- Email: [email protected]
Lysis by natural killer cells requires viral replication and glycoprotein expression
Immunology Letters, 13 (1986) 261-268 Elsevier Imlet 794
LYSIS BY NATURAL
KILLER CELLS REQUIRES VIRAL GLYCOPROTEIN EXPRESSION
REPLICATION
AND
Amichai SCHATTNER 1 and Bracha RAGER-ZISMAN 2 tDepartment of Virology, Weizmann Institute of Science, Rehovot, and 2The Department of Microbiology and Immunology, Faculty of Health Sciences, Ben Gurion University of the Negev, Beer Sheva, Israel (Received 28 May 1986) (Accepted 8 June 1986)
1. Summary Several lines of evidence suggest that natural killer cells (NK) have an important role in antiviral defense. To be thus effective, NK cells have to recognize and cause lysis of virusinfected cells. The mechanism underlying this interaction was investigated in a murine system using vesicular stomatitis virus (VSV). A large number of cell lines was screened to identify a permissive target for VSV infection and the B16 murine melanoma cells were chosen since VSV replicates in these cells without producing cytopathic effects 24 h after infection. Spontaneous lysis of B16 melanoma cells by spleen cells occurred only if the target cells were previously infected with VSV. Treatment of spleen cells with anti NK 1.1 or anti Thy 1.2 plus complement decreased the specific lysis by 50°70, therefore, the phenotype of the effector cells in this system corresponds to the NK cell subset. Immunofluorescent staining with polyclonal and monoclonal anti-VSV antibodies demonstrated that the viral glycoprotein G is abundantly expressed on the target cell surface. Treatment of Key words: natural killer cells - vesicular stomatitis virus viral glycoprotein - recognition
Abbreviations: VSV, vesicular stomatitis virus; NK, natural killer cells; CTL, cytotoxic T lymphocytes; MOI, multiplicity of infection; wt, wild type; G-protein, viral glycoprotein; Mprotein, viral matrix protein; TM, tunicamycin; pfu, plaque forming units.
the VSV-infected targets with tunicamycin prevented glycosylation of newly synthesized VSV glycoprotein G on the cell membrane. This treatment abrogated completely NK-mediated killing of the infected B16 melanoma cells. These results indicate that virus replication and membrane expression of glycosylated protein G are essential for recognition and lysis of VSV-infected targets by NK ceils.
2. Introduction Target recognition by NK cells appears to be a spontaneous non-specific and M H C unrestricted event [1-3], yet it is well established that definite target structures must be involved [3, 4]. Cells that are resistant to lysis by NK cells become highly susceptible when they are acutely or persistently infected with virus [5-7]. The possibility that virus encoded proteins were the major target structure was suggested as early as a decade ago [8], but conflicting data were also obtained [9] and the subject remains unclear. We have recently used temperature sensitive mutants of VSV to demonstrate that expression of two viral proteins is required for NK cell recognition [10]. The present data further supports the view that virus replication and glycoprotein expression in the infected target cell are essential for lysis by NK cells.
0165-2478 / 86 / $ 3.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
261
3. Materials and Methods
3.1. Mice 5-8 week old C B A / J or C57BL/6 mice from our own breeding colony were used in all experiments. 3.2. Viruses The Indiana (HRC) wild type (wt) serotype of vesicular stomatitis virus (VSV) was used. Stock virus was grown in primary chick embryo fibroblasts. VSV (wt) was kindly provided by Dr. P. I. Marcus, and propagated on vero cell monolayers at the permissive temperature of 31 °C [11]. 3.3. Cells Clone M2R derived from B16 mouse melanom a cells [12], obtained from Dr. J. Mather, Rockefeller University, New York was grown in vitro in Dulbecco's modified Eagle's medium (DME/F12, Gibco, Grand Island, NY), supplemented with 10O7o fetal calf serum (FCS), glutamine and antibiotics. The cells were recloned, the subclone most resistant to NK lysis being selected for use in these experiments. In addition, clone J774.16, a continuous macrophage-like cell line derived from the murine reticulum cell sarcoma J774.16 was also used. These cells were propagated and cultured as previously described [13]. 3.4. Tunicamycin (TM) TM (Calbiochem, Calbiochem-Behring Corp., La Jolla, CA) was added at a final concentration of 0.1, 0.5 or 1 # g / m l to target cells 2 h before virus infection. 3.5. Ant&era Polyclonal rabbit anti-VSV, anti-VSV-G protein and anti-VSV-M protein sera were generously provided by Dr. J. Holland. Monoclonal antiVSV-G ascites was a gift from Drs. R. Weiss and H. Koprowski. 3.6. Immunofluorescent staining Indirect immunofluorescence was used to visualize viral antigen expression. The specific antisera and FITC conjugated antibody (the lat262
ter purchased from Cappel, Cochranville, PA 19330) were absorbed with an equal volume of uninfected B16 mouse melanoma cells for 30 min at 4 °C and subsequently centrifuged (15,000 rpm) to remove cells and aggregates. For cytoplasmic immunofluorescent staining, cells were grown in Lab-Tek tissue culture chamber slides (Lab-Tek Div., Miles Labs Inc., Nagerville, IL 60540) until nearly confluent. Monolayers were washed, air-dried and fixed in cold acetone. Unfixed monolayers were used for membrane fluorescence. 3.7. Preparation of effector cells Spleens from C B A / J or C57BL/6 mice were passed through a fine Sellector sieve (Bellco, Inc., Vineland, NH), washed 3 times with phosphatebuffered saline (PBS) and resuspended in medium at a concentration of 1 × 107 cells/ml. 3.8. Treatment with monoclonal anti-Thy 1.2 Monoclonal anti-Thy 1.2 antibody was purchased from New England Nuclear, Boston, MA. Spleen cells (5 × 107 cells/ml) were incubated with the antibody (final dilution 1:1,000) and rabbit low toxicity complement (c'; Low Tox, Cedarlane, Ontario, Canada; final dilution 1:12) for 45 min at 37 °C in 5°70 CO2. 3.9. Treatment with alloantiserum N K 1.1 Mouse anti NKI.1 [14] was kindly provided by Dr. S. Pollack (Seattle, WA). Spleen cells at 3 × 107 cells were incubated with anti NKI.1 (final dilution 1:30) and rabbit c' (final dilution 1:30) for 45 min at 37 °C. Treatment of C57BL/6 mouse spleen cells with this antiserum reduced the splenic cytotoxicity against YAC-1 cells by 50%. Rabbit c' alone had no effect. 3.10. Virus infection and labeling of target cells Virus infected target cells were prepared by pelleting 5 × 10 6 cells and resuspending the pellet in 0.5 ml of virus suspension and 100 #Ci of NAz51CrO4 at multiplicities of infection (MOI) of 0.1-0.001. After 2 h incubation at 31 °C, v i r u s - c e l l mixtures were washed twice in medium and then seeded at 2×104 cells per well in 96 well flat bottom microtiter plates (Linbro, 76-003-05) and incubated overnight at 37 °C.
3.11. Cytotoxicity assay 51Cr-Iabeled monolayers of uninfected or VSVinfected target cells were washed again with medium to remove nonadherent and dead cells. Spleen cell suspensions (0.2 ml) were then added to the target cells, centrifuged 5 min at 600 rpm and incubated at 37°C for 8 - 1 0 h. After incubation, plates were centrifuged, 100/~1 of supernatants were harvested for each well and then 100 #1 of 2 N HC1 were added to each well for estimation of maximal 51Cr release. The specific 51Cr release was calculated as described previously [61. 3.12. Analysis of glycosylated and non-
Table 1 Replication of VSV in BI6 m e l a n o m a cells and in a continuous macrophage (M~b) like cell line J774.16 Input virus (pfu)
5×104 5x103 5 x 102 5 × 101
MOI
0.1 0.1 0.001 0.0001
Virus yields (pfu/ml) B16 melanoma
M~
3×105 5x105 1 x 104 3 × 103
6 3 1 4.5
J774.16 ×107 ×107 x 107 x 106
Virus yields in supernatants were measured 24 h after infection. Virus titers are expressed in plaque forming units (pfu) per 1 ml. Virus titrations were performed as described in Materials and Methods.
glycosylated G protein B16 m e l a n o m a cells were pretreated with 1 # g / m l TM for 2 h at 37 °C, and were infected with VSV (wt) at an M O I of 10-20. After adsorption for 2 h at 31 °C, [3H]leucine was added. Cells were incubated for an additional 16 h at 37 °C and then solubilized in sodium dodecyl sulphate (SDS) buffer. Cell extracts were immunoprecipitated and analyzed on polyacrylamide gel electrophoresis (SDS-PAGE) [15].
4. Results
4.1. Replication of VSV in B16 melanoma cells
and in a continuous macrophage-like cell line J774.16 Since VSV is a lytic virus and rapidly inhibits host cell protein synthesis, it was important to find target cells which will be permissive for VSV but will not undergo extensive lysis during the incubation period with the effector cells. A large number of cell lines were screened. Table 1 exemplifies susceptibility to infection with VSV of two cell lines, a macrophage-like cell line J774.16 which is highly susceptible to VSV, and B16 m e l a n o m a which were permissive but less susceptible to infection with this virus. The macrophage cell line J774.16 was chosen since VSVinfected macrophages have t~een previously used as targets in the study of mechanisms of killing by C T L [16]. Monolayers of J774.16 and B16 m e l a n o m a were infected at a range of multiplici-
ties of 0.1-0.0001, J774.16 was found to be extremely susceptible to VSV. After 24 h, complete cell destruction occurred at multiplicities of 10 -1 --10 -3. High yields of VSV ranging from 6x107 plaque forming units (pfu)/ml to 4.5 x 106 p f u / m l were measured in supernatants of infected cultures. When B16 melanoma cells were infected with the same multiplicities of VSV, no cytopathic changes were detected after 24 h while VSV replication occurred as determined by virus yields in supernatants of the infected cultures. Levels of spontaneous 5~Cr release from VSV infected B16 cells were < 30°7o in all experiments. B16 m e l a n o m a cells were therefore chosen as target cells for our studies. 4.2. Expression of VSV antigens in B16 melano-
ma cells As shown above, VSV replicates in B16 melan o m a cells without producing cytopathic effects 24 h after infection. It was important however, to determine whether VSV proteins are produced and expressed by the infected cells within this period. As shown in Fig. 1, the B16 melanoma cells expressed high amounts of VSV proteins based on the intensity of the fluorescent staining and the number of positive cells ( 7 0 - 8 0 % ) . A polyclonal rabbit anti-VSV serum reacted strongly with the VSV-infected B16 melanoma cells showing both membrane and cytoplasmic fluorescence (Fig. 1A, B). In contrast, the monoclonal anti 263
V S V - G ascites s h o w e d o n l y m e m b r a n e , b u t n o c y t o p l a s m i c b i n d i n g (Fig. 1C), w h i l e t h e anti-VSVM p r o t e i n sera s h o w e d a reverse p a t t e r n w i t h cytoplasmic binding but no membrane fluoresc e n c e (Fig. 1D). 4.3. Lysis of B16 melanoma cells infected by VSV V S V - i n f e c t e d o r u n i n f e c t e d B16 m e l a n o m a cells were l a b e l e d w i t h 51Cr, i n c u b a t e d w i t h s u s p e n sions o f s p l e e n cells a n d t h e s p e c i f i c 51Cr release was d e t e r m i n e d . Target cell lysis in this t y p e o f s y s t e m m a y b e m e d i a t e d by N K cells o r o t h e r eff e c t o r cells m e d i a t i n g n a t u r a l c y t o t o x i c i t y [6, 17]. By v a r y i n g t h e t i m e o f i n f e c t i o n w i t h VSV a n d t h e M O I , it was f o u n d t h a t viral p r o t e i n s y n t h e sis was a n o b l i g a t o r y r e q u i r e m e n t for lysis o f V S V - i n f e c t e d targets. A s h o r t (3 h) i n f e c t i o n t i m e at a h i g h m u l t i p l i c i t y o r a l o n g (18 h) i n f e c t i o n t i m e u s i n g U.V. i n a c t i v a t e d VSV leaves t h e B16 m e l a n o m a t a r g e t resistant to lysis (Table 2).
Table 2 Requirement of VSV replication for lysis of B16 melanoma cells by NK cells Time of infection (h)
MOI
% 51Cr specific release
FA a
Uninfected 3 18 18 18
10 0.01 0.0001 U.V. inactivated
5 _+4 0 23 _+4 7_+2 0
0 0 80 5 - 10 0
a Percent positive cells. BI6 melanoma cells were infected with VSV (wt) for different periods of time and different MOI as indicated. The 5~Cr-release assay was performed at 37 °C at an effector to target ratio of 100:1 with an incubation period of 8 h. C57BL/6 spleen cells were used. Results are the average of three experiments. The mean spontaneous release was 24%.
Fig. 1. lmmunofluorescent staining of BI6 melanoma cells infected with ts mutants of VSV. (A) Membrane and (B) cytoplasmic staining of B16 cells infected with VSV wt at 37 °C with rabbit anti-VSV antibody. (C) Membrane staining of B16 cells infected with VSV at 37 °C with monoclonal anti-VSV G antibodies. (D) Cytoplasmic staining of BI6 cells infected with VSV at 37 °C with rabbit anti-VSV M antibodies. Magnification ×400. 264
However, a long (18 h) infection time with 1,000-fold less virus results in specific lysis. T h e effect was d o s e - d e p e n d e n t a n d w h e n less virus was used for infection, n o significant lysis occurred. At this low M O I , the percent o f FA positive cells decreased from 800/0 to less t h a n 100/0. 4.4. Characterization of effector cells Table 3 s u m m a r i z e s the effect of p r e t r e a t m e n t o f the effector ceils in this system with specific antisera to m e m b r a n e markers. Treatment with a n t i - N K 1.1 + c' or with anti-Thy 1.2 + c' decreased the specific lysis o f VSV-B16 m e l a n o m a target cells by 50%. Therefore, the p h e n o t y p e o f the p r e d o m i n a n t effector cell in this system corresponds to the N K I a n d N K x subsets previously described [18]. 4.5.
I n order to c o n f i r m this observation, we analyzed the effects of T M t r e a t m e n t o n the glycosylation of VSV G after infection of B16 m e l a n o m a cells. Cells were pretreated with TM, infected with VSV (wt), harvested 18 h post infection, i m m u n o p r e c i p i t a t e d with rabbit anti-VSV G protein a n t i s e r u m a n d analyzed o n SDS-PAGE gels. Results shown in Fig. 2 indicate that indeed only nonglycosytated G was synthesized in the presence of TM. The non-glycosylated G was present o n the cell m e m b r a n e as seen by imm u n o f l u o r e s c e n t surface staining with m o n o c l o n a l a n t i - G antibody. As shown in Fig. 3, A
13
Requirement for glycosylation of G proteins for NK cell recognition of virus infected targets
It has been recently established that the a n t i b i otic, t u n i c a m y c i n (TM) inhibits the synthesis of N-acetyl g l u c o s a m i n e - l i p i d intermediates a n d thereby prevents glycosylation o f newly synthesized VSV glycoprotein [19]. P r e t r e a t m e n t o f VSV-infected cells with T M almost totally inhibits infectious virus p r o d u c t i o n , a n d cells grown in the presence o f T M express n o n glycosylated G protein o n their cell surface [20].
Table 3 Phenotype characterization of the effector cells Pretreatment
% specific lysis (% decrease)
None c' alone anti-NK 1.1 + c' anti-Thy 1.2 + c'
23.9_+ 1.4 22.6_+2.9 11.5___1.0 11.1+ 1.7
(0) (52) (53)
C57BL/6 spleen cells were pretreated with anti-NK 1.1 antibodies or with anti Thy 1.2 antibodies as described in Materials and Methods. The 51Cr-release assay was performed at 37 °C at an effector to target cell ratio of 100:1 with an incubation period of 8 - 10 h. Target cells were B16 melanoma infected with VSV (wt). Results are the average of triplicates of three experiments. The mean spontaneous release was 24%.
Fig. 2. SDS-PAGE electrophoresis of precipitates from VSVinfected B16 melanoma cells with or without TM. B16 melanoma cells were pretreated with 1 ~g/ml TM for 2 h at 37 °C and infected with VSV at a MOI of 10-20. [3H]Leucinewas added 2 h later and cells were incubated for an additional 16 h at 37 °C. Cell extracts were immunoprecipitated with rabbit anti-VSV antiserum. (A) G protein of TM-treated VSVinfected cell lysates. (B) Nonglycosylated G protein of TMtreated VSV-infectedcell lysates. 265
100 8O •-
60
..421
-~
40 20
0
0.5
1.0
Tunicamycin (Mg) Fig. 3. Inhibition of lysis of VSV-infected cells by TM. B16 m e l a n o m a cells were pretreated with TM final concentrations of 1.0 ~g or 0.5 /~g, infected with VSV and incubated for 16 11 at 37 °C. The cytotoxicity assays were performed at 3 7 ° C with an effector to target ratio 100:1. S p o n t a n e o u s release did not exceed 30°7o. Results are the average of triplicates of 3 experiments. Uninfected B16 cells ( z~); VSV infected ( • ); YAC-1 cells ( n ).
TM inhibited the killing of B16 melanoma cells infected with VSV (wt), indicating that appropriate glycosylation was also required for NK cell lysis. In contrast, incubation of YAC-1 target cells in parallel under the same conditions with TM had no effect on the susceptibility of these cells to NK cell lysis. This result independently confirms the importance of appropriately processed G protein in recognition by NK cells.
5. Discussion
Target cell recognition by NK cells appears to be a spontaneous event which is nonspecific in the immunological sense, since xenogeneic and allogeneic targets can be lysed, and since there is no M H C restriction [3]. Yet, many observers have noted that different tumor cells have different susceptibilities to lysis by NK and that an acute or persistent virus infection dramatically augments this susceptibility [ 4 - 7 ] . These results indicate a highly "selective" recognition by the NK cells for certain target structures that are apparently associated with the infecting virus. However, the nature of these structures remains unclear. 266
In order to analyze the mechanism by which virus infection converts resistant cells to susceptible targets for NK, we have used murine B16 melanoma cells and vesicular stomatitis virus (VSV). Studies by Zinkernagel et al. [16] who also used VSV infected targets to study the mechanism of recognition by cytotoxic T lymphocytes (CTL), have shown that the VSV glycoprotein was the major viral encoded determinant that must be expressed for recognition by CTL. Experiments published by other groups which were carried out in the presence of the antibiotic, tunicamycin, indicated that glycosylation of the G protein was not an absolute requirement for recognition and lysis of VSV-infected target cells by CTL [21]. We have applied this model to study recognition of target structures by NK cells; however, a basic technical difficulty, not encountered in the CTL experiments, had to be overcome. While the virus infected targets are lysed within 4 h by CTL, incubation periods of 8 h or longer are generally required in order to obtain significant lysis of virus infected targets by NK [6]. However, since VSV is a lytic virus, extended incubation periods lead to unacceptably high spontaneous 51Cr-release from virus-infected cells. Furthermore, a murine target cell had to be identified that would be permissive to VSV infection, but would not be destroyed by the replicating virus. After screening a large number of target cells, we found that the B16 melanoma cells were indeed an appropriate cell line. Immunofluorescent staining confirmed that the VSV infected cells expressed high amounts of VSV proteins and that a high proportion of the cells were positive. Our results indicate that replication of virus was a prerequisite for target lysis by NK, since uninfected B16 melanoma cells were resistant to lysis, and infection with U.V. inactivated virus did not significantly alter susceptibility. Similarly, infection with a high MOI for a short time insufficient for de novo VSV protein synthesis, did not render the cells susceptible. However, when target cells were infected with virus at a MOI of 10 -2 for 18 h, they became NK sensitive with a more than four-fold increase in specific lysis. In order to obtain statistically significant killing, and yet no unacceptably high
backgrounds, a compromise had to be made resulting in spontaneous release values of 1 5 - 3 0 % over 8 h, and levels of specific cytotoxicity seldom exceeding 30%. The 50% decrease in specific lysis following treatment of the spleen cells with anti-NK 1.1 or anti-Thy 1.2 antibodies confirms that N K cells are the main cell type responsible (Table 3). Having established that intracellular VSV replication was essential for target cell recognition by NK, we proceeded to investigate the viral proteins involved. Since differences in recognition are usually associated with alterations in the outer cell membrane, and since the viral glycoprotein G was abundantly found on the B16 melanoma cell membrane (Fig. lc), we examined the effect of tunicamycin on NK cell lysis. Cells pretreated with tunicamycin and infected with VSV were analyzed by immunoprecipitation and SDS-PAGE gels and showed only nonglycosylated G protein on the cell membrane (Fig. 2). This absence of glycosylated G protein from the target cell membrane was associated with a marked inhibition of NK-mediated cytotoxicity. Thus, VSVinfected NK sensitive targets were no longer recognized and killed when the normally glycosylated G protein was not expressed on the membrane (Fig. 3). Of interest is the observation that tunicamycin treatment of a standard NK target, YAC-1 cells, had no effect on killing, indicating that its target structures were unaffected by this antibiotic. It has been previously shown that decreased glycoprotein expression on the surface of an infected cell correlates directly with the ability of that cell to resist killing by both CTL and cytotoxic antibody and complement [22, 23]. This in fact may be the mechanism underlying some instances of persistent infections in vivo [24]. Our findings suggest that decreased membrane expression of glycosylated viral protein may also be related to NK resistance (Fig. 3). It is possible that in the absence of glycosylated G protein, the virus infected cell evades recognition and binding by the NK cell, and that the interference is not at the lytic phase itself. We also found that killing could not be inhibited by antibodies against VSV, and that antibodies to interferon sufficient to neutralize inter-
feron produced within the assay system itself [6, 10], did not affect the degree of killing observed (unpublished results). This indicates that the interferon produced is not involved in the lysis observed in this system. In summary, our observations support the role of viral encoded glycoproteins expressed on the outer cell membrane in target recognition by NK. Furthermore, the presence of an identical transmembrane protein G which lacks its sugar residues was insufficient for NK recognition and killing. Additional support for our theory that the membrane glycoprotein is not the sole target structure in this system comes from our experiments which show that VSV replication was required in order to obtain lysis by NK. Recent studies with VSV G plasmids indeed suggest that the VSV M protein which is a submembranal protein may contribute to the formation of the target structure [10]. Another possible mechanism which would also be in line with our observations would be a change in the putative host cell membrane structure induced by the viral glycoprotein. Several observations suggest an inverse correlation between H-2 expression and NK sensitivity [23, 26]. A decrease in H-2 expression in VSV infected cells [27] could explain the increased sensitivity to NK, and these changes and their relationship to viral glycoproteins will be studied by us in the future.
Acknowledgements Supported by USPHS Grants A1099807 and Al10702, Grant 1006 from the National Multiple Sclerosis Society and Grant BC301 from the American Cancer Society. A.S. is a fellow of the Camp David Institute of International Health. B.R.Z. is a recipient of an ICRF Career Development Award.
References [1] Herberman, R. B. and Ortaldo, J. R. (1981) Science 214, 24. [2] Welsh, R. M. Jr. (1978) J. Immunol. 121, 1631. [3] Rager-Zisman, B. and Bloom, B. R. (1982) Springer Sem. Immunopathol. 4, 397.
267
[4] Schattner, A. and Duggan, D. (1985) Am. J. Hematol. 18, 435. [5] Welsh, R. M. (1981) Antiviral Res. 1, 5. [6] Minato, N., Bloom, B. R., Jones, C., Holland J. and Reid, L. M. (1979) Mechanism of rejection of virus persistently infected tumor cells by athymic nude mice. J. Exp. Med. 149, 1117. [7] Schattner, A., Rager-Zisman, B. and Bloom, B. R. (1985) Cell Immunol. 90, 103. [8] Kiessling, R., Klein, E. and Wigzell, H. (1975) Eur. J. lmmunol. 5, 112. [9] Becker, S., Fenyo, E. M. and Klein, E. (1976) Eur. J. Immunol. 6, 882. [10] Moller, J., Rager-Zisman, B., Quan, P. C., Schattner, A., Panish, D., Rose, J. K. and Bloom, B. R. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 2456. [11] Lodish, E H. and Weiss, R. A. (1979) J. Virol. 30, 177. [12] Mather, J. P. and Sato, G. H. (1979) Exp. Cell Res. 124, 215. [13] Rager-Zisman, B., Kunkel, M., Tanaka, Y. and Bloom, B. R. (1982) Infect. Immun. 36, 1229. [14] Pollack, S. B., Tam, M. R., Nowinsky, R. C. and Emons, S. L. (1979) J. Immunol. 23, 1818. [15] Birrer, M. J., Bloom, B. R. and Udem, S. (1981) Virol. 108, 381. [16] Zinkernagel, R. M., Althage, A. and Holland, J. (1978)
268
J. Immunol. 121,744. [17] Stutman, O., Lattime, E. C. and Figarella, E. C. (1981) Fed. Proc. 40, 2699. [18] Minato, N. L., Reid, M. and Bloom, B. R. (1981) J. Exp. Med. 154, 750. [19] Leavitt, R., Schlessinger, S. and Kornfeld, S. (1977) J. Virol. 21, 375. [20] Gibson, R., Leavitt, R., Kornfeld, S. and Schlessinger, S. (1978) Cell 13, 671. [21] Harris, D. T., Hale, A. H. and Lefrancois, L. (1981) J. lmmunol. 126, 1914. [22] Welsh, R. M. and Oldstone, M. B. A. (1977) J. Exp. Med. 145, 1449. [23] Oldstone, M. B. A. and Fujinami, R. S. (1982) in: Virus Persistence, Symp. 33, Soc. Gen. Microbiol. (B. W. J. Mahy, A. C. Minson and G. K. Darby, Eds.) p. 185, Cambridge University Press, Cambridge. [24] Oldstone, M. B. A. and Buchmeier, J. (1982) Nature 300, 360. [25] Gidlund, M., Orn, A., Pattengale, P. K., Jansson, M., Wigzell, H. and Nilsson, K. (1981) Nature 292, 848. [26] Karre, K. (1985) in: Mechanisms of Cytotoxicity by Natural Killer Cells (R. B. Herberman and D. Callewaert, Eds.) p. 81, Academic Press, Orlando. [27] Hecht, T. and Summers, D. E (1972) J. Virol. 10, 578.
LYSIS BY NATURAL
KILLER CELLS REQUIRES VIRAL GLYCOPROTEIN EXPRESSION
REPLICATION
AND
Amichai SCHATTNER 1 and Bracha RAGER-ZISMAN 2 tDepartment of Virology, Weizmann Institute of Science, Rehovot, and 2The Department of Microbiology and Immunology, Faculty of Health Sciences, Ben Gurion University of the Negev, Beer Sheva, Israel (Received 28 May 1986) (Accepted 8 June 1986)
1. Summary Several lines of evidence suggest that natural killer cells (NK) have an important role in antiviral defense. To be thus effective, NK cells have to recognize and cause lysis of virusinfected cells. The mechanism underlying this interaction was investigated in a murine system using vesicular stomatitis virus (VSV). A large number of cell lines was screened to identify a permissive target for VSV infection and the B16 murine melanoma cells were chosen since VSV replicates in these cells without producing cytopathic effects 24 h after infection. Spontaneous lysis of B16 melanoma cells by spleen cells occurred only if the target cells were previously infected with VSV. Treatment of spleen cells with anti NK 1.1 or anti Thy 1.2 plus complement decreased the specific lysis by 50°70, therefore, the phenotype of the effector cells in this system corresponds to the NK cell subset. Immunofluorescent staining with polyclonal and monoclonal anti-VSV antibodies demonstrated that the viral glycoprotein G is abundantly expressed on the target cell surface. Treatment of Key words: natural killer cells - vesicular stomatitis virus viral glycoprotein - recognition
Abbreviations: VSV, vesicular stomatitis virus; NK, natural killer cells; CTL, cytotoxic T lymphocytes; MOI, multiplicity of infection; wt, wild type; G-protein, viral glycoprotein; Mprotein, viral matrix protein; TM, tunicamycin; pfu, plaque forming units.
the VSV-infected targets with tunicamycin prevented glycosylation of newly synthesized VSV glycoprotein G on the cell membrane. This treatment abrogated completely NK-mediated killing of the infected B16 melanoma cells. These results indicate that virus replication and membrane expression of glycosylated protein G are essential for recognition and lysis of VSV-infected targets by NK ceils.
2. Introduction Target recognition by NK cells appears to be a spontaneous non-specific and M H C unrestricted event [1-3], yet it is well established that definite target structures must be involved [3, 4]. Cells that are resistant to lysis by NK cells become highly susceptible when they are acutely or persistently infected with virus [5-7]. The possibility that virus encoded proteins were the major target structure was suggested as early as a decade ago [8], but conflicting data were also obtained [9] and the subject remains unclear. We have recently used temperature sensitive mutants of VSV to demonstrate that expression of two viral proteins is required for NK cell recognition [10]. The present data further supports the view that virus replication and glycoprotein expression in the infected target cell are essential for lysis by NK cells.
0165-2478 / 86 / $ 3.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
261
3. Materials and Methods
3.1. Mice 5-8 week old C B A / J or C57BL/6 mice from our own breeding colony were used in all experiments. 3.2. Viruses The Indiana (HRC) wild type (wt) serotype of vesicular stomatitis virus (VSV) was used. Stock virus was grown in primary chick embryo fibroblasts. VSV (wt) was kindly provided by Dr. P. I. Marcus, and propagated on vero cell monolayers at the permissive temperature of 31 °C [11]. 3.3. Cells Clone M2R derived from B16 mouse melanom a cells [12], obtained from Dr. J. Mather, Rockefeller University, New York was grown in vitro in Dulbecco's modified Eagle's medium (DME/F12, Gibco, Grand Island, NY), supplemented with 10O7o fetal calf serum (FCS), glutamine and antibiotics. The cells were recloned, the subclone most resistant to NK lysis being selected for use in these experiments. In addition, clone J774.16, a continuous macrophage-like cell line derived from the murine reticulum cell sarcoma J774.16 was also used. These cells were propagated and cultured as previously described [13]. 3.4. Tunicamycin (TM) TM (Calbiochem, Calbiochem-Behring Corp., La Jolla, CA) was added at a final concentration of 0.1, 0.5 or 1 # g / m l to target cells 2 h before virus infection. 3.5. Ant&era Polyclonal rabbit anti-VSV, anti-VSV-G protein and anti-VSV-M protein sera were generously provided by Dr. J. Holland. Monoclonal antiVSV-G ascites was a gift from Drs. R. Weiss and H. Koprowski. 3.6. Immunofluorescent staining Indirect immunofluorescence was used to visualize viral antigen expression. The specific antisera and FITC conjugated antibody (the lat262
ter purchased from Cappel, Cochranville, PA 19330) were absorbed with an equal volume of uninfected B16 mouse melanoma cells for 30 min at 4 °C and subsequently centrifuged (15,000 rpm) to remove cells and aggregates. For cytoplasmic immunofluorescent staining, cells were grown in Lab-Tek tissue culture chamber slides (Lab-Tek Div., Miles Labs Inc., Nagerville, IL 60540) until nearly confluent. Monolayers were washed, air-dried and fixed in cold acetone. Unfixed monolayers were used for membrane fluorescence. 3.7. Preparation of effector cells Spleens from C B A / J or C57BL/6 mice were passed through a fine Sellector sieve (Bellco, Inc., Vineland, NH), washed 3 times with phosphatebuffered saline (PBS) and resuspended in medium at a concentration of 1 × 107 cells/ml. 3.8. Treatment with monoclonal anti-Thy 1.2 Monoclonal anti-Thy 1.2 antibody was purchased from New England Nuclear, Boston, MA. Spleen cells (5 × 107 cells/ml) were incubated with the antibody (final dilution 1:1,000) and rabbit low toxicity complement (c'; Low Tox, Cedarlane, Ontario, Canada; final dilution 1:12) for 45 min at 37 °C in 5°70 CO2. 3.9. Treatment with alloantiserum N K 1.1 Mouse anti NKI.1 [14] was kindly provided by Dr. S. Pollack (Seattle, WA). Spleen cells at 3 × 107 cells were incubated with anti NKI.1 (final dilution 1:30) and rabbit c' (final dilution 1:30) for 45 min at 37 °C. Treatment of C57BL/6 mouse spleen cells with this antiserum reduced the splenic cytotoxicity against YAC-1 cells by 50%. Rabbit c' alone had no effect. 3.10. Virus infection and labeling of target cells Virus infected target cells were prepared by pelleting 5 × 10 6 cells and resuspending the pellet in 0.5 ml of virus suspension and 100 #Ci of NAz51CrO4 at multiplicities of infection (MOI) of 0.1-0.001. After 2 h incubation at 31 °C, v i r u s - c e l l mixtures were washed twice in medium and then seeded at 2×104 cells per well in 96 well flat bottom microtiter plates (Linbro, 76-003-05) and incubated overnight at 37 °C.
3.11. Cytotoxicity assay 51Cr-Iabeled monolayers of uninfected or VSVinfected target cells were washed again with medium to remove nonadherent and dead cells. Spleen cell suspensions (0.2 ml) were then added to the target cells, centrifuged 5 min at 600 rpm and incubated at 37°C for 8 - 1 0 h. After incubation, plates were centrifuged, 100/~1 of supernatants were harvested for each well and then 100 #1 of 2 N HC1 were added to each well for estimation of maximal 51Cr release. The specific 51Cr release was calculated as described previously [61. 3.12. Analysis of glycosylated and non-
Table 1 Replication of VSV in BI6 m e l a n o m a cells and in a continuous macrophage (M~b) like cell line J774.16 Input virus (pfu)
5×104 5x103 5 x 102 5 × 101
MOI
0.1 0.1 0.001 0.0001
Virus yields (pfu/ml) B16 melanoma
M~
3×105 5x105 1 x 104 3 × 103
6 3 1 4.5
J774.16 ×107 ×107 x 107 x 106
Virus yields in supernatants were measured 24 h after infection. Virus titers are expressed in plaque forming units (pfu) per 1 ml. Virus titrations were performed as described in Materials and Methods.
glycosylated G protein B16 m e l a n o m a cells were pretreated with 1 # g / m l TM for 2 h at 37 °C, and were infected with VSV (wt) at an M O I of 10-20. After adsorption for 2 h at 31 °C, [3H]leucine was added. Cells were incubated for an additional 16 h at 37 °C and then solubilized in sodium dodecyl sulphate (SDS) buffer. Cell extracts were immunoprecipitated and analyzed on polyacrylamide gel electrophoresis (SDS-PAGE) [15].
4. Results
4.1. Replication of VSV in B16 melanoma cells
and in a continuous macrophage-like cell line J774.16 Since VSV is a lytic virus and rapidly inhibits host cell protein synthesis, it was important to find target cells which will be permissive for VSV but will not undergo extensive lysis during the incubation period with the effector cells. A large number of cell lines were screened. Table 1 exemplifies susceptibility to infection with VSV of two cell lines, a macrophage-like cell line J774.16 which is highly susceptible to VSV, and B16 m e l a n o m a which were permissive but less susceptible to infection with this virus. The macrophage cell line J774.16 was chosen since VSVinfected macrophages have t~een previously used as targets in the study of mechanisms of killing by C T L [16]. Monolayers of J774.16 and B16 m e l a n o m a were infected at a range of multiplici-
ties of 0.1-0.0001, J774.16 was found to be extremely susceptible to VSV. After 24 h, complete cell destruction occurred at multiplicities of 10 -1 --10 -3. High yields of VSV ranging from 6x107 plaque forming units (pfu)/ml to 4.5 x 106 p f u / m l were measured in supernatants of infected cultures. When B16 melanoma cells were infected with the same multiplicities of VSV, no cytopathic changes were detected after 24 h while VSV replication occurred as determined by virus yields in supernatants of the infected cultures. Levels of spontaneous 5~Cr release from VSV infected B16 cells were < 30°7o in all experiments. B16 m e l a n o m a cells were therefore chosen as target cells for our studies. 4.2. Expression of VSV antigens in B16 melano-
ma cells As shown above, VSV replicates in B16 melan o m a cells without producing cytopathic effects 24 h after infection. It was important however, to determine whether VSV proteins are produced and expressed by the infected cells within this period. As shown in Fig. 1, the B16 melanoma cells expressed high amounts of VSV proteins based on the intensity of the fluorescent staining and the number of positive cells ( 7 0 - 8 0 % ) . A polyclonal rabbit anti-VSV serum reacted strongly with the VSV-infected B16 melanoma cells showing both membrane and cytoplasmic fluorescence (Fig. 1A, B). In contrast, the monoclonal anti 263
V S V - G ascites s h o w e d o n l y m e m b r a n e , b u t n o c y t o p l a s m i c b i n d i n g (Fig. 1C), w h i l e t h e anti-VSVM p r o t e i n sera s h o w e d a reverse p a t t e r n w i t h cytoplasmic binding but no membrane fluoresc e n c e (Fig. 1D). 4.3. Lysis of B16 melanoma cells infected by VSV V S V - i n f e c t e d o r u n i n f e c t e d B16 m e l a n o m a cells were l a b e l e d w i t h 51Cr, i n c u b a t e d w i t h s u s p e n sions o f s p l e e n cells a n d t h e s p e c i f i c 51Cr release was d e t e r m i n e d . Target cell lysis in this t y p e o f s y s t e m m a y b e m e d i a t e d by N K cells o r o t h e r eff e c t o r cells m e d i a t i n g n a t u r a l c y t o t o x i c i t y [6, 17]. By v a r y i n g t h e t i m e o f i n f e c t i o n w i t h VSV a n d t h e M O I , it was f o u n d t h a t viral p r o t e i n s y n t h e sis was a n o b l i g a t o r y r e q u i r e m e n t for lysis o f V S V - i n f e c t e d targets. A s h o r t (3 h) i n f e c t i o n t i m e at a h i g h m u l t i p l i c i t y o r a l o n g (18 h) i n f e c t i o n t i m e u s i n g U.V. i n a c t i v a t e d VSV leaves t h e B16 m e l a n o m a t a r g e t resistant to lysis (Table 2).
Table 2 Requirement of VSV replication for lysis of B16 melanoma cells by NK cells Time of infection (h)
MOI
% 51Cr specific release
FA a
Uninfected 3 18 18 18
10 0.01 0.0001 U.V. inactivated
5 _+4 0 23 _+4 7_+2 0
0 0 80 5 - 10 0
a Percent positive cells. BI6 melanoma cells were infected with VSV (wt) for different periods of time and different MOI as indicated. The 5~Cr-release assay was performed at 37 °C at an effector to target ratio of 100:1 with an incubation period of 8 h. C57BL/6 spleen cells were used. Results are the average of three experiments. The mean spontaneous release was 24%.
Fig. 1. lmmunofluorescent staining of BI6 melanoma cells infected with ts mutants of VSV. (A) Membrane and (B) cytoplasmic staining of B16 cells infected with VSV wt at 37 °C with rabbit anti-VSV antibody. (C) Membrane staining of B16 cells infected with VSV at 37 °C with monoclonal anti-VSV G antibodies. (D) Cytoplasmic staining of BI6 cells infected with VSV at 37 °C with rabbit anti-VSV M antibodies. Magnification ×400. 264
However, a long (18 h) infection time with 1,000-fold less virus results in specific lysis. T h e effect was d o s e - d e p e n d e n t a n d w h e n less virus was used for infection, n o significant lysis occurred. At this low M O I , the percent o f FA positive cells decreased from 800/0 to less t h a n 100/0. 4.4. Characterization of effector cells Table 3 s u m m a r i z e s the effect of p r e t r e a t m e n t o f the effector ceils in this system with specific antisera to m e m b r a n e markers. Treatment with a n t i - N K 1.1 + c' or with anti-Thy 1.2 + c' decreased the specific lysis o f VSV-B16 m e l a n o m a target cells by 50%. Therefore, the p h e n o t y p e o f the p r e d o m i n a n t effector cell in this system corresponds to the N K I a n d N K x subsets previously described [18]. 4.5.
I n order to c o n f i r m this observation, we analyzed the effects of T M t r e a t m e n t o n the glycosylation of VSV G after infection of B16 m e l a n o m a cells. Cells were pretreated with TM, infected with VSV (wt), harvested 18 h post infection, i m m u n o p r e c i p i t a t e d with rabbit anti-VSV G protein a n t i s e r u m a n d analyzed o n SDS-PAGE gels. Results shown in Fig. 2 indicate that indeed only nonglycosytated G was synthesized in the presence of TM. The non-glycosylated G was present o n the cell m e m b r a n e as seen by imm u n o f l u o r e s c e n t surface staining with m o n o c l o n a l a n t i - G antibody. As shown in Fig. 3, A
13
Requirement for glycosylation of G proteins for NK cell recognition of virus infected targets
It has been recently established that the a n t i b i otic, t u n i c a m y c i n (TM) inhibits the synthesis of N-acetyl g l u c o s a m i n e - l i p i d intermediates a n d thereby prevents glycosylation o f newly synthesized VSV glycoprotein [19]. P r e t r e a t m e n t o f VSV-infected cells with T M almost totally inhibits infectious virus p r o d u c t i o n , a n d cells grown in the presence o f T M express n o n glycosylated G protein o n their cell surface [20].
Table 3 Phenotype characterization of the effector cells Pretreatment
% specific lysis (% decrease)
None c' alone anti-NK 1.1 + c' anti-Thy 1.2 + c'
23.9_+ 1.4 22.6_+2.9 11.5___1.0 11.1+ 1.7
(0) (52) (53)
C57BL/6 spleen cells were pretreated with anti-NK 1.1 antibodies or with anti Thy 1.2 antibodies as described in Materials and Methods. The 51Cr-release assay was performed at 37 °C at an effector to target cell ratio of 100:1 with an incubation period of 8 - 10 h. Target cells were B16 melanoma infected with VSV (wt). Results are the average of triplicates of three experiments. The mean spontaneous release was 24%.
Fig. 2. SDS-PAGE electrophoresis of precipitates from VSVinfected B16 melanoma cells with or without TM. B16 melanoma cells were pretreated with 1 ~g/ml TM for 2 h at 37 °C and infected with VSV at a MOI of 10-20. [3H]Leucinewas added 2 h later and cells were incubated for an additional 16 h at 37 °C. Cell extracts were immunoprecipitated with rabbit anti-VSV antiserum. (A) G protein of TM-treated VSVinfected cell lysates. (B) Nonglycosylated G protein of TMtreated VSV-infectedcell lysates. 265
100 8O •-
60
..421
-~
40 20
0
0.5
1.0
Tunicamycin (Mg) Fig. 3. Inhibition of lysis of VSV-infected cells by TM. B16 m e l a n o m a cells were pretreated with TM final concentrations of 1.0 ~g or 0.5 /~g, infected with VSV and incubated for 16 11 at 37 °C. The cytotoxicity assays were performed at 3 7 ° C with an effector to target ratio 100:1. S p o n t a n e o u s release did not exceed 30°7o. Results are the average of triplicates of 3 experiments. Uninfected B16 cells ( z~); VSV infected ( • ); YAC-1 cells ( n ).
TM inhibited the killing of B16 melanoma cells infected with VSV (wt), indicating that appropriate glycosylation was also required for NK cell lysis. In contrast, incubation of YAC-1 target cells in parallel under the same conditions with TM had no effect on the susceptibility of these cells to NK cell lysis. This result independently confirms the importance of appropriately processed G protein in recognition by NK cells.
5. Discussion
Target cell recognition by NK cells appears to be a spontaneous event which is nonspecific in the immunological sense, since xenogeneic and allogeneic targets can be lysed, and since there is no M H C restriction [3]. Yet, many observers have noted that different tumor cells have different susceptibilities to lysis by NK and that an acute or persistent virus infection dramatically augments this susceptibility [ 4 - 7 ] . These results indicate a highly "selective" recognition by the NK cells for certain target structures that are apparently associated with the infecting virus. However, the nature of these structures remains unclear. 266
In order to analyze the mechanism by which virus infection converts resistant cells to susceptible targets for NK, we have used murine B16 melanoma cells and vesicular stomatitis virus (VSV). Studies by Zinkernagel et al. [16] who also used VSV infected targets to study the mechanism of recognition by cytotoxic T lymphocytes (CTL), have shown that the VSV glycoprotein was the major viral encoded determinant that must be expressed for recognition by CTL. Experiments published by other groups which were carried out in the presence of the antibiotic, tunicamycin, indicated that glycosylation of the G protein was not an absolute requirement for recognition and lysis of VSV-infected target cells by CTL [21]. We have applied this model to study recognition of target structures by NK cells; however, a basic technical difficulty, not encountered in the CTL experiments, had to be overcome. While the virus infected targets are lysed within 4 h by CTL, incubation periods of 8 h or longer are generally required in order to obtain significant lysis of virus infected targets by NK [6]. However, since VSV is a lytic virus, extended incubation periods lead to unacceptably high spontaneous 51Cr-release from virus-infected cells. Furthermore, a murine target cell had to be identified that would be permissive to VSV infection, but would not be destroyed by the replicating virus. After screening a large number of target cells, we found that the B16 melanoma cells were indeed an appropriate cell line. Immunofluorescent staining confirmed that the VSV infected cells expressed high amounts of VSV proteins and that a high proportion of the cells were positive. Our results indicate that replication of virus was a prerequisite for target lysis by NK, since uninfected B16 melanoma cells were resistant to lysis, and infection with U.V. inactivated virus did not significantly alter susceptibility. Similarly, infection with a high MOI for a short time insufficient for de novo VSV protein synthesis, did not render the cells susceptible. However, when target cells were infected with virus at a MOI of 10 -2 for 18 h, they became NK sensitive with a more than four-fold increase in specific lysis. In order to obtain statistically significant killing, and yet no unacceptably high
backgrounds, a compromise had to be made resulting in spontaneous release values of 1 5 - 3 0 % over 8 h, and levels of specific cytotoxicity seldom exceeding 30%. The 50% decrease in specific lysis following treatment of the spleen cells with anti-NK 1.1 or anti-Thy 1.2 antibodies confirms that N K cells are the main cell type responsible (Table 3). Having established that intracellular VSV replication was essential for target cell recognition by NK, we proceeded to investigate the viral proteins involved. Since differences in recognition are usually associated with alterations in the outer cell membrane, and since the viral glycoprotein G was abundantly found on the B16 melanoma cell membrane (Fig. lc), we examined the effect of tunicamycin on NK cell lysis. Cells pretreated with tunicamycin and infected with VSV were analyzed by immunoprecipitation and SDS-PAGE gels and showed only nonglycosylated G protein on the cell membrane (Fig. 2). This absence of glycosylated G protein from the target cell membrane was associated with a marked inhibition of NK-mediated cytotoxicity. Thus, VSVinfected NK sensitive targets were no longer recognized and killed when the normally glycosylated G protein was not expressed on the membrane (Fig. 3). Of interest is the observation that tunicamycin treatment of a standard NK target, YAC-1 cells, had no effect on killing, indicating that its target structures were unaffected by this antibiotic. It has been previously shown that decreased glycoprotein expression on the surface of an infected cell correlates directly with the ability of that cell to resist killing by both CTL and cytotoxic antibody and complement [22, 23]. This in fact may be the mechanism underlying some instances of persistent infections in vivo [24]. Our findings suggest that decreased membrane expression of glycosylated viral protein may also be related to NK resistance (Fig. 3). It is possible that in the absence of glycosylated G protein, the virus infected cell evades recognition and binding by the NK cell, and that the interference is not at the lytic phase itself. We also found that killing could not be inhibited by antibodies against VSV, and that antibodies to interferon sufficient to neutralize inter-
feron produced within the assay system itself [6, 10], did not affect the degree of killing observed (unpublished results). This indicates that the interferon produced is not involved in the lysis observed in this system. In summary, our observations support the role of viral encoded glycoproteins expressed on the outer cell membrane in target recognition by NK. Furthermore, the presence of an identical transmembrane protein G which lacks its sugar residues was insufficient for NK recognition and killing. Additional support for our theory that the membrane glycoprotein is not the sole target structure in this system comes from our experiments which show that VSV replication was required in order to obtain lysis by NK. Recent studies with VSV G plasmids indeed suggest that the VSV M protein which is a submembranal protein may contribute to the formation of the target structure [10]. Another possible mechanism which would also be in line with our observations would be a change in the putative host cell membrane structure induced by the viral glycoprotein. Several observations suggest an inverse correlation between H-2 expression and NK sensitivity [23, 26]. A decrease in H-2 expression in VSV infected cells [27] could explain the increased sensitivity to NK, and these changes and their relationship to viral glycoproteins will be studied by us in the future.
Acknowledgements Supported by USPHS Grants A1099807 and Al10702, Grant 1006 from the National Multiple Sclerosis Society and Grant BC301 from the American Cancer Society. A.S. is a fellow of the Camp David Institute of International Health. B.R.Z. is a recipient of an ICRF Career Development Award.
References [1] Herberman, R. B. and Ortaldo, J. R. (1981) Science 214, 24. [2] Welsh, R. M. Jr. (1978) J. Immunol. 121, 1631. [3] Rager-Zisman, B. and Bloom, B. R. (1982) Springer Sem. Immunopathol. 4, 397.
267
[4] Schattner, A. and Duggan, D. (1985) Am. J. Hematol. 18, 435. [5] Welsh, R. M. (1981) Antiviral Res. 1, 5. [6] Minato, N., Bloom, B. R., Jones, C., Holland J. and Reid, L. M. (1979) Mechanism of rejection of virus persistently infected tumor cells by athymic nude mice. J. Exp. Med. 149, 1117. [7] Schattner, A., Rager-Zisman, B. and Bloom, B. R. (1985) Cell Immunol. 90, 103. [8] Kiessling, R., Klein, E. and Wigzell, H. (1975) Eur. J. lmmunol. 5, 112. [9] Becker, S., Fenyo, E. M. and Klein, E. (1976) Eur. J. Immunol. 6, 882. [10] Moller, J., Rager-Zisman, B., Quan, P. C., Schattner, A., Panish, D., Rose, J. K. and Bloom, B. R. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 2456. [11] Lodish, E H. and Weiss, R. A. (1979) J. Virol. 30, 177. [12] Mather, J. P. and Sato, G. H. (1979) Exp. Cell Res. 124, 215. [13] Rager-Zisman, B., Kunkel, M., Tanaka, Y. and Bloom, B. R. (1982) Infect. Immun. 36, 1229. [14] Pollack, S. B., Tam, M. R., Nowinsky, R. C. and Emons, S. L. (1979) J. Immunol. 23, 1818. [15] Birrer, M. J., Bloom, B. R. and Udem, S. (1981) Virol. 108, 381. [16] Zinkernagel, R. M., Althage, A. and Holland, J. (1978)
268
J. Immunol. 121,744. [17] Stutman, O., Lattime, E. C. and Figarella, E. C. (1981) Fed. Proc. 40, 2699. [18] Minato, N. L., Reid, M. and Bloom, B. R. (1981) J. Exp. Med. 154, 750. [19] Leavitt, R., Schlessinger, S. and Kornfeld, S. (1977) J. Virol. 21, 375. [20] Gibson, R., Leavitt, R., Kornfeld, S. and Schlessinger, S. (1978) Cell 13, 671. [21] Harris, D. T., Hale, A. H. and Lefrancois, L. (1981) J. lmmunol. 126, 1914. [22] Welsh, R. M. and Oldstone, M. B. A. (1977) J. Exp. Med. 145, 1449. [23] Oldstone, M. B. A. and Fujinami, R. S. (1982) in: Virus Persistence, Symp. 33, Soc. Gen. Microbiol. (B. W. J. Mahy, A. C. Minson and G. K. Darby, Eds.) p. 185, Cambridge University Press, Cambridge. [24] Oldstone, M. B. A. and Buchmeier, J. (1982) Nature 300, 360. [25] Gidlund, M., Orn, A., Pattengale, P. K., Jansson, M., Wigzell, H. and Nilsson, K. (1981) Nature 292, 848. [26] Karre, K. (1985) in: Mechanisms of Cytotoxicity by Natural Killer Cells (R. B. Herberman and D. Callewaert, Eds.) p. 81, Academic Press, Orlando. [27] Hecht, T. and Summers, D. E (1972) J. Virol. 10, 578.