- Email: [email protected]
Persistent poliovirus infection: Establishment and maintenance involve distinct mechanisms
VIROLOGY
188,398-408 (1992)
Persistent Poliovirus Infection: Establishment and Maintenance Involve Distinct Mechanisms S. BORZAKIAN, T. COUDERC, Y. BARBIER, G. AT-TAL, I. PELLETIER, AND F. COLBERE-GARAPIN’ Unit6 de Virologie Mgdicale, Received August
lnstitut Pasteur, 75724 Paris cedex 15, France 15, 1991; accepted
October 21, 7991
Mutants of poliovirus (PV) with highly modified biological properties can be selected in vitro in cells of neural origin. Mutations accumulate in the genome of type 1 PV strains selected in human neuroblastoma cells, modifying cell specificity and conferring to the virus the ability to persist in such nonneural cells as HEp-2c (Pelletier et a/,, virology 180,729 1991). With this cell system, we have both parent lytic strains and persistent PV mutants; these were used to study the mechanisms of the establishment and maintenance of the persistent infection. We found that a persistent infection was established when the lytic potential of the virus was reduced; this involved both an early and a late event of the virus cycle for the type 1 mutants. In contrast, maintenance of the infection did not correlate with the reduced lytic potential of the viruses, but rather with the selection of mutant cell populations of various phenotypes. Two cell lines, representative of two phenotypes, were studied in greater detail. In the first one, HEpS32 (cl7), the PV receptor was not detected by cytofluorometry and viral genomes were detected by in situ hybridization in 2% of the cells. In the second cell line, HEp-S31 (cl1 8), 97% of the cells expressed the PV receptor, viral genomes were detected in 9-l 0% of the cells, and viral antigens in 5-10% of the cells. With this cell line, the cure of the culture or, alternatively, the lysis of the majority of cells, could be induced under specific culture conditions. We propose a model involving an equilibrium between an abortive and a lytic infection to explain the properties of cells persistently infected with PV. Q 1992 Academic Press. Inc.
ruses, such as Theiler’s virus and echovirus 6, interferon and defective particles, respectively, were demonstrated to play a role in in vitro persistence (Roos et a/., 1982; Gibson and Rigthand, 1985). A mechanism of viral interference was proposed to explain the establishment of a persistent infection in HeLa cells cotransfected with a PV viral genome and a subgenomic replicon (Kaplan et a/., 1989). However, in the examples cited above, the mechanisms of persistence were difficult to elucidate because the properties of defined plaque-purified mutants with a persistent phenotype in vitro could not be compared to those of the corresponding parent lytic viral strains. We have recently developed an in vitro model to study the ability of PV to persist in human cell cultures: wild-type as well as attenuated strains of PV can establish a persistent infection in human neuroblastoma cells, and PV mutants with altered biological properties can be isolated from these cells (Colbere-Garapin et al., 1989). The persistent virus variants, which have a highly mutated genome, also have a modified cell specificity and can induce a secondary persistent infection in cells of nonneural origin (Pelletier et a/., 1991). We have chosen to study the mechanisms of in vitro persistence in nonneural HEp-2c cells, because in this cell line, the wild-type and the attenuated PV strains are lytic, while the mutants derived from them by selection in neuroblastoma cells are persistent. We report
INTRODUCTION Poliovirus (PV) is an enterovirus of the family Picomaviridae, that causes paralytic poliomyelitis. In the natural course of either clinical or subclinical infection, virus is present in the pharynx for 1-2 weeks and persists in the intestine for several weeks (reviewed by Melnick, 1990). A late post-polio syndrome has been described: new deterioration of motor neurons emerges an average of 30 years after the initial infection (Campbell et al., 1969). One of the hypotheses proposed to account for this syndrome is a persistent PV infection (Dalakas, 1986). Sharief et a/. (199 1) recently reported an intrathecal immune response against poliovirus in many patients with the post-polio syndrome. It thus appears important to know whether PV can persist in neurons or in other cells. Many picornaviruses establish a persistent infection in viva and/or in vitro (Brahic and Stroop, 1981; Carp, 1981; Roos et al., 1982; Vallbracht et a/., 1984; Gercel et a/., 1985; De la Torre et al., 1985; Colbere-Garapin et a/., 1989). The carrier state of foot-and-mouth disease virus (FMDV) in BHK21 cells has been studied in detail, and it was shown that the coevolution of cells and virus contributes to the maintenance of persistence (De la Torre et a/., 1989a). For other picornavi’ To whom requests for reprints should be addressed. 0042.6822l92
$3.00
Copyright Q 1992 by Academic Press, Inc. All rights of reproduction in any form resewed.
398
PERSISTENT
POLIOVIRUS
TABLE 1
Virus type
PV straina
1 3 1 1 3 3 3 3 3
Sl s3 s12 s12 531 S32 S32 532 S32
IMR-Sl IMR-S3 HEp-S12 c-HEp-S 126 HEp-S31 (~118) HEp-S32 (~17) HEp-S32 (cl+39”’ HEp-S32 (~18) c-HEp-S32’
Virion production + + + + + + -
a PV strain used to establish the persistent infection. S12 is a PV mutant isolated from Sl-persistently infected IMR-32 neuroblastoma cells; 531 and S32 are PV mutants isolated from S3 persistently infected IMR-32 cells. ’ c-HEp-St 2 and c-HEp-S32 were spontaneously cured of virus one year and 5 weeks after the primary infection, respectively. ’ HEp-S32 (cl7)-39” was cured after 3 passages at 39”.
here that the establishment of PV persistence in HEp2c cells depended on the virus strain-several steps of the virus cycle were involved for type 1 mutants-and on the multiplicity of infection. The maintenance of PV persistence, on the contrary, did not depend on the reduced lytic phenotype of the mutant viruses, but involved the selection of host cells in which PV multiplication was restricted. MATERIALS
399
assay was performed on HEp-2c and LM cells according to Couderc et al. (1990).
PV PERSISTENTLYINFECTEDCELL LINES Persistently infected cells (cell clone)
INFECTION
AND METHODS
Virus and cells PV type 1 Mahoney and Sabin 1 (Sl) strains and PV type 3 Leon 37 and Sabin 3 (S3) strains were used. The persistent PV mutants Sl 1 and S12 were isolated from IMR-32 cells persistently infected with Sl while mutants S31 and S32 were isolated from IMR-32 cells persistently infected with S3 (Pelletier ef al., 1991). The PV persistently infected cell lines used in this study are given in Table 1. Human HEp-2c and murine thymidine kinase negative (tk-) LM cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% newborn calf serum. Thymidine kinase negative LM cells transformed by plasmid pAG60, which carries the gene of resistance to the antibiotic G418, were designated LM/pAGGO and were grown in the same medium supplemented with G418 (Geneticin, GIBCO) (Colb&e-Garapin et al., 1981). The human neuroblastoma cell line IMR-32 (Tumilowicz et al., 1970) was grown in DMEM with 10% fetal calf serum. Except when specified, virus infectivity was titered on IMR-32 cells by an endpoint micromethod. The virus binding
Effect of the multiplicity of infection on the establishment of a PV persistent infection and resistance of PV persistently infected cells to superinfection HEp2c cells (1 04) in wells of 96-well plates were infected with the PV persistent Sll or S12 mutant or with the parent Sl strain diluted 1O-fold from 1O4to 0.1 ID50 (the infectious doses were determined with HEp2c cells). Seven days after infection, the percentage of wells demonstrating cytopathic effects (c.p.e.) and growing cells surviving infection was determined for each virus dilution. The resistance of PV persistently infected cells to superinfection was tested with a heterotypic wild-type or persistent virus at a multiplicity of 10 ID50 per cell. The progeny of the superinfecting virus was titered in the presence of a polyclonal serum neutralizing the endogenous persisting virus. Analysis
of RNA by Northern
blot and dot blot
Cytoplasmic RNA was extracted either from persistently infected cells or from cells 5 hr after infection at a multiplicity of 0.1 ID50 per cell, as described (Pelletier et al., 1991). Viral RNA was either run on a 1% agarose-O.66 M formaldehyde gel and transferred to .a nylon membrane, or directly spotted onto the membrane using a Manifold dot blot apparatus (Bethesda Research Laboratories). Blots were hybridized with a 32P-labeled probe produced by random primer synthesis (Feinberg and Vogelstein, 1983) with the cloned Sl cDNA carried by plasmid pV(Sl)IC-O(T) (Kawamura et al., 1989) or the cloned S3 cDNA carried by plasmid FLC3 (Cann et a/., 1983), as described (Pelletier et al., 1991). The two plasmids were kindly provided by A. Nomoto (Tokyo) and 1. Almond (NIBSC, London), respectively. Indirect
immunofluorescence
and cytofluorometry
The presence of the PV receptor was tested by fluorescent labeling of cell membranes. Cells were washed with DMEM and incubated for 1 hr at 0” with the monoclonal antibody (mAb) 280 (kindly provided by P. Minor, NIBSC, London), which specifically binds to the PV receptor (Minor et al., 1984). Cells were then washed with DMEM and stained with FITC-labeled anti-mouse immunoglobulin at 0” for 1 hr, washed again, and fixed with formaldehyde (10%) in PBS. The cell fluorescence was either examined directly by fluorescence microscopy, or analyzed by cytofluorometty with a flow cytometer (Cytofluorograf IIS, Ortho-Diagnostic Systems), equipped with an argon ion laser
400
BORZAKIAN
ET AL.
(Lexer) and a computer station (MP2150, Ortho-Diagnostic Systems). Cytoplasmic immunofluorescence was performed with an anti-Sabin 3 (D particles) rabbit serum as described (Colbere-Garapin et a/., 1988, 1989). Briefly, cells were fixed in 3% paraformaldehyde, treated with 0.5% Triton X100, washed, incubated in 50 mM NH,CI, and washed again. Fixed cells were successively incubated with the first antibodies, then with FITC-labeled anti-rabbit immunoglobulin, each for 60 min at 37”. In situ hybridization Cells were fixed and their permeability was increased as described (Couderc eta/., 1989). The probe consisted of the complete cDNA of Sabin 3 isolated from the FLC3 plasmid (Cann et a/., 1983) and labeled with [35S]dCTP or C3H]dCTP(Amersham) to a specific activity of 4 X 1O* or 2 X 1O7d.p.m./pg, respectively. Probe concentration was 0.1 to 1 rig/J in hybridization mixture consisting of 50% deionized formamide, 10% dextran sulfate (Pharmacia), 10 mM Tris-HCI, pH 8, 1 mM EDTA, 0.6 M NaCI, 1X Denhardt’s medium, 100 pg/ml of denatured sonicated calf thymus DNA, 50 PgI ml of poly(A), and 300 pg/ml of total RNA extracted from uninfected HEp-2c cells. Hybridization was carried out for 48 hr at 25”, and the slides were washed as described (Couderc eta/,, 1989) with minor modifications. Briefly, slides were dipped in 2X SSC once for 5 min at 25” and twice for 30 min at 50”. Slides were further washed in medium containing 50% formamide, 0.6 M NaCI, 10 mM Tris-HCI, pH 7.4, and 1 mM EDTA at 25“ once for 1 hr, then overnight. Slides were then dipped in 0.1 X SSC for 30 min at 25”, dehydrated in ethanol baths, coated with NTB-2 Kodak emulsion, and stored at 4” for 1 to 72 hrwith the 35S-labeled probe and for 1 to 21 days with the 3H-labeled probe. After development, cells were counterstained with Giemsa. Somatic cell hybridization Equal numbers of PV persistently infected HEp-2c cells (virus+, tk+, G418-sensitive) and LM/pAGGO cells (virus-, tk-, G418-resistant) were seeded together or in separate flasks so that cells were confluent 4 hr later when they became attached. Cells were then washed and treated with 50% polyethylene glycol 1000 (Merck) in DMEM for 1 min at 20”. Cells were washed twice again and fed with DMEM containing 10% newborn calf serum. Twenty-four hours after treatment, a selective medium containing hypoxanthin, aminopterin, thymidine, and G418 was added to the cells as described (Colbere-Garapin et a/., 1981).
-1
0
1
log
2
3
4
ID50/well
FIG. 1. Effect of the multiplicity of infection on the establishment of a PV persistent infection in HEp-2c cells. HEp-2c cells (104) in wells of 96-well plates were inoculated with Sl 1 (-) or S12 (. . +) at the indicated infectious doses (determined with HEp-2c cells). Seven days after infection, the percentage of wells demonstrating cytopathic effects and growing cells was determined for each virus dose.
RESULTS Establishment of PV persistence in HEp-2c cells The persistent PV mutants selected in neuroblastoma cells do not lyse HEp-2c cells efficiently (ColbereGarapin et al,, 1989). Experiments were performed to determine more precisely the effect of the multiplicity of infection (m.o.i.) on the probability of cells to survive infection, despite c.p.e. As shown in Fig. 1, when HEp2c cells were infected with Sl 1 or S12 virions, some wells contained growing cells 7 days after infection despite the presence of other cells with c.p.e. In control plates in which cells were infected under the same conditions with a lytic strain such as Sl, no growing cells were detected 7 days after infection in wells containing cells with c.p.e., independently of the m.o.i. With the persistent viruses, the proportion of wells with c.p.e. and growing cells increased when the m.o.i. decreased (Fig. 1). Growing cells surviving infection were observed when a m.o.i. equal to or lower than 0.1 ID50/HEp-2c cell was used. Both Sl and S12 viruses were persistent in IMR-32 cells, while only S12 was persistent in HEp-2c cells. Under the conditions of a single-step virus growth cycle, the accumulation of viral proteins and virions in HEp-2c cells was similar for both persistent and lytic type 1 PV strains (Pelletier et al., 1991). However, as shown above, the high m.o.i. required for a single-step growth cycle is not compatible with the establishment
PERSISTENT
POLIOVIRUS
TABLE 2 AVERAGENUMBER OF VIRAL GENOME EQUIVALENTSIN CELLS INFECTED WITH PERSISTENTMUTANTS OR CONTROL Sl STRAINS
Cells
Virus
Time after infection
HEp-2c HEp-2c IMR-32 IMR-32
Sl s12 Sl s12
5 5 5 5
HEp-S12 HEp-S31 (cl1 8) HEp-S32 (~17) IMR-St
s12 s31 S32 Sl
>l >l >l >l
hr hr hr hr year year year year
Virus genome equivalents per infected cell 139,300 39,600 18,000 311,500 10,500 41,000 3,100 3,300
a The data are the mean values of four independent experiments. In PV persistently infected HEp cells, however, as mentioned under Results, the average number of viral genome equivalents per cell depended on specific culture conditions, such as the cell incubation temperature or the frequency of cell subculturing.
of a persistent infection. Thus, to find out whether or not the early steps of virus multiplication are involved in the establishment of the persistent infection in HEp-2c cells, we compared the accumulation of viral RNA during the first 5 hr after infection at a multiplicity of 0.1 ID50 per cell, for a type 1 persistent mutant (S12) and for the parent Sl strain. The results, shown in Table 2, indicated that 5 hr after infection, the average number of genome equivalents per infected cell was 3.5 times lower for the persistent S12 virus than for the parent lytic Sl strain in HEp-2c cells. This difference was amplified with time: there was 25-fold less viral RNA 20 hr after infection with S12 than with Sl (not shown). Furthermore, 5 hr after infection, the amount of S12 RNA was 7.8 times lower in HEp-2c than in IMR-32 cells, while the amount of Sl RNA was 7.7 times higher in HEp-2c than in IMR-32 cells. This result is probably related to the fact that S12 was selected in IMR-32 cells (Pelletier et al., 1991) while Sl was usually grown in HEp-2c cells. Therefore, the establishment of PV persistence in HEp-2c cells depended both on whether the viral strain was persistent or lytic and on the m.o.i. The reduced potential of persistent PV strains to lyse HEp-2c cells probably involved both an early step of virus infection, resulting in reduced viral RNA accumulation, as shown here (Table 2) and a late event corresponding to the delayed and inefficient virus liberation from HEp-2c cells, as shown previously (Pelletier et al., 1991). The early step of virus infection more probably concerned early biosynthetic events rather than early entry events, since the efficiency of attachment of [35S]methionine-labeled virus on HEp-2c cells was found to be identical for Sl and S12 (not shown).
401
INFECTION
To determine whether, upon coinfection, it is the persistent or the lytic phenotype that is dominant early in infection, lytic PV strains were titered in the absence or in the presence of a persistent mutant. The persistent mutants, even when present in large excess, did not interfere with the lytic effect of wild-type or attenuated PV strains (Table 3). Consequently, early in infection, the lytic phenotype was dominant. To determine whether this changed later on after infection, HEp-2c cells were first infected with S31 or S32 persistent mutants at a multiplicity of infection of 0.1 ID50 per cell, then challenged 1 to 30 days later with the type 3 wild-type Leon strain at the same multiplicity. When the superinfection occurred during the first 5 to 7 days after the primary infection, no cell survived infection with the Iytic strain. Later, cells became progressively resistant to superinfection, although during the first 30 days after the primary infection, more c.p.e. was observed in superinfected cells than in control cells infected with the persistent virus alone (not shown).
Maintenance of the PV persistent infection in HEp2c cells To determine the degree of resistance of PV persistently infected HEp-2c cells to superinfection, cultures
TABLE 3 TEST OF INTERFERENCEBETWEENPERSISTENTPV MUTANTS AND WILDTYPEOR ATTENUATEDPV STRAINSIN COINFECTEDHEP-2c CELLS PV inoculum Wild-type or attenuateda Sl Sl Mahoney Mahoney s3 s3 s3
Leon 37 Leon 37 Leon 37
+
+ + +
+ +
Persistent (ID50/well)*
Virus titer (log ID50/ml)
-
9.55 9.3 <1.3 10.05 10.05 9.55 9.05 9.3 <1.3 <1.3 7.92 7.8 8.05
S12(50) S12(50) S12(50) S31(500) S32(5) S31(500) S32(5) S31(500) S32(5)
a Serial 1 O-fold dilutions of wild-type or attenuated PV were titered in the absence or in the presence of persistent PV strains. * Constant doses of persistent virus (determined on IMR-32 cells and given in parentheses) were added to each well. The maximum dose which did not generate c.p.e. in HEp-2c cells was chosen for each persistent virus. This dose was higher for S31 than for S32 because S32 was much lytic on HEp-2c cells than S31 (Pelletier et al., 1991).
402
BORZAKIAN
persistently infected with PV type 1 or PV type 3 were superinfected with heterotypic wild-type or persistent PV. Superinfected cultures of cells continuously producing virus and mock-superinfected cultures had the same appearance under the microscope. The average production of the superinfecting wild-type or persistent virus, titered in the presence of antibodies neutralizing the heterotypic endogenous virus, was very low or undetectable (Table 4). In contrast, spontaneously cured c-HEp-S32 cells, which had been persistently infected by S32 (Table l), demonstrated c.p.e. upon superinfection, and the superinfecting virus multiplied in these cells almost as efficiently as in HEp-2c cells (Table 4). Since the establishment of the persistent infection in HEp-2c cells depended on the cell specificity, i.e., the reduced Iytic potential of the PV mutants in these cells, we looked at the cell specificity of the mutants reisolated from cultures of HEp cells persistently infected for 6 months. The results, shown in Table 5, indicated that the mutants selected in HEp-2c cells after 6 months of persistent infection had recovered the ability to lyse HEp-2c cells as efficiently as IMR-32 cells. Consequently, persistently infected HEp-2c cells were partially resistant to reinfection by the virus they generated, while uninfected HEp-2c cells were fully permissive to the same virus. Three attempts to establish a persistent infection in HEp-2c cells with PV mutants isolated 6 months postinfection from persistently infected HEp-2c cells were unsuccessful. The time during which infectious virus or viral RNA could remain strictly intracellular was determined by growing the cells in an excess of neutralizing antibodies, as described (Pelletier et a/., 1991) and by following virion production after extracellular virus had been TABLE 4 RESTRICTEDGROWTHOF A SUPERINFECTINGPV IN PERSISTENTLY INFECTEDHEP-2 CELLS Average production
Virus
HEp-2c
HEp-S12
Mahoney Sll Leon 37 S32
0 2700 1000 2000 ND
100 NDa ND 0.5* ND
of virus per cell per 24 hr in HEpS31 (cl1 8) 4.5 1 3b 1.3b ND ND
c-HEp-S32 0 ND ND 600 2000
’ Not determined. b This value represents only the progeny of the superinfecting virus, titered in the presence of antibodies neutralizing the heterotypic endogenous virus of persistently infected cells. Infectious titers were determined on IMR-32 cells by a micromethod and expressed in ID50.
ET AL. TABLE 5 REVERSIONOF PV MUTANTS IN PERSISTENTLYINFECTEDHEP-2c CELLS TO A PHENOTYPELMIC FOR HEP-2c CELLS Difference in titer (log ID50/ml) on IMR-32 and HEp-2c cells of the virus isolated
Viral strain
From persistently infected IMR-32 cells
After 6 months of persistent infection in HEp-2c cells
Sll 512 s31 S32
1.75 3.25 4 3.4
0.5 0.25 -1.25 0.75
completely neutralized. The results indicated that, in both HEp-2c and IMR-32 cells that had been persistently infected for 8 to 10 months, the persistent vi: ruses-or infectious genomes-could remain intracellular for as long as 10 days (not shown). When HEpS12 cells were grown for 20 days in the presence of neutralizing antibodies, then for 30 days in the absence of antibodies, intracellular viral RNA was no longer detected by Northern blotting. Since the maintenance of the persistent infection was clearly independent of the lytic potential of the mutants in HEp-2c cells, we tried to elucidate the mechanism of persistence in these cells by looking for the presence of the PV receptor at the surface of infected cells. The percentage of persistently infected cells expressing the PV receptor at their surface varied greatly with the cell clones. For example, in HEp-S32 (~17) the PV receptor was not detected either by immunofluorescence (not shown) or by cytofluorometry (Table 6); the results obtained with this clone were similar to those obtained with PV receptor-negative murine LM cells. In contrast, almost all the HEp-S31 (cl1 8) cells expressed the PV receptor, although the average fluorescence intensity per positive cell was only 40% of that of HEp-2c cells (Table 6, Fig. 2). For comparison, in PV persistently infected neuroblastoma cells, onethird to one-half of the cells expressed the PV receptor, and the average fluorescence intensity per positive cell was slightly lower than that of uninfected IMR-32 cells (Table 6). Thus, the reduced expression of the PV receptor may be responsible for, or may contribute to, the maintenance of the persistent infection in some clones, but it does not seem to be an absolute requirement for maintaining persistence. Total viral RNA synthesis was evaluated by dot blot analysis to determine whether viral replication is blocked in PV persistently infected clones. The average number of genome equivalents per cell varied
PERSISTENT
POLIOVIRUS
TABLE 6 DETECTIONBYCYTOFLUOROMETRYOFTHEPV RECEPTORATTHESURFACE OF PERSISTENTLYINFECTEDCELLS
Cells HEp-2cb HEp-2c LMC c-HEp-S12 c-HEp-S32 HEpS32 (~17) HEpS31 (~118) IMR-32’ IMR-32 LM” IMR-Sl IMR-S3
mAb 280
% of receptor + fluorescent cells
+ -
99.6 1.1 7.4 100 99.5 5 96.8
326 24 242 226 52 131
86.3 3.3 5 33 54.3
267 -
+ + + + + + + + +
Average fluorescence per positive cella
218 249
a In arbitrary units. b In the upper and lower part of the table, the conditions of cytofluorometry were optimized for HEp-2c and IMR-32 cells, respectively. ’ LM cells are PV receptor-negative mouse cells, used as a negative control.
greatly, between 3,100 and 41,000, according to the clone (Table 2). In situ hybridization experiments revealed the distribution of viral genomes in the cell population to be quite heterogeneous. In HEp-S31 (cl1 8) 9-10% of the cells harbored most of the viral genomes, and in half of the positive cells, the number of viral genomes was the same as that found in HEp-2c cells 4 hr after infection (Fig. 3). However, cells representative of all infection steps could be detected. Only 2% of the HEp-S32(cl7) cells had viral genomes as detected by in situ hybridization and these genomes were often localized in specific regions of the cytoplasm (Fig. 3F). In IMR-S3 cells, viral genomes were detected in 0.1% of the cells (Fig. 3H). The detection of viral antigens by indirect immunofluorescence on HEp-S31 (~118)cells confirmed the results of in situ hybridization experiments: the distribution of viral antigens was heterogeneous and 5 to 10% of cells were brilliantly stained (Fig. 4). This result indicated that, despite the presence of the PV receptor at the surface of 97% of HEp-S31 (cl1 8) cells, virus multiplication was blocked in the majority of cells. This block was, in fact, unstable. For example, when cells were fed but not subcultured or when they were split in a ratio greater than 1 in 4, all cells entered the lytic infection cycle, indicating that all of them were potentially permissive to PV. To determine whether this block was due to the lack of a host factor required for PV replication, or to a domi-
403
INFECTION
nant type of inhibition of PV multiplication, somatic cell hybrids were generated by fusion of persistently infected HEp-2c cells (thymidine kinase-positive, G418sensitive) and murine LM cells transformed by plasmid pAG60 (thymidine kinase-negative, G418-resistant) (Colbere-Garapin et al., 1981). The only factor limiting PV multiplication in murine cells is the absence of the virus receptor (Holland et al., 1959; Mendelsohn eta/., 1989). During the 10 days following fusion of HEp-S31 (cl1 8) with the murine cells, no increase in virus production was noted, as compared to the virus production of the clones alone, for which the same number of cells was treated under the same conditions (not shown). These results suggested that the mechanism of inhibition of PV multiplication in HEp-S31 cells was dominant. Spontaneous or induced cure of some cell lines HEp-2c cells infected with the persistent PV mutants were occasionally spontaneously cured of PV, between 5 weeks and 18 months after infection. The spontaneous cure of infected IMR-32 cells, however, was never observed (Colbere-Garapin eT a/., 1989). Some of the spontaneously cured HEp cell lines that we tested expressed at their surface the PV receptor (Table 6); they were free of viral RNA (Fig. 5) and they were permissive to superinfection (Table 4). The virion production by HEp-S32 (~17)cells highly depended on the temperature at which cells were incubated (Fig. 6). The virus produced by this clone was found to be more temperature-sensitive than the original S32 strain: its titer was 1.2 log lower at 39” than at 34”, while the titer of S32 was the same at both temperatures (Pelletier et al., 1991). Similar results were obtained with other cell lines: HEp-S12, HEp-S3l(cl18), and HEp-S32(cl8) (not shown). When HEp-S32 (~17) cells were grown for more than three passages at 39’, they were cured of virus. The lytic Mahoney and Sl strains multiplied in the cured cells without cytopathic effects, but the average virion production per cell per 24 hr was about 3 log lower in these cells than in HEp2c cells. These results indicated that cells partially resistant to PV infection had been selected during the persistent infection. As expected, the expression of the PV receptor was as low at the surface of the cured cells as on the parent HEp-S32 (~17)cells (not shown). DISCUSSION The PV persistent infection of HEp-2c cells has similarities with FMDV persistent infection of BHK21 cells, for which it was proposed that cell diversity may be an important element for long-term virus and cell survival (De la Torre et a/., 1989a). However, in the attempt to
404
FIG. 2. Detection of the PV receptor at the surface of cells by Indirect immunofluorescence with mAb 280 (Minor eta/., 1984). The PV receptor was detected on uninfected HEp-2c cells (A), HEp-S31 (cl1 8) cells (C, D), and c-HEp-S32 cells (F). The receptor was not detected on murine LM cells with mAb 280 (6) or on HEp-S31 (cl1 8) when a control mAb was used (E). The bar indicates 10 pm.
clarify the mechanisms of virus persistence, the PV system has one advantage over systems developed for other picornaviruses: both fully lytic and persistent mutants are available, allowing their phenotypes to be directly compared. In the present work, the virus hostcell interactions were studied during the establishment of the PV persistent infection, i.e., from the time of infection up to a few weeks after infection, and during the maintenance of the persistent infection, i.e., from 2 months to 2 years after infection. With this system, it was possible to distinguish the effects of mutations in the viral genome from those of modifications in the host cell phenotype. We have found that the attachment of the persistent mutant S12 and of the parent lytic Sl strain to HEp-2c cells was the same. At a low m.o.i., the accumulation of viral RNA during the first replication cycle, 5 hr after infection, was 3.5 times lower in HEp-2c cells infected with S12 than in cells infected with St. This difference was amplified with time. This result suggested that the difference between the mutant virus and its parent affects an early biosynthetic event in virus infection in
HEp-2c cells. At a high m.o.i., no delay in the early steps of virus multiplication was noticed for type 1 mutants in HEp-2c cells (Pelletier et al., 1991). It seems probable that the infection at a high multiplicity allowed a greater number of viral genomes to be replicated and expressed in these cells, leading to cell death, despite the delay in virus liberation. Therefore, both early steps of the viral cycle, as shown here, and late steps, as shown previously (Pelletier et a/., 1991), seem to be implicated in the establishment of PV persistence in HEp-2c cells. During this period, the lytic phenotype of wild-type and attenuated strains was dominant, indicating that viral functions altered by the mutations of persistent viruses did not interfere with the lytic virus cycle. In contrast to the establishment of PV persistence, the maintenance of the infection was not correlated to the reduced Iytic potential of the PV mutants, as shown by two kinds of unrelated experiments: 1) the persistent viruses reverted toward a lytic phenotype in HEp2c cells (Table 5) and 2) cells producing virus were resistant, even at a high m.o.i., to superinfection by a
PERSISTENT
POLIOVIRUS
INFECTION
405
FIG. 3. Detection of viral genomes by in situ hybridization in PV-infected cells. The PV genome was hybridized to a probe consisting of the complete S3 cDNA labeled with [35S]dCTP as described under Material and Methods. (A) Uninfected HEp-2c cells, (B) HEp-2c cells 4 hr after infection with S3, (C, D) HEp-S3l(cl18), (E, F) HEp-S32(cl7), (G) uninfected IMR-32 cells, (H) IMR-S3 cells. The exposure time was 4 hr for HEp cells and 8 hr for IMR cells. The bar indicates 3 pm.
heterotypic wild-type PV strain (Table 4). The maintenance of the persistent infection correlated with the selection of cells in which PV multiplication was restricted. Two PV persistently infected cell clones, representative of two different types of cell selection, were studied in greater detail: HEp-S32 (~17) and HEp-S31 (cl1 8). In the first clone, the expression of the PV receptor was similar to the background level, preventing virus from efficiently propagating through the extracel-
lular medium. Such a down regulation of the expression of the PV receptor has already been found in HeLa cell clones resistant to PV infection (Kaplan et al., 1989). These results are compatible with a carrier state of persistent infection. This model, however, seems too restrictive to account for the results obtained with HEp-S3 1 (cl 18). In our second clone, HEp-S3 1 (cl 18) almost 100% of the cells expressed the PV receptor, but the multiplica-
406
BORZAKIAN
ET AL
3~. 4. Detection of viral antigens by indirect immunofluorescence of HEp-S3 1 (cl 18). The amount of viral antigen marked with an anti-S3 I-abbit ser urn in HEp-S31 (cl1 8) cells varied among the cells (A-D). Uninfected HEp-2c cells (E) and HEp-2c cells 72 hr after infection with S31 (F) were USC?d as negative and positive control, respectively. The bar indicates 10 pm.
tion of the endogenous virus-as well as that of a superinfecting lytic virus-was found to be restricted in the majority of cells (Figs. 3 and 4, Table 4). Results of cell fusion experiments with murine cells suggested that the intracellular block was not due to the lack of a host factor required for PV replication, but to a dominant type of inhibition. Although this result was obtained with a clone-producing virus, it is in agreement with previous results obtained with hybrids of somatic hamster cells resistant to FMDV (De la Torre et a/., 198913). The intracellular block of HEp-S31 (cl1 8) was unstable under some conditions of cell culture in which all cells entered a lytic infection cycle. This suggested that, at least in HEp-S31 (cl1 8) there is an equilibrium between an abortive and a complete Iytic infection, rather than a bona fide carrier state of persistent infection.
The possibility to cure the cells by growing them at 39” allowed the role of viruses and that of cells in the virus-host cell interactions of an established persistent infection to be further distinguished. With HEpS31 (~17) it was shown by superinfecting the heatcured cells with fully lytic PV strains that the resistance of cells to the PV lytic effects was independent of the virus and depended only on the properties of the cells selected by the virus. Thus, while the establishment of PV persistence mainly depended on the virus strain and on the multiplicity of infection, the maintenance of persistence involved mainly the properties of the cells selected during the persistent infection. The spontaneous cure of cells infected with FMDV has been observed repeatedly and has been found to result from increasing deletions occurring in the virus genomes over the course of sequential passages, so
PERSISTENT
POLIOVIRUS
that after passage 65 virions were no longer detectable (De la Torre et al., 1985). in our PV system, when the infectious titer decreased before the complete cure, we did not observe deleted viral RNA species of defined sizes; however, this does not exclude the possibility of small deletions, undetectable by Northern blot. The spontaneous cure of cells occurred not after a specific number of passages, but apparently at random, between 5 weeks and 18 months after infection. The spontaneous cure could be explained if a few cells escaped virus infec?ion, or if the virus cycle aborted in some of the infected cells. Because the infectious titers in supernatants of p&sistently infected HEp-2c cells were on the order of 10’ ID50/ml, the probability that a few cells escaped infection after several months seems very low. The model that we propose of an equilibrium between an abortive and a lytic infection could help to explain the spontaneous cure of cells: if, for example, the equilibrium is correlated to the number of viral genomes in a cell, it could be displaced toward the abortive cycle when this number is low, resulting in the cure of cells. According to this model, the cured cells might be susceptible to a superinfecting virus, as we observed (Table 4). The level of expression of one or a few host factor(s) involved in PV multiplication would be responsible for the equilibrium between the abortive
407
INFECTION
0
2
4
6
6
days FIG. 6. The virion production of HEp-S32(cl7) depended on the temperature of incubation. HEp-S32 (~17) cells were subcultured and replicate plates were incubated at 34” (0) 37“ (0). or 39” (0). The virion production was titered at the indicated days. The results are the average of two independent experiments.
and the lytic infection. This model is also compatible with the results obtained with HEp-S32(cl7), the limiting factor for this clone probably being the PV receptor. The results of several experiments are in agreement with the model of an equilibrium between an abortive and a lytic infection: the period during which the virus or its genome can remain intracellular (10 days), the high percentage of HEp-S31 (cl1 8) cells demonstrating viral genomes (9-l 0%) and viral antigens (5-l 00/o),the spontaneous cure of some cell lines despite their susceptibility to virus infection, and the possibility of displacing the equilibrium toward the cure of cells (by in- . cubating them at 39”) or toward the lytic cycle (by feeding the cells without subculturing them, or growing them at a very low density). More experiments, however, are required to confirm this model and to elucidate its mechanism at the molecular level. In conclusion, we have shown that the establishment and the maintenance of a PV persistent infection involve distinct mechanisms. We propose that the expression of a host factor involved in PV multiplication and the stability of this phenotype may be responsible for an equilibrium between an abortive and a complete lytic virus cycle in PV persistently infected cells. ACKNOWLEDGMENTS
FIG. 5. Northern blot analysis of total viral RNAs present in four persistently infected cell lines used in this study: HEp-S32 (~17) (12.5 pg); HEpS32 (~18) (30 pg); c-HEp-S32 (30 pg) and IMR-S3 (17 pg). HEp-2c cell polyA+ RNA was used as a negative control. The probe consisted of 32-P labeled plasmid FLC3 (Cann et a/., 1983) which carries the entire S3 cDNA. The arrowhead indicates the position of viral RNA isolated from Leon 37 purified virions. The other two arrows indicate the position of 28s (6.3 kb) and 18s (2.4 kb) ribosomal RNAs.
We are grateful to Professor F. Horaud for continuous interest in this work and critical reading of the manuscript. We are indebted to Professor P. Minor for the monoclonal antibody 280, to Professor A. Nomoto for the pVS(1) IC-O(T) plasmid, and to Professor J. Almond for the FLC3 plasmid. We thank P. Metezeau and H. Kieffer for help in cytofluorometry. We thank K. Pepper for reviewing the manuscript. The generosity of the Fondation Marcel Merieux and of the Ihstitut de Recherches lnternationales Servier & Compagnie Developpement is acknowledged for the fellowships attributed to S.B. and Y.B., respectively. This work was supported by a grant from the lnstitut
408
BORZAKIAN
National de la Sante et de la Recherche to F.C.-G.
MBdicale (Grant 88 1006)
REFERENCES BRAHIC, M., and STROOP,W. G. (1981). Theiler’s virus persists in glial cells during demyelinating disease. Cell 26, 123-l 28. CAMPBELL, A. M. G., WILLIAMS, E. R., and PEARCE, J. (1969). Late motor neuron degeneration following poliomyelitis. Neurology 19, 1101-1106. CANN, A. J.. STANWAY, G., HAUPTMANN, R., MINOR, P. D., SCHILD, G. C., CLARKE, L. D., MOUNTFORD. R. C., and ALMOND, J. W. (1983). Poliovirus type 3: Molecular cloning of the genome and nucleotide sequence of the region encoding the protease and polymerase proteins. Nucleic Acids Res. 11, 1267-l 281. CARP, R. I. (1981). Persistent infection of human lymphoblastoid cells with poliovirus and development of temperature sensitive mutants. Intervirology 15, 49-56. COLB~RE-GARAPIN,F., HORODNICEANU,F.. KOURILSKY,P., and GAFIAPIN, A. (1981). A new dominant hybrid selective marker for higher eukaryotic cells. J. MO/. Biol. 150, l-l 4. COLB~RE-GARAPIN, F., CHRISTODOULOU, C., CRAINIC, R., GAFIAPIN, A. C., and CANDREA,A. (1988). Addition of a foreign oligopeptide to the major capsid protein of poliovirus. Proc. Nat/. Acad. Sci. USA 85,8668-8672. COLB~RE-GARAPIN,F., CHRISTODOULOU,C., CRAINIC, R., and PELLETIER, I. (1989). Persistent poliovirus infection of human neuroblastoma cells. Proc. Natl. Acad. Sci. USA 86, 7590-7594. COUDERC,T., CHRISTODOULOU,C.. KOPECKA,H., MARSDEN, S., TAFFS, L. F., CRAINIC, R., and HORAUD, F. (1989). Molecular pathogenesis of neuronal lesions induced by poliovirus type 1. /. Gen. Viral. 70, 2907-2918. COUDERC, T., BARZU, T., HORAUD, F., and CRAINIC, R. (1990). Poliovirus permissivity and specific receptor expression on human endothelial cells. virology 174, 95-102. DALAKAS, M. C. (1986). New neuromuscular symptoms in patients with old poliomyelitis: a three year follow-up study. Eur. Neurol. 25,381-387. DE LATORRE,.I. C., DAVILA, M., SOBRINO,F.. ORTIN, J., and DOMINGO, E. (1985). Establishment of cell lines persistently infected with footand-mouth disease virus. Virology 145, 24-35. DE LA TORRE, 1. C., MARTINEZ-SALAS, E., DIEZ, J., and DOMINGO, E. (1989a). Extensive cell heterogeneity during persistent infection with foot-and-mouth disease virus. J. Viral. 63, 59-63. DE LATORRE, J. C., DE LA LUNA, S., DIEZ, J., and DOMINGO, E. (198913). Resistance to foot-and-mouth disease virus mediated by trans acting cellular products. /. Viral. 63, 2385-2387.
ET AL FEINBERG,A. P., and VOGELSTEIN,B. (1983). A technique for radlolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132, 6-l 3. GERCEL, C., MAHAN, K. B., and HAMPARIAN, V. V. (1985). Preliminary characterization of a persistent infection of HeLa cells with human rhinovirus type 2. /. Gen. Viral. 66, 131-l 39. GIBSON, 1. P.. and RIGTHAND,V. F. (1985). Persistence of echovirus 6 in cloned human cells. J. L&o/. 54, 2 19-223. HOLLAND, J. J., MCLAREN. L. C., and SYLVERTON,J. T. (1959). The mammalian cell virus relationship. IV. Infection of naturally insusceptible cells with enterovirus ribonucleic acid. 1. Exp. Med. 110, 65-80. KAPLAN, G., LEW, A., and RACANIELLO, V. R. (1989). Isolation and characterization of HeLa cell lines blocked at different steps in the poliovirus life cycle. /. Viral. 63, 43-51. KAWAMURA, N., KOHARA, M., ABE, S., KOMATSU, T., TAGO, K., ARITA, M., and NOMOTO, A. (1989). Determinants in the 5’ noncoding region of poliovirus Sabin 1 RNA that influence the attenuation phenotype. 1. Viral. 63, 1302-l 309. MELNICK, J. L. (1990). Enteroviruses: Polioviruses, coxsackieviruses, echoviruses, and newer enteroviruses. In “Virology,” 2nd ed. (B. N. Fields, D. M. Knipe et a/.), Raven Press, New York. MENDELSOHN, C., WIMMER, E.. and RACANIELLO,V. R. (1989). Cellular receptor for poliovirus: molecular cloning, nucleotide sequence and expression of a new member of the immunoglobulin superfamily. Cell 56, 855-865. MINOR, P. D., PIPKIN, P. A., HOCKLEY, D., SCHILD, G. C., and ALMOND, J. W. (1984). Monoclonal antibodies which block cellular receptors of poliovirus. Virus Res. 1, 203-212. PELLETIER,I., COUDERC, T., BORZAKIAN,S., WYCKOFF. E., CRAINIC, R., EHRENFELD,E., and COLBERE-GARAPIN,F. (1991). Characterization of persistent poliovirus mutants selected in human neuroblastoma cells. Virology 180, 729-737. Roos. R. P., RICHARDS, 0. C., GREEN, J., and EHRENFELD,E. (1982). Characterization of a cell culture persistently infected with the DA strain of Theiler murine encephalomyelitis virus. J. Vkol. 43, 1 1 181122. SHARIEF, M. K.. HENTGES, R., and CIARDI. M. (1991). lntrathecal immune response in patients with the post-polio syndrome. N. Engl. J. Med. 325, 749-755. TUMILOWICZ, J. J., NICHOLS, W. W., CHOLON, J. J., and GREENE, A. E. (1970). Definition of a continuous human cell line derived from neuroblastoma. Cancer Res. 30, 211 O-2 118. VALLBRACHT, A., HOFMANN, L., WURSTER, K. G., and FLEHMIG, B. (1984). Persistent infection of human fibroblasts by hepatitis A virus. J. Gen. Virol. 65, 609-6 15.
188,398-408 (1992)
Persistent Poliovirus Infection: Establishment and Maintenance Involve Distinct Mechanisms S. BORZAKIAN, T. COUDERC, Y. BARBIER, G. AT-TAL, I. PELLETIER, AND F. COLBERE-GARAPIN’ Unit6 de Virologie Mgdicale, Received August
lnstitut Pasteur, 75724 Paris cedex 15, France 15, 1991; accepted
October 21, 7991
Mutants of poliovirus (PV) with highly modified biological properties can be selected in vitro in cells of neural origin. Mutations accumulate in the genome of type 1 PV strains selected in human neuroblastoma cells, modifying cell specificity and conferring to the virus the ability to persist in such nonneural cells as HEp-2c (Pelletier et a/,, virology 180,729 1991). With this cell system, we have both parent lytic strains and persistent PV mutants; these were used to study the mechanisms of the establishment and maintenance of the persistent infection. We found that a persistent infection was established when the lytic potential of the virus was reduced; this involved both an early and a late event of the virus cycle for the type 1 mutants. In contrast, maintenance of the infection did not correlate with the reduced lytic potential of the viruses, but rather with the selection of mutant cell populations of various phenotypes. Two cell lines, representative of two phenotypes, were studied in greater detail. In the first one, HEpS32 (cl7), the PV receptor was not detected by cytofluorometry and viral genomes were detected by in situ hybridization in 2% of the cells. In the second cell line, HEp-S31 (cl1 8), 97% of the cells expressed the PV receptor, viral genomes were detected in 9-l 0% of the cells, and viral antigens in 5-10% of the cells. With this cell line, the cure of the culture or, alternatively, the lysis of the majority of cells, could be induced under specific culture conditions. We propose a model involving an equilibrium between an abortive and a lytic infection to explain the properties of cells persistently infected with PV. Q 1992 Academic Press. Inc.
ruses, such as Theiler’s virus and echovirus 6, interferon and defective particles, respectively, were demonstrated to play a role in in vitro persistence (Roos et a/., 1982; Gibson and Rigthand, 1985). A mechanism of viral interference was proposed to explain the establishment of a persistent infection in HeLa cells cotransfected with a PV viral genome and a subgenomic replicon (Kaplan et a/., 1989). However, in the examples cited above, the mechanisms of persistence were difficult to elucidate because the properties of defined plaque-purified mutants with a persistent phenotype in vitro could not be compared to those of the corresponding parent lytic viral strains. We have recently developed an in vitro model to study the ability of PV to persist in human cell cultures: wild-type as well as attenuated strains of PV can establish a persistent infection in human neuroblastoma cells, and PV mutants with altered biological properties can be isolated from these cells (Colbere-Garapin et al., 1989). The persistent virus variants, which have a highly mutated genome, also have a modified cell specificity and can induce a secondary persistent infection in cells of nonneural origin (Pelletier et a/., 1991). We have chosen to study the mechanisms of in vitro persistence in nonneural HEp-2c cells, because in this cell line, the wild-type and the attenuated PV strains are lytic, while the mutants derived from them by selection in neuroblastoma cells are persistent. We report
INTRODUCTION Poliovirus (PV) is an enterovirus of the family Picomaviridae, that causes paralytic poliomyelitis. In the natural course of either clinical or subclinical infection, virus is present in the pharynx for 1-2 weeks and persists in the intestine for several weeks (reviewed by Melnick, 1990). A late post-polio syndrome has been described: new deterioration of motor neurons emerges an average of 30 years after the initial infection (Campbell et al., 1969). One of the hypotheses proposed to account for this syndrome is a persistent PV infection (Dalakas, 1986). Sharief et a/. (199 1) recently reported an intrathecal immune response against poliovirus in many patients with the post-polio syndrome. It thus appears important to know whether PV can persist in neurons or in other cells. Many picornaviruses establish a persistent infection in viva and/or in vitro (Brahic and Stroop, 1981; Carp, 1981; Roos et al., 1982; Vallbracht et a/., 1984; Gercel et a/., 1985; De la Torre et al., 1985; Colbere-Garapin et a/., 1989). The carrier state of foot-and-mouth disease virus (FMDV) in BHK21 cells has been studied in detail, and it was shown that the coevolution of cells and virus contributes to the maintenance of persistence (De la Torre et a/., 1989a). For other picornavi’ To whom requests for reprints should be addressed. 0042.6822l92
$3.00
Copyright Q 1992 by Academic Press, Inc. All rights of reproduction in any form resewed.
398
PERSISTENT
POLIOVIRUS
TABLE 1
Virus type
PV straina
1 3 1 1 3 3 3 3 3
Sl s3 s12 s12 531 S32 S32 532 S32
IMR-Sl IMR-S3 HEp-S12 c-HEp-S 126 HEp-S31 (~118) HEp-S32 (~17) HEp-S32 (cl+39”’ HEp-S32 (~18) c-HEp-S32’
Virion production + + + + + + -
a PV strain used to establish the persistent infection. S12 is a PV mutant isolated from Sl-persistently infected IMR-32 neuroblastoma cells; 531 and S32 are PV mutants isolated from S3 persistently infected IMR-32 cells. ’ c-HEp-St 2 and c-HEp-S32 were spontaneously cured of virus one year and 5 weeks after the primary infection, respectively. ’ HEp-S32 (cl7)-39” was cured after 3 passages at 39”.
here that the establishment of PV persistence in HEp2c cells depended on the virus strain-several steps of the virus cycle were involved for type 1 mutants-and on the multiplicity of infection. The maintenance of PV persistence, on the contrary, did not depend on the reduced lytic phenotype of the mutant viruses, but involved the selection of host cells in which PV multiplication was restricted. MATERIALS
399
assay was performed on HEp-2c and LM cells according to Couderc et al. (1990).
PV PERSISTENTLYINFECTEDCELL LINES Persistently infected cells (cell clone)
INFECTION
AND METHODS
Virus and cells PV type 1 Mahoney and Sabin 1 (Sl) strains and PV type 3 Leon 37 and Sabin 3 (S3) strains were used. The persistent PV mutants Sl 1 and S12 were isolated from IMR-32 cells persistently infected with Sl while mutants S31 and S32 were isolated from IMR-32 cells persistently infected with S3 (Pelletier ef al., 1991). The PV persistently infected cell lines used in this study are given in Table 1. Human HEp-2c and murine thymidine kinase negative (tk-) LM cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% newborn calf serum. Thymidine kinase negative LM cells transformed by plasmid pAG60, which carries the gene of resistance to the antibiotic G418, were designated LM/pAGGO and were grown in the same medium supplemented with G418 (Geneticin, GIBCO) (Colb&e-Garapin et al., 1981). The human neuroblastoma cell line IMR-32 (Tumilowicz et al., 1970) was grown in DMEM with 10% fetal calf serum. Except when specified, virus infectivity was titered on IMR-32 cells by an endpoint micromethod. The virus binding
Effect of the multiplicity of infection on the establishment of a PV persistent infection and resistance of PV persistently infected cells to superinfection HEp2c cells (1 04) in wells of 96-well plates were infected with the PV persistent Sll or S12 mutant or with the parent Sl strain diluted 1O-fold from 1O4to 0.1 ID50 (the infectious doses were determined with HEp2c cells). Seven days after infection, the percentage of wells demonstrating cytopathic effects (c.p.e.) and growing cells surviving infection was determined for each virus dilution. The resistance of PV persistently infected cells to superinfection was tested with a heterotypic wild-type or persistent virus at a multiplicity of 10 ID50 per cell. The progeny of the superinfecting virus was titered in the presence of a polyclonal serum neutralizing the endogenous persisting virus. Analysis
of RNA by Northern
blot and dot blot
Cytoplasmic RNA was extracted either from persistently infected cells or from cells 5 hr after infection at a multiplicity of 0.1 ID50 per cell, as described (Pelletier et al., 1991). Viral RNA was either run on a 1% agarose-O.66 M formaldehyde gel and transferred to .a nylon membrane, or directly spotted onto the membrane using a Manifold dot blot apparatus (Bethesda Research Laboratories). Blots were hybridized with a 32P-labeled probe produced by random primer synthesis (Feinberg and Vogelstein, 1983) with the cloned Sl cDNA carried by plasmid pV(Sl)IC-O(T) (Kawamura et al., 1989) or the cloned S3 cDNA carried by plasmid FLC3 (Cann et a/., 1983), as described (Pelletier et al., 1991). The two plasmids were kindly provided by A. Nomoto (Tokyo) and 1. Almond (NIBSC, London), respectively. Indirect
immunofluorescence
and cytofluorometry
The presence of the PV receptor was tested by fluorescent labeling of cell membranes. Cells were washed with DMEM and incubated for 1 hr at 0” with the monoclonal antibody (mAb) 280 (kindly provided by P. Minor, NIBSC, London), which specifically binds to the PV receptor (Minor et al., 1984). Cells were then washed with DMEM and stained with FITC-labeled anti-mouse immunoglobulin at 0” for 1 hr, washed again, and fixed with formaldehyde (10%) in PBS. The cell fluorescence was either examined directly by fluorescence microscopy, or analyzed by cytofluorometty with a flow cytometer (Cytofluorograf IIS, Ortho-Diagnostic Systems), equipped with an argon ion laser
400
BORZAKIAN
ET AL.
(Lexer) and a computer station (MP2150, Ortho-Diagnostic Systems). Cytoplasmic immunofluorescence was performed with an anti-Sabin 3 (D particles) rabbit serum as described (Colbere-Garapin et a/., 1988, 1989). Briefly, cells were fixed in 3% paraformaldehyde, treated with 0.5% Triton X100, washed, incubated in 50 mM NH,CI, and washed again. Fixed cells were successively incubated with the first antibodies, then with FITC-labeled anti-rabbit immunoglobulin, each for 60 min at 37”. In situ hybridization Cells were fixed and their permeability was increased as described (Couderc eta/., 1989). The probe consisted of the complete cDNA of Sabin 3 isolated from the FLC3 plasmid (Cann et a/., 1983) and labeled with [35S]dCTP or C3H]dCTP(Amersham) to a specific activity of 4 X 1O* or 2 X 1O7d.p.m./pg, respectively. Probe concentration was 0.1 to 1 rig/J in hybridization mixture consisting of 50% deionized formamide, 10% dextran sulfate (Pharmacia), 10 mM Tris-HCI, pH 8, 1 mM EDTA, 0.6 M NaCI, 1X Denhardt’s medium, 100 pg/ml of denatured sonicated calf thymus DNA, 50 PgI ml of poly(A), and 300 pg/ml of total RNA extracted from uninfected HEp-2c cells. Hybridization was carried out for 48 hr at 25”, and the slides were washed as described (Couderc eta/,, 1989) with minor modifications. Briefly, slides were dipped in 2X SSC once for 5 min at 25” and twice for 30 min at 50”. Slides were further washed in medium containing 50% formamide, 0.6 M NaCI, 10 mM Tris-HCI, pH 7.4, and 1 mM EDTA at 25“ once for 1 hr, then overnight. Slides were then dipped in 0.1 X SSC for 30 min at 25”, dehydrated in ethanol baths, coated with NTB-2 Kodak emulsion, and stored at 4” for 1 to 72 hrwith the 35S-labeled probe and for 1 to 21 days with the 3H-labeled probe. After development, cells were counterstained with Giemsa. Somatic cell hybridization Equal numbers of PV persistently infected HEp-2c cells (virus+, tk+, G418-sensitive) and LM/pAGGO cells (virus-, tk-, G418-resistant) were seeded together or in separate flasks so that cells were confluent 4 hr later when they became attached. Cells were then washed and treated with 50% polyethylene glycol 1000 (Merck) in DMEM for 1 min at 20”. Cells were washed twice again and fed with DMEM containing 10% newborn calf serum. Twenty-four hours after treatment, a selective medium containing hypoxanthin, aminopterin, thymidine, and G418 was added to the cells as described (Colbere-Garapin et a/., 1981).
-1
0
1
log
2
3
4
ID50/well
FIG. 1. Effect of the multiplicity of infection on the establishment of a PV persistent infection in HEp-2c cells. HEp-2c cells (104) in wells of 96-well plates were inoculated with Sl 1 (-) or S12 (. . +) at the indicated infectious doses (determined with HEp-2c cells). Seven days after infection, the percentage of wells demonstrating cytopathic effects and growing cells was determined for each virus dose.
RESULTS Establishment of PV persistence in HEp-2c cells The persistent PV mutants selected in neuroblastoma cells do not lyse HEp-2c cells efficiently (ColbereGarapin et al,, 1989). Experiments were performed to determine more precisely the effect of the multiplicity of infection (m.o.i.) on the probability of cells to survive infection, despite c.p.e. As shown in Fig. 1, when HEp2c cells were infected with Sl 1 or S12 virions, some wells contained growing cells 7 days after infection despite the presence of other cells with c.p.e. In control plates in which cells were infected under the same conditions with a lytic strain such as Sl, no growing cells were detected 7 days after infection in wells containing cells with c.p.e., independently of the m.o.i. With the persistent viruses, the proportion of wells with c.p.e. and growing cells increased when the m.o.i. decreased (Fig. 1). Growing cells surviving infection were observed when a m.o.i. equal to or lower than 0.1 ID50/HEp-2c cell was used. Both Sl and S12 viruses were persistent in IMR-32 cells, while only S12 was persistent in HEp-2c cells. Under the conditions of a single-step virus growth cycle, the accumulation of viral proteins and virions in HEp-2c cells was similar for both persistent and lytic type 1 PV strains (Pelletier et al., 1991). However, as shown above, the high m.o.i. required for a single-step growth cycle is not compatible with the establishment
PERSISTENT
POLIOVIRUS
TABLE 2 AVERAGENUMBER OF VIRAL GENOME EQUIVALENTSIN CELLS INFECTED WITH PERSISTENTMUTANTS OR CONTROL Sl STRAINS
Cells
Virus
Time after infection
HEp-2c HEp-2c IMR-32 IMR-32
Sl s12 Sl s12
5 5 5 5
HEp-S12 HEp-S31 (cl1 8) HEp-S32 (~17) IMR-St
s12 s31 S32 Sl
>l >l >l >l
hr hr hr hr year year year year
Virus genome equivalents per infected cell 139,300 39,600 18,000 311,500 10,500 41,000 3,100 3,300
a The data are the mean values of four independent experiments. In PV persistently infected HEp cells, however, as mentioned under Results, the average number of viral genome equivalents per cell depended on specific culture conditions, such as the cell incubation temperature or the frequency of cell subculturing.
of a persistent infection. Thus, to find out whether or not the early steps of virus multiplication are involved in the establishment of the persistent infection in HEp-2c cells, we compared the accumulation of viral RNA during the first 5 hr after infection at a multiplicity of 0.1 ID50 per cell, for a type 1 persistent mutant (S12) and for the parent Sl strain. The results, shown in Table 2, indicated that 5 hr after infection, the average number of genome equivalents per infected cell was 3.5 times lower for the persistent S12 virus than for the parent lytic Sl strain in HEp-2c cells. This difference was amplified with time: there was 25-fold less viral RNA 20 hr after infection with S12 than with Sl (not shown). Furthermore, 5 hr after infection, the amount of S12 RNA was 7.8 times lower in HEp-2c than in IMR-32 cells, while the amount of Sl RNA was 7.7 times higher in HEp-2c than in IMR-32 cells. This result is probably related to the fact that S12 was selected in IMR-32 cells (Pelletier et al., 1991) while Sl was usually grown in HEp-2c cells. Therefore, the establishment of PV persistence in HEp-2c cells depended both on whether the viral strain was persistent or lytic and on the m.o.i. The reduced potential of persistent PV strains to lyse HEp-2c cells probably involved both an early step of virus infection, resulting in reduced viral RNA accumulation, as shown here (Table 2) and a late event corresponding to the delayed and inefficient virus liberation from HEp-2c cells, as shown previously (Pelletier et al., 1991). The early step of virus infection more probably concerned early biosynthetic events rather than early entry events, since the efficiency of attachment of [35S]methionine-labeled virus on HEp-2c cells was found to be identical for Sl and S12 (not shown).
401
INFECTION
To determine whether, upon coinfection, it is the persistent or the lytic phenotype that is dominant early in infection, lytic PV strains were titered in the absence or in the presence of a persistent mutant. The persistent mutants, even when present in large excess, did not interfere with the lytic effect of wild-type or attenuated PV strains (Table 3). Consequently, early in infection, the lytic phenotype was dominant. To determine whether this changed later on after infection, HEp-2c cells were first infected with S31 or S32 persistent mutants at a multiplicity of infection of 0.1 ID50 per cell, then challenged 1 to 30 days later with the type 3 wild-type Leon strain at the same multiplicity. When the superinfection occurred during the first 5 to 7 days after the primary infection, no cell survived infection with the Iytic strain. Later, cells became progressively resistant to superinfection, although during the first 30 days after the primary infection, more c.p.e. was observed in superinfected cells than in control cells infected with the persistent virus alone (not shown).
Maintenance of the PV persistent infection in HEp2c cells To determine the degree of resistance of PV persistently infected HEp-2c cells to superinfection, cultures
TABLE 3 TEST OF INTERFERENCEBETWEENPERSISTENTPV MUTANTS AND WILDTYPEOR ATTENUATEDPV STRAINSIN COINFECTEDHEP-2c CELLS PV inoculum Wild-type or attenuateda Sl Sl Mahoney Mahoney s3 s3 s3
Leon 37 Leon 37 Leon 37
+
+ + +
+ +
Persistent (ID50/well)*
Virus titer (log ID50/ml)
-
9.55 9.3 <1.3 10.05 10.05 9.55 9.05 9.3 <1.3 <1.3 7.92 7.8 8.05
S12(50) S12(50) S12(50) S31(500) S32(5) S31(500) S32(5) S31(500) S32(5)
a Serial 1 O-fold dilutions of wild-type or attenuated PV were titered in the absence or in the presence of persistent PV strains. * Constant doses of persistent virus (determined on IMR-32 cells and given in parentheses) were added to each well. The maximum dose which did not generate c.p.e. in HEp-2c cells was chosen for each persistent virus. This dose was higher for S31 than for S32 because S32 was much lytic on HEp-2c cells than S31 (Pelletier et al., 1991).
402
BORZAKIAN
persistently infected with PV type 1 or PV type 3 were superinfected with heterotypic wild-type or persistent PV. Superinfected cultures of cells continuously producing virus and mock-superinfected cultures had the same appearance under the microscope. The average production of the superinfecting wild-type or persistent virus, titered in the presence of antibodies neutralizing the heterotypic endogenous virus, was very low or undetectable (Table 4). In contrast, spontaneously cured c-HEp-S32 cells, which had been persistently infected by S32 (Table l), demonstrated c.p.e. upon superinfection, and the superinfecting virus multiplied in these cells almost as efficiently as in HEp-2c cells (Table 4). Since the establishment of the persistent infection in HEp-2c cells depended on the cell specificity, i.e., the reduced Iytic potential of the PV mutants in these cells, we looked at the cell specificity of the mutants reisolated from cultures of HEp cells persistently infected for 6 months. The results, shown in Table 5, indicated that the mutants selected in HEp-2c cells after 6 months of persistent infection had recovered the ability to lyse HEp-2c cells as efficiently as IMR-32 cells. Consequently, persistently infected HEp-2c cells were partially resistant to reinfection by the virus they generated, while uninfected HEp-2c cells were fully permissive to the same virus. Three attempts to establish a persistent infection in HEp-2c cells with PV mutants isolated 6 months postinfection from persistently infected HEp-2c cells were unsuccessful. The time during which infectious virus or viral RNA could remain strictly intracellular was determined by growing the cells in an excess of neutralizing antibodies, as described (Pelletier et a/., 1991) and by following virion production after extracellular virus had been TABLE 4 RESTRICTEDGROWTHOF A SUPERINFECTINGPV IN PERSISTENTLY INFECTEDHEP-2 CELLS Average production
Virus
HEp-2c
HEp-S12
Mahoney Sll Leon 37 S32
0 2700 1000 2000 ND
100 NDa ND 0.5* ND
of virus per cell per 24 hr in HEpS31 (cl1 8) 4.5 1 3b 1.3b ND ND
c-HEp-S32 0 ND ND 600 2000
’ Not determined. b This value represents only the progeny of the superinfecting virus, titered in the presence of antibodies neutralizing the heterotypic endogenous virus of persistently infected cells. Infectious titers were determined on IMR-32 cells by a micromethod and expressed in ID50.
ET AL. TABLE 5 REVERSIONOF PV MUTANTS IN PERSISTENTLYINFECTEDHEP-2c CELLS TO A PHENOTYPELMIC FOR HEP-2c CELLS Difference in titer (log ID50/ml) on IMR-32 and HEp-2c cells of the virus isolated
Viral strain
From persistently infected IMR-32 cells
After 6 months of persistent infection in HEp-2c cells
Sll 512 s31 S32
1.75 3.25 4 3.4
0.5 0.25 -1.25 0.75
completely neutralized. The results indicated that, in both HEp-2c and IMR-32 cells that had been persistently infected for 8 to 10 months, the persistent vi: ruses-or infectious genomes-could remain intracellular for as long as 10 days (not shown). When HEpS12 cells were grown for 20 days in the presence of neutralizing antibodies, then for 30 days in the absence of antibodies, intracellular viral RNA was no longer detected by Northern blotting. Since the maintenance of the persistent infection was clearly independent of the lytic potential of the mutants in HEp-2c cells, we tried to elucidate the mechanism of persistence in these cells by looking for the presence of the PV receptor at the surface of infected cells. The percentage of persistently infected cells expressing the PV receptor at their surface varied greatly with the cell clones. For example, in HEp-S32 (~17) the PV receptor was not detected either by immunofluorescence (not shown) or by cytofluorometry (Table 6); the results obtained with this clone were similar to those obtained with PV receptor-negative murine LM cells. In contrast, almost all the HEp-S31 (cl1 8) cells expressed the PV receptor, although the average fluorescence intensity per positive cell was only 40% of that of HEp-2c cells (Table 6, Fig. 2). For comparison, in PV persistently infected neuroblastoma cells, onethird to one-half of the cells expressed the PV receptor, and the average fluorescence intensity per positive cell was slightly lower than that of uninfected IMR-32 cells (Table 6). Thus, the reduced expression of the PV receptor may be responsible for, or may contribute to, the maintenance of the persistent infection in some clones, but it does not seem to be an absolute requirement for maintaining persistence. Total viral RNA synthesis was evaluated by dot blot analysis to determine whether viral replication is blocked in PV persistently infected clones. The average number of genome equivalents per cell varied
PERSISTENT
POLIOVIRUS
TABLE 6 DETECTIONBYCYTOFLUOROMETRYOFTHEPV RECEPTORATTHESURFACE OF PERSISTENTLYINFECTEDCELLS
Cells HEp-2cb HEp-2c LMC c-HEp-S12 c-HEp-S32 HEpS32 (~17) HEpS31 (~118) IMR-32’ IMR-32 LM” IMR-Sl IMR-S3
mAb 280
% of receptor + fluorescent cells
+ -
99.6 1.1 7.4 100 99.5 5 96.8
326 24 242 226 52 131
86.3 3.3 5 33 54.3
267 -
+ + + + + + + + +
Average fluorescence per positive cella
218 249
a In arbitrary units. b In the upper and lower part of the table, the conditions of cytofluorometry were optimized for HEp-2c and IMR-32 cells, respectively. ’ LM cells are PV receptor-negative mouse cells, used as a negative control.
greatly, between 3,100 and 41,000, according to the clone (Table 2). In situ hybridization experiments revealed the distribution of viral genomes in the cell population to be quite heterogeneous. In HEp-S31 (cl1 8) 9-10% of the cells harbored most of the viral genomes, and in half of the positive cells, the number of viral genomes was the same as that found in HEp-2c cells 4 hr after infection (Fig. 3). However, cells representative of all infection steps could be detected. Only 2% of the HEp-S32(cl7) cells had viral genomes as detected by in situ hybridization and these genomes were often localized in specific regions of the cytoplasm (Fig. 3F). In IMR-S3 cells, viral genomes were detected in 0.1% of the cells (Fig. 3H). The detection of viral antigens by indirect immunofluorescence on HEp-S31 (~118)cells confirmed the results of in situ hybridization experiments: the distribution of viral antigens was heterogeneous and 5 to 10% of cells were brilliantly stained (Fig. 4). This result indicated that, despite the presence of the PV receptor at the surface of 97% of HEp-S31 (cl1 8) cells, virus multiplication was blocked in the majority of cells. This block was, in fact, unstable. For example, when cells were fed but not subcultured or when they were split in a ratio greater than 1 in 4, all cells entered the lytic infection cycle, indicating that all of them were potentially permissive to PV. To determine whether this block was due to the lack of a host factor required for PV replication, or to a domi-
403
INFECTION
nant type of inhibition of PV multiplication, somatic cell hybrids were generated by fusion of persistently infected HEp-2c cells (thymidine kinase-positive, G418sensitive) and murine LM cells transformed by plasmid pAG60 (thymidine kinase-negative, G418-resistant) (Colbere-Garapin et al., 1981). The only factor limiting PV multiplication in murine cells is the absence of the virus receptor (Holland et al., 1959; Mendelsohn eta/., 1989). During the 10 days following fusion of HEp-S31 (cl1 8) with the murine cells, no increase in virus production was noted, as compared to the virus production of the clones alone, for which the same number of cells was treated under the same conditions (not shown). These results suggested that the mechanism of inhibition of PV multiplication in HEp-S31 cells was dominant. Spontaneous or induced cure of some cell lines HEp-2c cells infected with the persistent PV mutants were occasionally spontaneously cured of PV, between 5 weeks and 18 months after infection. The spontaneous cure of infected IMR-32 cells, however, was never observed (Colbere-Garapin eT a/., 1989). Some of the spontaneously cured HEp cell lines that we tested expressed at their surface the PV receptor (Table 6); they were free of viral RNA (Fig. 5) and they were permissive to superinfection (Table 4). The virion production by HEp-S32 (~17)cells highly depended on the temperature at which cells were incubated (Fig. 6). The virus produced by this clone was found to be more temperature-sensitive than the original S32 strain: its titer was 1.2 log lower at 39” than at 34”, while the titer of S32 was the same at both temperatures (Pelletier et al., 1991). Similar results were obtained with other cell lines: HEp-S12, HEp-S3l(cl18), and HEp-S32(cl8) (not shown). When HEp-S32 (~17) cells were grown for more than three passages at 39’, they were cured of virus. The lytic Mahoney and Sl strains multiplied in the cured cells without cytopathic effects, but the average virion production per cell per 24 hr was about 3 log lower in these cells than in HEp2c cells. These results indicated that cells partially resistant to PV infection had been selected during the persistent infection. As expected, the expression of the PV receptor was as low at the surface of the cured cells as on the parent HEp-S32 (~17)cells (not shown). DISCUSSION The PV persistent infection of HEp-2c cells has similarities with FMDV persistent infection of BHK21 cells, for which it was proposed that cell diversity may be an important element for long-term virus and cell survival (De la Torre et a/., 1989a). However, in the attempt to
404
FIG. 2. Detection of the PV receptor at the surface of cells by Indirect immunofluorescence with mAb 280 (Minor eta/., 1984). The PV receptor was detected on uninfected HEp-2c cells (A), HEp-S31 (cl1 8) cells (C, D), and c-HEp-S32 cells (F). The receptor was not detected on murine LM cells with mAb 280 (6) or on HEp-S31 (cl1 8) when a control mAb was used (E). The bar indicates 10 pm.
clarify the mechanisms of virus persistence, the PV system has one advantage over systems developed for other picornaviruses: both fully lytic and persistent mutants are available, allowing their phenotypes to be directly compared. In the present work, the virus hostcell interactions were studied during the establishment of the PV persistent infection, i.e., from the time of infection up to a few weeks after infection, and during the maintenance of the persistent infection, i.e., from 2 months to 2 years after infection. With this system, it was possible to distinguish the effects of mutations in the viral genome from those of modifications in the host cell phenotype. We have found that the attachment of the persistent mutant S12 and of the parent lytic Sl strain to HEp-2c cells was the same. At a low m.o.i., the accumulation of viral RNA during the first replication cycle, 5 hr after infection, was 3.5 times lower in HEp-2c cells infected with S12 than in cells infected with St. This difference was amplified with time. This result suggested that the difference between the mutant virus and its parent affects an early biosynthetic event in virus infection in
HEp-2c cells. At a high m.o.i., no delay in the early steps of virus multiplication was noticed for type 1 mutants in HEp-2c cells (Pelletier et al., 1991). It seems probable that the infection at a high multiplicity allowed a greater number of viral genomes to be replicated and expressed in these cells, leading to cell death, despite the delay in virus liberation. Therefore, both early steps of the viral cycle, as shown here, and late steps, as shown previously (Pelletier et a/., 1991), seem to be implicated in the establishment of PV persistence in HEp-2c cells. During this period, the lytic phenotype of wild-type and attenuated strains was dominant, indicating that viral functions altered by the mutations of persistent viruses did not interfere with the lytic virus cycle. In contrast to the establishment of PV persistence, the maintenance of the infection was not correlated to the reduced Iytic potential of the PV mutants, as shown by two kinds of unrelated experiments: 1) the persistent viruses reverted toward a lytic phenotype in HEp2c cells (Table 5) and 2) cells producing virus were resistant, even at a high m.o.i., to superinfection by a
PERSISTENT
POLIOVIRUS
INFECTION
405
FIG. 3. Detection of viral genomes by in situ hybridization in PV-infected cells. The PV genome was hybridized to a probe consisting of the complete S3 cDNA labeled with [35S]dCTP as described under Material and Methods. (A) Uninfected HEp-2c cells, (B) HEp-2c cells 4 hr after infection with S3, (C, D) HEp-S3l(cl18), (E, F) HEp-S32(cl7), (G) uninfected IMR-32 cells, (H) IMR-S3 cells. The exposure time was 4 hr for HEp cells and 8 hr for IMR cells. The bar indicates 3 pm.
heterotypic wild-type PV strain (Table 4). The maintenance of the persistent infection correlated with the selection of cells in which PV multiplication was restricted. Two PV persistently infected cell clones, representative of two different types of cell selection, were studied in greater detail: HEp-S32 (~17) and HEp-S31 (cl1 8). In the first clone, the expression of the PV receptor was similar to the background level, preventing virus from efficiently propagating through the extracel-
lular medium. Such a down regulation of the expression of the PV receptor has already been found in HeLa cell clones resistant to PV infection (Kaplan et al., 1989). These results are compatible with a carrier state of persistent infection. This model, however, seems too restrictive to account for the results obtained with HEp-S3 1 (cl 18). In our second clone, HEp-S3 1 (cl 18) almost 100% of the cells expressed the PV receptor, but the multiplica-
406
BORZAKIAN
ET AL
3~. 4. Detection of viral antigens by indirect immunofluorescence of HEp-S3 1 (cl 18). The amount of viral antigen marked with an anti-S3 I-abbit ser urn in HEp-S31 (cl1 8) cells varied among the cells (A-D). Uninfected HEp-2c cells (E) and HEp-2c cells 72 hr after infection with S31 (F) were USC?d as negative and positive control, respectively. The bar indicates 10 pm.
tion of the endogenous virus-as well as that of a superinfecting lytic virus-was found to be restricted in the majority of cells (Figs. 3 and 4, Table 4). Results of cell fusion experiments with murine cells suggested that the intracellular block was not due to the lack of a host factor required for PV replication, but to a dominant type of inhibition. Although this result was obtained with a clone-producing virus, it is in agreement with previous results obtained with hybrids of somatic hamster cells resistant to FMDV (De la Torre et a/., 198913). The intracellular block of HEp-S31 (cl1 8) was unstable under some conditions of cell culture in which all cells entered a lytic infection cycle. This suggested that, at least in HEp-S31 (cl1 8) there is an equilibrium between an abortive and a complete Iytic infection, rather than a bona fide carrier state of persistent infection.
The possibility to cure the cells by growing them at 39” allowed the role of viruses and that of cells in the virus-host cell interactions of an established persistent infection to be further distinguished. With HEpS31 (~17) it was shown by superinfecting the heatcured cells with fully lytic PV strains that the resistance of cells to the PV lytic effects was independent of the virus and depended only on the properties of the cells selected by the virus. Thus, while the establishment of PV persistence mainly depended on the virus strain and on the multiplicity of infection, the maintenance of persistence involved mainly the properties of the cells selected during the persistent infection. The spontaneous cure of cells infected with FMDV has been observed repeatedly and has been found to result from increasing deletions occurring in the virus genomes over the course of sequential passages, so
PERSISTENT
POLIOVIRUS
that after passage 65 virions were no longer detectable (De la Torre et al., 1985). in our PV system, when the infectious titer decreased before the complete cure, we did not observe deleted viral RNA species of defined sizes; however, this does not exclude the possibility of small deletions, undetectable by Northern blot. The spontaneous cure of cells occurred not after a specific number of passages, but apparently at random, between 5 weeks and 18 months after infection. The spontaneous cure could be explained if a few cells escaped virus infec?ion, or if the virus cycle aborted in some of the infected cells. Because the infectious titers in supernatants of p&sistently infected HEp-2c cells were on the order of 10’ ID50/ml, the probability that a few cells escaped infection after several months seems very low. The model that we propose of an equilibrium between an abortive and a lytic infection could help to explain the spontaneous cure of cells: if, for example, the equilibrium is correlated to the number of viral genomes in a cell, it could be displaced toward the abortive cycle when this number is low, resulting in the cure of cells. According to this model, the cured cells might be susceptible to a superinfecting virus, as we observed (Table 4). The level of expression of one or a few host factor(s) involved in PV multiplication would be responsible for the equilibrium between the abortive
407
INFECTION
0
2
4
6
6
days FIG. 6. The virion production of HEp-S32(cl7) depended on the temperature of incubation. HEp-S32 (~17) cells were subcultured and replicate plates were incubated at 34” (0) 37“ (0). or 39” (0). The virion production was titered at the indicated days. The results are the average of two independent experiments.
and the lytic infection. This model is also compatible with the results obtained with HEp-S32(cl7), the limiting factor for this clone probably being the PV receptor. The results of several experiments are in agreement with the model of an equilibrium between an abortive and a lytic infection: the period during which the virus or its genome can remain intracellular (10 days), the high percentage of HEp-S31 (cl1 8) cells demonstrating viral genomes (9-l 0%) and viral antigens (5-l 00/o),the spontaneous cure of some cell lines despite their susceptibility to virus infection, and the possibility of displacing the equilibrium toward the cure of cells (by in- . cubating them at 39”) or toward the lytic cycle (by feeding the cells without subculturing them, or growing them at a very low density). More experiments, however, are required to confirm this model and to elucidate its mechanism at the molecular level. In conclusion, we have shown that the establishment and the maintenance of a PV persistent infection involve distinct mechanisms. We propose that the expression of a host factor involved in PV multiplication and the stability of this phenotype may be responsible for an equilibrium between an abortive and a complete lytic virus cycle in PV persistently infected cells. ACKNOWLEDGMENTS
FIG. 5. Northern blot analysis of total viral RNAs present in four persistently infected cell lines used in this study: HEp-S32 (~17) (12.5 pg); HEpS32 (~18) (30 pg); c-HEp-S32 (30 pg) and IMR-S3 (17 pg). HEp-2c cell polyA+ RNA was used as a negative control. The probe consisted of 32-P labeled plasmid FLC3 (Cann et a/., 1983) which carries the entire S3 cDNA. The arrowhead indicates the position of viral RNA isolated from Leon 37 purified virions. The other two arrows indicate the position of 28s (6.3 kb) and 18s (2.4 kb) ribosomal RNAs.
We are grateful to Professor F. Horaud for continuous interest in this work and critical reading of the manuscript. We are indebted to Professor P. Minor for the monoclonal antibody 280, to Professor A. Nomoto for the pVS(1) IC-O(T) plasmid, and to Professor J. Almond for the FLC3 plasmid. We thank P. Metezeau and H. Kieffer for help in cytofluorometry. We thank K. Pepper for reviewing the manuscript. The generosity of the Fondation Marcel Merieux and of the Ihstitut de Recherches lnternationales Servier & Compagnie Developpement is acknowledged for the fellowships attributed to S.B. and Y.B., respectively. This work was supported by a grant from the lnstitut
408
BORZAKIAN
National de la Sante et de la Recherche to F.C.-G.
MBdicale (Grant 88 1006)
REFERENCES BRAHIC, M., and STROOP,W. G. (1981). Theiler’s virus persists in glial cells during demyelinating disease. Cell 26, 123-l 28. CAMPBELL, A. M. G., WILLIAMS, E. R., and PEARCE, J. (1969). Late motor neuron degeneration following poliomyelitis. Neurology 19, 1101-1106. CANN, A. J.. STANWAY, G., HAUPTMANN, R., MINOR, P. D., SCHILD, G. C., CLARKE, L. D., MOUNTFORD. R. C., and ALMOND, J. W. (1983). Poliovirus type 3: Molecular cloning of the genome and nucleotide sequence of the region encoding the protease and polymerase proteins. Nucleic Acids Res. 11, 1267-l 281. CARP, R. I. (1981). Persistent infection of human lymphoblastoid cells with poliovirus and development of temperature sensitive mutants. Intervirology 15, 49-56. COLB~RE-GARAPIN,F., HORODNICEANU,F.. KOURILSKY,P., and GAFIAPIN, A. (1981). A new dominant hybrid selective marker for higher eukaryotic cells. J. MO/. Biol. 150, l-l 4. COLB~RE-GARAPIN, F., CHRISTODOULOU, C., CRAINIC, R., GAFIAPIN, A. C., and CANDREA,A. (1988). Addition of a foreign oligopeptide to the major capsid protein of poliovirus. Proc. Nat/. Acad. Sci. USA 85,8668-8672. COLB~RE-GARAPIN,F., CHRISTODOULOU,C., CRAINIC, R., and PELLETIER, I. (1989). Persistent poliovirus infection of human neuroblastoma cells. Proc. Natl. Acad. Sci. USA 86, 7590-7594. COUDERC,T., CHRISTODOULOU,C.. KOPECKA,H., MARSDEN, S., TAFFS, L. F., CRAINIC, R., and HORAUD, F. (1989). Molecular pathogenesis of neuronal lesions induced by poliovirus type 1. /. Gen. Viral. 70, 2907-2918. COUDERC, T., BARZU, T., HORAUD, F., and CRAINIC, R. (1990). Poliovirus permissivity and specific receptor expression on human endothelial cells. virology 174, 95-102. DALAKAS, M. C. (1986). New neuromuscular symptoms in patients with old poliomyelitis: a three year follow-up study. Eur. Neurol. 25,381-387. DE LATORRE,.I. C., DAVILA, M., SOBRINO,F.. ORTIN, J., and DOMINGO, E. (1985). Establishment of cell lines persistently infected with footand-mouth disease virus. Virology 145, 24-35. DE LA TORRE, 1. C., MARTINEZ-SALAS, E., DIEZ, J., and DOMINGO, E. (1989a). Extensive cell heterogeneity during persistent infection with foot-and-mouth disease virus. J. Viral. 63, 59-63. DE LATORRE, J. C., DE LA LUNA, S., DIEZ, J., and DOMINGO, E. (198913). Resistance to foot-and-mouth disease virus mediated by trans acting cellular products. /. Viral. 63, 2385-2387.
ET AL FEINBERG,A. P., and VOGELSTEIN,B. (1983). A technique for radlolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132, 6-l 3. GERCEL, C., MAHAN, K. B., and HAMPARIAN, V. V. (1985). Preliminary characterization of a persistent infection of HeLa cells with human rhinovirus type 2. /. Gen. Viral. 66, 131-l 39. GIBSON, 1. P.. and RIGTHAND,V. F. (1985). Persistence of echovirus 6 in cloned human cells. J. L&o/. 54, 2 19-223. HOLLAND, J. J., MCLAREN. L. C., and SYLVERTON,J. T. (1959). The mammalian cell virus relationship. IV. Infection of naturally insusceptible cells with enterovirus ribonucleic acid. 1. Exp. Med. 110, 65-80. KAPLAN, G., LEW, A., and RACANIELLO, V. R. (1989). Isolation and characterization of HeLa cell lines blocked at different steps in the poliovirus life cycle. /. Viral. 63, 43-51. KAWAMURA, N., KOHARA, M., ABE, S., KOMATSU, T., TAGO, K., ARITA, M., and NOMOTO, A. (1989). Determinants in the 5’ noncoding region of poliovirus Sabin 1 RNA that influence the attenuation phenotype. 1. Viral. 63, 1302-l 309. MELNICK, J. L. (1990). Enteroviruses: Polioviruses, coxsackieviruses, echoviruses, and newer enteroviruses. In “Virology,” 2nd ed. (B. N. Fields, D. M. Knipe et a/.), Raven Press, New York. MENDELSOHN, C., WIMMER, E.. and RACANIELLO,V. R. (1989). Cellular receptor for poliovirus: molecular cloning, nucleotide sequence and expression of a new member of the immunoglobulin superfamily. Cell 56, 855-865. MINOR, P. D., PIPKIN, P. A., HOCKLEY, D., SCHILD, G. C., and ALMOND, J. W. (1984). Monoclonal antibodies which block cellular receptors of poliovirus. Virus Res. 1, 203-212. PELLETIER,I., COUDERC, T., BORZAKIAN,S., WYCKOFF. E., CRAINIC, R., EHRENFELD,E., and COLBERE-GARAPIN,F. (1991). Characterization of persistent poliovirus mutants selected in human neuroblastoma cells. Virology 180, 729-737. Roos. R. P., RICHARDS, 0. C., GREEN, J., and EHRENFELD,E. (1982). Characterization of a cell culture persistently infected with the DA strain of Theiler murine encephalomyelitis virus. J. Vkol. 43, 1 1 181122. SHARIEF, M. K.. HENTGES, R., and CIARDI. M. (1991). lntrathecal immune response in patients with the post-polio syndrome. N. Engl. J. Med. 325, 749-755. TUMILOWICZ, J. J., NICHOLS, W. W., CHOLON, J. J., and GREENE, A. E. (1970). Definition of a continuous human cell line derived from neuroblastoma. Cancer Res. 30, 211 O-2 118. VALLBRACHT, A., HOFMANN, L., WURSTER, K. G., and FLEHMIG, B. (1984). Persistent infection of human fibroblasts by hepatitis A virus. J. Gen. Virol. 65, 609-6 15.