et al. in EAE in mice by intravenous admin-istration of MBP (27). Taking advantage of an approachsimilar to that described above for glucocorticosteroidhormones, we demonstrated that intravenous treatmentwith recombinant P2 protein can prevent AT-EAN andactive EAN of the Lewis rat, and causes a profoundincrease in T cell apoptosis in the sciatic nerve (Figure16) (112). With regard to induction of apoptosis in AT-EAN antigen therapy was at least as effective as steroidhormones. When both treatment strategies were applied,they acted synergistically and enhanced T cell apoptosisin situ (Figure 16). Under these circumstances even Tcell proliferation in the nervous system could beobserved which does not normally occur in untreatedEAN (118). Interestingly, the native P2 protein showeda much higher efficacy than the neuritogenic peptidespanning aa 53-78 which is recognized by these neuri-togenic T cell lines in vitro, supporting the concept thatantigen presentation must occur in situ (113) and maybe executed by activated Schwann cells in the PNS(114). Antigen therapy may prove useful as highly spe-cific treatment of autoimmune disorders of the nervoussystem. In ongoing investigations TNF-a has been iden-tified as the principal mediator of the effects of antigentherapy in situ (Weishaupt, Gold, submitted for publica-tion). The intracellular signalling mechanisms that areassociated with high-dose antigen therapy have not yetbeen elucidated.

Conclusions and perspectives The peripheral nervous system, due to the lack of a

strong anatomic barrier at strategic sites (most proxi-mally at the root exit zone and most distally at the motorterminals), is jeopardized by immunoinflammatoryresponses from the systemic immune compartment.Hence, tight control mechanisms are required to ensure

its viability. These are manifold and include the elabo-ration of various down-regulatory molecules by residentpartially immunocompetent cells such as Schwann cellsand endoneurial macrophages. Recently, elimination ofinvading autoaggressive T cells by apoptosis has beenidentified as a major process to contain neural inflam-mation and protect the integrity of peripheral nerve.Elucidation of the immunoregulatory circuitry in thePNS and its disturbance in human disease and experi-mental models will aid in the development of improvedtreatments for immune-mediated neuropathies.

AcknowledgementsWe are indebted to Dr. K.V. Toyka, Würzburg for

many stimulating discussions, and to our colleagues DrsS. Jung, R. Kiefer, B. Kieseier, S. Previtali, A.Weishaupt, G. Wohlleben and U.K. Zettl for theirinvaluable contributions.

The work of the authors is supported by grants fromthe Bundesministerium für Forschung, DeutscheForschungsgemeinschaft, the Gemeinnützige HertieStiftung, the Wilhelm-Sander Stiftung, the German MSsociety, funds from the State of Bavaria and the Karl-Franzens-Universität Graz.

References

1. Abromson-Leeman S, Bronson R, and Dorf ME.(1995)Experimental autoimmune peripheral neuritis induced inBALB/c mice by myelin basic protein-specific T cellclones. J Exp Med 182: 587-592

2. Adelman M and Linington C.(1992) Molecular mimicryand the autoimmune response to the peripheral nervemyelin PO glycoprotein. Neurochem Res 17: 887-891

3. Archelos JJ and Hartung H-P.(1997) The role of adhesionmolecules in multiple sclerosis: biology, pathogenesis andtherapeutic implications. Mol Med Today 3: 310-321

4. Archelos JJ, Jung S, Rinner W, Lassmann H, Miyasaka M,and Hartung HP.(1998) Role of the leukocyte-adhesionmolecule L-selectin in experimental autoimmuneencephalomyelitis. J Neurol Sci 159: 127-134

5. Archelos JJ, Mäurer M, Jung S, Miyasaka M, Tamatani T,Toyka KV, and Hartung H-P.(1994) Inhibition of experi-mental autoimmune neuritis by an antibody to the lym-phocyte functions-associated antigen-1. Lab Invest 70:667-675

6. Archelos JJ, Mäurer M, Jung S, Toyka KV, and HartungHP.(1993) Suppression of experimental autoimmune neu-ritis by antibodies to the intracellular adhesion molecule-1. Brain 116: 1043-1058

7. Archelos JJ, Previtali SC, and Hartung H-P.(1999) Therole of integrins in immune-mediated diseases of thenervous system. Trends Neurosci 20: 30-38

8. Armati PJ, Pollard JD, and Gatenby P.(1990) Rat andhuman Schwann cells in vitro can synthesize andexpress MHC molecules. Muscle Nerve 13: 106-116


9. Aumailley M and Gayraud B.(1998) Structure and biolog-ical activity of the extracellular matrix. J Mol Med 76: 253-265

10. Bai XF, Zhu J, Zhang GX, Kaponides G, Höjeberg B, Vander Meide PH, and Link H.(1997) IL-10 suppresses exper-imental autoimmune neuritis and down-regulates TH1-type immune responses. Clin Immunol Immunopathol 83:117-126

11. Bauer J, Bradl M, Hickey WF, Forss-Petter S, BreitschopfH, Linington C, Wekerle H, and Lassmann H.(1998) T-cellapoptosis in inflammatory brain lesions - Destruction of Tcells does not depend on antigen recognition. Am J Pathol153: 715-724

12. Beck RW, Cleary PA, Trobe JD, Kaufman DI, KupersmithMJ, Paty DW, Brown CH, and and the Optic Neuritis StudyGroup.(1993) The effect of corticosteroids for acute opticneuritis on the subsequent development of multiple scle-rosis. N Engl J Med 329: 1764-1769

13. Benoist C and Mathis D.(1998) The pathogen connection.Nature 394: 227-228

14. Bergsteinsdottir K, Kingston A, and Jessen KR.(1992) RatSchwann cells can be induced to express major histo-compatibility complex class II molecules in vivo. JNeurocytol 21: 382-390

15. Bergsteinsdottir K, Kingston A, Mirsky R, and JessenKR.(1991) Rat Schwann cells produce interleukin-1. JNeuroimmunol 34: 15-23

16. Bolin LM, Verity N, Silver JE, Shooter EM, and AbramsJS.(1995) Interleukin-6 production by Schwann cells andinduction in sciatic nerve injury. J Neurochem 64: 850-858

17. Bonetti B, Invernizzi F, Rizzuto N, Bonazzi ML, ZanussoGL, Chinaglia G, and Monaco S.(1997) T-cell-mediatedepineurial vasculitis and humoral-mediated microan-giopathy in cryoglobulinemic neuropathy. J Neuroimmunol73: 145-154

18. Boumpas DT, Paliogianni F, Anastassiou ED, and BallowJE.(1991) Glucocorticosteroid action on the immune sys-tem: molecular and cellular aspects. Clin Exp Rheumatol9: 413-423

19. Brann DW, Hendry LB, and Mahesh VB.(1995) Emergingdiversities in the mechanism of action of steroid hor-mones. J Steroid Biochem Molec Biol 52: 113-133

20. Bursch W, Paffe S, Putz B, Barthel G, and Schulte-Hermann R.(1990) Determination of the length of the his-tological stages of apoptosis in normal liver and in alteredhepatic foci of rats. Carcinogenesis 11: 847-853

21. Butcher EC and Picker LJ.(1996) Lymphocyte homingand homeostasis. Science 272: 60-66

22. Buttgereit F, Wehling M, and Burmester GR.(1998) A newhypothesis of modular glucocorticoid actions. ArthritisRheumat 41: 761-767

23. Chao DT and Korsmeyer SJ.(1998) BCL-2 family:Regulators of cell death. Annu Rev Immunol 16: 395-419

24. Chluba J, Steeg C, Becker A, Wekerle H, and EpplenJT.(1989) T cell receptor beta chain usage in myelin basicprotein-specific rat T lymphocytes. Eur J Immunol 19: 279-284

25. Cohen JJ, Duke RC, Fadok VA, and Sellins KS.(1992)Apoptosis and programmed cell death in immunity. AnnuRev Immunol 10: 267-293

26. Constable AL, Armati PJ, Toyka KV, and Hartung H-P.(1994) Production of prostanoids by Lewis rat Schwanncells in vitro . Brain Res 635: 75-80

26b. Conti G, Scarpini E, Rostami A, Livraghi S, Baron PL,Pleasure D, Scarlato (1998) Schwann cell undergoesapoptosis during experimental allergic neuritis (EAN). JNeurol Sci 161: 29-35

27. Critchfield JM, Racke MK, Zuniga Pflucker JC, CannellaB, Raine CS, Goverman J, and Lenardo MJ.(1994) T celldeletion in high antigen dose therapy of autoimmuneencephalomyelitis. Science 263: 1139-1143

28. Dashiell SM, Vanguri P, and Koski C.(1997) Inflammatorycytokines mediate dibutyryl cyclin AMP-induced C3expression in glial cells. Glia 20: 308-321

29. Dyck PJ, O'Brien PC, Oviatt KF, Dinapoli RP, Daube JR,Bartleson JD, Mokri B, Swift T, Low PA, and WindebankAJ.(1982) Prednisone improves chronic inflammatorydemyelinating polyneuropathy more than no treatment.Ann Neurol 11: 136-141

30. Enders U, Lobb R, Pepinsky RB, Hartung H-P, Toyka KV,and Gold R.(1998) The role of the very late antigen-4(VLA-4) and its counterligand vascular cell adhesion mol-ecule-1 (VCAM-1) in the pathogenesis of experimentalautoimmune neuritis (EAN) of the Lewis rat. Brain 121:1257-1266

31. Fabry Z, Raine CS, and Hart MN.(1994) Nervous tissueas an immune compartment: the dialect of the immuneresponse in the CNS. Immunol Today 15: 218-224

32. Gabriel CM, Hughes RAC, Moore SE, Smith KJ, andWalsh FS.(1998) Induction of experimental autoimmuneneuritis with peripheral myelin protein-22. Brain 121:1895-1902

33. Gavrieli Y, Sherman Y, and Ben Sasson SA.(1992)Identification of programmed cell death in situ via specificlabeling of nuclear DNA fragmentation. J Cell Biol 119:493-501

34. Gillen C, Jander S, and Stoll G.(1998) Sequential expres-sion of mRNA for proinflammatory cytokines and inter-leukin-10 in the rat peripheral nervous system:Comparison between immune-mediated demyelinationand Wallerian degeneration. J Neurosci Res 51: 489-496

35. Gold R, Hartung HP, and Lassmann H.(1997) T-cell apop-tosis in autoimmune diseases: termination of inflamma-tion in the nervous system and other sites with specializedimmune-defense mechanisms. Trends Neurosci 20: 399-404

36. Gold R, Oehlschlager M, Pepinsky RB, Toyka KV, andHartung H-P. Lipocortin-1 expression in the sciatic nerveof Lewis rats with experimental autoimmune neuritis.[Abstract] J Neurol 1994;241: S156

37. Gold R, Pepinsky RB, Zettl UK, Toyka KV, and HartungHP.(1996) Lipocortin-1 (annexin-1) suppresses activationof autoimmune T cell lines in the Lewis rat. JNeuroimmunol 69: 157-164


38. Gold R, Schmied M, Giegerich G, Breitschopf H, HartungHP, Toyka KV, and Lassmann H.(1994) Differentiationbetween cellular apoptosis and necrosis by the combineduse of in situ tailing and nick translation techniques. LabInvest 71: 219-225

39. Gold R, Schmied M, Tontsch U, Hartung HP, Wekerle H,Toyka KV, and Lassmann H.(1996) Antigen presentationby astrocytes primes rat T lymphocytes for apoptotic celldeath: A model for T cell apoptosis in vivo. Brain 119: 651-659

40. Gold R, Toyka KV, and Hartung HP.(1995) Synergisticeffect of gamma-IFN and TNF-a on expression of immunemolecules and antigen presentation by Schwann cells.Cell Immunol 165: 65-70

41. Gold R, Zielasek J, Kiefer R, Toyka KV, and HartungHP.(1996) Secretion of nitrite by Schwann cells and itseffect on T-cell activation in vitro . Cell Immunol 168: 69-77

42. Griffin JW and George R.(1993) Editorial. The residentmacrophages in the peripheral nervous system arerenewed from the bone marrow: new variations on an oldtheme. Lab Invest 69: 257-260

43. Griffin JW, George R, Lobato C, Tyor WR, Yan LC, andGlass JD.(1992) Macrophage responses and myelinclearance during Wallerian degeneration: relevance toimmune-mediated demyelination. J Neuroimmunol 40:153-165

44. Hafer-Macko C, Hsieh ST, Li CY, Ho TW, Sheikh K,Cornblath DR, McKhann GM, Asbury AK, and GriffinJW.(1996) Acute motor axonal neuropathy: an antibody-mediated attack on axolemma. Ann Neurol 40: 635-644

45. Hartung H-P, Schwenke C, Bitter-Suermann D, and ToykaKV.(1997) Guillain-Barré Syndrome: activated comple-ment components C3a and C5a in cerebrospinal fluid.Neurology 37: 1006-1009

46. Hartung H-P, Toyka KV, and Griffin JW.(1998) Guillain-Barre syndrome and chronic inflammatory demyelinatingpolyneuropathy. Clinical neuroimmunology. Antel J,Birnbaum G, and Hartung H-P (eds). Blackwell Science:Malden. pp: 294-306

47. Hartung HP, Pollard JD, Harvey GK, and Toyka KV.(1995)Immunopathogenesis and treatment of the Guillain-Barresyndrome - part I. Muscle & Nerve 18: 137-15348.

48. Hartung HP, Reiners K, Michels M, Hughes RAC,Heidenreich F, Zielasek J, Enders U, and Toyka KV.(1994)Serum levels of soluble E-selectin (ELAM-1) in immune-mediated neuropathies. Neurology 44: 1153-1158

49. Hartung HP, Schäfer B, Heininger K, Stoll G, and ToykaKV.(1988) The role of macrophages and eicosanoids inthe pathogenesis of experimental allergic neuritis. Serialclinical, electrophysiological, biochemical and morpholog-ical observations. Brain 111: 1039-1059

50. Hartung HP, Stoll G, and Toyka KV.(1993) Immune reac-tions in the peripheral nervous system. Peripheral neu-ropathy. 3rd ed.: Dyck PJ, Thomas PK, Griffin JW, Low PA,and Poduslo JF. (eds.) W.B. Saunders Company:Philadelphia. pp 418-444

50b. Hartung HP (1999) Editorial Commentary: Infections andthe Guillain-Barrè syndrome. J Neurol NeurosurgPsychiatry 66: 277

50c. Hartung HP, van der Meché FGA, Pollard JD (1998)Editorial Review: Guillain-Barrè syndrome, CIDP andother chronic immune-mediated neuropathies. Curr OpinNeurol 11: 497-513

51. Hartung HP, Willison H, Jung S, Pette M, Toyka KV, andGiegerich G.(1996) Autoimmune responses in peripheralnerve. Springer Semin Immunopathol 18: 97-123

52. Harvey GK, Gold R, Toyka KV, and Hartung HP.(1995)Nonneural-specific T lymphocytes can orchestrate inflam-matory peripheral neuropathy. Brain 118: 1263-1272

53. Ho TW, McKhann M, and Griffin JW.(1998) Humanautoimmune neuropathies. Ann Rev Neurosci 21: 187-226

54. Hua LL, Liu JSH, Brosnan CF, and Lee SC.(1998)Selective inhibition of human glial inducible nitric oxidesynthase by interferon-b: Implications for multiple sclero-sis. Ann Neurol 43: 384-387

55. Hughes PM, Wells GMA, Clements JM, Gearing AJH,Redford EJ, Davies M, Smith KJ, Hughes RAC, BrownMC, and Miller KM.(1998) Matrix metalloproteinaseexpression during experimental autoimmune neuritis.Brain 121: 481-494

56. Jacobs BC, van Doorn PA, Groeneveld JHM, Tio-GillenAP, and van der Meche FGA.(1997) Cytomegalovirusinfections and anti-GM2 antibodies in Guillain-Barre syn-drome. J Neurol Neurosurg Psychiatry 62: 641-643

57. Jander S, Pohl J, Gillen C, and Stoll G.(1996) Differentialexpression of interleukin-10 mRNA in Wallerian degener-ation and immune-mediated inflammation of the ratperipheral nervous system. J Neurosci Res 43: 254-259

58. Jung S, Huitinga I, Schmidt B, Zielasek J, Dijkstra CD,Toyka KV, and Hartung HP.(1993) Selective elimination ofmacrophages by dichlormethylene diphosphonate-con-taining liposomes suppresses experimental autoimmuneneuritis. J Neurol Sci 119: 195-202

59. Jung S, Schluesener HJ, Schmidt B, Fontana A, ToykaKV, and Hartung HP.(1994) Therapeutic effect of trans-forming growth factor-beta 2 on actively induced EAN butnot adoptive transfer EAN. Immunology 83: 545-551

60. Jung S, Toyka K, and Hartung HP.(1996) T cell directedimmunotherapy of inflammatory demyelination in theperipheral nervous system - Potent suppression of theeffector phase of experimental autoimmune neuritis byanti-CD2 antibodies. Brain 119: 1079-1090

61. Kerr JFR, Wyllie AH, and Currie AR.(1972) Apoptosis: abasic biological phenomenon with wide ranging implica-tions in tissue kinetics. Br J Cancer 26: 239-257

62. Kiefer R, Funa K, Schweitzer T, Jung S, Bourde O, ToykaKV, and Hartung HP.(1996) Transforming growth factor-b1in experimental autoimmune neuritis - Cellular localizationand time course. Am J Pathol 148: 211-223

63. Kiefer R, Gold R, Gehrmann J, Lindholm D, Wekerle H,and Kreutzberg GW.(1993) Transforming growth factorbeta expression in reactive spinal cord microglia andmeningeal inflammatory cells during experimental allergicneuritis. J Neurosci Res 36: 391-398


64. Kiefer R, Kieseier BC, Brück W, Hartung H-P, and ToykaKV.(1998) Macrophage differentiation antigens in acuteand chronic autoimmune polyneuropathies. Brain 121:469-479

65. Kieseier BC, Clements JM, Pischel HB, Wells GMA, MillerK, Gearing AJH, and Hartung HP.(1998) Matrix metallo-proteinases MMP-9 and MMP-7 are expressed in experi-mental autoimmune neuritis and the Guillain-Barre syn-drome. Ann Neurol 43: 427-434

66. Kingston AE, Bergsteinsdottir K, Jessen KR, Van derMeide PH, Colston MJ, and Mirsky R.(1989) Schwanncells co-cultured with stimulated T cells and antigenexpress major histocompatibility complex (MHC) class IIdeterminants without interferon-gamma pretreatment:synergistic effects of interferon-gamma and tumor necro-sis factor on MHC class II induction. Eur J Immunol 19:177-183

67. Kojima K, Berger T, Lassmann H, Hinze Selch D, Zhang Y,Gehrmann J, Reske K, Wekerle H, and Linington C.(1994)Experimental autoimmune panencephalitis and uveore-tinitis transferred to the Lewis rat by T lymphocytes spe-cific for the S100 beta molecule, a calcium binding proteinof astroglia. J Exp Med 180: 817-829

68. Koski CL.(1990) Characterization of complement-fixingantibodies to peripheral nerve myelin in Guillain-Barrésyndrome. Ann Neurol 27(Suppl.): 44-47

69. Koski CL.(1997) Mechanisms of Schwann cell damage ininflammatory neuropathy. J Infect Dis 176(Suppl.): 169-172

70. Koski CL, Estep AE, Sawant-Mane S, Shin ML,Highbarger L, and Hansch GM.(1996) Complement regu-latory molecules on human myelin and glial cells:Differential expression affects the deposition of activatedcomplement proteins. J Neurochem 66: 303-312

70b. Mix E, Zettl UK, Zielasek J, Hartung HP, Gold R (1999)Apoptosis induction by macrophage-derived reactive oxy-gen species in myelin-specific T cells requires cell-cellcontact. J Neuroimmunol (in press)

71. Lefcort F, Venstrom K, McDonald J, and ReichardtT.(1992) Regulation of expression of fibronectin and itsreceptor alpha5beta1, during development and regenera-tion of peripheral nerve. Development 116: 767-782

72. Leger OJP, Yednock TA, Tanner L, Horner HC, Hines DK,Keen S, Saldanha J, Jones T, Fritz LC, and BendigMM.(1997) Humanization of a ouse antibody againsthuman alpha-4 integrin: a potential therapeutic for thetreatment of multiple sclersosis. Hum Antibod 8: 3-16

73. Liblau R, Tisch R, Bercovici N, and McDevitt HO.(1997)Systemic antigen in the treatment of T-cell-mediatedautoimmune diseases. Immunol Today 18: 599-604

74. Linington C, Izumo S, Suzuki M, Uyemura K, MeyermannR, and Wekerle H.(1984) A permanent rat T cell line thatmediates experimental allergic neuritis in the Lewis rat invivo. J Immunol 133: 1946-1950

75. Linington C, Lassmann H, Ozawa K, Kosin S, andMongan L.(1992) Cell adhesion molecules of theimmunoglobulin supergene family as tissue-specificautoantigens: induction of experimental allergic neuritis(EAN) by PO protein-specific T cell lines. Eur J Immunol22: 1813-1817

76. Lisak RP, Skundric DS, Bealmear B, and RaghebS.(1997) The role of cytokines in Schwann cell damage,protection and repair. J Infect Dis 176(Suppl.): 173-179

77. Malipiero U, Frei K, Spanaus KS, Agresti C, Lassmann H,Hahne M, Tschopp J, Eugster HP, and Fontana A.(1997)Myelin oligodendrocyte glycoprotein-induced autoim-mune encephalomyelitis is chronic/relapsing in perforinknockout mice, but monophasic in Fas- and Fas ligand-deficient lpr and gld mice. Eur J Immunol 27: 3151-3160

78. Mason D.(1991) Genetic variation in the stress response:susceptibility to experimental allergic encephalomyelitisand implications for human inflammatory disease.Immunol Today 12: 57-60

79. Medawar P.(1948) Immunity to homologous grafted skin.III. The fate of skin homografts transplanted to the brain,to subcutaneous tissue, and to the anterior chamber ofthe eye. Brit J Exp Path 29: 58-69

80. Morfin R, Young J, Corpechot C, Egestad B, Sjovall J, andBaulieu EE.(1992) Neurosteroids: pregnenolone inhuman sciatic nerves. Proc Natl Acad Sci U S A 89: 6790-6793

81. Mousa SA and Cheresh DA.(1997) Recent advances incell adhesion molecules and extracellular matrix proteins:potential clinical implications. Drug Develop Today 2: 187-199

82. Nagata S and Golstein P.(1995) The Fas death factor.Science 267: 1449-1455

83. Neumann H, Boucraut J, Hahnel C, Misgeld T, andWekerle H.(1996) Neuronal control of MHC class IIinducibility in rat astrocytes and microglia. Eur J Neurosci8: 2582-2590

84. Ohmori K, Hong Y, Fujiwara M, and Matsumoto Y.(1992)In situ demonstration of proliferating cells in the rat centralnervous system in experimental autoimmuneencephalomyelitis. Lab Invest 66: 54-62

85. Panegyres PK, Faull RJ, Russ GR, Appleton SL, WangelAG, and Blumbergs PC.(1992) Endothelial cell activationin vasculitis of peripheral nerve and skeletal muscle. JNeurol Neurosurg Psych 55: 4-7

86. Pender MP, Nguyen KB, McCombe PA, and KerrJF.(1991) Apoptosis in the nervous system in experimen-tal allergic encephalomyelitis. J Neurol Sci 104: 81-87

87. Pollard JD.(1987) A critical review of therapies in acuteand chronic demyelinating polyneuropathies. Muscle &Nerve 10: 214-221

88. Pollard JD, Baverstock H, and McLeod JG.(1987) Class IIantigen expression and inflammatory cells in the Guillain-Barre syndrome. Ann Neurol 21: 337-341

89. Pollard JD, Westland KW, Harvey GK, Jung S, Bonner J,Spies JM, Toyka KV, and Hartung H-P.(1995) Activated Tcells of nonneural specificity open the blood-nerve barrierto circulating antibody. Ann Neurol 37: 467-475

90. Previtali S, Archelos JJ, and Hartung H-P.(1998)Expression of integrins in experimental autoimmune neu-ritis and Guillain-Barré syndrome. Ann Neurol 44: 611-621

91. Quattrini A, Previtali S, Feltri ML, Canal N, Nemni R, andWrabetz L.(1996) Beta1 integrin and other Schwann cellmarkers in axonal neuropathy. Glia 17: 294-306


92. Redford EJ, Smith KJ, Gregson NA, Davies M, Hughes P,Gearing AJH, Miller K, and Hughes RAC.(1997) A com-bined inhibitor of matrix metalloproteinase activity andtumour necrosis factor-a processing attenuates experi-mental autoimmune neuritis. Brain 120: 1895-1905

93. Roche PH, Figarella-Branger D, Daniel L, Bianco N, PelletW, and Pellsier JF.(1997) Expression of cell adhesion mol-ecules in normal nerves, chronic axonal neuropathies andSchwann cell tumors. J Neurol Sci 151: 127-133

94. Savill J, Fadok V, Henson P, and Haslett C.(1993)Phagocyte recognition of cells undergoing apoptosis.Immunol Today 14: 131-136

95. Sawant-Mane S, Piddlesden SJ, Morgan BP, Holers VM,and Koski CL.(1996) CD59 homologue regulates comple-ment-depedent cytolysis of rat Schwann cells. JNeuroimmunol 69: 63-71

96. Schmidt B, Stoll G, Hartung HP, Heininger K, Schafer B,and Toyka KV.(1990) Macrophages but not Schwann cellsexpress Ia antigen in experimental autoimmune neuritis.Ann Neurol 28: 70-77

97. Schmidt B, Toyka KV, Kiefer R, Full J, Hartung HP, andPollard J.(1996) Inflammatory infiltrates in sural nervebiopsies in Guillain-Barre syndrome and chronic inflam-matory demyelinating neuropathy. Muscle Nerve 19: 474-487

98. Schmidt J, Gold R, Hartung H-P, Zettl U, and ToykaKV.(1997) Corticosteroid pulse therapy in experimentalautoimmune encephalomyelitis (EAE) standard methyl-prednisolone therapy is not sufficient to augmentT cellapoptosis in situ. Neurology 48: A422-A423

99. Schmied M, Breitschopf H, Gold R, Zischler H, Rothe G,Wekerle H, and Lassmann H.(1993) Apoptosis of T lym-phocytes in experimental autoimmune encephalomyelitis.Evidence for programmed cell death as a mechanism tocontrol inflammation in the brain. Am J Pathol 143: 446-452

100. Schneider C, Dalakas MC, Toyka KV, Said G, Hartung H-P, and Gold R.(1999) T cell apoptosis is not augmented ininflammatory neuromuscular disorders associated withHIV-infection. Arch Neurol 56:79-83

101. Schneider C, Gold R, Dalakas MC, Schmied M,Lassmann H, Toyka KV, and Hartung H-P.(1996) MHCclass I mediated cytotoxicity does not induce apoptosis inmuscle fibers nor in inflammatory T cells: Studies inpatients with polymyositis, dermatomyositis, and inclusionbody myositis. J Neuropathol Exp Neurol 55: 1205-1209

102. Sessa G, Nemni R, Canal N, and Marchisio PC.(1997)Circulating fragments of myelin-associated a6b4 integrinin Guillain-Barre syndrome. J Neuroimmunol 80: 115-120

103. Stoll G, Jander S, Jung S, Archelos J, Tamatani T,Miyasaka M, Toyka KV, and Hartung HP.(1993)Macrophages and endothelial cells express intercellularadhesion molecule-1 in immune-mediated demyelinationbut not in Wallerian degeneration of the rat peripheralnervous system. Lab Invest 68: 637-644

104. Stoll G, Jung S, Jander S, van der Meide P, and HartungHP.(1993) Tumor necrosis factor-a in immune-mediateddemyelination and Wallerian degeneration of the ratperipheral nervous system. J Neuroimmunol 45: 175-182

105. Streilein JW.(1995) Unraveling immune privilege. Science270: 1158-1159

106. Tabi Z, McCombe PA, and Pender MP.(1994) Apoptoticelimination of V beta 8.2+ cells from the central nervoussystem during recovery from experimental autoimmuneencephalomyelitis induced by the passive transfer of Vbeta 8.2+ encephalitogenic T cells. Eur J Immunol 24:2609-2617

107. Tabi Z, McCombe PA, and Pender MP.(1995) Antigen-specific down-regulation of myelin basic protein-reactive Tcells during spontaneous recovery from experimentalautoimmune encephalomyelitis: further evidence of apop-totic deletion of autoreactive T cells in the central nervoussystem. Int Immunol 7: 967-973

108. Tontsch U and Rott O.(1993) Cortical neurons selectivelyinhibit MHC class II induction in astrocytes but not inmicroglial cells. Int Immunol 5: 249-254

109. Vass K, Hickey WF, Schmidt RE, and Lassmann H.(1993)Bone-marrow derived elements in the peripheral nervoussystem. An immunohistochemical and ultrastructuralinvestigation in chimeric rats. Lab Invest 69: 275-282

110. Vedeler C, Ulvestad E, Bjorge L, Conti G, Williams K,Mork S, and Matre R.(1994) The expression of CD59 innormal human nervous tissue. Immunology 82: 542-547

111. Vedeler CA, Matre R TI (1990) Peripheral nerve CR1 limitcomplement-mediated haemolysis. J Neuroimmunol 30:95-8

112. Weishaupt A, Gold R, Giegerich G, Hartung H-P, andToyka KV.(1997) Antigen therapy eliminates T-cell inflam-mation by apoptosis: Effective treatment of experimentalautoimmune neuritis with recombinant myelin protein P2.Proc Natl Acad Sci USA 94: 1338-1342

113. Wekerle H, Linington C, Lassmann H, and MeyermannR.(1986) Cellular immune reactivity within the CNS.Trends Neurosci 9: 271-277

114. Wekerle H, Schwab M, Linington C, and MeyermannR.(1986) Antigen presentation in the peripheral nervoussystem: Schwann cells present endogenous myelinautoantigens to lymphocytes. Eur J Immunol 16: 1551-1557

115. Wyllie AH, Kerr JFR, and Currie AR.(1980) Cell death: thesignificance of apoptosis. Int Rev Cytol 68: 251-306

116. Yi Hu Z, Bourreau E, Jung-Testas I, Robel P, and BaulieuEE.(1987) Neurosteroids: oligodendrocyte mitochondriaconvert cholesterol to pregnenolone. Proc Natl Acad SciUSA 84: 8215-8219

117. Yuki N.(1997) Molecular mimicry between gangliosidesand lipopolysaccharides of Campylobacter jejuni isolatedfrom patients with Guillain-Barre syndrome and MillerFisher syndrome. J Infect Dis 176: S150-S153

118. Zettl UK, Gold R, Toyka KV, and Hartung HP.(1996) In situdemonstration of T cell activation and elimination in theperipheral nervous system during experimental autoim-mune neuritis in the Lewis rat. Acta Neuropathol (Berl )91: 360-367

119. Zettl UK, Mix E, Zielasek J, Stangel M, Hartung HP, andGold R.(1997) Apoptosis of myelin-reactive T cellsinduced by reactive oxygen and nitrogen intermediates invitro . Cell Immunol 178: 1-8


Documents

Mechanisms of Immune Regulation in the Peripheral Nervous System