key: cord-0745494-t7vav41i
authors: Hamo, Ludwig; Stohlman, Stephen A.; Otto‐Duessel, Maya; Bergmann, Cornelia C.
title: Distinct regulation of MHC molecule expression on astrocytes and microglia during viral encephalomyelitis
date: 2007-06-28
journal: Glia
DOI: 10.1002/glia.20538
sha: c880191cc7b27ef7fbe0c7c36387d38c4721730d
doc_id: 745494
cord_uid: t7vav41i

The potential interplay of glial cells with T cells during viral induced inflammation was assessed by comparing major histocompatibility complex molecule upregulation and retention on astrocytes and microglia. Transgenic mice expressing green fluorescent protein under control of the astrocyte‐specific glial fibrillary acidic protein promoter were infected with a neurotropic coronavirus to facilitate phenotypic characterization of astrocytes and microglia using flow cytometry. Astrocytes in the adult central nervous system up‐regulated class I surface expression, albeit delayed compared with microglia. Class II was barely detectable on astrocytes, in contrast to potent up‐regulation on microglia. Maximal MHC expression in both glial cell types correlated with IFN‐γ levels and lymphocyte accumulation. Despite a decline of IFN‐γ concomitant to virus clearance, MHC molecule expression on glia was sustained. These data demonstrate distinct regulation of both class I and class II expression by microglia and astrocytes in vivo following viral induced inflammation. Furthermore, prolonged MHC expression subsequent to viral clearance implies a potential for ongoing presentation. © 2007 Wiley‐Liss, Inc.

MHC molecules, required for T cell activation and effector recognition, are not constitutively expressed on central nervous system (CNS) resident cells (Aloisi et al., 2000; Sedgwick and Hickey, 1997; Xiao and Link, 1998) . Restrictions on T cell surveillance are further imposed by the blood brain barrier and absence of classical lymphatic drainage (Aloisi et al., 2000; Hickey, 2001; Xiao and Link, 1998) . This quiescent state is rapidly overcome following CNS infections. During virus induced inflammation, microglia readily up-regulate both class I and class II MHC molecules in the presence of IFN-g secreted by T cells . However, in vivo detection of MHC molecules on astrocytes and oligodendrocytes during inflammation has been more difficult (Horwitz et al, 1999; Redwine et al., 2001; Stuve et al., 2002; Traugott et al., 1985) . Whereas class I expression was detected on oligodendroglia during a neuronotropic viral infection, it was not detected on astrocytes (Redwine et al., 2001) . Nevertheless, infection of cultured astrocytes with mouse hepatitis virus (MHV) induces MHC up-regulation (Gilmore et al., 1994; Lavi et al., 1989; Suzumura et al., 1986) . Furthermore, although IFN-g readily increases constitutive MHC class I expression and induces class II expression on cultured astrocytes derived from neonates (Fierz et al., 1985; Massa et al., 1993; Wong et al., 1985) , constitutive transgenic IFN-g expression was unable to induce class I or class II expression on astrocytes in vivo (Horwitz et al., 1999) . Systemic administration of IFN-g to mice also did not induce class II expression on astrocytes (Momburg et al., 1986) , whereas intracerebral injection did (Wong et al., 1984) . The recent unequivocal demonstration of class I expression on astrocytes in vivo following CNS adenoviral infection by confocal microscopy (Barcia et al., 2006) , in addition to indirect evidence for MHC restricted CD8 1 T cell function (Cabarrocas et al., 2003; Lin et al., 1997) suggests technical limits imposed by histological analysis (Lampson, 1995) may have underestimated MHC expression.

In contrast to the CD45 lo phenotype used to identify microglia, identification of astrocytes derived from the adult CNS has been hampered by the lack of a suitable surface marker. This study takes advantage of transgenic mice expressing the green fluorescent protein (GFP) under control of the astrocyte-specific glial fibrillary acidic protein (GFAP) promoter (Zhuo et al., 1997) to directly analyze surface molecule expression on GFP 1 astrocytes derived from the adult inflamed CNS by flow cytometry. Encephalomyelitis was induced by infection with a neurotropic coronavirus variant designated JHMV. Infection of adult mice induces an immune-mediated primary demyelination, and has served as a model for the human demyelinating disease, multiple sclerosis . JHMV infection induces encephalitis in most strains of mice, infecting ependymal cells, microglia, astrocytes, and oligodendrocytes, but rarely neurons (Wang et al., 1992) . Following a vigorous CNS inflammatory response, infectious virus is generally cleared by CD8 1 T cells within 2 weeks of infection . Control of virus infection in microglia and astrocytes is dependent on perforin-mediated cytolysis , implicating class I antigen presentation to virus-specific CD8 1 T cells. By contrast, viral clearance from oligodendrocytes appears independent of perforin-mediated cytolysis, but requires IFN-g (Bergmann et al., 2003; Parra et al., 1999) . Despite control of infectious virus by two anti-viral effector mechanisms, viral antigen and RNA persists in the absence of detectable infectious virus . Viral persistence appears to be restricted to astrocytes (Perlman and Ries, 1987) and oligodendrocytes (Lavi et al., 1984; Parra et al., 1999; Ramakrishna et al., 2003) . Loss of CD8 T cell effector function as well as poor MHC expression on glial subsets potentially contributes to the inability of T cells to control viral persistence (Ramakrishna et al., 2006) . MHC molecule expression on astrocytes was compared with microglia during viral encephalomyelitis in adult mice using flow cytometry. Astrocytes strongly up-regulated class I, albeit delayed compared with microglia. By contrast, class II remained undetectable on the majority of astrocytes, differing from the potent class II up-regulation on microglia. However, once induced, MHC molecule surface expression was maintained following virus clearance and down regulation of IFN-g. These novel data highlight the distinct abilities of microglia and astrocytes to up-regulate expression of class I and class II molecules in response to virus-induced inflammation and suggest that prolonged MHC expression may provide an opportunity for auto antigen presentation.

Mice expressing GFP under control of the human GFAP promoter (FVB/N-Tg[GFAPGFP]14Mes/J) (Zhuo et al., 1997) were purchased from The Jackson Laboratory (Bar Harbor, ME) and bred in an accredited animal facility at the Keck School of Medicine, University of Southern California, Los Angeles, CA. All procedures were carried out in compliance with Institutional Animal Care and Use Committee approved protocols. Six-to seven-week-old mice of both sexes were injected in the left cerebral hemisphere (i.c.) with 30 lL of endotoxinfree PBS containing 500 plaque forming units (PFU) of the 2.2-v1 mAb-derived variant of JHMV (Fleming et al., 1986) . Clinical symptoms were evaluated daily based on a 0-4 scale as previously described (Fleming et al., 1986) . Briefly, mice were graded as: 0 5 healthy; 1 5 ruffled hair and hunchbacked appearance; 2 5 reduced mobility and inability to upright; 3 5 paralysis and wasting; 4 5 death. To determine virus replication within the CNS one half brains from individual mice were homogenized in 4 mL sterile phosphate buffer saline (pH 7.4) using ice cold Tenbroeck tissue homogenizers. Homogenates were centrifuged at 500g for 7 min at 4°C. Supernatants were frozen at 270°C for determination of infectious virus by plaque assay carried out on monolayers of the DBT astrocytoma cell line as previously described (Fleming et al., 1986; Lin et al., 1997) .

For isolation of CD45 hi bone marrow derived CNS infiltrating cells, pooled brains (n 5 3-4 per time point) were homogenized and centrifuged at 500g for 7 min at 4°C as described earlier for virus supernatants. Cell pellets were resuspended in PBS and adjusted to 30% Percoll (Pharmacia, Uppsala, Sweden). Following addition of a 1 mL 70% Percoll underlay and centrifugation at 800g at 4°C for 25 min cells were isolated from the 30%/70% interface, washed in RPMI medium, and resuspended in PBS containing 0.5% BSA. Astrocytes were isolated using proteolytic digestion as previously described for oligodendrocytes (Gonzales et al., 2005) . Brains from 3 to 4 mice per time point were combined, rinsed in RPMI supplemented with 25 mM HEPES (pH 7.2) and 1% fetal calf serum (FCS), and placed into 100 mm plates. After finely mincing in 1 mL RPMI supplemented with HEPES and 1% fetal FCS with razor blades, 4 mL of a prewarmed 0.25% trypsin solution was added per brain, and digestions carried out in 50 mL tubes at 37°C for 30 min with intermittent agitation. Digestion was terminated by addition of 5 mL of ice cold RPMI containing 20% FCS per brain. Following centrifugation at 500g for 7 min at 4°C, cells were resuspended in RPMI supplemented with 25 mM HEPES containing 30% Percol, and concentrated by centrifugation onto a 70% Percol cushion as described earlier.

Brains were isolated from newborn FVB/N-Tg-[GFAPGFP]14Mes/J mice and glial cells cultured as described (Hindinger et al., 2005) . After removal of meninges and superficial blood vessels frontal lobes were digested with 0.25% trypsin in PBS containing 0.02% EDTA for 20 min at 37°C and dissociated by trituration. Cells from 3 to 5 mice were added to 75 cm T flasks containing Dulbecco's modified minimal essential medium (Irvine Scientific, Santa Ana, CA) supplemented with 10% FCS and penicillin (1000 units/mL)/streptomycin (1000 lg/mL). Medium was replaced after 12-16 h at 37°C and subsequently at 3 day intervals. Cells were subcultured once after 7-10 days and used at confluency. Mixed glial cultures were incubated in the presence or absence of 1 ng/mL recombinant murine IFN-g (BD PharMingen, San Diego, CA) for 48 h. MHC class I and class II expression was analyzed following trypsinization (0.025% trypsin/EDTA) of monolayers by flow cytometry.

CNS derived cells were incubated at 4°C for 20 min with anti-FcIII/IIR mAb (2.4G2) (BD PharMingen) and 2% mouse serum to block nonspecific binding. Surface expression was determined using mAb for CD4 (clone GK1.5), CD8 (53-6.7), CD11b (M1/70), CD45 (30-F11), NK1.1 (PK135), MHC class II (M5-114) (all from BD PharMingen), MHC class I (mAb 28-14-8) (eBiosciences, San Diego, CA), and F4/80 (A3-1, Serotec, Raleigh, NC) conjugated with APC, PE, or PerCP. Isotype control mAb for class I and class II were mouse IgG 2a (mAb 28-14-8) (eBiosciences, San Diego, CA) and rat IgG 2b ,j (BD Phar-Mingen), respectively. Cells were incubated with Ab for 30 min at 4°C and washed twice before analysis on a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA). Astrocytes were defined as GFP 1 CD45 2 cells, microglia as CD45 lo cells, and infiltrating cells as CD45 hi . Data were analyzed using CellQuest-Pro (Becton Dickinson) or Flowjo software (Tree Star, Ashland, OR).

IFN-g was measured in supernatants from brain homogenates of infected mice by ELISA. Supernatants were prepared in PBS as described earlier and stored at 270°C. Briefly, 96-well plates coated overnight with 100 ng of IFN-g mAb (R4-6A2) in carbonate buffer (0.1 M, pH 9.5) at 4°C were washed 3 times with PBS containing 0.05% Tween 20 (Sigma) and blocked with PBS containing 10% FCS for 1 h at RT. Serial twofold diluted brain supernatants or IFN-g standard (BD PharMingen) were added in duplicate. Samples were incubated at RT for 2 h, and washed prior to addition of biotinylated anti-IFN-g mAb (XMG1.2) and avidin-horseradish peroxidase. Color was developed following addition of tetramethylbenzidine Reagent (BD PharMingen) and absorbance measured at 405 nm using a Microplate Autoreader model 680, with data analyzed using Microplate Manager 5.2 software (Bio-Rad, Hercules, CA).

Brains and spinal cords were examined for inflammation, demyelination, and distribution of viral Ag. Tissues were fixed for 3 h in Clark's solution (75% ethanol, 25% glacial acetic acid) prior to embedding. Sections were stained with hematoxylin and eosin (H&E) to determine inflammation and luxol fast blue (LFB) to determine demyelination. Distribution of viral Ag was determined by immunoperoxidase staining (Vectastain-ABC kit; Vector Laboratory, Burlingame, CA) using the anti-JHMV mAb J.3.3 specific for the carboxyl terminus of the viral nucleocapsid protein as the primary Ab and horse anti-mouse as secondary Ab (Vector Laboratory). Sections were scored for inflammation, demyelination, and viral Ag in a blinded fashion. Representative fields were identified based on average scores of all sections in each experimental group.

JHMV infection elicits an acute encephalomyelitis, accompanied by demyelination that resolves into a chronic demyelinating disease characterized by persistence of viral antigen and RNA in BALB/c (H-2 d ) and C57BL/6 (H-2 b ) mice . To confirm susceptibility and similar pathogenesis in FVB/N (H-2 q ) mice expressing GFP in astrocytes, virus replication within the CNS and clinical symptoms were monitored until 21 days post infection (p.i.) (see Fig. 1 ). Virus replication in the CNS of infected transgenic FVB/N mice peaked at 5 days p.i. and then declined to undetectable levels in all mice by 21 days p.i. (Fig. 1A) . These data are consistent with the kinetics of acute JHMV replication in the CNS of both H-2 d and H-2 b mice Ramakrishna et al., 2002) , with the exception of residual infectious virus at day 14 p.i., which is not observed in the latter strains. Clinical symptoms in infected FVB/N mice were observed as early as day 5 p.i., and all mice were symptomatic by day 10 p.i. (Fig. 1B) . Clinical symptoms reached a plateau between days 13 and 18 p.i. and gradually declined thereafter, with the majority of mice exhibiting essentially complete clinical recovery (Fig. 1B) . Although overall onset of symptoms was similar to BALB/c and C57BL/6 mice, the severity in infected FVB/ N mice was reduced from an average score of 2.5 (Lin Parra et al., 1999; Ramakrishna et al., 2002) to 1.5 in FVB/N mice. Despite the reduced clinical symptoms, JHMV infection induced a prominent inflammatory response in both brain (data not shown) and spinal cord (see Fig. 2 ), as well as vigorous demyelination predominantly in the spinal cord (see Fig. 2 ).

The reduced clinical symptoms in FVB/N mice suggested the possibility that the inflammatory response differed from those characterized in other strains of infected mice. Therefore, the composition and kinetics of cells entering the CNS of the infected FVB/N mice were analyzed by flow cytometry. Infiltrating CD45 hi cells peaked at 7 days p.i. (Fig. 3A) , declined by day 10 p.i. and persisted out to day 21 p.i. Initial infiltrates comprised mainly innate components (Fig. 3B) . Although macrophages dominated, neutrophils and NK cells were also recruited into the infected CNS. Neutrophils peaked early and rapidly disappeared by day 5 p.i., while infiltration of NK cells was maximal between days 5 and 7 p.i. and gradually decreased thereafter. As the number of total infiltrates increased, the percentage of the macrophages declined and T cells became prominent (Fig. 3C) . While total T cells constituted 3 3 10 5 and 10 3 10 5 of the infiltrates at days 5 and 7, respectively, NK cells remained relatively constant at 2.5 3 10 5 . CD4 1 T cells consistently outnumbered CD8 1 T cells by at least 2:1 throughout the course of infection. The increased ratio of CD4 1 T cells to CD8 1 T cells distinguishes the pathogenesis in FVB/N mice from both BALB/c and C57BL/6 mice, in which CD8 1 T cells prevail by day 8 p.i. Nevertheless, JHMV infection of FVB/N mice overall induces an acute encephalomyelitis closely resembling the pathogenesis in previously described mouse strains . 

JHMV infection induces both MHC class I and class II expression on microglia by 6 days p.i. and expression correlates with maximal T cell infiltration and IFN-g secretion . Whereas expression of MHC class II on microglia is strictly IFN-g dependent, class I is induced, albeit at lower levels and more transiently in the absence of IFN-g . To compare regulation of MHC molecule expression on astrocytes to microglia by flow cytometry GFP 1 astrocytes were isolated from the adult CNS. Mechanical preparation of CNS derived cells, used to characterize CD45 hi infiltrating cells and CD45 lo microglia, did not yield appreciable GFP 1 glia. Therefore, a mild trypsin digestion was employed to release astrocytes from brain tissues as previously described for isolation of oligodendrocytes Ramakrishna et al., 2006) . This approach typically resulted in 3-15% GFP 1 CD45 2 astrocytes within total viable cells recovered from the brains and was thus used to monitor MHC expression patterns on microglia and astrocytes throughout the course of JHMV infection (see Fig. 4 ).

Class I expression on CD45 lo microglia was directly compared with class I 1 CD45 hi cells (Fig. 4B ) and increases in class I surface expression over time depicted relative to na€ ıve microglia (Figs. 4B,C). Microglia from na€ ıve mice exhibited low amounts of constitutive expression, as indicated by the shift in class I staining relative to isotype control. This baseline class I expression on microglia from na€ ıve mice contrasts the inability to detect surface class I on microglia from na€ ıve BALB/c or C57BL/ 6 mice (data not shown), but may reside in reactivity of the class I 28-14-8 mAb. A substantial population of microglia up-regulated MHC class I by day 3 p.i. (Fig. 4C) . Expression levels increased further by day 5 p.i., reached maximal levels by day 7 p.i. and remained high out to 21 days p.i. Class I expression on microglia exceeded that of the majority of infiltrating cells by day 10 p.i. These data are consistent with upregulation of MHC class I expression on microglia derived from infected BALB/c mice . No MHC class I expression was detected on astrocytes derived from uninfected adult mice (Fig. 4D) . Following infection up-regulation of MHC class I on astrocytes lagged compared with microglia. Although class I expression was barely detectable by day 3 p.i., (22% class I 1 astrocytes vs. 63% microglia), it gradually increased thereafter. By day 7 p.i. nearly 100% of astrocytes expressed class I and expression was maintained through 21 days p.i., similar to microglia.

To confirm the paucity of MHC class II expression by astrocytes in vivo (Horwitz et al., 1999; Redwine et al., 2001) , class II expression on astrocytes was compared with microglia by flow cytometry. Neither microglia nor astrocytes from na€ ıve mice expressed detectable levels of class II (Figs. 5A,B) . Microglia revealed a defined population expressing class II by day 5 p.i., which increased prominently by day 7 p.i. (Fig. 5A) . By day 10 p.i. all microglia expressed high levels of class II and expression was sustained through day 21 p.i. By contrast, expression of MHC class II on astrocytes was tenuous (Fig. 5B) . Although a slight shift in fluorescence was evident throughout days 7-21 p.i. on 20% of astrocytes, the extent of class II expression was very low compared with microglia and limited to a small subset.

To confirm IFN-g inducible MHC expression patterns observed in astrocyte cultures derived from other mouse strains (Aloisi et al., 2000; Dong and Benveniste, 2001; Hindinger et al., 2005; Stuve et al., 2002) , MHC expression was also examined on microglia and astrocytes within mixed glial cultures prepared from neonatal FVB/ N mice. Consistent with numerous previous reports, MHC class I was constitutively expressed on cultured astrocytes and microglia and the level of expression was increased in both cell types by IFN-g treatment (see Fig. 6 ). The descrepency of constitutive class I expression may reside in prolonged growth in culture or in the immature nature of these glia compared with those derived and analyzed ex vivo from adult mice. MHC class II was below the level of detection in untreated cultures, but up-regulated by IFN-g on both microglia and astrocytes, although upregulation was more tentative on astrocytes (see Fig. 6 ). Thus, no strain dependent alterations in MHC expression were detected on the GFP 1 astrocytes derived from the transgene expressing FVB/N mice.

Initial up-regulation of MHC molecules by day 5 p.i. coincides with infiltration of NK cells, CD8 1 , and CD4 1 T cells, all potential sources of IFN-g. IFN-g is a potent antiviral cytokine that also enhances MHC class I and class II expression on microglia during JHMV infection . IFN-g protein levels were thus measured at 3-17 days p.i. to directly gauge IFN-g secretion in the inflamed CNS. No IFN-g was detected at day 3 p.i. IFN-g peaked rapidly between days 5 and 7 p.i. and the levels declined dramatically by day 10 p.i. (see Fig. 7 ). Overall, IFN-g levels detected within the CNS coincided Fig. 5 . MHC class II expression on microglia and astrocytes. CNS cells from uninfected and infected mice were analyzed for MHC class II expression on microglia and astrocytes as depicted in Fig. 4 . Histograms show kinetics of class II expression on microglia (A; R1 gate) and astrocytes (B; R2 gate). Rat IgG2b was used as isotype control Ab for setting thresholds for positive staining. Numbers within histograms represent percentage of class II 1 glia; numbers in brackets depict MFI of the class II 1 population. Plots are representative of 3-4 experiments. Fig. 6 . MHC class I and class II expression on IFN-g treated astrocytes. Mixed glial cultures from neonatal FVB/N mice were analyzed for CD45 and class I (top panel) and class II (bottom panel) expression after treatment with or without 1 ng/mL IFN-g for 48 h. Cells were gated on microglia (CD45 1 ; left panel) or astrocytes (CD45 2 ; right panel). Microglia comprised 30% of the culture. Microglia MHC expression on untreated cells is indicated by solid gray lines and on IFN-g treated cells by solid black lines. Mouse IgG2a and Rat IgG2b, respectively, were used as isotype control Ab (dashed lines); control staining overlaps in untreated and treated cells. Numbers within histograms represent percentage of MHC 1 glia to the right (IFN-g treated) or above (untreated) the respective population; numbers in brackets reveal MFI in the MHC 1 population.

with the robust up-regulation of class I on astrocytes (Fig. 4A ) and class II on microglia at day 5 p.i. (Fig. 5A) . The high level of IFN-g present at day 5 p.i. suggests that NK and T cells secreted abundant IFN-g per cell, as T cell numbers were comparable to NK cell numbers at day 5 p.i., but significantly lower than T cells at day 7 p.i. Although a contribution by NK cells cannot be ruled out, the high virus antigen load, in addition to robust MHC molecule surface expression at day 5 p.i., supports the concept that MHC molecule-TCR interactions are responsible for IFN-g release (Slifka et al., 1999; Liu and Whitton, 2005) . Similarly, the abrupt decline in IFN-g at day 10 p.i., coincident with the control of virus replication, may reflect limited TCR engagement by MHC molecules presenting viral antigen, despite sustained MHC expression. Alternatively, upregulation of inhibitory ligands may block sustained IFN-g expression (Barber et al., 2006; Suvas et al., 2006) .

The contribution of astrocytes to antigen presentation during inflammatory responses in the adult CNS is controversial. Although both MHC class I and class II molecule expression is readily induced in astrocytes cultured from neonates, MHC surface expression on astrocytes in the inflamed CNS has been difficult to confirm in adults. In the present study a neurotropic coronavirus infection was used to induce CNS inflammation in mice expressing GFP under the GFAP promoter, thus facilitating analysis of astrocytes by flow cytometry. The virus variant used replicates prominently in microglia/macrophages, astrocytes and oligodendrocytes, but rarely in neurons, and induces a lymphocyte dependent sublethal demyelinating disease . Whereas MHC class I up-regulation was readily demonstrable on astrocytes following infection, class II expression was tentative at best, despite strong expression on microglia. Delayed up-regulation of MHC class I on astrocytes compared with microglia further suggested that expression by astrocytes is more constrained compared with microglia. Early class I up-regulation on microglia following virus infection is consistent with IFN-g independent induction and is likely mediated by virus induced type I interferons (Rempel et al., 2004; Zhou et al., 2005) . The apparent delay of astrocytes to up-regulate class I expression in vivo may be due to an inherent unresponsiveness to IFNa/b; however, virus infection itself and type I interferons enhance class I expression on cultured, neonatal astrocytes (Kraus et al., 1992; Massa et al., 1986) . The robust up-regulation of class I on astrocytes and both class I and class II on microglia coincident with maximal IFNg detection implicates IFN-g as a driving force for antigen presentation. T cells are the most likely source of IFN-g, as the absence of NK cells does not alter MHC molecule expression patterns following JHMV infection in C56BL/ 6 mice (Zuo et al., 2006) . The increased CNS infiltration of CD4 1 T cells, compared with CD8 1 T cells, in response to JHMV infection is unique to infected FVB/N mice and suggests that CD4 1 T cells may make a prominent contribution to IFN-g secretion in the local environment. However, neither class I or class II restricted viral epitopes have been mapped in this haplotype, preempting experimental verification.

Class I expression on astrocytes is consistent with a perforin mediated mechanism of viral clearance from this cell type Lin et al., 1997; Parra et al., 1999) and the capacity of astrocytes to perpetuate class I restricted CD8 1 T cells (Sedgwick et al., 1991) . Interactions of class I expressing astrocytes with CD8 1 T cells in vivo are further supported by selective destruction of a subset of the astrocytes expressing a neo-autoantigen following transfer of autoantigen specific TCR transgenic CD8 1 T cells (Cabarrocas et al., 2003) and by immunological synapse formation between virally infected astrocytes and CD8 1 T cells (Barcia et al., 2006) . The inability to detect class I on astrocytes following infection with a related neurotropic JHMV variant, which infects both neurons and glial cells may reside in limitations imposed by histological analysis, although class I is detected on oligodendrocytes, microglia, and neurons (Redwine et al., 2001) . Therefore, the difference in astrocyte responsiveness may also reflect a unique cytokine milieu induced by each virus. Neurotropic coronavirus variants induce both differential kinetics and magnitude of type I and type II IFN mRNA (Parra et al., 1997; Remple et al., 2004) , which may affect class I heavy chains as well as the genes involved in antigen processing and presentation (Strehl et al., 2005) . The kinetics of class I and class II up-regulation on microglia in JHMV infected FVB/N mice used in this study are similar to those in BALB/c and C57BL/6 mice Zuo et al., 2006) suggesting different genetic backgrounds do not suffice to explain these contradictory results.

The apparent impairment in astrocytes to express class II molecules during acute viral encephalomyelitis, despite the presence of abundant IFN-g, is reminiscent of the inability to detect class II in this cell type by immunohistological analysis (Horwitz et al., 1999; Redwine et al., 2001) . Similarly, class II expression was essentially absent from astrocytes in EAE lesions of wt mice (Stuve et al., 2002) . CIITA is essential for class II gene transcription in astrocytes (LeibundGut-Landmann et al., 2004; Waldburger et al., 2001) . However, although transgene driven constitutive CIITA results in class II expression on astrocytes in non-inflamed CNS tissue, only a very limited number of class II expressing astrocytes are present in EAE lesions in these mice (Stuve et al., 2002) . Similarly, although class II positive astrocytes have been detected in lesions during relapsing EAE (Sakai et al., 1986) , the paucity of class II up-regulation on astrocytes compared with microglia, demonstrates in vivo cell type specific regulation. These data support that CIITA expression is selectively suppressed in astrocytes during acute inflammation, even in the presence of IFN-g. Transcriptional silencing may be due to epigenetic regulation or active suppression (Gresser et al., 2000; Landmann et al., 2001; Massa, 1993; Tooze et al., 2006) . CIITA is regulated by numerous factors including IFN-g, TGF-b, IL-1, IL-4, and IL-10 (Dong and Benveniste, 2001; Lei-bundGut-Landman et al., 2004; Rohn et al., 1996) . While ineffective IFN-g signaling by astrocytes cannot be excluded, prominent, yet delayed, class I expression suggests this pathway is intact. Furthermore, JHMV inflammation primarily constitutes a Th1-type response, as indicated by high IFN-g protein levels and low IL-4 and IL-10 gene transcription (Parra et al., 1997) negating a role for anti-inflammatory cytokines. In vitro, other mediators inhibiting class II expression on astrocytes include TGF-b, IL-1b, IFN-b, glutamate, vasoactive intestinal peptide, norepinephrine, nitric oxide, and direct contact with neurons (Dong et al., 2001 ). An interesting candidate may be IFN-b which is rapidly induced following JHMV infection (Zhou et al., 2005) , and acts downstream of CIITA transcription to block class II expression (Lu et al., 1995) .

In summary the data demonstrate MHC class I up-regulation, and the absence of detectable class II expression, on the vast majority of astrocytes during virus induced inflammation. More stringent class I regulation by astrocytes, compared with microglia, may reside in reduced type I IFN and/or IFN-g responsiveness or additional factors known to enhance IFN-g mediated effects such as TNF-a. Although class I and class II expression correlate with IFN-g levels, maintenance of class I surface expression on astrocytes and microglia, even after clearance of infectious virus, suggest that abrogation of IFN-g secretion does not lead to immediate MHC molecule down regulation. Similarly, sustained class II expression on microglia suggests a low turnover rate of MHC molecules. Sustained MHC expression may provide an opportunity to activate autoantigen specific T cells. The expression of GFP in astrocytes enables future enrichment of this cell type for comparative gene profiling and additional insights into the regulation of antigen presentation by both distinct glial subsets and at distinct stages of inflammation.

Glia-T cell dialogue

Restoring function in exhausted CD8 T cells during chronic viral infection

In vivo mature immunological synapses forming SMACs mediate clearance of virally infected astrocytes from the brain

Inverted immunodominance and impaired cytolytic function of CD8 1 T cells during viral persistence in the central nervous system

Coronavirus infection of the CNS: Host-virus standoff

Perforin-mediated effector function within the central nervous system requires IFN-g-mediated MHC up-regulation

Effective and selective immune surveillance of the brain by MHC class I-restricted cytotoxic T lymphocytes

Immune function of astrocytes

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

Pathogenicity of antigenic variants of murine coronavirus JHM selected with monoclonal antibodies

Coronavirus induction of class I major histocompatibility complex expression in murine astrocytes is virus strain specific

Expression of a dominant negative IFN-g receptor on mouse oligodendrocytes

Tumor necrosis factor a and interleukin-1 a inhibit through different pathways interferon-g-induced antigen presentation, processing and MHC class II surface expression on astrocytes, but not on microglia

Basic principles of immunological surveillance of the normal central nervous system

Astrocyte expression of a dominant-negative IFN-g receptor

Detailed in vivo analysis of interferon-g induced major histocompatibility complex expression in the the central nervous system: Astrocytes fail to express major histocompatibility complex class I, II molecules

Augmentation of major histocompatibility complex class I, ICAM-1 expression on glial cells following measles virus infection: Evidence for the role of type-1 interferon

Interpreting MHC class I expression and class I/ class II reciprocity in the CNS: Reconciling divergent findings

Maturation of dendritic cells is accompanied by rapid transcriptional silencing of class II transactivator (CIITA) expression

Persistence of mouse hepatitis virus A59 RNA in a slow virus demyelinating infection in mice as detected by in situ hybridization

Induction of MHC class I antigens on glial cells is dependent on persistent mouse hepatitis virus infection

Mini-review: Specificity and expression of CIITA, the master regulator of MHC class II genes

Mouse hepatitis virus is cleared from the central nervous systems of mice lacking perforinmediated cytolysis

Cutting edge: Re-evaluating the in vivo cytokine responses of CD8 1 T cells during primary and secondary viral infections

Interferon (IFN) b acts downstream of IFN-ginduced class II transactivator messenger RNA accumulation to block major histocompatibility complex class II gene expression and requires the 48-kD DNA-binding protein, ISGF3-g

Specific suppression of major histocompatibility complex class I, class II genes in astrocytes by brain-enriched gangliosides

Viral particles induce Ia antigen expression on astrocytes

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

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

Kinetics of cytokine mRNA expression in the CNS following lethal and sublethal coronavirus-induced encephalomyelitis

IFN-g is required for viral clearance from central nervous system oligodendroglia

The astrocyte is a target cell in mice persistently infected with mouse hepatitis virus

Vaccine induced memory CD8 1 T cells cannot prevent virus reactivation

Control of central nervous system viral persistence by neutralizing antibody

Mechanisms of central nervous system viral persistence: The critical role of antibody and B cells

In vivo expression of major histocompatibility complex molecules on oligodendrocytes and neurons during viral infection

Differential regulation of innate and adaptive immune responses in viral encephalitis

Regulation of class II MHC expression

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

Antigen presentation in the central nervous system

Major histocompatibility complex-expressing nonhematopoietic astroglial cells prime only CD8 1 T lymphocytes: Astroglial cells as perpetuators but not initiators of CD4 1 T cell responses in the central nervous system

Rapid on/off cycling of cytokine production by virus-specific CD8 1 T cells

Interferon-g, the functional plasticity of the ubiquitin-proteasome system, and MHC class I antigen processing

The role of the MHC class II transactivator in class II expression and antigen presentation by astrocytes and in susceptibility to central nervous system autoimmune disease

Qa-1b and CD94-NKG2a interaction regulate cytolytic activity of herpes simplex virus-specific memory CD8 1 T cells in the latently infected trigeminal ganglia

Coronavirus infection induces H-2 antigen expression on oligodendrocytes and astrocytes

Repression of IFN-g induction of class II transactivator: A role for PRDM1/Blimp-1 in regulation of cytokine signaling

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

Selective abrogation of major histocompatibility complex class II expression on extrahematopoietic cells in mice lacking promoter IV of the class II transactivator gene

Sequential infection of glial cells by the murine hepatitis virus JHM strain (MHV-4) leads to a characteristic distribution of demyelination

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

Interferon-g induces the expression of H-2 and Ia antigens on brain cells

Immune regulation within the central nervous system

Expression of matrixmetalloproteinases and their tissue inhibitor during viral encephalitis

Live astrocytes visualized by green fluorescent protein in transgenic mice

Mouse hepatitis virus pathogenesis in the central nervous system is independent of IL-15 and natural killer cells

Authors thank Emanuel Dimacali and the USC Norris Cancer Center Flow Cytometry Core for assistance with flow cytometry. Authors also thank Wen Wei and Ernesto Baron for assistance with the histopathology and Eva Boruka for maintenance of the homozygous FVB/N-Tg[GFAPGFP]14Mes/J colony.