key: cord-0005549-rhji4io9
authors: Popko, Brian; Corbin, Joshua G.; Baerwald, Kristine D.; Dupree, Jeffrey; Garcia, Annie M.
title: The effects of interferon-γ on the central nervous system
date: 1997
journal: Mol Neurobiol
DOI: 10.1007/bf02740619
sha: 42c6dba50d91c61087f1f89b3cab0049a64cc811
doc_id: 5549
cord_uid: rhji4io9

Interferon-gamma (IFN-γ) is a pleotropic cytokine released by T-lymphocytes and natural killer cells. Normally, these cells do not traverse the blood-brain barrier at appreciable levels and, as such, IFN-γ is generally undetectable within the central nervous system (CNS). Nevertheless, in response to CNS infections, as well as during certain disorders in which the CNS is affected, T-cell traffic across the blood-brain barrier increases considerably, thereby exposing neuronal and glial cells to the potent effects of IFN-γ. A large portion of this article is devoted to the substantial circumstantial and experimental evidence that suggests that IFN-γ plays an important role in the pathogenesis of the demyelinating disorder multiple sclerosis (MS) and its animal model experimental allergic encephalomyelitis (EAE). Moreover, the biochemical and physiological effects of IFN-γ are discussed in the context of the potential consequences of such activities on the developing and mature nervous systems.

Interferon-gamma (IFN-?) was originally discovered by Wheelock (1965) as an activity that interfered with viral replication. Tlymphocytes and natural killer cells are the only cells known to be capable of expressing this cytokine (reviewed in Trinchieri and Perussia, 1985) . Because under normal condi-Under these conditions, neuronal and glial cells that normally do not encounter IFN-',v are exposed to the potent, pleotropic effects of this cytokine. There is considerable evidence that suggests that the effects of IFN-?~ are important in a variety of CNS disorders (reviewed in Ransohoff and Benveniste, 1996) . In this article, we discuss various molecular, biochemical, cellular, and physiological properties of IFN-~, and its evoked response. Moreover, the potential role this cytokine plays in CNS disorders is discussed.

A considerable amount of information has been learned concerning the molecular biology and biochemistry of IFN-~, since the human and mouse cDNAs were cloned over a decade ago (Gray et al., 1982; Goeddel, 1982, 1983) . The human and mouse proteins are 143 and 134 amino acids in length, respectively, and are encoded for by four exons that reside on 6 kb of DNA. Interestingly; the human and mouse proteins share only limited homology, approx 40% at the amino acid level, and do not bind to the other's receptor to an appreciable extent (reviewed by Farrar and Schreiber, 1993) . Both proteins contain two glycosylation sites (Gray and Goeddel, 1982; Ealick et al., 1991) , although glycosylation is not essential for IFN-'I activity (Kelker et al., 1983) . One factor that may play a role in regulating the biological activity of IFN-y is message stability, since the IFN-y mRNA contains an AU rich sequence in the 3'-untranslated region that has been shown to reduce [he mR_NA halfqife of other cytokines dramatically (Shaw and Kamen, 1986) . Biologically active IFN-~/from both human and mouse is a noncovalently bound homodimer that contains two receptor binding sites .

The IFN-1 receptor is a multimeric receptor complex composed of a ligand binding subunit (a-chain) (Aguet et al., 1988 ) and a transmembrane, accessory factor (~-chain) (Soh et al., 1994; Hemmi et al., 1994) . The c~-chain is a ubiq-uitously expressed, glycosylated, cell-surface protein that binds IFN-~ , with high affinity (reviewed by Farrar and Schreiber, 1993) . The extent of glycosylation is variable depending on cell type and accounts for the range of sizes of the a-chain, between 80 and 95 kDa. On binding, IFN-7 induces (r-chain dimerization, which is believed to be critical to the initiation of the signal transduction cascade (Greenlund et al., 1993; Fountoulakis et al., 1992) . This feature of the IFN-7 receptor has been exploited to generate mutant forms that confer IFN-3, unresponsiveness to cells in a dominant manner (Dighe et al., 1993 (Dighe et al., , 1995 . The crystal structure of the complex between IFN-y and the a-chain of its receptor has recently been reported (Walter et al., 1995) . The ~-chain of the receptor, which associates with the or-chain in an IFN-,f-dependent manner, is essential for the induction of the IFN-7 signaling cascade (Bach et al., 1996) .

Recently, considerable progress has been made in the elucidation of the IFN-,~-induced signal transduction pathway (Sadowski et al., 1993; Darnell et al., 1994; Shuai, 1994; Vilcek and Oliveira, 1994; David, 1995; Schindler, 1995) . The Janus kinases JAK1 and JAK2 become phosphorylated following binding of IFN-7 to its receptor, with which the JAK kinases associate (Muller et al., 1993; Wafting et al., 1993) . The JAK1 kinase appears to associate with the a-chain, and the JAK2 kinase associates with the [3-chain (Sakatsume et al., 1995) . Following activation, the JAK kinases phosphorytate the transcriptional factor known as signal transducer and activator of transcription (STAT-I~z), w~hich binds to a specific DNA element referred to as ~,-activation site (GAS) . STAT-la binding results in the rapid transcriptional induction of genes containing the GAS element. For example, the genes encoding the gaunylate binding protein and the IgG Fc receptor are induced in this way: Other genes that are stimulated by IFN-~/, such as the class I and class II molecules of the major histocompatibility complex (MHC), require a longer period (several hours) before transcriptional induction is detected.

These genes lack the GAS element and are likely activated through IFN-?-induced intermediates (Benveniste and Benos, 1995) .

(Summarized in Table 1) IFN-? plays a critical role in the regulation of the immune response (Young and Hardy, 1995) . For this reason, it is often referred to as immune interferon. IFN-y facilitates the stimulation of the ~1 subpopulation of T-lymphocytes, which control cell-mediated immunity and which, in fact, express IFN-y; whereas it impedes the stimulation and activity of the Th2 subpopulation of T-cells, :which express IL-4 and are involved in regulating humoral immu-ni~ (Gajewski et al., 1989; Swain et al., 1991; Paul and Seder, 1994; Seder and Paul, 1994; Reiner and Seder, 1995) . Cytotoxic T-cells and natural killer cells are also activated by IFN-7 (Trinchieri and Perussia, 1985) . IFN-? also plays a role in regulating antibody production, promoting an IgG2a response through a direct effect on B-cells (Snapper and Paul, 1987) . importantly, IFN-? is a potent stimulator of expression of the antigen-presenting components of the immune system. MHC class I and II molecules are dramatically upregulated in the presence of IFN-'y on a variety of cell types (reviewed in Vilcek et al., 1985; Benveniste and Benos, 1995) .

IFN-? is also a powerful effector molecule of the inflammatory response (reviewed in Schreiber and Celada, 1985) . IFNq,-stimulated macrophages release a number of inflammatory cytokines, including tumor necrosis factor-alpha (TNF-a), IL-1, IL-6, and IL-8. In addition, IFN-? increases the microbicidal activities of macrophages through the induction of the synthesis of hydrogen peroxide and nitric oxide . IFN-? is also a potent stimulator of the expression of the Fc receptors for IgG immunoglobins on macrophages (Erbe et al., 1990) . All of these actions of IFN-? on macrophages serve to stimulate the cytocidal activity of these cells. 

Multiple sclerosis (MS) is the most common human demyelinating disease that affects the CNS. Although the etiology of MS has not been established, it is widely believed that immunological mechanisms are involved (reviewed in Martin et al., 1992; Raine, 1994a,b) . Re CNS is immunologically distinct owing to the presence of the blood~brain barrier (Crone, 1986) , which inhibits entry of both immune cells and humoral factors. In MS patients, this barrier is disrupted (Martin et al., 1992) allowing increased access of blood components (cellular and noncellular) to the CNS. The enhanced permeability of the blood-brain barrier is thought to be a contributing factor in immunemediated demyelinating disorders (reviewed in Poser, 1987 Poser, , 1993 . Importantly, there is an increased infiltration of T-lymphocytes (CD4+ and CD8+) into the CNS of MS patients (Martin et al., 1992; Utz and McFarland, 1994) , with CD4+ (T-helper) T-cells being predominant at the sites of active demyelLnation (Booss et al., 1983; Traugott et al., 1983) . T-cell responses to a variety of myelin antigens, including myelin basic protein (MBP) and proteolipid protein (PLP), have been identified in MS patients (reviewed in Miller et al., 1995) .

Experimental allergic encephalomyelitis (EAE) is the primary animal model used in the study of MS (reviewed in Alvord et al., 1984; Zamvil and Steinman, 1990; Martin and McFarland, 1995) . EAE can be induced in a variety of laboratory animal species by immunization with either CNS tissue, myelin, or

Volume i4, 1997 myelin components. As demonstrated through passive transfer studies, EAE is a T-cell (CD4+, Thl) mediated disease model (Zamvil and Steinman, 1990; VoskuhI et al., 1993) . A relapsing/remitting form of EAE is induced at high incidence in SJL mice immunized with an encephalitogenic epitope of PLP (Tuohy et al., 1988a (Tuohy et al., ,b, 1989 (Tuohy et al., , 1992 McRae et al., 1992) . This model of EAE exhibits many of the clinical, pathological, and immunological feat~ares of MS.

The T-cells that infiltrate into the CNS in MS and EAE are activated and produce IFN-I, which is believed to play a role in the pathogenesis of irnmune-rnediated demyelinating disorders (reviewed in Hartung et al., 1992; Olsson, 1992; Panitch, 1992 ). There is considerable circumstantial evidence that suggests that IFN-~, is a key mediator of inflammation in MS and EAE. Many of the pathological events Observed in MS and EAE, such as increased MHC expression, macrophage activation, increased expression of leukocyte adhesion molecules on blood-brain barrier endothelial cells (Olsson, 1992) , and reactive gliosis (Balasingam et al., 1994) , are consistent with the known effects of IFN-~,. IFN-7 is detected in active MS lesions (Traugott and Lebon, 1988) , and increased numbers of IFN-~,-secreting Tlymphocytes are present in the cerebrospinal fluid of MS patients (Olsson et al., 1990) . Beck et al. (198ff) reported a cor~Iation between the onset of clinical symptoms of MS and mitogenstimulated IFN-y production. Moreover, EAE can be transferred to naive animals using IFN-~,-secreting Tqymphocytes isolated from animals experiencing EAE (Zamvil and SteLnman, 1990) , and a significant increase in the level of IFN-?, mRNA is Observed in the CNS of animals with severe EAE compared to animals with mild EAE (Renno et al., 1994) .

Perhaps the best direct evidence for the deleterious effects of IFN-y in MS came inadvertently from a small clinical trial in the early 1980s. Of 18 MS patients receiving IFN-~/, seven experienced exacerbations during the 4-wk treatment period, which was significantly higher than the pretreatment exacerbation rate (Panitch et at., 1987; Panitch, 1992) . Not only did this study clearly eliminate IFN-~/as a potential therapeutic agent for MS, but it also demonstrated that IFN-~, activity can lead to a wo~,ens of tlne disease course, suggesting a role for this cytokine in normal disease progression.

There are, however, contradictory experimental data concerning the role of IFN-~, in EAE. The administration of antibodies to IFN-~, e~anced the severity of EAE in mildly susceptible mice (Billiau et al., 1988; Voorthuis et al., 1990; Duong et al., 1992) , and resulted in EAE susceptibility in resistant strains (Duong et al., 1994) . Moreover, treatment of SJL/J mice, which are very sensitive to EAE, with IFN-~, results in enhanced survival (Billiau et al., 1988) . Nevertheless, the protective effect of IFN-~, is not observed when EAE is passively transferred. Voorthuis et al. (1990) reported that intraventricular administration of IFN-~/ into EAE-induced rats resulted in complete suppression of clinical signs, and Willenborg et al. (1995) , using a viral delivery system, suggested that IFN-y had no effect on disease. Recently, Ferber et al. (1996) examined the EAE susceptibility of B10.PL mice that contained an experimentally inactivated IFN-~, gene (Dalton et al., 1993) . B10.PL mice are normally very susceptible to MBP-induced EAE, and the IFN-?, ka~ockout mice appear equally sensitive to disease induction. Although morphological data were not presented, this work dearly demonstrates that IFN-~, is not essential for EAE induction, at least not in B10.PL mice.

To understand better the effects of IFN-u on CNS cells in vivo, in the absence of other T-cellderived factors, a variety of approaches have been used to deliver this cytokine to the CNS. The direct injection of IFN-~, into various animal models has been shown to upregulate the

Volume 14, 1997 expression of both MHC class I and class II glycoproteins on various neuronal celt types that do not normally express these proteins at detectable levels. Wong et al. (1984) demonstrated a dramatic increase in the expression of the MHC class I glycoprotein in astrocytes, neurons, oligodendrocytes and microglia following the delivery of IFN-y to the CNS of 2-dold mice. Direct injection of IFN-y into the CNS has also been demonstrated to increase the expression of MHC class I on rat endothelial and ependymal cells and class II on microglia, ependymal, and perivascular cells (Sethna and Lampson, 1991) . Continuous iv infusion of IFN-~f into Lewis rats for a period of 3 d induced MHC class II expression on microglia, ependymal and endothelial cells of large blood vessels (Steiniger and van der Meide, 1988) . These findings support the earlier work of Momburg et at. (1986) who reported that the iv infusion of IFN-y into mice resulted in class II expression in many organs, including the brain, where "round cells in the vicinity of meningeal blood vessel," presumably microglia, were class II immunoreactive. IFN-y has also been suggested to play a role in the breakdown of the blood-brain barrier and the recruitment of inflammatory cells into the CNS. Seth_ha and Lampson (1991) reported that a single injection of IFN-y into the rat basal ganglia induced the infiltration of CD4+ T-cells into the perivascular space and monocytes/ macrophages, and presumptive natural killer cells into the brain parenchyma. These authors suggested that such cellular recruitment may be facilitated by IFN-y-induced changes in the endothelial cells, resulting in an alteration in the permeability of the blood-brain barrier. Moreover, Simmons and Willenborg (1990) , using direct injection of IFN-~. into the spinal cord of rats, reported an inflammatory response similar in pattern to that observed in early EAE with the accumulation of inflammatory cells in the meninges and around spinal cord veins also suggestive of a compromised blood-brain barrier. Tjuvajev et aI. (1995) reported an IFN-y-induced breakdown of the blood-brain barrier in an in vivo brain tumor model, and Huynh and Dorovini-Zis (1993) demonstrated that brain endothelial cells undergo dramatic morphological, functional, and permeability changes in response to IFN-y, suggesting that this cytokine alters bloodbrain barrier function. Steffen et al. (1994) demonstrated increased expression of vascular cell adhesion molecules (VCAM) and intercellular adhesion molecule-1 (ICAM-1) on the brain endothelium of SJL mice experiencing EAE, and antibodies to these molecules inhibited binding of peripheral lymphocytes to inflamed brains. Furthermore, McCarron et al. (1993) showed that IFN-y significantly increased the adhesion of MgP-specific encephatitogenic T-ceils to mouse cerebrovascular endothelial cells, and that this effect correlated with the induction of ICAM-1 expression by the endothelial cells.

The studies discussed above consistently indicate a role for IFN-y in the inducement of expression of MHC and cell adhesion molecules, and the recruitment of inflammatory cells; however, the infusion of IFN-y into the CNS has also been implicated in a variety of other processes. For example, Yong et al. (1991) reported that the direct administration of IFN-~/ into the adult mouse brain following corticectomy resulted in an increase in trauma-initiated gliosis. Brosnan et al. (1989) , using the visual pathway of the rabbit, demonstrated that the injection of IFN-y significantly reduced neural conduction and induced subtle axonal pathology. In addition, the simultaneous injection of IFN-7 and a myelin/oligodendrocyte glycoprotein antibody significantly delayed cervical somatosensory evoked potentials, and caused massive demyelination in the spinal cord (Vass et al., 1992) . Vass et al. (1992) also reported an IFN-y-induced increase in MHC expression, which accompanied spinal cord demyelination.

In complement with various delivery systems previously discussed, we have used a transgenic approach to examine the effects of IFN-y on the developing nervous system, (Corbin et al., 1996) . Transgenic mice were generated in which the expression of IFNq was

Volume 14, I997

placed under the transcriptional control of the MBP gene (MBP/IFN-y transgenic animals). Transgenic mice generated with tl-ds construct have a shaking/shivering phenotype that is similar to that observed in naturally occurring mouse models of hypomyelination (e.g., sb_iverer, jimpy, quaking), and these transgenic animals have dramatically less CNS myelin than control animals. Reactive gliosis and increased macrophage/microglia F4/80 immunostaining were also observed. Additionally, MHC class I and class II mRNA levels were increased in the CNS of MBP/IFN-y transgenic mice, and the increase in MHC class I mRNA expression was detected in both white and gray matter regions. Surprisingly, cerebelfar granule cell migration was also disrupted in these animals. These results Strongly support the hypothesis that IFN-~, is a key effector molecule in irnmtme-mediated demyelinating disorders, and suggest that the presence of this cytokine may also disrupt the cytoarchitecture of the developing nervous system. 

The mechanism by which IFN-7 exerts its action within the CNS remains unresolved. One idea concerning the potential cellular site of action of IFN-y in CNS demyelinating disorders is that IFN-1 activates macrophages/ microglial cells, which in turn mediate cytotoxic effects on oligodendr~ytes (Fig. 1) . TNF-(z, which has previously been shown to be toxic to oligodendrocytes in vitro, is expressed by IFN-y-activated macrophages/microglia (Selmaj et al., 1991) . More recent in vitro experiments suggest that the microglial cell toxicity to otigodendrocytes is mediated through cellsurface expression of TNF-cz and the production of nitric oxide (Merrill et al., 1993) . Nitric oxide is very labile, such that to elicit their cytotoxic effects, the activated microglia need to be in intimate contact with the target oligodendrocytes. Oligodendrocytes undergo necrotic cell death following exposure to nitric oxide (Mitrovic et al., 1995) . In support of the hypoth~ esis that nitric oxide is important in the pathogenesis of immune-mediated demyelinating diseases, it has been shown that nitric oxide synthase mRNA levels are increased in mice experiencing EAE (Koprowski et al., 1993) , and NADPH-diaphorase histochemical staining, a marker for nitric oxide production, as well as nitric oxide synthase rnRNA levels were found to be elevated in the brains of MS patients (B6 et al., 1994) .

IFN-y may also have a direct deleterious effect on oligodendrocytes (Fig. I) , which have been shown to express IFN-7 receptors (Torres et al., 1995) . We have examined the effects of IFN-7 on oligodendrocytes using the MOCH-1 cell line, which was derived from an oligodendroglioma that developed in a transgenic line of mice and expresses a variety of features of oligodendrocytes (Hayes et al., 1992; Li et aL, 1995) . When MOCH-1 cells are cultured in medium containing low concentrations of

Volume 14, 1997 serum (1% fetal bovine serum [FBS]) they extend multiple, thin, branched processes and have phase-bright cell bodies, which are typi ~ cal features of cultured primary oligodendrocytes (reviewed in Popko et al., 1994) . Furthermore, these cells are immunoreactive against antibodies to galactocerebroside, a glycolipid marker of mature oligodendrocytes, and abundantly express myelin protein genes (Hayes et al., 1992) . MOCHq cells cultured in 1% FBS or chemically defined medium do not express detectable levels of the astrocyte marker glial fibrillary acidic protein (GFAP). When MOCH-1 cells are cultured in 1% FBS medium containing IFN-7. However, they transform to flat, astrocyte-like cells with enlarged cell bodies and appear more adhesive ~ In addition to these morphological changes, IFN-1 stimulates the expression of abundant levels of the astrocyte marker GFAP, as well as slightly elevated levels of MBP and PLP mRNA. IFN-7 also induces an enormous increase of MHC class I expression, as well as a comparatively modest increase of MHC class II mRNA expression in MOCHq cells. The morphological and molecular changes MC)CH-1 cells undergo in response to IFN-7 suggest that IFN-7 may induce a direct effect on oligodendrocytes, or a subpopulafion of these cells, in vivo .

Recent studies further support the hypothesis that IFN-y has a direct, deleterious effect on oligodendrocytes. Using a morphological assay, oligodendrocytes in vitro were shown to undergo apoptotic cell death at very high frequency following exposure to relatively low concentrations of IFN-7 (Vartanian et al., 1995) . The addition of rnicroglia to the culture did not significantly increase the frequency of otigodendrocyte death, further supporting a direct effect of the cytokine. Interestingly, Vartanian et aI. (1995) also detected apoptotic oligodendrocytes, along with IFN-7, in the advancing margin of active MS plaques, but not in the chronic portion of the plaque or in adjacent unaffected white matter. Agresti et al. (1996) have also observed deleterious effects of IFN-y on cultured oligodendrocytes. In the absence of cell death, they observed a decrease in the ability of oligodendrocyte progenitors to divide and differentiate, as well as changes in myelin protein gene mRNAs and a general metabolic depression. These effects were shown to be fully reversible following cytokine withdrawal (Agresti et al., 1996) .

The harmful effects of IFN-y on myelination might be mediated through the induction of MHC expression in oligodendrocytes, which normally do not express these molecules at an appreciable level. In support of this possibility, transgenic mice in which MHC class I expression has been targeted to oligodendrocytes are severely hypomyelinated (Turnley et al., 1991b; Yoshioka et aL, 1991; Turnley and Morahan, 1995) . This hypomyelh~ation occurs in the absence of immune celt infiltration, suggesting that MHC class I expression in oligodendrocytes is sufficient to disrupt myelination. Recently; Power et al. (1996) examined these transgenic animals in more detail and demonstrated that cell-surface expression of MHC class I in oligodendrocytes of these mice was minimal Moreover, immunocytochemistry for f~2-microglobulin, which is required for cellsurface expression of MHC class I, revealed that expression of ~2-microglobulin was undetectable in oligodendrocytes and limited to microglia in Lhe CNS. Power et al. (1996) suggest that the accumulation of MHC class I protein m the cytoplasm of the oligodendrocytes, owing to the absence of f~2-microglobulin, may interfere with normal myelin protein trafticking, leading to the observed hypomyelination in these mice.

In vitro, oligodendrocytes at all stages of differentiation can be induced by IFN-7 to express cell-surface MHC class I (Suzumura et al., 1986; Turnley et al., 1991a; Massa et al., 1993) . Nevertheless, it is unclear what the consequences of oligodendroglial expression of MHC class I are. As mentioned previously, CD8+ T-cells are present in the CNS of MS patients (Martin et al., 1992; Utz and McFarland, 1994) . It is, however, unknown whether oligodendrocytes in vivo can present antigen to these cells.

In contrast, many investigators have searched for, but failed to demonstrate, IFN-7mediated induction of MHC class II on oligodendrocytes in culture (Wong et al., 1984; Suzumura et al., 1986; Turnley et al., 1991a; Satoh et al., 1991b) , suggesting that oligodendrocytes are refractory to MHC class II induction. Calder et al. (1988) demonstrated that oligodendrocyte precursors (O-2A progenitor cells) can be induced to express MHC class II by IFN-~/, but that maturation into oligodendrocytes leads to loss of inducibility. Nevertheless, Bergsteinsdottir et al. (1992) were able to induce MHC class II expression in a subset of mature oligodendrocytes in culture with IFN-7 in the presence of the synthetic glucocorticoid dexamethasone. Although gtucocorticoids are normal CNS constituents (Birmingham et al., 1984; Kumar et al., 1989) , the in vivo significance of these in vitro observations is unclear.

Another activity of IFN-T that might be relevant to the pathogenesis of immune-mediated demyelinating disorders concerns the ability of this cytokine to stimulate the expression of Mat-shock (stress) proteins. These are ubiquitously expressed proteins that are believed to play a role as chaperones in normal protein synthesis and degradation (reviewed in Jindal, 1996) . The expression of these proteins increases following cell stress, and this response is thought to be protective. Heat-shock proteins are highly conse~rved across divergent species and are very irnmunogenic. It has been demonstrated that there is increased heat shock protein expression in MS and EAE lesions, and importantly, that there is an immune response to [he stress proteins in these disorders (Selmaj et al., 1992; Prabhaker et al., 1994; Gao et aL, 1995; van Noort et al., 1995) . It has also been demonstrafed that 1FN-'f is capable of potentiating the induction of stress protein synthesis (Morange et al., 1986; Ferm et al., 1992) , such that this effect might facilitate the induction of the immune response to the heat-shock proteins in immune-mediated disorders. The characterLzation of the heat-shock response in these disorders should provide further insight into the role these proteins play in the pathogenesis of disease.

A potential direct effect of IFN-? on oligodendrocytes in immune-mediated demyelinating disorders is consistent with recent studies that examined biopsy specimens from patients with acute MS (Rodriguez and Scheithauer, 1994) . They observed degeneration of inner myelin loops with preservation of outer loops and axons in early plaques, as well as areas in which well-myelinated and demyelinated axons were in close proximity. These observations suggest that the early pathological lesions that occur in MS are compatible with a primary disturbance in the myelinating function of oligodendrocytes, not a nonspecific inflammatory reaction that affects the myelin sheath, as might occur if microglia were activated subsequently to phagocytose the myelin sheath.

Astrocytes, which have been shown to express the receptor for IFN-7 (Rubio and de Felipe, 1991) , respond to the presence of this cytokine in a variety of ways. As mentioned previously, MHC class I expression on astrocytes increases substantially in the presence of IFN-7 (Wong et al., 1984; Mauerhoff et al., 1988; Massa et al., 1993) . This expression appears functional in that astrocytes are susceptible to lysis by cytotoxic T-cells in an antigen-specific manner (Skias et al., 1987) . IFNq,-induced MHC class II expression by astrocytes has also been demonstrated in vitro and in vivo (Fontana et aL, 1984; Fierz et al., 1985; Pulver et al., 1987) , although fine in vivo levels are low relative to that detected on microglial cells (reviewed in Benveniste, 1992) . The biological significance of the antigen-presenting potential of astrocytes remains to be resolved in vivo.

There is also evidence that suggests that IFN-7 plays a role in the reactive astrogliotic response, particularly in immune-mediated disorders. Reactive astrogliosis is the response in which astrocytes express increased levels of GFAP, undergo hypertrophy, and proliferate (Eng, 1988) . There are dramatically elevated levels of GEAP expression in the CNS of the

Volume I4, 1997 MBP/IFN-7 transgenic animals, although it is unclear whether this represents a direct response of astrocytes to I~-7 or an indirect response elicited by the myelin abnormalities that occur in these mice (Corbin et al., 1996) . Moreover, as mentioned, Yong et al. (1991) showed that IF1N<7, as well as a number of other cytokines (Balasingam et al., 1994) , increased trauma-induced gliosis in mice. These investigators also demonstrate that IFN-? promotes the proliferation of adult human astrocytes in vitro (Yong et al., 1991) . Nevertheless, in a follow-up study, Yong et al. (1992) were unable to reproduce these results using mouse astrocytes, suggesting that there are species differences in the astrocytic response to IFN-7. IFN-7 has also been shown to increase the expression of intercellular adhesion molecules on astrocytes. Human fetal (Frohman et al., 1989) and adult (Satoh et al., 1991a) astrocytes, as well as mouse astrocytes (Satoh et al., 1991b) , have been shown to express ICAM-1 in the presence of IFN-7. Such expression might facilitate the interaction of these cells with lymphocytes. IFN-~,-stimulated human and rat astrocytes have also been shown to express VCAM-1, possibly suggesting a role in lymphocytic infiltration across the blood-brain barrier into the CNS (Rosenman et al., I995) .

Although IFN-7 readily induces the expression of MHC molecules in astrocytes, oligodendrocytes, and resident CNS microglia, !FN-7-induced expression of MHC molecules in neurons is under tighter control. Early evidence demonstrating the ability of brain cells, including neurons, to express MHC molecules includes experiments in which IFN-y was either added to mixed brain cultures or i~ected directly into the CNS (Wong et al., 1984 (Wong et al., , 1985 . As detected by immunohistochemical analysis, the addition of IFN-7 to brain cells in culture results in 9% of neurons expressing MHC class I, whereas direct injection into the CNS results in approx 15% of neurons expressing MHC class I. Moreover, the response of the OLB21 neuronal cell line to IFN-? has been investigated (Joly and Oldstone, 1992) . The OLB21 cell line was developed by retroviral-mediated oncogene delivery to cells from the olfactory bulb (Ryder et al., 1990) . In these cells, IFN-7 stimulates the expression of MHC class I both at the mRNA level and on the cell surface. Furthermore, in the presence of IFN-7, the peptide transporters HAM 1 and HAM 2, which are involved in the processing of class I antigens, are induced in these cells, whereas an increase in the expression of ~2-microglobulin also occurs (Joly and Oldstone, 1992) . In other studies, cell lines derived from two medulloblastomas, which are immunoreactive against neurofilament antibodies, but unreactive against GFAP and S-100 antibodies, have also been shown to empress MHC class I and class II antigens following exposure to IFN-y (Tamura et al., 1989) . Although these studies further demonstrated that cells of neuronal origin are intrinsically capable of expressing MHC molecules and suggest that this expression may serve a functional role, the physiological conditions under which functional neuronal MHC expression can occur have not yet been fully elucidated. Toward this goal, Neumann et al. (1995) have combined whole-cell patch-clamp recording techniques with single-cell RT-PCR analysis to demonstrate that in the presence of IFN-7, both MHC class I and 132-microglobulin are significantly more inducible on electrically silent neurons compared to neurons spontaneously firing action potentials. This result suggests that functional MHC class I expression may occur only in neurons that are no longer biologically active. This would then allow these cells to be recognized and killed by cytotoxic T-lymphocytes, thus permitting active virus to be cleared from inactive neurons. This result also suggests that active neurons may not possess the potential to express functional MHC, thus allowing neurons to survive at the cost of continued viral infection (Joly et al., 1991) . Future investigation in this area will likely yield additional insight into our understanding of the functional interactions between neurons and cells of the immune system.

Although the majority of the studies addressing the effect of IFN-~, on neurons have been directed at examining the expression of MHC glycoproteins, several studies have reported other IFN-~,-induced neuronal consequences. As previously discussed, Brosnan et al. (1989) reported a significant reduction in neural conduction and modest axonal pathology in the visual system of the rabbit following injection of IFN-,f. Recently, McMillian et al. (1995) reported that the treatment of rat primary basal forebrain mixed cultures with IFN-7 decreased the number of choline acetyl transferase (CHAT) immunopositive neurons, whereas the number of ChAT immunonegative neurons was unaffected. This selective neuronal killing was shown to be mediated through IFN-~,-induced microglia activation. Moreover, Birdsall (1991) demonstrated that following exposure to IFN-~[, two human neuroblastoma cell lines, but not cultured human cortical neurons, expressed ICAM. IFN-y also enhanced TNF-el-induced binding of these cells to neutrophils. Finally, Collins (1995) demonstrated that susceptibility to viral infection of a human cerebral cortical neuronal cell line was dramatically increased following treatment with human IFN-~, and that this increased susceptibility was owing to membrane expression of HLA class I molecules.

IFN--f has also been shown to affect neuronal differentiation in vitro. Chang et al. (1990) demonstrated that IFN-~, retarded the death of cultured rat sympathetic neurons following nerve growth factor (NGF) deprivation. Moreover, IFN-y has been shown to potentiate NGFstimulated differentiation of PC12 cells (Improta et al., 1988) . Barish et al. (1991) also demonstrated that IFN-y increased the differentiation of cultured cortical and hippocampal neurons. Nevertheless, it is unlikely that the developing nervous system is exposed to appreciable levels of IFN-7 under normal conditions, such that the in vivo relevance of these observations is unclear.

In contrast, there is in vivo evidence that suggests that IFN-~, might have a detrimental effect on the developing nervous system. As mentioned earlier, transgenic mice in which IFN-y was targeted to the CNS demonstrate a disruption in cerebellar granule cell migration (Corbin et al., 1996) . During development, cerebellar granule cells migrate from the external granule cell layer (EGL), through the molecular layer, to the internal granule cell layer (IGL) along processes extended by radial gliaI cells (Chuong, 1990) . In transgenic mice expressing IFN-? within the CNS, many granule cells are still observed in the EGL and in the molecular layer. This observation is reminiscent of other mouse mutants (e.g., weaver, staggerer) in which the interaction between granule cells and radial glia is disrupted (Chuong, 1990) . Perhaps the ectopic expression of IFN-~, in the cerebellum of transgenic mice has a deleterious effect on Lhe interaction between granule cells and radial glia, thus disrupting the migration of these cells from the EGL to the IGL. Further investigation is warranted to characterize the potentially deleterious developmental effects of IFN-g on the CNS in more detail.

Recently, Meda et al. (1995) suggested a role for IFN-y in Alzheimer's disease. The senile plaques that develop in Alzheimer's patients are characterized by the extracellular deposits of [3-amyloid protein. Meda et al. (1995) demonstrated that IFN-y acted synergistically with fJ-amyloid in the activation of microglia, which produce TNF-o~ and nitric oxide. In coculture experiments, it was shown that microglia activated by IFNq, and ~-amyloid caused neuronal injury. Clearly, these are exciting findings that lead to the speculation that IFN-~/might also be involved in other neurodegenerative disorders.

IFN-? is a multifunctional cytokine involved in regulating a variety of immune responses. This cytokine also has a multitude of effects in the CNS (Table 2) . Although normally excluded from the CNS, this cytokine is present during many disorders in which the CNS is affected, including the most common human demyelinating disease, MS. The presence of Table 2 Effects of IFN-? on the CNS Disruption of myelination Exacerbation of reactive gliosis Disruption of the blood-brain barrier Recruitment of inflammatory cells from the periphery Induction of MHC expression Stimulation of macrophage/microglial cells Disruption of cerebellar granule cell migration this cytokine in the CNS leads to an abundance of pathological effects, including decreased myelination, increased gliosis, disruption of blood-brain barrier func~on, increased MHC expression, stimulation of macrophage/microglial activation, and disruption of cerebellar granule cell migration. Taken as a whole, these observations demonstrate that IFN-? can either directly or indirectly mediate many of the CNS pathological effects that are observed in disorders in which immune cell traffic in the CNS increases. Understanding the mechanism(s) by which IFN-~,r exerts its actions on CNS function will undoubtedly prove useful toward generating rational therapeutic approaches for the treatment of CNS disorders in which this cytokine plays a harmful role.

We thank numerous colleagues for many helpful discussions that guided the formulation of the ideas presented here. We also thank Kathy Toews for assistance in preparing the manuscript. The IFN-? work in Brian Popko's laboratory has been supported by the National Institutes of Health (NIH) research grant ROI NS34939, and the National Multiple Sclerosis Society research grant RG 2089. J. D. is supported by a training grant HD07201 from the NIH. B. P. is a recipient of an NIH Research Career Development Award (NS 01637) from NINDS.

Reversible inhibitory effects of interferon-y and tumouar necrosis factor-o: on oligodendroglial lineage cell proliferation and differentiation in vitro

Molecular cloning and expression of the human interferon-? receptor

Experimental Allergic Encephalomyelitis: A Usefid Model for Multiple Sclerosis

Ligand-induced assembly and activation of the gamma interferon receptor in intact cells

Reactive astrogliosis in the neonatal mouse brain and its modulation by cytokines, f Neurosci

Gamma-interferon promotes differentia* tion of cultured cortical and hippocampal neurons

Increased production of interferon gamma and tumor necrosis factor precedes clinical manifestation in multiple sclerosis: do cytokines trigger off exacerbations?

Inflammatory cytokines within the central nervous system: sou~, traction, and mechanism of action

TNF-~ and IFN-7-mediated signal transduction pathways: effects on glial cell gene expression and function

In the presence of dexamethasone, y interferon induces rat oligodendrocytes to express major histcycompatibiliw complex class II molecules

Enhancement of experimental allergic encephalomyelitis in mice by antibodies against IFN-y

Induction of ICAM-I on human neural cells and mechanisms of neutrophilmediated injury

Localization of aldosterone and corticosterone in the central nervous system, assessed by quantitative autoradiography

Induction of nitric oxide synthase in demyelinating regions of multiple sclerosis brains

Immunohistochemical analysis of T-lymphocyte subsets in the central nervous system in chronic progressive multiple sclerosis

Preliminary studies of cytokine-induced functional effects on the visual pathways in the rabbit

The differentiation of 0-2A progenitor cells into oligodendrocytes is associated with a loss of inducibility of Ia antigens

Interferon suppresses sympathetic neuronal cell death caused by nerve growth factor deprivation. L Neurochem

Differential roles of multiple adhesion molecules in cell migrafion: granule cell migration in cerebellum

Interferon g potentiates human coronavirus ~43 infection of neuronal cells by modulation of HLA class I expression

Targeted CNS expression of interferon-gamma in transgenic mice leads to hypomyelination, reactive gliosis, and abnormal cerebellar development

The blood-brain barrier; a modified fight epithelium, in The Blood-Brain Barrier in Health and Disease (Suckling

Multiple defects of immune cell function in mice with disrupted interferon-y genes

I994) Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins

Transcription factors in interferon signaling

Inhibition of cellular responsiveness to interferon-y (IFNT) induced by overexpression of inactive forms of the IFN7 recepton

Tissue-specific targeting of cytokine unresponsiveness in transgenic mice

Effect of anti-interferon-y and anti-interleukin-w monoclonal antibody treatment on the development of actively and passively induced experimental allergic encephalomyelitis in the SJL/J mouse

I994) Effect of anti-interferon-gamma monoclonal antibody treatment on the development of experimental allergic encephalomyelitis in resistant mouse strains

Three-dimensional structure of recombinant human IFN-y

Regulation of g!ial intermediate filaments in astrogliosis

The effect of cytokines on the expression and function of FC receptors for IgG on human myeloid cells

Nervous tissue as an im_mune compartment: the dialect of the immune response in the CNS

The molecular cell biology of interferon-y and its receptor

Mice with a disrupted interferon-y gene are susceptible to the induction of experimental autoimmune encephalomyelitis (EAE)

Induction of human hsp60 expression in monocytic cell lines

Astrocytes as antigen-presenting cells. I. Induction of Ia antigen expression on astrocytes by T cells via immune interferon and its effect on antigen presentation

Astrocytes present myelin basic protein to encephalitogenic T-cell lines

Stoichiometry of interaction between interferon y and its receptor

The induction of intercellular adhesion molecule 1 (ICAM-1) expression on human fetal astrocytes by interferon-gamma, tumor necrosis factor alpha, lymphotoxin, and interleukin-l: relevance to intracerebral antigen presentation

Regulation of T-cell activation: differences among T cell subsets

Experimental autoimmune encephalomyelitis. Qualitative and semiquanfitative differences in heat shock protein 60 expression in the central nervous system. 1~ lmmunol

I982) Structure of the human immune interferon gene

Cloning and expression of murine immune interferon cDNA

Expression of human immune interferon cDNA in E. coli and monkey ceils

In terferon-y induces receptor dimerization in solution and on cells

Inflammatory mediators in demyelinating disorders of the CNS and PNS

Expression of the neu oncogene under the transcriptional control of the myelin basic protein gene in transgenic mice: generation of transformed glial cells

I994) A novel member of the interferon receptor family complements functionality of the murine interferon "I receptor in human cells

Migration of hematogenous cells through the blood-brain barrier and the initiation of

Effects of interferon-gamma on primary cultures of human brain microvesset endothelial cells

IFN-gamma facilitates NGF-induced neuronal differentiation in PCt2 cells

Heat shock proteins: applications in health and disease

Neuronal cells are deficient in loading peptides onto MHC class I molecules

!) Viral persistence in neurons explained by lack of major histocompatibility class I expression

Effects of glycosidase treatment on the physicochemical properties and biological actiw ity of human interferon-)

In vivo expression of inducible nitric oxide synthase in experimentally induced neurologic diseases

Differential regulation of oligodendrocyte markers by glucocorticoids: post-h-anscriptional regulation of both proteolipid protein and myelin basic protein and transcriptional regulation of glycerol phosphate dehydrogenase

Morphological and molecular response of the MOCH-1 oligodendrocyte cell line to serum and Lnterferon-y: possible implications for demyelinating disorders

Immunological aspects of experimental allergic encephalomyelitis and multiple sclerosis

Immunological aspects of demyelinating diseases

Cell type-specific regulation of major histocompatibility complex (MHC) class I gene expression in astrocytes, oligodendrocytes, and neurons

Differential expression and regulation of major histocompatibility complex (MHC) products in neural and glial cells of the human fetal brain

Selective killing of cholinergic neurons by microglial activation in basal forebrain mixed neuronal/glial cultures

Cytokine-regulated adhesion between encephalitogenic T lymphocytes and cerebrovascular endothelial cells

Induction of active and adoptive relapsing experimental autoimmune encephalomyelitis (EAE) using an encephalitogenir epitope of proteolipid protein

Activation of microglial cells by beta-amyloid protein and interferon-gamma

Microglial cell cytotoxicity of oligodendrocytes is mediated through nitric oxide

Evdution of the T-cell d~ng the course of experimental imanune-mediated demyelinating disea~s

Nitric oxide induces necrotic but not apoptotic cell death in oligodendrocytes

Differential expression of Ia and Ia-associated invariant chain in mouse tissue after in vivo treatment with IFN-7

Interferon pretreatment lowers the threshold for maximal heat-shock response in mouse cells

The protein tyrosine kinase JAK1 complements defects in interferon-alpha/beta andgamma signal transduction

Induction of MHC class I genes in neurons

Cytokines in neuroinflammatory disease: role of myelin autoreactive T cell production of interferon-gamma

Autoreactive T lymphocytes in multiple sclerosis determined by antigen induced secretion of interferon-gamma

Interferons in multiple sclerosis

Treatment of multiple sclerosis with game, ha interferon: exacerbations associated with activation of the immune system

Lymphocyte responses and cytokines

MOCHq cells: an oligodendrocyte cell line generated using a transgenic approach

Trauma and multiple sclerosis

The pathogenesis of multiple sclerosis. Additional considerations

Major histocompatibility complex class I expression in oligodendrocytes induces hypomyelination in transgenic mice

Heat shock protein immunoreactivity in CSF: correlation with oligoclonal banding and demyelinating disease~ Neurology

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

The Dale E. McFarlin memorial lecture: the immunology of the multiple sclerosis lesion

Presidential address: multiple sclerosis: immune system molecule expression in the central nervous system

T helper cell differentiation in immune response

Cytokine production by cells in cerebrospinal fluid during experimental allergic encephalomyelitis in SJL/J mice

Ultrastructure of multiple sclerosis

Cytokineinduced expression of vascular cell adhesion molecule-1 (VCAM-t) by astrocytes and astrocytoma cell lines

Demonstration of the presence of a specific interferon-~, receptor on murine astrocyte cell surface

Establishment and characterization of multipotent neural cell lines using retrovirus vectormediated oncogene transfer

A common nuclear signal transduction pathway activated by growth factor and cytokine receptors

The Jak kinases differentially associate with the ot and [5 (accessory factor) chains of the interferon I receptor to form a functional receptor unit capable of activating STAT transcription factors

Cytokine-induced expression of intercellular adhesion molecule-1 (ICAM-1) in cultured human oligodendrocytes and astrocytes

Expression and induction of intercellular adhesion molecules (ICAMs) and major histocompatibility complex (MHC) antigens on cultured murine oligodendrocytes and astrocytes

Molecular characterization of interferon gamma as a macrophage activating factor

Acquisition of lymphokine-producing phenotype by CD4+ T cells

Cytokine cytotoxicity against oligodendrocytes. Apoptosis induced by lymphotoxin

Expression of heat shock protein-65 by oligodendrocytes in vivo and in vitro: Implications for multiple sclerosis

Immune modulation within the brain: recruitment of inflammatory cells and increased major histocompatibility antigen expression following intracerebral injection of interferon-gamma

A conserved AU sequence from the 3 ~ untranslated region of GM-CSF ml-hNA mediates selective mRNA degradation

I994) Interferon-activated signal transduction to the nucleus

A single phosphotyrosine residue of Stat91 required for gene activation by interferongamma

Direct injection of cytokines into the spinal cord causes autoimmune encephalomyelitis-like inflammation

Susceptibility of astrocytes to class I MHC antigen-specific cytotoxicity

Interferongamma and B cell stimulatory factor-1 reciprocally regulate Ig isotype production

Identification and sequence of an accessory factor required for activation of the human interferon %, receptor

Evidence for involvement of ICAM-1 and VCAM-1 in lymphocyte interaction with endothelium in experimental autoimmune encephalomyelitis in the cen~al nervous system in the SJL/J mouse

Rat ependyma and microglia cells express class II MHC antigens after intravenous infusion of recombinant gamma interferon

The expression of MHC antigens on oligodendrocytes: induction of polymorphic H-2 expression by lymphokines

Helper T-cell subsets: phenotype, function and the role of lymphokines in regulating their development

Expression of major histocompatibility complex on human medulloblastoma cells with neuronal differentiation

RG-2 glioma growth attenuation and severe brain edema caused by local production of interleuk~u-2 and interferon-gamma

Expression of interferon-~[ receptors on murine oligodendrocytes and its regulation by cytokines and mitogens

Deu, non~stration of ct, t3, and ~[ interferon in active chronic multiple sclerosis lesions

Multiple sclerosis: distribution of T-cell subsets within active chronic lesions~

Immune interferon: a pleiotropic lymphokine with multiple effects

Myelin proteotipid protein-induced experimen--tal allergic encephalomyelitis

A synthetic peptide from mye

Identification of an encephalitogenic determinant of myelin proteolipid protein for SJL mice

I992) Myelin proteolipid protein: minimum sequence requirements for active induction of autoimmune encephalomyelitis in SWR/J and SJL/J mice

Dysmyelination in class I MHC transgenic mice

Regulation of MHC molecules on MBP positive oligodendrocytes in mice by IFN-7 and TNF-a

Dysmyelialation in transgenic mice resulting from expression of class I histocompatibility molecules in oligodendrocytes

The role ofT cells in multiple sclerosis: implications for therapies targeting the T cell receptor

The small heat-shock protein alpha B-crystalline as candidate autoantigen in multiple sclerosis

Interferon-y-induced oligodendrocyte cell death: implications for the pathogenesis of multiple sclerosis

Interferon-1 potentiates anti~ body-mediated demyelination in vivo

Recent progress in the elucidation of interferon-gamma actions: molecular biology and biological functions

Interferon gamma: a 1}nnphokine for all seasons

Suppression of experimental allergic encephalomyelitis by intraventricular administration of interferon-gamma in Lewis rats

T helper 1 (Tftl) functional phenotype of human myelin basic protein-specific T lyrnphocytes

Crystal structure of a complex between interferon-gamma and its soluble high-affinity receptor

Complementation by the protein tyrosine kinase JAK2 of a mutant cell line defective in the interferon-gamma signal transduction pathway

Interferon-like virus-inhibitor induced in human leukocytes by phytohernagglutinin

Cytokines and murine autoimmune encephalomyelitis: inhibition or enhancement of disease with antibodies to select cytokines, or by delivery of exogenous cytokines using a recombinant vaccinia virus system

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

Interferon-gamma induces the expression of H-2 and la antigens on brain ceils

R (1991) y-Interferon promotes proliferation of adult human astrocytes in vitro and reactive gliosis in the adult mouse brain in vivo

Differential proliferafive response of human and mouse astrocytes to gala-interferon

~1) Try_ ~nsgewic mouse model for central nervo~is system demyelknation

Role of interferon-y in immune cell regulation

The T lymphocyte in experimental allergic encephalomyelitis