key: cord-0009968-7omkcwro
authors: Fontana, Adriano; Frei, Karl; Bodmer, Stefan; Hofer, Erhard
title: Immune‐Mediated Encephalitis: On the Role of Antigen‐Presenting Cells in Brain Tissue
date: 2006-04-28
journal: Immunol Rev
DOI: 10.1111/j.1600-065x.1987.tb00532.x
sha: 3e5625149849daca66ed0082029b99a209b4b539
doc_id: 9968
cord_uid: 7omkcwro

nan

The observation of impaired rejection of xenografts implanted into the brain parenchyma led to the description of the brain as an "immune-privileged site". Given an intact blood brain barrier, lymphocyte traffic through the central nervous system (CNS) tissue is not very likely to occur. However, an early and effective elimination of antigens, e.g. neurotropic viruses, is required also within the CNS. Indeed., intact effector functions of the immune system within the brain are evidenced by 1) the demonstration of T lymphocyte infiltrations in viral and autoimmune encephalitis (Traugott et al. 1981 (Traugott et al. , 1982 , 2) the successful transfer of experimental autoimmune encephalitis (EAE) with myelin basic protein (MBP)-specificT cells (Paterson 1960 , Ben-Nun et al. 1981 , 3) the synthesis of immunoglobulins within the CNS in some forms of encephalitis (Tourtelotte & Ma 1978) and 4) the prevention of brain damage in certain experimental viral diseases by treatment with imjnunosuppressive drugs (Lipton & Dal Canto 1976) . In order to fulfil the functions of the immune system within the CNS, i.e. recognition and elimination of antigens, the immune system may require regulatory elements at the interface to the CNS. As will be discussed in the following sections, the proposed regulatory system may propagate immune functions in the brain tissue but may also restrict immune reactions to an absolute minimum in order to spare the vital neuronal system. Dysfunction of the regulatory system may contribute to the development of immune-mediated encephahtis. Secretion of immunosuppressive factors may alter immunoreactivity within the CNS. A peptide with T cell suppressor activity has recently been purified from conditioned medium of cultured glioblastoma cells (see section 6).

The formation of a dense astro-glial scar is a prominent feature of highly inflammatory CNS wounds. The hypertrophy and hyperplasia of astrocytes may play an important role in the provision of neurotropic factors for mature neurons (for review see Lindsay 1986 ). However, it could also reflect participation of astrocytes in immune-mediated reactions. Indeed, upon treatment with lipopolysaccharide (LPS, E.coli)-cultured astrocytes established from newborn mice were found to secrete interleukin-1 (IL-1) (Fontana et al. 1982) . IL-1-like factors were also identified in conditioned medium of C-6 rat glioma cells and human glioblastoma cells (Fontana et al. 1983 (Fontana et al. , 1984a . Furthermore, IL-1 was detected in brain extracts obtained either from mice after intraperitoneal injection of LPS or from Lewis rats injected with encephalitogenic myelin basic protein (MBP)-specific T cell lines (Fontana 1984b , Fierz & Fontana 1986a . Besides astrocytes, microglial cells also secrete lL-1 (Giulian et al. 1986 ). Depending upon the species, as well as the strain of animals used for cell cultures, the capacity of IL-1 production by microglial cells is variable compared to IL-1 production by astrocytes (personal observation). The significance of intracerebral synthesis of IL-1 may be fundamental to intracerebral T cell activation as IL-I enhances production of interleukin-2 (IL-2) and expression of IL-2 receptors on T cells. To evaluate the capacity of astrocytes to function as antigen-presenting cells (APC), astrocytes of Lewis rats were cocultured with syngeneic MBP-specific, la-restricted T cell line cells of Lewis rat origin. Astrocytes clearly stimulated the proliferation of the T cells, the process being antigen-specific and restricted to the major histocompatibility complex (MHC) (Fontana et al. 1984c) . During such cocultivation of T cells and astrocytes, the latter were induced by the interacting preactivated T cells to express la antigens (Fontana et al. 1984c , Fierz et al. 1985 . Independently of our studies, Hirsch et al. (1983) demonstrated that interferon-y (IFN->')-containing supernatants of lectin-stimulated spleen cells induced murine astrocytes to express la antigens. Similar effects of recombinant IFN-y were shown later on murine astrocytes by Wong et al. (1984) and Fierz et al. (1985) and on human astrocyte cell lines by Pulver et al. (1987) . The dependence of astrocytes as antigenpresenting cells on la-inducing signals such as IFN-y was also clearly demonstrated when taking astrocytes as stimulator cells and, instead of T cell lines, using either unprimed resting lymph node T cells as responders in syngeneic or allogeneic lymphocyte reactions or la-restricted beef insuhn-specific T hybridomas which do not secrete IFN-y but release IL-2 after antigen-specific activation. In both assays, only IFN-y-treated but not untreated astrocytes served as APC , Frb et al. 1986 ).

These data have been confirmed by Takiguchi & Frelinger (1986) who have demonstrated recently that IFN-y-treated BIO.A astrocytes were effective in presenting antigen to T cell hybrids. Untreated astrocytes or IFN-y-treated oligo-dendrocytes were not able to function as APC. In this regard, it is of interest that Suzumura et at. (1986a) failed to show la antigen expression on IFN-ytreated mouse oligodendrocytes. Since in the rat optic nerve a glial progenitor cell being A2B5-positive is committed to become either an oligodendrocyte or a type 2 (fibrous) astrocyte (Raff et al. 1983) it is pertinent to ask whether both oligodendrocytes and the type 2 astrocytes are uninducible to express la antigens. The la-positive astrocytes would then be mainly type 1 (protoplasmic) astrocytes which develop from an own progenitor cell population. Staining of Lewis rat astrocyte cultures for type 1 astrocyte-specific Ran-2 antigen revealed that 90% of all cells were positive and 95% of the cells were glial fibrillary acidic protein (GFAP)-positive astrocytes (Massa et al. 1987) . Since, upon IFN-y treatment, up to 50% of the cultured cells become la-positive, Ran-2 antigen-positive type I astrocytes have to be inducible. Studies are needed to test for la antigen expression on A2B5-positive progenitor cells and on type 2 astrocytes. When using a Percoll density gradient method to establish glial cell cultures from adult human brain, however, MHC class II antigens were identified not only on astrocytes but also on some oligodendrocytes (Kim et al. 1985) .

For an understanding of the pathogenesis of immune-mediated encephalitis in viral infection, the report of Massa et al. (1986) is of particular relevance. When infected with coronavirus, astrocytes established from brains of newborn Lewis rats became la antigen-positive. Anti-IFN-y antibodies did not abrogate the lainducing effect of the virus. Virus-neutralizing antibodies, however, blocked la antigen induction by the virus. UV-inactivated virus was found to be as effective as infective virus. The possibility of IFN-y-independent mediators being released by cells binding to the virus has to be clarified in the future.

Unlike la antigens, MHC class I antigens are expressed on cultured murine astrocytes ). This does not reflect the in vivo situation, since brain cells, including astrocytes, usually express only very low levels of class I antigens: cell suspensions prepared from brains of newborn mice showed less than 1% of H-2K-positive cells (Wong et al. 1984) . However, after intracerebral injection of IFN-y about 50% of brain cells expressed detectable H-2K antigens (Wong et al. 1984) . Very low amounts of class 1 antigens were also detected in human brain tissue (Williams et al. 1980 ). The expression of H-2K molecules on our cultured astrocytes may be due to stimulatory components of the culture medium (e.g. from fetal calf serum) which enhance expression of H-2K. When taking cultured astrocytes not being pretreated with IFN-y, haptenated H-2Kpositive astrocytes were able to support growth of hap ten-specific cytotoxic T cells ). More recent studies by Skias et al. (1987) have provided further support for the expression of MHC class I antigens on astrocytes: mouse astrocytes in vitro are susceptible targets of MHC class I-restricted cytotoxicity by cytotoxic T cells. In vivo expression of MHC class I is induced on both astrocytes and oligodendrocytes during neurotropic mouse hepatitis virus (strain A 59) infection which leads to a chronic demyelinating CNS disease (Suzumura etal. 1986b) .

In contrast to MHC class I, class II molecules are only expressed on certain types of cells, namely on B cells, monocytes, and dendritic cells. The tissue expression of la reflects the distribution of these cells in the different organs, e.g. la-positive dendritic cells being observed in the interstitial connective tissues of all the organs tested (heart, liver, thyroid, pancreas, skin, kidney, ureter, skeletal, muscle and bladder) except brain (Hart & Fabre 1981) . When examining the brains of BIO.A mice, small nxunbers (< 1 %) of la-positive cells were identified (Ting et al. 1981) . Studies on human tissue performed by Hauser et al. (1983) have demonstrated that 1-2% of human brain cells express class II antigens of the MHC. Investigating normal rat brain tissue sections by using immunofluorescence techniques, we could only fmd occasional la-positive cells in the meninges, especially around blood vessels, but the CNS parenchyma inside of the limiting membrane was free of la-positive cells (Fierz & Fontana 1986b ). Therefore, besides the blood brain barrier, also the absence of MHC class II antigens in the brain puts an additional severe limit on the expression of immune functions within the CNS.

Direct injections of IFN-y into the brains of mice induced la antigens on astrocytes but not on oligodendrocytes or neurons, indicating that astrocytes have the potential to express MHC class II antigens in vivo (Wong et al. 1984) . Intravenous treatment of BIO.BR mice with IFN-y results in a dramatic increase of la antigen expression throughout the body, including de novo expression on capillary endothelial cells, KupfTer cells, epithelial cells of the intestinal tract and kidney tubular epithelium. However, neurons, astrocytes, oligodendrocytes, microglial cells and even the capillaries of the brain were not inducible for la antigen by this form of treatment, which may be due to poor penetration of IFNy through the intact blood brain barrier (Momburg et al. 1986 ). Patients with multiple sclerosis being treated with intravenous infusions of IFN-y responded with exacerbations of the disease (Panitch et al. 1987) . This effect could be due to induction of la antigens on brain endothelial cells or intraparenchymal antigenpresenting cells by IFN-y penetrating through the altered blood brain barrier in these patients.

The observation of MHC class Il-positive cells in brain tissue during immunemediated diseases would provide another approach towards the understanding of the role of local APC. Using a modified peroxidase-antiperoxidase technique, Traugott et al. (1985) demonstrated la expression on endothelial cells and astrocytes in brain lesions in multiple sclerosis (MS). The highest density ofla-positive astrocytes in active chronic MS was at the lesion edge and within the zone of adjacent normal-appearing white matter (Traugott et al. 1985) . In brain sections of the 3 patients studied by Hofman et al. (1986) , the majority of the la-bearing tissue cells stained for the astrocyte marker GFAP as shown by double staining. In a recent immunoelectron microscopic study on CNS demyelination induced by Theiler's murine encepbalomyelitis virus, the majority of la-positive glial cells in susceptible BIO.S and B10.ASR2 mice had the morphologic characteristics of astrocytes (Rodriguez et al. 1987) . Some la-positive oligodendrocytes were detected as well. In the uninfected mice or in virus-infected resistant B10.S(9R) mice only occasional la-positive microglial cells were observed, the astrocytes and oligodendrocytes being la-negative. In SJL mice with acute or chronic relapsing EAE induced by MBP-specific T cell lines, the I-A'-positive cells in the parenchyma were identified on serial sections to be astrocytes (Sakai et al. 1986 ). However, when investigating Lewis rats with EAE, Matsumoto et al. (1986) failed to detect la-positive astrocytes and Hickey et al. (1985) only demonstrated low numbers of such cells. In detailed studies using electron microscope immune cytochemistry neither astrocytes nor endothelial cells were found to be positive in Lewis rats with EAE induced by immunization with MBP or injection of MBP-specific T cell line cells ). The same investigators identified some la-positive astrocytes in chronic progressive EAE lesions ).

In general it is not entirely clear why in some systems the la-positive astrocytes have been described as a hallmark of the immuno-histological picture (Traugott et al. 1985 , Sakai etal. 1986 , Hofman etal. 1986 , Rodriguez et al. 1987 ) and in others no or only low numbers of la-positive astrocytes were noted (Hickey et al. 1985 , Matsumoto et al. 1986 ). These observations can be interpreted as an indication that the capacity of astrocytes to function as APC in vivo may be relevant in one disease or disease model, such as viral immunemediated encephalitis or multiple sclerosis, but not in other circumstances, such as acute EAE in Lewis rats. However, as pointed out by Vass et al. (1986) , technical problems may account for controversial results in regard to in vivo demonstration of la antigens on endothelial cells and astrocytes. As severe demyelination in multiple sclerosis has been found to be associated with destruction of astrocytes (Itoyama et al. 1985 ) and la-restricted T cell-mediated cytotoxicity of la-bearing astrocytes has been demonstrated (Sun & Wekerle 1986) , it may well be that negative staining for la-positive astrocytes is due to loss of such cells as a consequence of interactions with cytotoxic la-restricted T cells.

Following the work of Pasteur on treatment of rabies, Anjeszky, whilst attempting to immunize dogs against rabies by the inoculation of normal nervous tissue.

found (in 1900) that the animals developed paralysis and convulsions. In 1933, Rivers et al. observed an encephalitis with myelin destruction in monekys that had been injected with rabbit brain extract. Morgan (1946) and Kabat et al. (1946) were the first to mix brain homogenates with adjuvants for immunization to get reproducible EAE in a high percentage of immunized animals. In EAE, antibodies to myelin components have been detected and T cell mediation is wellproven (see Paterson 1973 , Raine 1983 ).

The study of EAE has been aided by the development of MBP-specific T lymphocyte lines, which have been obtained from rats or mice immunized with MBP in complete Freund's adjuvant. In vivo, the MBP-specific T cell lines are able to induce acute or even a chronic relapsing EAE (Ben-Nun et al. 1981 , Zamvil et al. 1985 ). EAE appears to be strain-specifie since Brown-Norway (BN) rats and several mouse strains, e.g. BALB/c or C57BL/6 mice, are relatively resistant, whereas Lewis rats and SJL mice are fully susceptible (Linthicum & Frelinger 1982 , Montgomery & Ranch 1982 . Previous studies comparing EAEsusceptible and -resistant strains, Lewis and BN rats, respectively, have indicated that Lewis rat MBP-specific T cells recognize an epitope in the 68-88 peptide sequence of MBP, whereas the immunodominant epitope for Ihe T cells of BN rats was located outside of that peptide, probably in the 43-67 sequence (Berand et al. 1986 ). The identification of different epitopes by T lymphocytes of EAEsusceptible and -resistant rats cannot, however, fully explain different expression of disease, since in EAE-resistant PVG rats immunization with MBP gives rise to development of T cells which recognize the encephalitogenic 68-88 peptide (Ben-Nun et al. 1982) . Thus, the nature of the underlying regulatory events and mechanisms operating in the pathogenesis of immune-mediated encephalitis are not clear, and the contribution of non-MHC genetic factors which influence different expression of EAE is not understood.

In order to localize the physiological compartment conveying susceptibility to mice for EAE induction, hematopoietic radiation chimeras were prepared between susceptible SJL and resistant BIO.S strains. A low incidence of disease was observed in BIO.S^B.IO.S and SJL-»B10.S chimeras and a high incidence in B10.S-*SJL chimeras challenged with SJL spinal cord homogenate (Korngold et al. 1986 ). As the donor origin of the lymphoid compartment does not correlate with disease susceptibihty, one can speculate on a contribution of the CNS itself in disease susceptibility.

A CNS-derived influence on different EAE susceptibilities may be explained by our in vitro experiments. Astrocytes derived from susceptible strains (Lewis rats or SJL mice) express much higher levels of la upon treatment with IFN-y compared with astrocytes established from EAE-resistant strains (BN rats or BALB/c mice) (Massa et al. 1987) . At least 1 gene responsible for la hyperinduction is located outside the rat RT-1 or the mouse MHC locus, as the cells derived from animals eongeneic at the RT-i or MHC locus of the resistant strain, but with background genes of the susceptible strain, exhibit intermediate levels of la as compared to fully resistant and susceptible strains. As shown in Fig. 1 , the IFN-y-induced hyperinduction of la in EAE-susceptible animals is astrocytespecific, since both peritoneal macrophages and microglial cells of susceptible and resistant strains exhibit identical profiles of la induction, From these data we suggest a different role for microglia compared to astrocytes in at least some forms of immune-mediated encephalitis, the astrocytes being the initial trigger for intracerebral T cell activation and dictating development of disease.

In tissue sections of adult mice, microglial cells express the macrophage-specific antigen F4/80 and are positive for Fc IgG l/2b and type-three complement receptors (Perry et al. 1985) . In the electron microscope immunocytochemistry studies of Vass et al. (1986) resting microglia in rat tissue did not show la-reactivity. Microglial Hyperinducibility of la antigens on astrocyles of EAE-susceptible SJL mice. Aslrocytes and microglial cells were seeded in 24-well plates al a density of 500 000 ceils per well as described previously (Massa etal. 1987 , Frei etal. 1987 ). The cells were stimulated with IFN-y{\2 U/ml) for 72 h. Thereafter a cell radioimmunoassay was performed in tbe culture wells incubating step-wise for 1 h with monoclonai anti-I-A' (K25-8.7) antibodies and 200 ^I/well of a '--I-labelled anti-mouse immunoglobulin F(ab')2 fragment (corresponding to 200000 counts). Thereafter the monolayer was washed, the cells lysed with 5nNaOH and counted in a y-counter. Data represent the mean CPM ±SD of duplicate cultures. ĉ ells are thought to be of hematogenous origin and to bave the functional properties typical of tissue macrophages. In an attempt to characterize microglial cells in vitro, we bave established a procedure to isolate microglial c«lls from glial cell cultures obtained from newborn mice (Frei et al. 1986 ). Principally, the cells floating in the medium after shaking the glial cell cultures on a rotary shaker for 1 h were collected and seeded in flasks. After 2 h the non-adherent cells were discarded. The adherent cells showed all the characteristics of macrophages/microglial cells: the cells were phagocytic, contained non-specific esterase activity and expressed Fc (IgG I /IgG 2b) and type-three complement receptors. Morphologically both tbe ameboid and the ramified types of microglia were present. Upon stimulation with LPS the cells released substantial amounts of tumor necrosis factor-a and when being primed witb iFN->' and then exposed to LPS tbe cells became cytotoxic for tumor cells (P-815 cells) (Fig. 2) . As sbown by Giulian et al. (1986) , using microglial cells from newborn rats, tbe cells also secrete IL-1. Furtbermore, after stimulation with IFN-y tbe microglial cells became la-positive (see section 4) and functioned as APC . Similar findings have been made independently by Suzumura et al. (1987) demonstrating la induction on cultured murine mieroglial cells by IFNy treatment. Microghal cells have been found to grow very well in vilro when added on astrocyte monolayers (Frei et al. 1986 ). Astrocytes release an interleukin-3 (IL-3)-like factor which induces growth of microglial cells, peritoneal macrophages and IL-3-dependent cell lines such as 32 DCL cells. Furtbermore, the 33-kd factor secreted by astrocytes as well as by C-6 rat glioma ceils induced the expression of 20-a-bydroxy-steroid dehydrogenase in nu/nu spleen cells (Frei et al. 1985) . Production of this IL-3-like factor by astrocytes in vivo may account for the expansion of both resident microglial cells and monocytes having invaded the brain parenchyma in inflammatory lesions. Besides IL-3, the granulocyte-macrophage colony stimulating factor (GM-CSF) as well as the maerophage colony stimulating factor (M-CSF) promoted microglial cell growth.

Based on 1) the expression of Fc and type-three complement receptors, 2) tbe phagocytic activity, 3) the inducible secretion of TNF-a. 4) the tumor eytotoxicity and 5) tbe growth activity of GM-CSF and M-CSF, microglial cells harvested from brain cell cultures belong to tbe macropbage lineage. This is also supported by the genetic studies comparing the inducibility of la antigens on cultured astrocytes, microglial cells and macrophages derived from EAE-susceptible orresistant animals: the microglial cells behaved like macrophages but not like astrocytes (see section 4).

Several recent reviews have covered in depth the subject of impaired cellular immunity in patients witb ghoblastoma (for review see Fontana et al. 1987) . As (v/v) for their cytolytic activity on L-M cells according to described procedures . In analogy to murine TNF-a (Haranaka et al. 1986 ) the TNF activity of microglial cells (•) was found in a high molecular weight form (M, 150000) as well as in a low molecular weight form with an estimated Mt of 39 000. Using a polyclonal rabbit anti-murineTNF-(i antiserum (final dilution of 1:100) both high and low molecular weight TNF activities produced by microglial cells were completely neutralized (o). In contrast, no TNF activity was detected in LPS-treated astrocyte cultures (•). TNF activity was expressed as the percentage of viable cells relative to an untreated control and represents mean values of triplicate L-M cell cultures (5 x W cells/well). The AcA 54 column was calibrated by using the following markers: BSA, bovine serum albumin (67000); OVA. ovalbumin (45000); CHYM, chymotrypsinogen (25 000); CYT, cytochrome c (12300). V,, indicates the void volume.

a nondialyzabie inhibitory factor has been detected in the cystic fluid of glioblastoma and in patient serum before but not after tumor removal, it can be suggested that gliobiastoma cells release immunosuppressive factors. In 1984, we identified in the conditioned medium of cultured human glioblastoma cells a factor which interferes with T cell activation in vitro (Fontana et al. 1984a , Schwyzer & Fontana 1985 . This factor, termed glioblastoma-derived T cell suppressor factor (G-TsF), has been successfully purified to homogeneity . Purified G-TsF at concentrations of 4 x 10"" M was found to inhibit the growth promoting effect of IL-2 on IL-2-dependent T cells and the proliferative response of thymocytes stimulated with lectins.

Aminoterminal sequence analysis of G-TsF ) demonstrated that 12 out of the first 25 amino acids are identical to human transforming growth factor-^? (TGF-^, Derynck et al. 1985) . This was recently confirmed by isolating a cDNA clone coding for G-TsF from human glioblastoma cells. In addition, sequencing of the cDNA proved cysteine residues in positions 7, 15 and 16, where blanks had been obtained in the aminoterminal sequence of the purified peptide (de Martin et al. 1987) . Thus, there is actually a 60% sequence homology between the first 25 amino acids of G-TsF and TGF-^ (Fig. 3) . Based on this sequence homology, G-TsF is a new member of the merging TGF-f amily which also comprises inhibin from ovarian follicular fluid (Mason et al. 1985) and the testicular glycoprotein MuUerian inhibiting substance (Cate et al. 1986) , both factors being active in the endocrine system. Presumably G-TsF is the human analog of TGF-^2 and CIF B isolated recently from porcine platelets (Cheifetz et al. 1987 ) and bovine bone (Seyedin et al. 1987) , respectively. There is complete identity between the first 25 aminoterminal amino adds of all three peptides. Besides their structural relationship, TGF-^ and G-TsF share the property that they inhibit IL-2-mediated growth of IL-2-dependent T cell lines (Fontana et al. 1984 , Kehrl et al. 1986 ). The precise mechanism of this action of G-TsF on T cells is not yet clear. In analogy to TGF-^, one could assume that binding of G-TsF to a specific receptor triggers events resulting in reduced expression of certain growth-stimulating genes which function during the Go/Gl phase of the cell cycle. An inhibition of c-myc gene expression was shown in endothelial cells treated with TGF-^ (Takehare et al. 1987 ). However, it is possible that G-TsF and TGF-^ share a common set of receptors with probably different affinities for the two peptides. If released in vivo by the tumor cells, G-TsF may prevent local expansion of tumor-specific cytotoxic T cells, lymphocyte-activated killer cells and natural killer cells. In accordance with this assumption, the tumor infiltrating T cells harvested from fresh glioblastoma were found to be unresponsive to T cell mitogens in vitro (Miescher et al. 1986 ).

Our recent isolation of a cDNA clone for G-TsF will make large quantities of recombinant G-TsF available in the near future and provide the necessary materials to study the effects in tumor, transplant and autoimmune models in vivo.

There is no doubt that tissue damage in the course of some forms of encephalitis such as post-measles encephalitis or multiple sclerosis is due to an immunemediated response. While the participation of humoral (auto-)antibodies cannot be ruled out, the lesions may represent a T cell-mediated cellular immune reaction initiated by activated T helper cells. The blood brain barrier (BBB) plays an obvious role as a primary barrier separating the blood compartment from the brain, so lowering the chances of their mutual communication and restricting the intracerebral invasion of the cellular elements of the immune system. However, in the course of a systemic, e.g. antiviral, immune response antigen-specific T cells may adhere to brain endothelial cells, a process which may be enhanced by IL-1 and which may depend on the presence of antigen and MHC class II molecules induced by IFN-y on the endothelial cell surface (McCarron et al. 1985) . The antigen may be blood-derived or transported from inside the brain to the vessel wall. There may be an important role of brainderived chemotactic factors (IL-1?, leukotrienes?) in guiding T cells and monocytes through the BBB. Penetration of the lymphocytes through the BBB may be facilitated by the elaboration of endoglycosidases by activated T cells (Na-parstek et al. 1984) . Having arrived in the brain tissue, the future of the intruding T cells may depend on 1) the presence of the antigen towards the T cells have been sensitized, 2) the capacity of the T cells to release IFN-y and 3) the amount of la being induced by IFN-y on antigen-presenting cells within the tissue. As it has been shown that a virus is able to induce la antigens by IFNy-independent mechanisms, the release of IFN-y by infiltrating T cells may not be altogether critical. Both astrocytes and microglial cells can function as APC. However, we feel that astrocytes play the primary role within the brain tissue as I) astrocytes outnumber microglial cells by far in the adult brain and are therefore more immediately accessible for antigen-presenting functions and 2) in vitro studies demonstrate that the amount of la expressed on IFN-y-treated astrocytes correlates with susceptibility to immune-mediated encephalitis in vivo. The T helper cells once activated may 1) mediate la-restricted T cell cytolysis of la-bearing astrocytes which would cause disruption of support functions provided by astrocytes, including known metabolic cooperation between astrocytes, neurons, and oligodendrocytes (Sun & Wekerle 1986) , 2) release IL-2 which leads to further intracerebral T cell activation and 3) secrete factors like IFN-y which enhance various macrophage functions involved in tissue injury. IL-2 and other T cell products may even contribute to remyelination and development of astrocyte scarring, since IL-2 receptors have been identified on oligodendrocytes (Merrill et al. 1984 , Saneto et al. 1986 , and T cell products as well as IL-1 have been found to stimulate growth of astrocytes (Fontana et al. 1981 , Giulian & Lachman 1985 . Recruitment and expansion of cells of the macrophage lineage may be achieved by chemotactic factors for blood monocytes released in the tissue and by T cell-derived macrophage colony stimulating factors (GM-CSF. IL-3) or by an IL-3-like factor produced by activated astrocytes. These growth factors may be active on resident microglial cells as well as on infiltrating monocytes.

One can envisage that development and/or chronicity of immune-mediated encephalitis also depends on virus persistence, being favored by the low expression of MHC class I and II antigens in normal brain tissue, by genetically determined hyperinducibility of MHC class II antigens on astrocytes and by the development of autoreactive T cells that have been passively recruited during the course of the virus-induced inflammatory process. In brain tumors, especially malignant glioblastoma which expresses MHC class I and II molecules (Carrel et al. 1982) , both macrophages/microglial ceils and T cells infiltrating the tumor tissue have been identified. However, the activation of the T cells may be limited due to the ghoblastoma cell-derived T cell suppressor factor (G-TsF) which interferes with T cell growth. During development, the secretion of G-TsF by non-malignant brain cells, e.g. glioblasts, may protect the fetal nervous system -with its not yet functional blood brain barrier -from immune-mediated injury.

Ueber Immunisiening gegen Wut mit normaler Nervensubstanz

Experimental autoimmune encephalomyelitis (EAE) in genetically resistant rats: PVG rats resist active induction of EAE but are susceptible to and can generate EAE efTcctor T cell lines

The rapid isolation of clonable antigenspecific T lymphocyte lines capable of mediating autoimmune encephalomyelitis

Experimental autoimmune encephalomyelitis mediated by T lymphocyte lines: genotype of antigen-presenting cells influences immunodominant epitope of basic protein

Expression of HLA-DR and common acute lymphobiastic leukemia antigens on glioma cells

Isolation of tbe bovine and human genes for Muellerian Inhibiting Substance and expression of the human gene in animal cells

The transforming growth factor-^ system, a complex pattern of cross-reactive ligands and receptors

Human transforming growth factor-ĉ DNA sequence and expression in tumor cell lines

Accessory cells and the activation and expression of different T cell functions. In: Regulation of immune gene expression

Astrocytes as antigenpresenting cells. I. Induction of la antigen expression on astrocytes by T cells via immune interfcron and its effect on antigen presentation

An immunological function for astrocytes

The role of astrocytes in the interaction between the immune and nervous system

) lmmuneregulatory factors secreted by astro cytes and glioblastoma cells

Astrocytes as antigen-presenting cells. II. Unlike H-2K-dependent cytotoxic T cells, H-2I-restricted T cells are only stimulated in the presence of interfero n-y

Astrocytes present myelin basic protein to encephalitogenic T-cell lines

E. {1984a) Gliobiastoma cells release interleukJn-1 and factors inhibiting intcrleuk.in-2 mediated effects

Production of prostaglandin E and interleukin 1-like factors by cultured astrocytes and C-6 glioma cells

Biological and biochemical characterization of an interleukin l-like factor from rat C-6 glioma cells

Glia cell stimulating factor (GSF): A new lymphokine. Part 2. Cellular sources and partial purification of human GSF

Synthesis of interieukin-1/endogenous pyrogen in the brain of endotoxin treated mice: A step in fever induction?

Astrocytes of the brain synthesize interieukin-3 like factors

Astrocyte-derived interleukin-3 as a growth factor for microglia cells and peritoneal macrophages

Antigen presentation and tumor cytotoxicity by interferon-y treated microglial cells

Interleukin-1 of the central nervous system is produced by ameboid microglia

Interleukin-1 stimulation of astroglial proliferation after brain injury

Purification, characterization and antitumor activity of n on recombinant mouse tumor necrosis factor

Demonstration and eharacterization of Ia-positivc dendritic cells in the interstitial connective tissues of rat heart and other tissues, but not brain

Immunohistochemical staining of human brain with monoclonal antibodies that identify lymphocytes, monocytes and the la antigen

Expression of la molecules by astrocytes during acute experimental allergic encephalomyelitis in the Lewis rat

Expression of la antigens by cultured astrocytes treated with Interferon-j

Immunoregulatory molecules and IL-2 receptors identified in multiple sclerosis brain

Spinal cord multiple sclerosis lesions in Japanese patients. Schwann cell remyelination occurs in areas that lack glial fibrillary acidic protein

t946) The rapid production of acute disseminated encephalomyelitis in rhesus monkeys by injection of heterologous and homologous brain Ussue with adjuvants. VII. The effect of cortisone

The production of TGF-^ by human T lymphocytes and its potential role in the regulation of T cell growth

Expression of la antigens on the surface of human oligodendrocytes and astrocytes in culture

Acute experimental allergic encephalo-myelitis in radiation bone marrow chimeras between high and low susceptible strains of mice

Characterization of inflammatory infiltrates in experimental allergic encephalomyelitis

Reactive gliosis

Acute autoimmune encephalomyelitis in mice. H. Susceptibility is controlled by the combination of H-2 and histamine sensitization genes

Theiler's virus-induced demyelination: Prevention by immunosuppression

Complementary DNA for human glioblastoma-derived T cell suppressor factor, a novel member of the transforming-^ gene family

Complementary DNA sequences of ovarian folHcular fluid inhibin show precursor structure and homology with transforming growth factor fi

Viral particles induce la antigen expression on astrocytes

Hyperinducibility of la antigen on astrocytes correlates with strain-specific susceptibility to experimental autoimmune encephalomyelitis

Immunohistochemical analysis of the rat central nervous system during experimental allergic encephalomyelitis, with special reference to Ia-positive cells with dendritic morphology

Presentation of myelin basic protein by murine cerebral vascular endothelial cells

Proliferation of astroglia and oligodendroglia in response to human T cellderived factors

Functional properties of tumor-infiltrating and blood lymphocytes in patients with solid tumors: effects of tumor cells and their supernatants on proliferative responses of lymphocytes

Difierential expression of la and la-associated invariant chain in mouse tissues after in vivo treatment with interferon-y

Experimental allergic encephalomyelitis (EAE) in mice: primary control of EAE susceptibility is outside the H-2 complex

Allergic encephalomyelitis in monkeys in response to injection of normal monkey cord

Activated T lymphocytes produce a matrix-degrading heparan sulfate endoglykosidase

I960) Transfer of allergic encephalomyelitis in rats by means of lymph node cells

Multiple sclerosis: An immunologic reassessment

Immunohistochemical localization of macrophages and microglia in the adull and developing mouse brain

Cultured human fetal astrocytes can be induced by interferon-y to express HLA-DR

A giial progenitor ceU that develops in vitro into an astrocyte or an oligodendrocyle depending on culture medium

Multiple sclerosis and chronic relapsing EAE: Comparative ultrastructural neuropathology

Observations on attempts to produce acute disseminated encephalomyelitis in monkeys

Immune response gene products (la antigens) on glial and endothelial cells in virus-induced demyehnation

) la expression in chronic relapsing experimental allergic encephalomyelitis induced by long-term cultured T cell lines in mice

Interleukin-2 mediates the inhibiton of oligodendrocyte progenitor cell proliferation in vitro

Partial purification and biochemical characterization of a T cell suppressor factor produced by human glioblastoma cells

Cartilage-inducing factor-B is a unique protein structurally and functionally related to transforming growth factor-B

Susceptibility of astrocytes to class I MHC antigen-specific cytotoxicity

) la-restricted encephalitogenic T lymphocytes mediating EAE lyse autoantigen-presenting astrocytes

MHC antigen expression on bulk isolated macrophage-microglia from newborn mouse brain: induction of la antigen expression by interferon-)

The expression of MHC anligens on oligodendrocytes: induction of polymorphic H-2 expression by lympbokines

TGF-;5 inhibition of endothelial cell proliferation: alteration of EGF binding and EGF-induced growth-regulatory (competence) gene expression

Induction of antigen presentation ability in purified cultures of astroglia by interferon-)

Expression and synthesis of murine immune response-associated (la) antigens by brain cells

Multiple sclerosis: The blood-brain barrier and the measurement of de novo central nervous system IgG synthesis

On the presence of la-positive endothelial cells and astrocytes in multiple sclerosis lesions and its relevance to antigen presentation

Autoimmune encephalomyelitis: simultaneous localization of T and B ceils in the target organ

Acute experimental autoimmune encephaiomyeiitis: T-and B-cell distribution within the target organ

The distribution of la antigen in the lesions of rat aculc experimental allergic encephalomyelitis

Distribution and quantitation of HLA-ABC and DR (la) antigens on human kidney and other tissues

Inducible expression of H-2 and la antigens on brain cells

T cell suppressor factor from human gliobiastoma cells is a 12.5 kd protein closely related to transforming growth factor-/?

T-cell clones specific for myelin basic protein induce chronic relapsing paralysis and demyelination

This work has been supported by grants from the Swiss National Science Foundation , the National Multiple Sclerosis Society and the Swiss Cancer League.