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

Abstract

Neurotropic strains of mouse hepatitis viruses (MHV) such as MHV-A59 (A59) and MHV-4 (JHMV) cause acute and chronic encephalomyelitis and demyelination in susceptible strains of mice and rats. They are widely used as models of human demyelinating diseases such as multiple sclerosis (MS), in which immune mechanisms are thought to participate in the development of lesions in the central nervous system (CNS). The effects of MHV infection on target cell functions in the CNS are not well understood, but A59 has been shown to induce the expression of MHC class I molecules in glial cells after in vivo and in vitro infection. Changes in class I expression in infected cells may contribute to the immunopathogenesis of MHV infection in the CNS. In this communication, a large panel of MHV strains was tested for their ability to stimulate class I expression in primary astrocytes in vitro. The data show that the more hepatotropic strains, such as MHV-A59, MHV-1, MHV-2, MHV-3, MHV-D, MHV-K, and MHV-NuU, were potent inducers of class I expression in astrocytes during acute infection, measured by radioimmunoassay. The Kb molecule was preferentially expressed over Db. By contrast, JHMV and several viral strains derived from it did not stimulate the expression of class I molecules. Assays of virus infectivity indicated that the class I-inducing activity did not correlate with the ability of the individual viral strain to replicate in astrocytes. However, exposure of the viruses or the supernatants from infected astrocytes to ultraviolet light abolished the class I-inducing activity, indicating that infectious virus is required for class I expression. These data also suggest that class I expression was induced directly by virus infection, and not by the secretion of a soluble substance into the medium by infected astrocytes. Finally, analyses of A59/JHMV recombinant viral strains suggest that class I-inducing activity resides in one of the A59 structural genes.

Full text

Page 1

Coronavirus Induction of Class I Major
Histocompatibility Complex Expression in Murine
Astrocytes Is Virus Strain Specific
By Wendy Gilmore,* Jorge Correale,* and Leslie P. Weiner*r
From the Departments of *Neurology and $Microbiology, University of Southern California
School of Mec~ine, Los Angeles, California 90033
Summary
Neurotropic strains of mouse hepatitis viruses (MHV) such as MHV-A59 (A59) and MHV-4
(JHMV) cause acute and chronic encephalomyelitis and demyelination in susceptible strains of
mice and rats. They are widely used as models of human demyelinating diseases such as multiple
sclerosis (MS), in which immune mechanisms are thought to participate in the development
of lesions in the central nervous system (CNS). The effects of MHV infection on target cell
functions in the CNS are not well understood, but A59 has been shown to induce the expression
of MHC class I molecules in glial cells after in vivo and in vitro infection. Changes in class
I expression in infected cells may contribute to the immunopathogenesis of MHV infection in
the CNS. In this communication, a large panel of MHV strains was tested for their ability to
stimulate class I expression in primary astrocytes in vitro. The data show that the more hepatotropic
strains, such as MHV-A59, MHV-1, MHV-2, MHV-3, MHV-D, MHV-K, and MHV-NuU,
were potent inducers of class I expression in astrocytes during acute infection, measured by
radioimmunoassay. The K b molecule was preferentially expressed over D b. By contrast, JHMV
and several viral strains derived from it did not stimulate the expression of class I molecules.
Assays of virus infectivity indicated that the class I-inducing activity did not correlate with the
ability of the individual viral strain to replicate in astrocytes. However, exposure of the viruses
or the supernatants from infected astrocytes to ultraviolet light abolished the class I-inducing
activity, indicating that infectious virus is required for class I expression. These data also suggest
that class I expression was induced directly by virus infection, and not by the secretion of a
soluble substance into the medium by infected astrocytes. Finally, analyses of A59/JHMV
recombinant viral strains suggest that class I-inducing activity resides in one of the A59 structural
genes.
M
ical, and respiratory diseases in a wide variety of mammalian
species (1-7). Although some MHV strains, such as A59 and
MHV-3, can cause both gastrointestinal and neurological dis-
ease in mice and rats, many tend to induce pathology that
is restricted to either system, prompting their classification
as hepatotropic or neurotropic. There has been considerable
interest in the study of the more neurotropic strains of MHV,
such as JHMV and A59, because of their ability to produce
demyelinating lesions that resemble the demyelinating plaques
observed in the human neurological disease, multiple scle-
rosis (MS). In addition, MHV and human coronavirus iso-
ouse hepatitis viruses (MHV) 1 are members of the
Coronaviridae, which cause gastrointestinal, neurolog-
1 Abbreviations used in thispaper: CNS, central nervous system; GFAP, glial
fibrillary acidic protein; i.c., intracerebral; M, membrane; MHV, mouse
hepatitis virus; m.o.i., multiplicity of infection; MS, multiple sclerosis;
N, nucleocapsid; p.i., post infection; S, spike; VSV, vesicular stomatitis virus.
1013
lates are capable of inducing demyelination in primates (8),
and coronavirus RNA and antigen have been detected in
demyelinating lesions in the brains of MS patients (9, 10).
As Coronaviridae, all MHV strains are enveloped viruses
that contain a single-stranded, positive-sense ILNA genome
of 31-kb that is expressed as 7 or 8 mRNAs encoding both
structural and nonstructural proteins (11). The structural pro-
teins have been well characterized, and include the nucleo-
capsid (N) protein, which interacts with the viral RNA, and
two envelope glycoproteins, the spike (S) and membrane (M)
proteins. The S protein forms the virion surface spikes and
is responsible for binding the cellular MHV receptor, inducing
cell-to-cell fusion and providing a target for neutralizing an-
tibodies. It is the least conserved of the structural proteins
among the MHV strains (12-16). M interacts with the N
protein-R.NA comply, and may be involved in virus assembly.
It is the most conserved structural protein among the MHV.
A third envelope glycoprotein, hemagglutinin-esterase (HE),
has been identified in MHV-S and several JHMV isolates,
J. Exp. Med. y The Rockefeller University Press i 0022-1007/94/09/1013/11 $2.00
Volume 180 September 1994 1013-1023

Page 2

and may participate in the pathogenesis of JHMV infection
in the central nervous system (CNS) (11, 17). Nonstructural
proteins include an RNA-dependent RNA polymerase, and
three additional proteins that do not appear to be required
for viral replication, but whose functions and biological ac-
tivities have not been identified.
After intracerebral (i.c.) inoculation in mice and rats, the
parent strains of both JHMV and A59 cause an acute en-
cephalomyelitis and demyelination from which a varying
number of animals survive to exhibit chronic demyelination
(1-7, 11). The severity of the disease varies with the individual
virus strain and with the age, genetic background, and im-
mune status of the infected host. The neuropathological fea-
tures of acute infection can be somewhat distinguished from
those of chronic disease by a prevalence of gray matter in-
volvement in which virus replicates in neurons, oligodendro-
cytes, and astrocytes to cause extensive CNS damage (18-21).
In chronic disease, lesions are more prevalent in the white
matter and consist of primary demyelination with axonal
sparing (2, 11). Infectious virus is rarely isolated during chronic
JHMV disease, though viral RNA can be detected by reverse
transcriptase (RT)-PCR in the brains of infected mice as late
as 2 yr post infection (p.i., Fleming, J., personal communi-
cation). Viral antigen and RNA tend to be restricted to as-
trocytes in chronic infection (2, 22-25), though there is evi-
dence that neurons are also affected (25). This pattern of white
matter involvement is also characteristic of infection with at-
tenuated or mutant strains of A59 (26, 27) or JHMV (25,
27-29), which cause demyelination in the relative absence
of encephalitis. In vitro, JHMV and A59 cause a lytic infec-
tion in primary oligodendrocyte cultures that is rapidly self-
limiting, but both virus strains can establish a productive and
relatively nonlytic, acute, and long-term chronic infection in
mixed glial cell cultures and in cultures enriched for astro-
cytes (30-32). These data support early reports that acute
demyelination is the result of virus-mediated oligodendro-
cyte death (2, 11), but more importantly, they also suggest
that the astrocyte plays an important role in acute encepha-
litis and chronic demyelination.
Currently, there is very little known about the ability of
MHV strains to influence specific activities of the host cells
that they infect. Early reports indicate that A59 infection en-
hances the expression of MHC class I molecules in primary
cultures of murine glial cells (33) and that the enhancement
also occurs in the brain after i.c. infection in C57B1/6 mice
(34). Since class I molecules play a key role in the interaction
between infected target cells and CD8 + cytotoxic T cells
(CTL; 35, 36), any change in their expression after virus in-
fection has potential implications for the outcome of the in-
fection. Cells in the CNS generally express little, if any, con-
stitutive class I antigen in vivo, but astrocytes have been
consistently reported to express both class I and class II an-
tigens in vitro after the addition of IFN-3' and/or TNF-ot
to the culture medium (37-42). Several reports indicate that
astrocytes are capable of acting as antigen-presenting cells for
in vitro antigen- or a11ospecific CD4 § and CD8 § T cell re-
sponses (43, 44), which suggests that they also have poten-
tial to participate in immune responses occurring within the
confines of the CNS. Since clearance of infectious JHMV from
the CNS requires class I-restricted CD8 + T cells (11,
45-47), it is of particular interest to characterize the effect
of MHV infection on class I expression in one of its principal
cellular targets.
In this communication, we have examined the ability of
a panel of MHV strains to induce class I expression in cul-
tures enriched at least 95% for astrocytes. The data indicate
that class I expression occurred in response to all the MHV
strains tested except JHMV and virus strains derived from
JHMV. Additional studies show that class I expression re-
quires the presence of infectious virus. Finally, the testing
of A59/JHMV recombinant viral strains suggests that the
class I-inducing activity resides in the 3' end of the A59 ge-
nome, possibly in one of the genes encoding the structural
proteins.
Materials and Methods
Primary Astrocyte Cultures. Astrocytes were isolated from mixed
glial cell cultures prepared from the brains of newborn C57B1/6
mice (Bantin and Kingman, Fremont, CA) at postnatal day 0-3
according to McCarthy and deVellis (48). Briefly, single cell sus-
pensions were prepared from cerebri dissected free of brain stems
and cerebelli, plated at 3-5 brains per T-75 flask and allowed to
grow to confluence at 12-15 d in vitro. Culture medium consisted
of DMEM/Ham's F12 (1:1; JRH Biosciences, Lenexa, KS) sup-
plemented with 10% FCS (Gemini Bioproducts, Inc., Calabasas,
CA), 15 mM Hepes, 2.5 mM t-glutamine, and penicillin/strep-
tomycin (100 U/ml-100 ~g/ml). At confluence the cultures were
mechanically shaken to dislodge microglia and oligodendroglia,
resulting in preparations enriched 95% or greater for cells staining
for glial fibrillary acidic protein (GFAP). Immunoperoxidase or im-
munofluorescent staining (described below) revealed that the cell
preparations contained ,-2-4% of cells of microglia/macrophage
lineage, expressing F4/80, T-200, the mouse equivalent of human
common leukocyte antigen (CD45), and/or Mac-1 surface markers.
Coronavirus Strains and Infection. Table 1 presents a summary
of the MHV strains used and the principal types of diseases they
cause in mice. The derivation, propagation, and sources of MHV-
A-59, JHM-DL, JHM-DS, MHV-1, MHV-2, MHV-3, MHV-D,
MHV-K, and MHV-Nuu have been described (12). The neutrali-
zation-resistant MHV-4 strains 2.2-V-1 and 2.2/7.2-V-2 (28, 29) were
the kind gift of Dr. John Fleming (University of Wisconsin,
Madison, WI). The development and properties of JHM-X and
the JHM/A59 recombinant MHV strains have been reviewed (49).
All virus strains were propagated using the murine astrocytoma
DBT as previously described (6). Virus titers were determined for
each virus preparation using DBT as indicator cells.
Infectious Center Assays. The number of infected astrocytes was
determined on day 3 p.i. to provide a measure of the relative efficiency
of infection by several MHV strains. Briefly, astrocytes were tryp-
sinized into single cell suspensions and, after washing, were plated
on DBT cell monolayers at 0.1, 1, 10, 100, and 1,000 cells/60 mm
petri dish. Cells were allowed to attach for 1 h at 37~ before the
addition of agarose at 0.6% in IkPMI 1640 supplemented with
2% heat-inactivated FCS, 20 mM Hepes, and penicillin/strep-
tomycin. Plaques were counted after 48 h incubation at 37~
Virus Inactivation by Exposure to Ultraviolet Light. Inactivation
of virus was accomplished by exposure to UV light under a UVP
transilluminator (UVP Inc.,; San Gabriel, CA) at 7 mW/cm Z for
1014 Coronavirus Induction of Class I MHC Expression Is Virus Strain Specific

Page 3

Table 1. MHV Strains and Their Characteristics
Virus strain Characteristics
JHM-DL (MHV-4)
JHM-DS (MHV-4)
JHM-X (MHV-4)
2.2-V-1 (MHV-4)
2.2/7.2-V-2 (MHV-4)
MHV-A59
MHV-NuU
MHV-K
MHV-D
MHV-1
MHV-2
MHV-3
Lethal encephalitis, demyelination
Demyelination, encephalitis
Lethal encephalitis, demyelination
Nonlethal demyelination
Nonpathogenic CNS infection
Lethal encephalitis, demyelination,
hepatitis
Nonpathogenic enteric infection
Nonlethal myeloproliferative disease,
hepatitis
Nonlethal hepatitis, enteritis
Nonlethal mild hepatitis
Lethal hepatitis
Lethal hepatitis, choroidoependymitis
30 min. Virus inactivation was confirmed as lack of virus infec-
tivity in plaque assays using DBT cells.
Antibodies. To identify class I molecules in primary astrocyte
cultures, mAbs specific for K b (AF6.88.5.3), DbL d (28-14-8S), K d
(SF1-1.1.1), and D e (34-5-8S) were obtained from the American
Type Culture Collection (ATCC HB 158, HB 27, HB 159, and
FIB 102, respectively; Rockville, MD) and used as tissue culture
supernatants. Optimal antibody concentrations were determined
by ILIA or FACS | (Becton Dickinson & Co., Mountain View, CA)
analyses in preliminary experiments using either astrocytes or lym-
phocytes from the appropriate haplotype. The percentage of GFAP-
positive astrocytes was determined in the enriched cultures using
polyclonal anti-GFAP antibody (rabbit antibovine GFAP; Dakopatts,
Glostrup, Denmark). Cells of macrophage/monocyte lineage were
identified using a mixture of mAb specific for Mac-1 (hybridoma
M1/70.15, ATCC TIB 128), F4/80 (ATCC HB 198), and T-200,
which recognizes all cells of bone marrow origin (hybridoma
M1/9.3.4HL.2, ATCC TIB 122). The possible presence of oligoden-
drocytes was identified using polyclonal rabbit antigalactocerebro-
side (Gal C, a gift from M. Smith, Stanford University, Stanford,
CA). Finally, viral antigen was identified using J.3.3, a mAb specific
for the N protein of JHMV that crossreacts with all of the MHV
strains used in this study (12).
RtA.
The expression of class I molecules was measured in as-
trocytes cultured in flat-bottomed 96-well plates at a density of 10 4
cells/well and infected with various virus strains at a multiplicity
of infection (m.o.i.) of 1-2. Mock infected cells served as controls.
On day 3 or 5 p.i., cells were washed twice using a wash buffer
of 0.3% BSA in 0.1 M PBS before the addition of 50/A of the
appropriate mAbs in triplicate. After a 60 rain incubation at room
temperature, cells were washed three times, followed by the addi-
tion of 20,000 cpm/well of 12SI-labeled protein A (30 /~Ci//zg;
ICN Biomedicals, Costa Mesa, CA). At the end of a second 60-
min incubation, cells were extensively washed in PBS to remove
unbound radiolabeled protein A and detached with 0.5% trypsin-
0.2% EDTA (JRH Biosciences). Data are presented as cpm bound
radioactivity + SD, or percent increase in class I expression in in-
fected cells compared with that in uninfected controls, corrected
for background binding in the absence of antibody. A result was
considered positive when the percent increase in expression was
100% or greater. Nonspecific staining was identified by the inclu-
sion of Kd-specific mAb SF1-1.1.1. In some experiments, superna-
tants from infected cells were substituted for virus and RIA per-
formed to detect class I 3 d later.
Immunoperoxidase and Immunofluorescent Staining. Cell pheno-
types and the number of infected ceUs were determined by im-
munoperoxidase staining in astrocytes cultured in tissue culture
chamber slides. Avidin-biotin immunoperoxidase staining kits (Vec-
tastain; Vector Laboratories Inc., Burlingame, CA) were used ac-
cording to the manufacturer's instructions as previously described
(21). Cells were infected with virus 1 d after plating and fixed on
day 3 or 5 p.i. in acetone/methanol (1:1). The number of positively
stained cells in each well was determined in three fields/well at a
magnification of 20x and reported as the percentage of the number
of total cells in the same fields. Background staining was deter-
mined in cells stained in the absence of primary antibody.
Class I expression was also evaluated by FACS | using an indirect
immunoftuorescent staining procedure. Briefly, single cell suspen-
sions of astrocytes were adjusted to 105 to 106 cells/tube before the
addition of class 1-specific primary antibodies. FITC-labeled goat
anti-mouse IgG F(ab)2 was used as secondary antibody (Jackson
ImmunoResearch Laboratories, West Grove, PA). Fluorescence data
were collected on 3 x 103 to 104 viable cells, indicated by forward
light scatter intensity, using a FACStar cell sorter (Becton Dick-
inson & Co.). Again, background fluorescence was determined in
cells stained in the absence of primary antibody.
Detection oflnterferons. The possible presence oflFN in the su-
pernatants of infected astrocytes was determined by bioassay in a
vesicular stomatitis virus (VSV) neutralization assay using vital dye
uptake as a spectrophotometrically measured endpoint in L929 cells
(50, 51). Supernatants were collected from infected and control unin-
fected astrocytes on day 3 and 7 p.i., UV irradiated, and added in
triplicate to L929 cells plated at 10 s cells/well in 96-well plates.
After a 24-h incubation at 37~ and 7% CO2, supernatants were
aspirated and cells infected with VSV (Indiana strain) at approxi-
mately 10 tissue culture infectious doses (TCIDs0)/well. Control
wells included uninfected L929 cells (cell control) and VSV-infected
L929 cells cultured in the absence of interferon (virus control). At
24-36 h p.i., when CPE was maximum in the control wells, cells
were stained for 2 h at 37~ with 0.004% neutral red. They were
then lysed to release retained dye using 0.01 M HC1 in 30% eth-
anol, and ODs measured at 540 nm in an automated microELISA
reader (Dynatech Laboratories, Inc., Chantilly, VA). IFN concen-
trations were deduced by comparison of sample ODs with those
of standard concentrations of recombinant mouse IFN-3' (Genzyme
Corp., Cambridge, MA) and murine IFN-c~//3 (Lee Biomolecular,
San Diego, CA).
Immunoneutralization was also used to identify the possible pres-
ence of IFN in UV-irradiated supernatants of A59qnfected astro-
cytes. For this purpose, rat monoclonal anti-mouse IFN-3' anti-
body (hybridoma R4-6A2; ATCC HB 170) or rabbit polyclonal
anti-mouse IFN-a/3 (Lee Biomolecular) was added to the astro-
cytes before the addition of A59, and class I expression measured
by RIA on day 3 p.i. Anti-IFN-3' (20-60/zg/ml) completely in-
hibited the protective activity of 10 U/ml of recombinant mouse
IFN-3', whereas 10-30 U/ml anti-IFN-a/r
tralizing 20 U/ml purified mouse IFN-c~//~.
Detection of TNE TNF activity was measured in a cytotox-
icity assay using actinomycin D-treated L929 fibroblasts as targets,
again with vital dye uptake as a spectrophotometric endpoint. Serial
2-fold dilutions of UV-irradiated supernatants, or 10-fold dilutions
effective in neu-
1015 Gilmore et al.

Page 4

of murine recombinant TNF-ot (Genzyme Corp.), as a standard
were added to L929 monolayers in 96-well plates in the presence
of actinomycin D (8 #g/ml; Sigma Chemical Co., St. Louis, MO).
After an 18-h incubation at 37~
added as described for the IFN assay, and OD read at 540 nm.
7% CO2, neutral red was
Results
MHV-A59, but Not JHM-DL, induces MHC Class I Ex-
pression in Primary Cultures of Murine Astrocytes. To inves-
tigate the effects of coronavirus infection on MHC class I
expression in primary astrocytes~ cultures were routinely used
at 20-30 d in vitro, or 6-12 d after mechanical shaking to
remove oligodendrocytes and microglia. Fig. 1 shows that
A59 induced abundant class I expression in astrocytes, mea-
sured by RIA using the Kb-specific mAb AF6.88.5.3. Class
I expression was routinely measured on days 3 and 5 p.i. after
infection at an m.o.i, of 1, though it was detectable within
24-48 h p.i. K b was preferentially expressed over D b after
A59 infection. This was not due to a lack of inducibility of
the D b molecule, since both K b and D b were expressed spon-
taneously with time in culture, and also in response to 100
U/ml of IFN-3/(data not shown). Under these conditions,
K b was expressed earlier than D b, and usually at significantly
higher levels. In infected astrocytes, the class I-inducing ac-
tivity of A59 was consistently 30-50% more potent than
the single dose of 100 U/ml of IFN-3' (Fig. 2). Finally, FACS |
analyses revealed that K b was expressed on '~50% of the as-
trocytes in the infected cultures.
By contrast, the large plaque morphology variant of JHMV
designated JHM-DL did not stimulate class I at all, or at
best, stimulated minimal expression at 10-20% over that ob-
served in uninfected ceUs (Fig. 1 and 2). These data indicate
that the ability to induce class I expression is not a general
property of MHV strains.
Figure 1. MHC class I is induced in primary murine astrocytes by A59,
but not JHM-DL. Class I expression was measured on days 3 and 5 p.i.
using mAb specific for K b, D b, and D a as described in Materials and
Methods. Data are represented as mean cpm _+ SD of triplicate determi-
nations.
Figure 2. MHC class I (K b) induction in primary astrocytes by A59
relative to that of IFN-% Recombinant IFN-3' was added to astrocytes
for 48 h at 100 U/m1 before R.IA at days 3 and 5 p.i. Data represent mean
cpm _+ SD of triplicate determinations.
Virus Strains Related to or Derived from JHM-DL Do Not
Stimulate Class I Expression. To determine whether the lack
of ability to stimulate class I expression in astrocytes is specific
for JHMV, a panel of virus strains derived from wild-type
JHMV or from JHM-DL were tested. The results indicate
that none of the JHMV strains isolated in our laboratories
stimulated class I (Table 2). In addition, JHM-X, which was
derived from JHMV passage in Japan and has been shown
to have extensive deletions in the S gene (11), did not affect
dass I expression. By contrast, all of the viral strains in a
panel of more hepatotropic isolates of MHV, including MHV-
NuU, MHV-K, MHV-D, MHV-1, MHV-2, and MHV-3,
were effective stimulators of class I expression. The magni-
tude of stimulation, expressed as percent change relative to
uninfected cells, varied with the individual viral strain 2-10-
fold over that in uninfected cells. A59 and MHV-2 were the
most potent inducers of the class I K b molecule, and both
inconsistently induced a low-level expression of D b. It
should be mentioned that the Kd-specific mAb SF1-1.1.1 oc-
casionally yielded a low level of nonspecific binding; how-
ever, it was significantly lower than the specific binding ob-
served with the Kb-specific antibody.
Studies with Recombinant A59/JHMV Strains Map the Genes
Responsible_for Class I Expression to the 3' End of the Genom~ To
determine whether a specific A59 gene or genes are respon-
sible for the class I-inducing activity, a panel of recombinant
viruses between A59 and JHMV that retain various portions
of A59 sequences were tested. The results, represented in Fig.
3 and summarized in Fig. 4, indicate that all strains that re-
tain A59 sequences at the 3' end of the recombinant genome
were potent class I inducers. Thus, CA13 and CA43, which
retain •30-40% JHMV sequences at the 3' end, were com-
pletely devoid of class I-inducing activity. This end of the
MHV genome contains the genes encoding the structural
1016 Coronavirus Induction of Class I MHC Expression Is Virus Strain Specific

Page 5

Table 2. MHV Strains Differ in Their Ability to Induce MHC Class I Expression in Murine Astrocytes*
Experiment Virus strain K b D b Dd/K d
1 JHM-DL
JHM-DS
JHM-X
2.2-V-1
2.2/7.2-V-2
A59
10
7
0
0
0
322
9 0/-~
0/-
0/-
0/-
0/-
0/-
10
25
0
0
0
2 JHM-DL
A59
A59H
MHV-2
7 29
110
0/-
0/-
0/-
0/-
958
1,042
822
0
0
3 A59
MHV-NuU
MHV-K
MHV-D
MHV-1
MHV-3
359
331
281
257
258
369
-
-
-
-
-
-
-/0
-/65
-/13
-/93
-/0
-/40
4 JHM-DL
A59H
MHV-2
17
786
955
11
8
94
0/-
0/-
0/-
* MHC Class I expression was measured by RIA on day 3 p.i. (m.o.i. = 1) using mAbs AF6.88.5.3 (Kb), 28-14-8S (Db), 34-5-8S (Dd), and SFI-I.I.1
(Kd). Values are expressed as percent change relative to uninfected ceUs stained with the indicated mAb. Positive results were defined as an increase
in class I expression greater than or equal to 100% over that observed in uninfected cells. Descriptions of assay procedures are included in Materials
and Methods.
* (-) Not tested.
proteins S, M, and N in addition to two nonstructural pro-
teins, which suggests that the class I-inducing activity re-
sides in one of these proteins. Often, though not exclusively,
the most potent inducers of class I were strains that retain
the highest percentage of A59 sequences at the 3' end, while
those retaining the least A59 character, or exhibiting more
than one crossover site for recombination, were significantly
less potent. In addition, it was observed in three out of four
experiments that RL1 was •50% less potent than EL3, as
illustrated in Fig. 3 B. EL3 retains slightly more A59 sequences
in the S gene than RL1. This observation suggests that class
I-inducing activity may involve S. However, it was not pos-
sible to attribute the class I-stimulating activity to an in-
dividual gene or gene product, since none of the recombinants
had crossover sites that sufficiently isolated individual A59
from JHMV genes. However, it was clear that ifJHMV se-
quences were retained at the 3' end, class 1-inducing activity
was abolished, and that 5' retention of A59 sequences did
not salvage it.
The Ability_ of MHV to Induce Class I Is Not Dependent on
Replication Efficiency. In our laboratory, A59 grows to higher
titer in vitro than the JHMV strains in DBT cells, often ex-
ceeding JHMV growth 10-20-fold. In addition, recombinant
viruses containing A59 leader predominate over those con-
taining JHMV leader, suggesting that A59 leader provides
a growth advantage (52). Thus, it was important to deter-
mine whether the lack of class I inducing activity by JHMV
was due to poor replication efficiency in primary astrocytes.
For this purpose, the number of astrocytes staining for viral
antigen was determined by immunoperoxidase staining, and
the yield of infectious viral particles/cell was measured in in-
fectious center assays. The data, presented in Table 3, indi-
cate that on day 3 p.i., 53% of the JHM-DL-infected astro-
cytes were positive for viral antigen, compared with 45%
of cells infected with A59. In addition, cells infected with
the recombinant virus CA13, which retains 3' JHMV and
5' A59 sequences and does not stimulate class I expression,
showed 48% antigen positive staining. Cells infected with
CA13 produced the highest level of infectious virus at 0.51
plaques/cell, in spite of its inability to stimulate class I ex-
pression (Table 3). By contrast, EL3, which was an effective
class I inducer, replicated at relatively low efficiency at 0.08
plaques/cell. Thus, class 1-inducing activity was not a func-
tion of the ability of the virus to infect or replicate in pri-
1017 Gilmore et al.

Page 6

.E
E
20000
=1
i
10000
8
4000
3000
2OOO
1OOO
A
z -~ 1 uJ cc
a
mary astrocytes, nor did it appear to depend on the presence
of A59 leader.
Class 1-inducing Activity Is a Direct Consequence of A59 Infec-
tion.
It has been reported that class I induction in A59-
infected glial cell cultures is not a direct consequence of in-
fection, but is instead due to the release of a soluble factor
into the medium that requires continual virus production (33,
53, 54). The presence of a non-IFN-like soluble factor was
demonstrated in supernatants from A59-infected glial cells
that had been exposed to UV light to inactivate the virus.
Similarly, our data show that UV-inactivated A59 and A59-
like recombinant coronavirus strains are not able to induce
class I activity in purified astrocytes (Table 4). Virus inactiva-
tion in the UV-treated virus preparations was confirmed by
plaque assay using DBT cells (data not shown). However,
class I-inducing activity was not detected in the supernatants
collected on days 3 and 7 p.i. from A59-infected astrocytes
and exposed to UV light to inactivate the virus (Fig. 5), while
supernatants that were not exposed to UV light and thus,
contained infectious virus, were able to induce class I. Induc-
tion was not inhibited by the addition of antibodies specific
for IFN-'y or IFN-ot/~ to the astrocytes before A59 infec-
tion (Table 5). Finally, there was no evidence of the presence
of TNF-ot in UV-treated supernatants of A59-infected as-
trocytes (data not shown). These data suggest that class I
induction is a direct consequence of A59 infection itself, and
does not occur indirectly as a result of the release of a soluble
class I inducer into the astrocyte medium.
Figure 3. MHC class I (K b) induction in primary astrocytes infected
by A59 and recombinant A59/JHMV coronavirus strains. Data represent
mean _+ SD of triplicate determinations. Data are also included for unin-
fected astrocytes exposed to IFN-y (100 U/ml) for 48 h before assay on
day 3 p.i. A and B are two experiments that are representative of four
experiments yielding similar results.
Discussion
In this communication, we report that the ability of A59
to stimulate MHC class I expression during acute infection
in primary murine astrocytes, previously observed by Suzumura
et al. (33, 53, 54), is not an inherent property of MHV strains.
It is interesting to note that the strains that did not stimulate
class I are derivatives of the highly neurotropic MHV-4, or
JHMV strain, whereas most of the MHV strains that did
are the more hepatotropic strains. However, class I-inducing
Table 3. Efficiency of Infection of Coronavirus Strains at
3 dp.i.
Percent
cells infected*
No.
Virus strain plaques/cell*
JHM-DL
A59
B1
EL3
RL1
CA13
53
45
40
41
42
48
0.20
0.19
0.18
0.08
0.14
0.51
* Determined by immunoperoxidase staining using an antibody specific
for the nucleocapsid protein of JHMV.
Measured in infectious center assays as described in Materials and
Methods.
L
131
C1
EL3
RL1 III
ILfi I
RL2 I
IL27
CA13 I
CA43 I
Figure 4.
Class I-inducing ability is indicated for each strain.
I (Pol. protease) 2 ItN S 4 5 M N
I
I
Symbols: I::::s:~. A59; nmm. JHM
Class I
§
§
§
§
Genetic structure of the recombinant A59/JHMV strains used.
1018 Coronavirus Induction of Class I MHC Expression Is Virus Strain Specific

Page 7

Table 4.
Induction of MHC Class I by A59 and A59-1ike Recombinant Coronavirus Stains Requires Infectious Virus*
Experiment Virus strain K b D b D d
A59 301 -
-
0
7 A59 UV* 0
A59 363 0
2
0
0 A59 UV 8
A59
B1
EL3
4,091
2,544
2,007
1,610
0
0
0
0
0
0
0
0
0
0
0
0
0
0
R.L1
B1 UV
EL3 UV
RL1 UV
0
0
0
* MHC class I was measured by RIA on day 3 p.i. (m.o.i. = 1-2) using mAbs as defined in the legend for Fig. 1. Values are expressed as percent
change relative to uninfected cells stained with the appropriate mAb.
* UV refers to virus inactivated by exposure to ultraviolet light as described in Materials and Methods.
ability could not be exclusively attributed to selective tissue
tropism, since A59 and MHV-3, which readily establish both
CNS and hepatic infections, were potent inducers of class
I expression in astrocytes. Lack of class I-inducing activity
11000 j
9000
[] Kb
[] Db
8000
[] Kd
7000
E 6000
5000
8 4000
3000
2000
1000
Uninf
I b I
A59 Day 3 Day 7
Figure 5.
induced by UV-treated supernatants collected on days 3 and 7 p.i. from
MHV-A59 infected astrocytes. Supernatants were added to fresh astrocyte
cultures and incubated for 3 d before RIA. Controls included superna-
tants from MHV-A59-infected cells that had not been UV treated.
Nonspecific class I expression was measured using a Ka-specific mAb. Data
are expressed as mean cpm of triplicate determinations, with error bars
indicating SDs.
Relative to uninfected control cultures, MHC class I is not
1019 Gilmore et al.
in JHMV was also not due to poor replication efficiency in
astrocytes, since JHMV infected the same percentage of cells
and produced similar levels of infectious virus as the class
I-inducing MHV strains (Table 3). Thus, some other MHV
characteristic must be responsible for class I-inducing activity,
or the lack thereof.
Since the expression of class I genes is regulated by inter-
ferons (55), it seemed possible that class I induction might
Table 5.
Antibodies to IFNs*
Class I Induction by MHV-A59 Is Not Inhibited by
Virus
strain Experiment Antibody K b D a
1 JHM-DL
MHV-A59
-
-
- 22
280
0
0
MHV-A59
MHV-A59
IFN-3~*
IFN-a/3 s
230
335
0
0
2 JHM-DL
MHV-A59
MHV-A59
-
-
- 35
338
282
30
31
0 IFN-y II
MHV-A59 IFN-a/3 338 4
* Antibodies to IFNs were added before the addition of virus, and class
I expression measured by RIA on day 3 p.i. Values are expressed as per-
cent change relative to uninfected cells stained with the appropriate mAb.
* Anti-mouse IFN-3' added at 60/~g/ml.
S Anti-mouse IFN-cd3 added at 10 U/ml.
II Anti-mouse IFN-~/added at 20/~g/ml.

Page 8

correlate with the ability of MHV strains to stimulate IFN
production. This possibility gains credibility from reports
that JHMV is a notoriously poor inducer of interferon after
in vitro (56) and in vivo (57) infection. However, Suzumura
et al. (53) did not detect IFNs in UV-treated supernatants
from A59-infected mixed glial cells in spite of the presence
of class I-inducing activity, and our findings indicate that
A59-induced class I expression was not inhibited by antibodies
to IFN-a/3 or -3' (Table 4). Unlike Suzumura et al. (53),
we did not find that class I-inducing activity was retained
after UV-inactivation of A59 or the A59-1ike recombinant
MHV strains (Table 4), and it was not detected in superna-
rants from A59-infected astrocytes that were UV-treated to
inactivate infectious virus (Table 5 and Fig. 5). In addition,
we have recently found that class I expression is no longer
observed when A59 replication is controlled by the culture
of persistently infected astrocytes in the presence of MHV-
specific polyclonal antiserum, in spite of the persistence of
viral protein and RNA (58). Overall, our data suggest that
class I induction by MHV in astrocytes requires infectious
virus and does not involve the secretion of a soluble factor.
In this respect, it is interesting to note that the class I-in-
ducing activity originally reported by Suzumura et al. (33,
53) was also dependent on continuous virus production in
glial cell cultures prepared from the brains of infected mice
(54). However, it is possible that our lack of detection of
class I-inducing activity in the supernatants could reflect cul-
ture, or other technical conditions.
Although it was not possible to attribute class I-inducing
activity to a single gene or gene product using the A59/JHMV
recombinant virus strains, the data clearly indicate that class
I-inducing activity maps to the 3' end of the A59 genome,
which contains the genes encoding the primary structural
proteins S, M, and N, and two poorly understood nonstruc-
tural proteins. However, there are some indications that S
is the most likely candidate. For example, recombinant strains
which retain more A59 sequences in the S gene tended to
be more potent than those retaining less A59, or moreJHMV
character. In addition, the S gene shows the greatest anti-
genic divergence of the structural proteins between A59 and
JHMV (12-16) and so might be more likely to exhibit differ-
ences in activity. This possibility is supported by data indi-
cating that A59 andJHMV differ considerably in their ability
to cause receptor-independent fusion and syncytium forma-
tion (59). In addition, preliminary data suggest that A59 and
JHMV may show selective binding to the cellular receptors
for MHV (60), which are members of the carcinoembryonic
antigen (CEA) family of proteins (61).
The mechanism by which A59, MHV-3, and the A59-
like recombinant MHV strains induce class I expression in
astrocytes is currently unknown. Massa et al. (62, 63) have
shown that MHC class I promoter activity is upregulated
in astrocytes treated with IFN-y and that the upregulation
is associated with an increase in the binding activities of the
MHC class I regulatory element (MHC-CRE) and the IFN
consensus sequence (ICS). Whether or not these regulatory
elements are engaged by A59 infection remains to be deter-
mined.
Perhaps one of the more interesting findings in these studies
is the prevalence of K b expression over that of D b after MHV
infection, and to a lesser extent, IFN-3" treatment. Differen-
tial modulation of H-2K and H-2D molecules has been
reported to follow in vitro JHMV infection in mouse brain
endothelial cells, and the data indicate that the nature of the
regulation was dependent on the mouse strain and did not
necessarily correlate with susceptibility to JHMV-induced
CNS disease (64). Thus, K was decreased and D increased
in endothelial cells from susceptible BALB/c mice, but both
were upregulated in susceptible B10.S and (BALB/c x SJL)F1
and resistant SJL endothelial cells. Differential expression of
K and D molecules has not always been observed after virus
infection of neural cells; both K and D molecules appear to
be equally enhanced by West Nile virus infection in astro-
cytes from CBA/H mice (65), and the murine neuroblastoma
C1300 does not show differential K and D expression when
persistently infected with measles virus (66). However, the
possibility that differential modulation of class I molecules
may be linked to CNS disease, or may be a marker of suscep-
tibility to virus-induced CNS disease, is suggested by studies
using Theiler's murine encephalomyelitis virus (TMEV; 67,
68). After i.c. TMEV inoculation, resistant B10 mice showed
minimal class I expression in the CNS that did not differ
between K and D, whereas susceptible B10.Q and B10.RBQ
mice showed a greater increase in D expression compared with
that of K (67).
The inability of JHMV to stimulate class I expression in
C57BL/6 astrocytes in the current studies was not reported
for astrocytes from BALB/c, CXJ-8, SJL, and B10.S mice by
Joseph et al. (69). In these experiments, JHMV infection at
an m.o.i, of 0.1 was followed by a two- to threefold increase
in the expression of H-2K molecules, measured by FACS |
analyses, but astrocytes from C57BL/6 mice were not tested.
The disparity in our findings may again reflect differences
in the regulation of class I expression that are mouse strain
specific. However, it is also possible that the disparity may
be due to the cellular composition of the astrocyte cultures
used for infection. In their report, Joseph et al. (69), did not
indicate whether or not they purified the astrocytes from mixed
glial cell cultures. In our hands, mixed glial cells that have
not been subjected to mechanical shaking for purification of
astrocytes show up to 30% contamination by microglia, or
cells staining for macrophage/monocyte surface antigens (data
not shown). Microglial cells readily express class I molecules
both in vivo and in vitro (70-72), and so might be expected
to express class I after JHMV infection. The cultures used
in our experiments routinely show 95% or greater purity,
containing 2-4% microglial cells, and "~50% express class
I after A59 infection. Thus, class I expression in these cul-
tures occurs primarily in astrocytes, with little, if any contri-
bution by microglia. Finally, it should also be mentioned that
class I expression in astrocytes, like that in endothelial cells,
does not appear to be linked to susceptibility to JHMV-induced
1020 Coronavirus Induction of Class I MHC Expression Is Virus Strain Specific

Page 9

CNS disease, since both BALB/c and C57BL/6 develop en-
cephalitis and demyelination in spite of their differences in
class I inducibility after JHMV infection. In addition, astro-
cytes from SJL mice express class I after JHMV infection in
spite of being resistant to CNS disease.
The different effect of A59 and JHMV on class I expres-
sion in astrocytes suggests that class I may play different roles
in their pathogenesis. As discussed by Maudsley et al. (73),
the classical picture of MHC expression after virus infection
is that of an increase that is most likely mediated by the re-
lease of IFNs by infiltrating immune cells, and is thought
to facilitate the ability of T cells to recognize the infected
cells and control or eliminate the infection. Unfortunately,
clearance of virus from infected tissue by CTLs is often ac-
companied by significant cellular destruction, which is not
well tolerated in tissues with limited regenerative capacity,
especially the CNS. Thus, the ability of a virus to stimulate
class I expression in host cells may facilitate a rapid CTL re-
sponse that in turn accelerates tissue destruction originally
begun by virus-mediated cell lysis (74). This may be the case
for A59 infection, since the A59 strains used in these experi-
ments cause a severe encephalitis after i.c. inoculation that
begins on day 4 p.i., and results in death of the majority of
infected mice by day 6-7 p.i. (our unpublished data). How-
ever, it is not known if the infiltration of T cells into the
CNS after i.c. A59 infection is associated with the onset of
encephalitis. By contrast, JHMV causes a similar clinical dis-
ease, but the onset of encephalitis is delayed in comparison
with that of A59, beginning on day 6-7 p.i. (21). Death occurs
on day 9-11 at a similar incidence. The onset of encephalitis
coincides with the appearance of immune cells in the CNS,
which, after JHM-DS infection, peaks on days 7-9 p.i. (75).
It is possible that class I expression is upregulated at this time
by the secretion of interferon by the infiltrating immune cells,
facilitating JHMV-specific, CTL-mediated tissue destruction.
Thus, the difference in the ability of JHMV and A59 to stimu-
late class I expression in astrocytes may contribute to the speed
of disease progression in the CNS. It is interesting to note
that Fazakerley et al. (25) reported that the JHMV variant,
V5A13.1, is neuroattenuated relative to parental JHMV by
its slower rate of spread in the CNS of BALB/c mice. Variant
V5A13.1 differs from parental JHMV by the deletion of 142
amino acids in one of the subunits of the S protein (16, 76),
suggesting that S plays an important role in the rate of virus
spread in the CNS. In this respect, it would be of interest
to test the ability of neuroattenuated A59 strains to stimu-
late class I expression in astrocytes.
In conclusion, we present data indicating that significant
differences exist among MHV strains in their ability to stimu-
late class I expression in murine astrocytes, and that these
differences may contribute to their ability to cause CNS dis-
ease. Since class I expression has been reported on astrocytes
in the brains of MS patients (77), the data have implications
for understanding the role of class I in human CNS disease.
We wish to thank Michael Lai for the recombinant virus strains and Steve Stohlman, Michael Lai, and
Minnie McMiUan for helpful critiques of the data and manuscript. We are also grateful for help with
the manuscript provided by Sonia Garcia.
This work was supported by National Institutes of Health grant NS-18146-12.
Address correspondence to Dr. Wendy Gilmore, Department of Neurology, USC School of Medicine,
1333 San Pablo Street, Los Angeles, CA 90033.
Received for publication 30 March 1994 and in revised form 7 June 1994.
References
1. Bailey, O.T., A.M. Pappenheimer, F.S. Cheever, andJ.B. Daniels.
1949. A murine virus (JHM) causing disseminated en-
cephalomyelitis with extensive destruction of myelin. J. Exp.
Med. 90:195.
2. Weiner, L.P. 1973. Pathogenesis of demyelination induced by
a mouse hepatitis virus OHM virus). Arch. Neurol. 28:298.
3. Robb, J.A., C.W. Bond, andJ.L. Liebowitz. 1979. Pathogenic
murine coronaviruses. III. Biological and biochemical charac-
terization of temperature-sensitive mutants of JHMV. Virology.
94:385.
4. Hirano, N., N. Goto, T. Ogawa, K. Ono, K. Murakami, and
K. Fujiwara. 1980. Hydrocephalus in suckling rats infected in-
tracerebrally with mouse hepatitis virus MHV-A59. Microbiol.
Immunol. 24:825.
5. Sorenson, O., D. Perry, and S. Dales. 1980. In vivo and in vitro
models of demyelinating diseases. Arch. Neurol. 37:478.
6. Stohlman, S.A., and L.P. Weiner. 1981. Chronic central nervous
system demyelination in mice after JHM virus infection. Neu-
rology. 31:38.
7. Wege, H., R. Watanabe, and V. ter Meulen. 1984. Relapsing
subacute demyelinating encephalomyelitis in rats during the
course of eoronavirus JHM infection. J. Neuroimmunol. 6:325.
8. Murray, R.S., G.Y. Cai, K. Hoel, J.Y. Zhang, K.F. Soike, and
G.F. Cabirac. 1992. Coronavirus infects and causes demyelina-
tion in primate central nervous system. Virology. 188:274.
9. Gerdes, J.C., I. Klein, B.L. DeVald, and J.S. Burks. 1981.
Coronavirus isolates SK and SD from multiple sclerosis patients
are serologically related to routine coronaviruses A59 and JHM
1021 Gilmore et al.

Page 10

and human coronavirus OC43, but not to human coronavirus
229E. J. Virol. 38:231.
10. Murray, R.S., B. Brown, D. Brian, and G.F. Cabirac. 1992.
Detection of coronavirus RNA and antigen in multiple scle-
rosis brain. Ann. Neurol. 31:525.
11. Lai, M.M.C., and S.A. Stohlman. 1992. Molecular basis ofneu-
ropathogenicity of mouse hepatitis virus. In Molecular Patho-
genesis. R.P. Roos, editor. Humana Press, Totowa, NJ. p. 319.
12. Fleming, J.O., S.A. Stohlman, R.C. Harmon, M.M.C. Lai,
J.A. Frelinger, and L.P. Weiner. 1983. Antigenic relationships
of murine coronaviruses: analysis using monoclonal antibodies
to JHM (MHV-4) virus. Virology. 131:296.
13. Gilmore, W., J.O. Fleming, S.A. Stohlman, and L.P. Weiner.
1987. Characterization of the structural proteins of the mu-
rine coronavirus strain A59 using monoclonal antibodies
(42532). Proc. Soc. Exp. Biol. Med. 185:177.
14. Luytjes, W., L. Sturman, P. Bredenbeek, J. Charite, B.A.M.
van der Zeijst, M.C. Horzinek, and W.J.M. Spaan. 1987. Pri-
mary structure of the glycoprotein E2 of coronavirus MHV-
A59 and identification of the trypsin cleavage site. Virology.
161:479.
15. Schmidt, I., M. Skinner, and S. Siddell. 1987. Nucleotide se-
quence of the gene encoding the surface projection glycopro-
rein of coronavirus MHV-JHM. J. Gen. Virol. 68:47.
16. Parker, S.E., T.M. Gallagher, and M.J. Buchmeier. 1989. Se-
quence analysis reveals extensive polymorphism and evidence
of deletions within the E2 glycoprotein gene of several strains
of murine hepatitis virus. Virology. 173:664.
17. Yokomori, K., S.C. Baker, S.A. Stohlman, and M.M.C. Lai.
1992. Hemagglutinin-esterase (HE)-specific monoclonal anti-
bodies alter the neuropathogenicity of mouse hepatitis virus.
J. Virol. 66:2865.
18. Fleury, H.J.A., R.D. Sheppard, M.B. Bornstein, and C.S.
Raine. 1980. Further ultrastructural observations of virus mor-
phogenesis and myelin pathology in JHM virus encephalomye-
litis. Neuropath. App. Neurobiol. 6:165.
19. Knobler, R.L., M. Dubois-Dalcq, M.V. Haspel, A.P. Clay-
smith, P.W. Lampert, and M.B.A. Oldstone. 1981. Selective
localization of wild type and mutant mouse hepatitis virus
(JHM strain) antigens in CNS tissue by fluorescence, light and
electron microscopy. J. Neuroimmunol. 1:81.
20. Fishman, P.S., J.S. Gass, and P.T. Swoverland. 1985. Infection
of the basal ganglia by a murine coronavirus. Science (Wash.
DC). 229:877.
21. Wang, F.-I., D.R. Hinton, W. Gilmore, M.D. Trousdale, and
J.O. Fleming. 1992. Sequential infection of glial cells by the
murine hepatitis virus JHM strain (MHV-4) leads to a charac-
teristic distribution of demyelination. Lab. Invest. 66:744.
22. Haspel, M.V., P.W. Lampert, and M.B.A. Oldstone. 1978.
Temperature-sensitive mutants of mouse hepatitis virus pro-
duce a high incidence of demyelination. Proc. Natl. Acad. Sci.
USA. 75:4033.
23. Dubois-Dalcq, M.E., E.W. Doller, M.V. Haspel, and K.V.
Holmes. 1982. Cell tropism and expression of mouse hepatitis
viruses (MHV) in mouse spinal cord cultures. Virology. 119:317.
24. Perlman, S., and D. Ries. 1987. The astrocyte is a target cell
in mice persistently infected with mouse hepatitis virus, strain
JHM. Microtx Pathog. 3:309.
25. Fazakerley, J,K., S.E. Parker, F. Bloom, and M.J. Buchmeier.
1992. The V5A13.1 envelope glycoprotein deletion mutant of
mouse hepatitis virus type-4 is neuroattenuated by its reduced
rate of spread in the central nervous system. Virology. 187:178.
26. van Berlo, M.F., G. Wolswijk, J. Calafat, M.J.M. Koolen, M.C.
Horzinek, and B.A.M. van der Zeijst. 1986. Restricted repli-
1022
cation of mouse hepatitis virus A59 in primary mouse brain
astrocytes correlates with reduced pathogenicity.J. Virol. 58:426.
27. Koolen, M.J., S. Love, W. Wouda, J. Calafat, M.C. Horzinek,
and B.A.M. van der Zeijst. 1987. Induction of demyelination
by a temperature-sensitive mutant of the coronavirus MHV-
A59 is associated with restriction of viral replication in the
brain. J. Gen. Virol. 68:703.
28. Fleming, J.O., M.D. Trousdale, S.A. Stohlman, and L.P. Weiner.
1987. Pathogenic characteristics of neutralization-resistant vari-
ants of JHM coronavirus (MHV-4). Adv. Exp. Med. Biol.
218:333.
29. Wege, H., M. Koga, R. Watanabe, K. Nagashima, and V.
ter Meulen. 1983. Neurovirulence of murine coronavirus JHM
temperature-sensitive mutants in rats. Infect. Immun. 39:1316.
30. Massa, P.T., H. Wege, and V. ter Meulen. 1986. Analysis of
murine hepatitis virus OHM strain) tropism toward Lewis rat
glial cells in vitro. Type 1 astrocytes and brain macrophages
(microglia) as primary glial cell targets. La~ Invest. 55:318.
31. Lavi, E., A. Suzumura, M. Hirayama, M.K. Highkin, D.M.
Dambach, D.H. Silberberg, and S.R. Weiss. 1987. Coronavirus
mouse hepatitis virus (MHV)-A59 causes a persistent, produc-
tive infection in primary glial cell cultures. Microtx Pathog. 3:79.
32. van Berlo, M.F., R. Warringa, G. Wolswijk, and M. Lopes-
Cardozo. 1989. Vulnerability of rat and mouse brain cells to
murine hepatitis virus 0HM-strain): studies in vivo and in vitro.
Glia. 2:85.
33. Suzumura, A., E. Lavi, S.R. Weiss, and D.H. Silberberg. 1986.
Coronavirus infection induces H-2 antigen expression on
oligodendrocytes and astrocytes. Nature (Lond.). 232:991.
34. Gombold, J.L., and S.R. Weiss. 1992. Mouse hepatitis virus
A59 increases steady-state levels of MHC mRNAs in primary
glial cell cultures and in the murine central nervous system.
Microbiol. Pathog. 13:493.
35. Yewdell, J.W., and J.R. Bennink. 1992. Cell biology of an-
tigen processing and presentation to major histocompatibility
complex class I molecule-restricted T lymphocytes. Adv. Im-
munol. 52:1.
36. Rammensee, H., K. Falk, and O. R.otzschke. 1993. Peptides
naturally presented by MHC class I molecules. Annu. Rev. Irn-
munol. 11:213.
37. Schnitzer, J., and M. Schachner. 198. Expression ofThy-1, H-2
and NS-4 cell surface antigens and tetanus toxin receptors in
early postnatal and adult mouse cerebellum. J. Neuroimmunol.
1:429.
38. Wong, G.H.W., P.F. Bartlett, I. Clark-Lewis, F. Battye, and
J.W. Schrader. 1984. Inducible expression of H-2 and Ia an-
tigens on brain cells. Nature (Lond.). 310:688.
39. Wong, G.H.W., P.F. Bartlett, I. Clark-Lewis, J.L. McKimm-
Breschkin, and J.W. Schrader. 1985. Interferon- 7 induces the
expression of H-2 and Ia antigens on brain cells. J. Neuroim-
munol. 7:255.
40. Massa, P.T., E.P. Cowan, B.-Z. Levi, K. Ozato, and D.E.
McFarlin. 1989. Genetic regulation of class I major histocom-
patibility complex (MHC) antigen induction on astrocytes.J.
Neuroimmunol. 24:125.
41. Satoh, J., S.U. Kim, L.f. Kastrukoff, and T. Takei. 1991. Ex-
pression and induction of intercellular adhesion molecules
(ICAMs) and major histocompatibility complex (MHC) an-
tigens on cultured murine oligodendrocytes and astrocytes. J.
Neurosci. Res. 29:1.
42. Lavi, E., A. Suzumura, D.M. Murasko, E.M. Murray, D.H.
Silberberg, and S.R. Weiss. 1988. Tumor necrosis factor in-
duces expression of MHC class I antigens on mouse astrocytes.
Ann. NY Acad. Sci. 540:488.
Coronavirus Induction of Class I MHC Expression Is Virus Strain Specific

Page 11

43. Fontana, A., P. Erb, H. Pircher, R. Zinkernagel, E. Weber,
and W. Fierz. 1986. Astrocytes as antigen-presenting cells. II.
Unlike H-2K-dependent cytotoxic T cells, H-2 Ia restricted
T cells are only stimulated in the presence of interferon-y, j.
NeuroimmunoL 12:15.
44. Sedgwick, J.D., K. MrBner, S. Schwender, and V. ter Meulen.
1991. Major histocompatibility complex-expressing non-
hematopoietic astroglial cells prime only CD8 § T lympho-
cytes: astroglial cells as perpetuators but not initiators of
CD4 § T cell responses in the central nervous system.J. Exp.
Med. 173:1235.
45. Sussman, M.A., K.A. Shubin, S. Kyuwa, and S.A. Stohlman.
1989. T-cell-mediated clearance of mouse hepatitis virus strain
JHM from the central nervous system. J. ViroL 63:3051.
46. Williamson, J.S.P., and S.A. Stohlman. 1990. Effective clear-
ance of mouse hepatitis virus from the central nervous system
requires both CD4 + and CD8 § T cells. J. ViroL 64:4589.
47. Yamaguchi, K., N. Goto, S. Kyuma, M. Hayami, and Y.
Toyoda. 1991. Protection of mice from a lethal coronavirus in-
fection in the central nervous system by adoptive transfer of
virus-specific T cell clones. J. NeuroimmunoL 32:1.
48. McCarthy, K.D., and J. de Vellis. 1980. Preparation of sepa-
rate astroglial and oligodendroglial cell cultures from rat cere-
bral tissue. J. Cell Biol. 85:890.
49. Lai, M.M.C. 1992. RNA recombination in animal and plant
viruses. Microbiot. Rev. 56:61.
50. Finter, N.B., 1969. Dye uptake methods for assessing viral
cytopathogenicity and their application to interferon assays.
j. Gen. ViroL 5:419.
51. Brennan, G.L., and L.H. Kronenberg. 1983. Automated bio-
assay of interferons in micro-test plates. Biotechniques. 1:79.
52. Makino, S., J.G. Keck, S.A. Stohlman, and M.M.C. Lai. 1986.
High-frequency KNA recombination of murine coronaviruses.
J. ViroL 57:729.
53. Suzumura, A., E. Lavi, S. Bhat, D. Murasko, S.R.. Weiss, and
D.H. Silberberg. 1988. Induction of glial cell MHC antigen
expression in neurotropic coronavirus infections: characteriza-
tion of the H-2-inducing soluble factor elaborated by infected
brain cells. J. ImmunoL 140:2068.
54. Lavi, E., A. Suzumura, E.M. Murray, D.H. Silberberg, and
S.R. Weiss. 1989. Induction of MHC class I antigens on glial
cells is dependent on persistent mouse hepatitis virus infec-
tion. j. NeuroimmunoL 22:107.
55. Revel, M., and J. Chebath. 1986. Interferon-activated genes.
Trends Biochem. Sci. 11:166.
56. Stohlman, S.A., A.Y. Sakaguchi, and A. Hiti. 1978. Interferon
production and activity in mouse neuroblastoma cells. Arch.
ViroL 57:91.
57. Collins, A.K., L.A. Tunison, and R.L. Knobler. 1983. Mouse
hepatitis virus type 4 infection of primary glial cultures from
genetically susceptible and resistant mice. Infect. Immun. 40:1192.
58. Correale, J., S. Li, L.P. Weiner, and W. Gilmore. The effect
of persistent mouse hepatitis virus infection on MHC Class
I expression in murine astrocytes. J. Neurosci. Res. In press.
59. Gallagher, T.M., M.J. Buchmeier, and S. Perlman. 1992. Cell
receptor-independent infection by a neurotropic murine corona-
virus. Virology. 191:517.
60. Yokomori, K., and M.M.C. Lai. 1992. Mouse hepatitis virus
utilizes two carcinoembryonic antigens as alternative receptors.
J. Virol. 66:6194.
61. Williams, R.K., G.-S. Jian, and K.V. Holmes. 1991. Receptor
for mouse hepatitis virus is a member of the carcinoembryonic
antigen family of glycoproteins. Proc. Natl. Acad. Sci. USA.
88:5533.
1023 Gilmore et al.
62. Massa, P.T., S. Hirschfeld, B.-Z. Levi, L.A. Quigly, K. Ozato,
and D.E. McFarlin. 1992. Expression of major histocompati-
bility complex (MHC) class I genes in astrocytes correlates with
the presence of nuclear factors that bind to constitutive and
inducible enhancers. J. NeuroimmunoL 41:35.
63. Massa, P.T., K. Ozato, and D.E. McFarlin. 1993. Cell type-
specific regulation of major histocompatibility complex (MHC)
class I gene expression in astrocytes, oligodendrocytes, and
neurons. Glia. 8:201.
64. Joseph, J., R.L. Knobler, F.D. Lublin, and M.N. Hart. 1989.
Differential modulation of MHC class I antigen expression on
mouse brain endothelial cells by MHV-4 infection.J. Neuroim-
munol. 22:241.
65. Liu, Y., N. King, A. Kesson, K.V. Blanden, and A. Mullbacher.
1989. Flavivirus infection up-regulates the expression of class
I and class II major histocompatibility antigens on and enhances
T cell recognition of astrocytes in vitro.J. Neuroimmunol. 21:157.
66. Gopas, J., D. Itzhaky, Y. Segev, S. Salzberg, B. Trink, N. Isakov,
and B. Rager-Zisman. 1992. Persistent measles virus infection
enhances histocompatibility complex class I expression and im-
munogenicity of murine neuroblastoma cells. Cancer ImmunoL
Immunother. 34:313.
67. Altintas, A., Z. Cai, L.R. Pease, and M. Rodriguez. 1993.
Differential expression of H-2K and H-2D in the central nervous
system of mice infected with Theiler's virus, j. Immunol.
151:2803.
68. Lindsley, M.D., A.K. Patick, N. Prayoonwiwat, and M.
Rodriguez. 1992. Coexpression of class I major histocompati-
bility antigen and viral RNA in central nervous system of mice
infected with Theiler's virus: a model for multiple sclerosis.
Mayo Clin. Proc. 67:829.
69. Joseph, J., K.L. Knobler, F.D. Lublin, and M.N. Hart. 1990.
Regulation of MHC class I and II antigens on cerebral en-
dothelial cells and astrocytes following MHV-4 infection. Adv.
Exp. Med. Biol. 276:579.
70. Hickey, W.F., and H. Kimura. 1988. Perivascular microglia
cells of the CNS are bone-marrow derived and present antigen
in vivo. Science (Wash. DC). 239:290.
71. Suzumura, A., S.G.E. Mezitis, N.K. Gonatas, and D.H. Sil-
berberg. 1987. MHC antigen expression on bulk-isolated
macrophage-microglia from newborn mouse brain: induction
of Ia antigen expression by y-interferon. J. NeuroimmunoL
15:263.
72. Sedgwick, J.D., S. Schwender, H. Imrich, K. Dorries, G.W.
Butcher, and V. ter Meulen. 1991. Isolation and direct charac-
terization of resident microglial cells from the normal and
inflamed central nervous system. Proc. Natl. Acad. Sci. USA.
88:7438.
73. Maudsley, J., and J.D. Pound. 1991. Modulation of MHC an-
tigen expression by viruses and oncogenes. ImmunoL Today.
12:429.
74. Oldstone, M.B.A. 1989. Viral persistence. Ceil 56:517.
75. WiUiamson, J.S.P., K.C. Sykes, and S.A. Stohlman. 1991. Char-
acterization of brain-infiltrating mononuclear cells during in-
fection with mouse hepatitis virus strain JHM.J. NeuroimmunoL
32:199.
76. Gallagher, T.M., S.E. Parker, and M.J. Buchmeier. 1990.
Neutralization-resistant variants of a neurotropic coronavirus
are generated by deletions within the amino-terminal half of
the spike glycoprotein. J. Virol. 64:731.
77. Traugott, U. 1987. Multiple sclerosis: relevance of class I and
class II MHC-expressing cells to lesion development. J. Neu-
roimmunot. 16:283.