key: cord-326688-a1djgqpa
authors: Dubois-Dalcq, Monique E.; Doller, Elizabeth W.; Haspel, Martin V.; Holmes, Kathryn V.
title: Cell tropism and expression of mouse hepatitis viruses (MHV) in mouse spinal cord cultures
date: 1982-06-30
journal: Virology
DOI: 10.1016/0042-6822(82)90092-7
sha: 
doc_id: 326688
cord_uid: a1djgqpa

Abstract Mouse hepatitis viruses (MHV) are coronaviruses which cause various infections in mice affecting lung, intestine, liver, and other organs as well as the central nervous system. The replication of three different MHV strains was studied in mouse dissociated spinal cord cultures containing differentiated neurons and nonneuronal cells (NN) (including astrocytes). Cell tropism and maturation of each virus strain was analyzed by immunolabeling methods using antisera to the virion or to purified membrane glycoproteins (E1 and E2) and by electron microscopy (EM). Wt-JHM, which causes acute encephalitis in mice, produces acute cytopathic changes in both neurons and NN cells. In neurons, virions mature in smooth ER cisternae closely associated to the Golgi apparatus. As judged by EM, fewer virions are produced by neurons than NN cells and neurons do not fuse or stain for E2 as do NN cells. NN cells contain large inclusions made of nucleocapsid strands. A temperature-sensitive mutant of JHM, Ts8-JHM, which causes demyelination in mice, infects NN cells but not neurons. Infected NN cells synthesize E1 and E2, and contain large inclusions but few mature virions, even at permissive temperatures. These inclusions appear granular and rarely contain nucleocapsid strands in contrast to wt-JHM infection. NN cells infected with this mutant also display numerous membrane whorls. The hepatotropic strain A59 lacks tropism for neurons and primarily infects NN cells, thus resembling ts8-JHM. Infected NN cells become loaded with intracytoplasmic virions which are secreted from the cells. E1 can only be detected in the perinuclear area of these cells while E2 rapidly spreads throughout the cytoplasm. The cytoplasm of A59 infected NN cells frequently contains large tubular structures often in the lumen of the RER. In conclusion, in primary CNS cultures consisting of neurons and NN cells: (1) wt-JHM replicates in both neurons and NN cells but has different effects on these cells; (2) Ts8-JHM exhibits no productive infection of neurons, and in NN cells appears to be defective in assembly and to stimulate membrane synthesis; (3) A59 also shows tropism restricted to NN cells which produce many viruses and display differential distribution of the two virion glycoproteins. Thus, in the absence of the immune system, the MHV strains assayed exhibit differences in viral tropism, cytopathic changes, and viral assembly in CNS cells, and these differences may account for the different disease patterns.

(3) A59 also shows tropism restricted to NN cells which produce many viruses and display differential distribution of the two virion glycoproteins. Thus, in the absence of the immune system, the MHV strains assayed exhibit differences in viral tropism, cytopathic changes, and viral assembly in CNS cells, and these differences (Knobler et al., 1981b) or a demyelinating disease, depending upon the age of the animal at the time of inoculation and the virus passage level (Haspel et al., 19'78; Herndon et al., 1975; Lampert et al., 1973; Stohlman and Weiner, 1981; Weiner, 1973) . In contrast, a chemically induced temperature-sensitive mutant of wt-JHM, ts8-JHM, rarely induces fatal encephalomyelitis in the same mouse strains but regularly leads to demyelination (Haspel et al., 1978) . In viva electron microscopic and immunocytochemical studies have shown that neurons do not appear to support replication of ts&JHM, while oligodendrocytes do (Knobler et al., 1981a) . Finally, a hepatotropic MHV strain, A59 (Sturman and Takemoto, 1972) , can also produce chronic CNS disease in CBH mice (K. V. Holmes and P. 0. Leinikki, unpublished observations.) A59 is presently the best characterized coronavirus strain: it has three virion-associated polypeptides which have been isolated and purified Holmes et al., 1981 Sturman and Holmes, 1977; Sturman et al., 1980) . The membrane glycoprotein El appears to be involved in virus assembly, while the peplomeric glycoprotein Ez is required for infectivity and virus-induced cell fusion. Monospecific antibodies have been raised against these proteins as well as to detergent-disrupted virions (Sturman et al., 1980) .

We have compared the tropism and maturation of two JHM virus strains with that of the A59 strain in CNS cells, using immunolabeling and electron microscopy. Our results suggest that specific properties of each virus strain and host factors present in differentiated nerve cells act in concert to control the features and outcome of MHV CNS infection at the target cell level.

Cells. Cultures of dissociated CNS cells were obtained from spinal cord dissected from 12-to 13-day-old C57B1/6 mouse embryos as described before (Faulkner et al., 1979; Hooghe-Peters et al., 1979) . C57B1/6 mice are permissive for the JHM strains. Cells were grown on 12-mm glass coverslips or 35-mm petri dishes coated with collagen and supplemented with medium as described (Faulkner et al., 1979; Hooghe-Peters et al., 1979) . Fibroblast growth was inhibited by a 2-day antimitotic treatment with a mixture of FudR and uridine after 1 week in vitro. Cultures were used after 2 to 3 weeks of in vitro maturation when neurons and astrocytes had acquired their mature stage (Figs. la and b). Three cell types were identified in these cultures. The neurons represent 5 to 10% of the cells; they are either isolated or in clusters and show a characteristic morphology with long processes as seen by phase microscopy ( Fig. la) (Faulkner et al., 1979; Hooghe-Peters et al., 1979) . The nonneuronal (NN) cells are comprised of at least two cell types as judged on the basis of positive or negative immunoreactivity for glial fibrillary acidic protein (GFAP), a specific protein of astrocytes (Bignami et al., 1980) . GFAP+ cells represent 50 to 70% of the NN cells; these astrocytes have a pleiomorphic shape, contain radiating fibrils, and form groups under the neurons. GFAP-cells probably represent primary fibroblastic cells of CNS origin (meningeal or perivascular).

Oligodendrocytes were not detected in these cultures.

VimLses. The A59 virus, a hepatotropic strain of mouse hepatitis virus, was used as described before (Sturman and Takemoto, 1972; Sturman and Holmes, 1977; Sturman et al., 1980) . The origin and propagation of the wt and ts8 strains of JHM have been published (Haspel et al., 1978) . The multiplicity of infection of 10 plaqueforming units (PFU)/cell was used with A59, whereas a multiplicity of infection of 0.3 to 1 PFU/cell was used with both JHM strains (the JHM stock viruses had much lower titers). Viral adsorption was carried out for 1 hr at 37" as previously described (Knobler et al., 1981a) . Cultures were incubated at 37", although in a few experiments paired cultures were also incubated at 34". The cultures were observed daily with a Zeiss inverted phase scope. Viral infectivity of supernatant fluids was determined by plaque assay in L2 or 17CL 1 cells as described (Sturman and Takemoto, 1972) . Antiserum and immunolubeling.

Four different rabbit antisera were used. Rabbit antiserum against GFAP was kindly provided by Dr. Rebecca Pruss; the rabbit had been injected with a suspension of human spinal cord. When reacted with cord proteins separated by agar gel electrophoresis, immune reactivity of this serum was found to be restricted to GFAP. Preparation and characterization of monospecific polyclonal rabbit antisera to the isolated El and Ez glycoproteins of A59 virion envelope has been previously described (Sturman et al., 1980) . Rabbit antisera directed against gradient-puri-fied NP40-disrupted A59 virions were also prepared. This hyperimmune serum immunoprecipitated the three major structural polypeptides of both A59 and JHM from radiolabeled infected cells. For fluorescent antibody (FA) staining, dissociated CNS cells grown on glass coverslips were fixed in a mixture of 5% acetic acid and 95% ethyl alcohol at -20" (for GFAP staining) or in acetone (for El or Ez staining) before being incubated for 1 hr in various dilutions of antisera. In some instances, living cells were incubated with antiviral sera at 4' or room temperature in order to detect only antigens on the cell surface. The second reagent used to detect bound antibodies was either protein Afluorescein (Pharmacia, Piscataway, N. J.) or goat anti-rabbit IgG coupled to fluorescein. This conjugate was absorbed with liver powder or 17CL 1 cells. For immunoperoxidase labeling, staining was performed as described before (Knobler et al., 1981a) . Briefly, the cells were fixed in 4% formaldehyde with 0.3 to 0.5% glutaraldehyde for 20 min, frozen and thawed in a sucrose-glycerol mixture in order to render the cell membrane permeable to antibody molecules. The cells were then allowed to react with antibody to the virion or to glycoprotein El or Ez. Protein A-peroxidase was used as a second layer and cells were further incubated in the peroxide-diaminobenzidine substrate. Immediate observations by light microscopy were done on a Zeiss photomicroscope III using a water immersion phase contrast (40X) objective for immunoperoxidase or an oil immersion (40x) objective for fluorescence.

When double labeling with antisera to both GFAP and viral antigens was performed, glass coverslip cultures of CNS cells were fixed with acid-alcohol at -20". After washing with serum containing medium, the coverslips were incubated successively in: (1) rabbit antiserum to GFAP, (Faulkner et al., 1979) , glass coverslips covered with cells were used for scanning electron microscopy @EM), whereas 35mm petri dishes were used for transmission electron microscopy (TEM) studies. Cells were usually fixed with Karnovsky fixative consisting of a mixture of 1.25% glutaraldehyde and 1% formaldehyde in 0.06 M cacodylate buffer at pH 7.2. For SEM studies, cells were then postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 2 hr, and dehydrated in methanol, critical point dried in COz in a Balzer apparatus, coated under high vacuum with gold-palladium alloy, and observed in an ETEC Autoscan at 20 kV and a tilt angle of 45'. For TEM studies, monolayers in petri dishes that had been labeled by the immunoperoxidase method (as described above) or fixed in situ with the Karnovsky fixative, were postfixed with osmium tetroxyde 1% in cacodylate buffer for 30 min, stained with uranyl acetate at pH 5, progressively dehydrated in ethanol, and embedded in Epon 812. The technique for circling and drilling out immunolabeled cells or syncytia from the embedded monolayer has been published in detail (Knobler et al., 1981a) . Sections were cut with a diamond knife and examined in Philips 201 or Philips 400-T electron microscopes. Thin sections were usually counterstained with uranyl acetate and lead salts, and examined at 80 kV. proteins between 16 and 18 hr. Titers of JHM was examined successively in inreleased virus were always lower than ti-fected neurons and NN cells. Infected neuters of CNS cells infected with A59 strain rons showed diffuse immunofluorescent (Table 1) PI, the organization of the cytoplasm of infected neurons was not altered since they had an extensive rough ER system and a normal distribution of other organelles. Budding and free virions with electron-lucent centers were detected in cisternae of smooth membranes closely associated with the Golgi apparatus (Figs.  2a, b) . Virus was also found occasionally in small vacuoles within dendrites (Fig.  2~ ) as seen in infected mouse spinal cord in tivo (Knobler et al., 1981a) . At 48 hr PI, the cytoplasm of infected neurons showed disorganization of RER membranes and vacuolization.

Scattered budding virions were seen in association with intracytoplasmic membranes as described before (Knobler et al., 1981a) . Fewer cytoplasmic vacuoles containing virions were detected in thin sections of infected neurons than in giant cells formed by NN cells (illustrated in Figs. 3 and 4) . Similarly, clusters of extracellular virions were rarely found around infected neurons, whereas they were numerous around NN cells (Fig. 4~ ). This correlates with previous in viva observations on infected neurons of the spinal cord (Knobler et al., 1981a) .

Infected NN cells contained both El and Ez glycoproteins at 16 hr PI. These proteins were predominantly localized in the perinuclear area of giant cells (Fig. 3, these fused NN cells revealed nucleocapsid loosely folded nucleocapsid strands apstrands accumulating in large inclusions proximately 8 nm in diameter were seen and numerous virions budding in vacu-in close apposition to the ER membrane oles (Fig. 3) . Before budding occurred, (Figs. 4a and b) . During budding, the nu-cleocapsid was closely apposed to the virion membrane of the bud which had acquired definite spikes (Fig. 4b) . Virions were budding from RER as well as smooth ER. Using antiserum against purified A59 virions, MHV antigens were detected in all membranes directly associated with virion .formation and in the virion surface (data not shown). Numerous free virions were released by giant cells and formed clusters firmly attached to infected cell surface and heavily labeled after immunoperoxidase staining (Fig. 4~) . Free virions in the extracellular space resembled those seen in vacuoles and were 80 to 100 nm in diameter (Fig.  4c, inset) .

Neurons showed no cytopathic changes or viral antigens after inoculation with this strain. However, scattered foci of NN cells slowly developed fusion over 3 to 5 days and contained viral antigen. Both GFAP+ and GFAP-cells were infected as demonstrated by double labeling for viral-and cell-specific antigens. The glycoproteins, Ez and El, were detected in the perinuclear area of giant cells at 24 hr and stayed in these regions throughout infec: tion. Slow progression of infection led to disintegration of the monolayer after 5 to 7 days. NN cells infected with ts8-JHM released at least 10 times less virus than cells infected with wt-JHM (Table l) , even at the permissive temperature (34') (data not shown).

By EM, inclusions present in NN cells infected with ts&JHM were markedly different from those observed in wt-JHM infection of NN cells. Wt-JHM inclusions contained nucleocapsid strands 8 nm in diameter as well as crescent-shaped corelike structures (37-40 nm in diameter), usually with an electronlucent center (Fig.  5a) . Ts8-JHM inclusions contained numerous dense granules (20 to 30 nm in diameter). Strands of nucleocapsid could not be clearly identified in these inclusions (Fig.  5b ). In addition, rows of dense granules (18 nm in diameter)

were seen aligned along cytoplasmic tubules of ts&JHM-in-fected NN cells (Fig. 5~) . The possibility that these granules represent ribosomes is not excluded. NN cells infected with ts8-JHM showed less intracellular and extracellular virions than in wt-JHM infection (Fig. 4~) . NN cells infected with ts&JHM displayed numerous membrane whorls sometimes organized in myelin-like figures (Fig. 6) . However, our cultures do not normally contain oligodendrocytes or show myelination.

These membrane whorls surrounded cytoplasmic compartments containing ER, dense granules, and, occasionally, budding virions (Fig. 6) . Some whorls were made of 40 pairs of membranes with an interspace of 12 to 20 nm. Immunoperoxidase labeling of viral antigens did not reveal staining of membrane whorls.

As with ts8-JHM, no infected neurons were found in CNS cultures inoculated with A59 and the infection appeared restricted to NN cells. Mild fusion of NN cells was observed and infection lasted for 5 to 10 days. Infected GFAP+ cells progressively retracted their fibrillar processes. A59-infected cultures produced 10 times more infectious virus than those infected with wt-JHM (Table 1 ) even though wt-JHM replicated in both neurons and NN cells. Infected astrocytes were wrapped around healthy, apparently uninfected neurons as seen by EM (Fig. 7) . These astrocytes were very rich in intermediate filaments and contained numerous viral particles in their ER system (Fig. 7) . The inset of Fig. 7 shows a virion in a vacuole close to the cell surface. The glycoprotein Ez was detected as early as 8 hr PI and localized throughout the cytoplasm of infected NN cells, while El, first detected at 16 hr, remained predominately perinuclear (Figs. 8a and b) . Such differential localization of these proteins has been observed in infected 17CL 1 fibroblasts (E. W. Doller et al., submitted for publication). Interestingly, virus maturation in perinuclear cisternae was frequently noted by EM (Fig. 8~) the lumen of the RER (Fig. 8d) . These tubular structures had a higher density than cellular microtubules. A59-infected NN cells also showed some membrane whorls but these were smaller than those observed in ts&JHM infection of NN cells (Fig. 7) . DISCUSSION The present study describes how three different MHV strains interact in vitro with cultured neurons and NN cells, including astrocytes, isolated from mouse spinal cords. Previous studies have demonstrated that persistent infection with the JHM strain could be obtained in cloned cell lines of glial and neuronal origin (Lucas et al., 1977 (Lucas et al., ,1978 Stohlman and Weiner, 1978) . Here, we have used primary cultures of differentiated nerve cells, an in vitro system that seems closer to the in viva situation. Our observations are summarized in Table 2 .

Wt-JHM is the only strain able to replicate in mature neurons as well as NN cells and this correlates well with in viva observations (Knobler et al., 1981a) . Wt-JHM never produces fusion of neurons as do other viruses , but rather induces neuronal rounding and lysis with little virus release. Simultaneously, extensive fusion and virus production occur in the nearby NN cells. Thus wt-JHM has very different effects in these two CNS cell types. In addition, the peplomeric glycoprotein Ez was not detected in infected neurons but was observed in NN cells. Thus, JHM virions produced in neurons may either lack Ez or have an altered E2 no longer reacting with antisera used. Virions lacking peplomers and E2 are formed in A59-infected 17CL 1 cells treated with tunicamycin (Holmes et al., 1981 . Thus it is possible for buds lacking Ez to form but such buds are known to be noninfectious (Holmes et al., 1981 . The possible differences in E2 glycoprotein synthesis between neurons and NN cells infected with wt-JHM should be further studied provided that purified neuron cultures could be obtained (Knobler et al., 1982) .

The other two MHV strains, ts8-JHM and A59, infect only NN cells. Similarly, in weanling mice, ts&JHM does not appear to replicate in neurons (Knobler et al., 1981a) . One possible explanation for this restricted tropism is that ts8-JHM and A59 may lack some structural aspects of the surface glycoprotein E2 necessary for its interaction with virus receptors on neurons. Alternately, the virus may have infected the neurons but intracellular factors of an unknown nature may have blocked its replication.

In a study of restricted tropism of VSV ts mutants, it was suggested that "secondary" virulence genes may control viral replication in cells through adequate or inadequate interactions of virus specific polymerase complexes with some specific host cell factor (Preble et al., 1980) .

and A59 is closely similar to that observed in fibroblast cell lines and liver cell cultures (David-Ferreira and Manaker, 1965 ) but more stages are captured in primary CNS cells. TEM analysis of these stages reveals that a close interaction between a folded nucleocapsid strand and a virus-modified ER membrane is occurring during the budding process, as is the case for other enveloped RNA viruses. There is a close association of the virus maturation sites with the Golgi system. Virus contained in vacuoles is "secreted" since these vacuoles eventually fuse with the plasma membrane . This release process is also observed with other enveloped viruses undergoing intracellular maturation such as herpes simplex (Rodriguez and Dubois-Dalcq, 1978) . A59 infected CNS cultures produce more virus than wt-JHM infected ones, even though A59 infects only NN cells in a slow, progressive way. In addition, differential but not in wt-JHM-infected NN cells. In A59 infections, the membrane glycoprotein El appears to remain in the Golgi area while the peplomeric glycoprotein E2 migrates rapidly to the plasma membrane ( Figs. 8a and b) . In contrast, the E2 synthesized in NN cells by the JHM strains did not disperse rapidly but remained near the cell center (Fig. 3, inset) . Possibly the signals for migration of the E2 glycoprotein are different for the JHM and A59 strains of MHV. These differences may account, at least in part, for the different yields of virus from cells infected with wt-JHM or A59.

Ts8-JHM-infected NN cells showed viral inclusions different from those observed in wt-JHM infection. These alterations suggest a defect in the ribonucleoprotein, resulting in defective assembly of t&3-JHM occurring even at permissive temperatures. Such an alteration was not observed in fibroblastoid cell lines but only in CNS cells (Dubois-Dalcq et al., unpublished observations) .

Thus the host cell appears to influence expression and maturation of this virus. In addition, the cytopathic changes produced by ts&JHM in NN cells were slowly progressive and resulted in the accumulation of abnormal whorls of cytoplasmic membranes. Such accumulation was not observed in fibroblastoid cell lines. In vivo, oligodendrocytes infected with JHM strains appear to lay down more myelin lamellae (Powell and Lampert, 1975) . Thus membrane proliferation induced by MHV infection may be a specific event occurring in differentiated glial cells.

Analysis of RNA and protein synthesis of ts8-JHM might elucidate the molecular basis of the ability of this mutant to produce chronic demyelination in vivo. There is no evidence yet that the host immune response plays a critical role in the induction of virus persistence in vivo since immunosuppression does not alter the course of the disease (Weiner, 1973) . Rather, our in vitro study demonstrates differences in viral tropism and cytopathic changes between neurons and NN cells which may be correlated with differences in disease patterns in vivo. Our study also points to the advantages of using cultures of specialized nerve cells to elucidate virus-nerve cell interactions.

Advances in Cellular Neurobiology

An electron microscope study of the development of mouse hepatitis virus in tissue culture cells

Antibody-induced modulation of rhabdovirus infection of neurons in vitro

Defective interfering particles modulate VSV infection of dissociated neuron cultures

19'78). Temperature-sensitive mutants of mouse hepatitis virus produce a high incidence of demyelination

Mouse hepatitis virus-induced recurrent demyelination

Evolution of a coronavirus during persistent infection in vitro

Analysis of the functions of coronavirus glycoproteins by differential inhibition of synthesis with tunicamycin

Tunicamycin resistant glycosylation of a coronavirus glycoprotein: Demonstration of a novel type of viral glycoprotein. %o.!ogv 11

Visna virus induced fusion of nerve cells in vitro

Host and virus factors associated with CNS cellular tropism leading to encephalomyelitis or demyelination induced by the JHM strain of mouse hepatitis virus

Selective localization of wild type and mutant mouse hepatitis virus (JHM strain) antigens in CNS tissue by fluorescence, light and electron microscopy

Mouse hepatitis virus type 4 (JHM strain) induced fatal central nervous system disease. I. Genetic control and the murine neuron as the susceptible site of disease

Mechanism of demyelination in JHM virus encephalomyelitis. Electron microscopic studies

In vivo and in vitro models of demyelinating disease: tropism of the JHM strain of murine hepatitis virus for cells of glial origin

In vivo and in vitro models of demyelinating diseases. II. Persistence and host-regulated thermo-sensitivity in cells of neural derivation infected with mouse hepatitis and measles viruses

Oligodendrocytes and their myelin-plasma membrane connections in JHM mouse hepatitis virus encephalomyelitis

Neurovirulence mutant of vesicular stomatitis virus with an altered target cell tropism in vivo

Chronic measles virus infection of mouse nerve cells in vitro

Intramembrane damages occurring during maturation of herpes simplex virus type I: Freeze-fracture study

Stability of neurotropic mouse hepatitis virus (JHM strain) during chronic infection of neuroblastoma cells

Chronic central nervous system demyelination in mice after JHM virus infection

Enhanced growth of a murine coronavirus in transformed mouse cells

Characterization of a coronavirus. II. Glycoproteins of the viral envelope: Tryptic peptide analysis

Isolation of coronavirus envelope glycoproteins and interaction with the viral nucleocapsid

Pathogenesis of demyelination induced by a mouse hepatitis virus (JHM virus)

We wish to thank Ray Rusten,