Control of Lymphocytic Choriomeningitis Virus Infection in Granzyme B Deficient Mice

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Virology 305, 1–9 (2003) doi:10.1006/viro.2002.1754

Allan J. Zajac,*,† John M. Dye,* and Daniel G. Quinn*,1

*Department of Microbiology and Immunology, Loyola University Chicago, Maywood, Illinois 60153; and †Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294 Received July 11, 2001; returned to author for revision October 5, 2001; accepted August 30, 2002

We have investigated whether granzyme B (GzmB)is required for effective cytotoxic T lymphocyte (CTL)mediated control of lymphocytic choriomeningitis virus (LCMV)infection. Clearance of LCMV from tissues of GzmB-deficient (GzmB Ϫ)mice following intraperitoneal infection with LCMV was impaired compared with control mice; however, the virus was ultimately eliminated. The impaired clearance of LCMV in GzmBϪ mice was not due to a deficiency in the generation of LCMV-specific T cells. In addition, CTL from LCMV-infected GzmBϪ mice efficiently lysed virus-infected cells in vitro, but were deficient in their ability to induce rapid DNA fragmentation in target cells. We examined whether the development of protective immunity against intracranial (i.c.)rechallenge with LCMV was compromised in GzmB Ϫ mice. We found that clearance of LCMV from the brain following secondary i.c. infection also was slower in the absence of GzmB; however, the virus was ultimately eliminated and the mice survived. Our data indicate that clearance of LCMV is delayed in the absence of GzmB expression, but that other CTL effector molecules can compensate for the absence of this granule constituent in vivo. © 2002 Elsevier Science (USA) Key Words: granzyme B; LCMV; cytotoxic T lymphocyte.

INTRODUCTION pathway of target cell killing is frequently used by CD4 CTL (Graubert et al., 1997; Zajac et al., 1996) but can be T cells play a critical role in the immune response also utilized by CD8 CTL (Ka¨gi et al., 1994b; Lowin et al., against viral infections. In general, the antiviral effects of 1994; Kojima et al., 1994). The major pathway of cytotox- T cells are mediated through the release of antiviral icity utilized by CD8 CTL, however, is the secretory path- cytokines or via direct killing of virus-infected cells by way. In this pathway, the release of perforin from gran- cytotoxic T lymphocytes (CTL) (reviewed by Doherty and ules contained within activated CTL leads to the forma- Ahmed, 1997; Zinkernagel, 1996). Control of infection with tion of polyperforin pores in the membrane of the target noncytopathic viruses, such as lymphocytic choriomeningitis virus (LCMV), is crucially dependent on CD8 CTL. cell, resulting in loss of membrane integrity. The CTL Infection of normal mice with LCMV induces high levels granules also contain enzymes called granzymes. The of virus-specific CD8 CTL activity and the virus is cleared major granzymes released by CD8 CTL are granzyme A within 8–10 days following infection. By contrast, mice (GzmA) and GzmB. Of these, GzmB has been shown to deficient in CD8 T cells fail to control LCMV and become bind to a specific receptor on target cell membranes persistently infected (Leist et al., 1987; Moskophidis et (Motyka et al., 2000) and to induce rapid apoptosis of al., 1987; Zajac et al., 1995). Since clearance of acute target cells in vitro (Simon et al., 1997). Upon entering the infection with LCMV is solely dependent on CD8 CTL target cell, GzmB is thought to proteolytically activate (Leist et al., 1987; Moskophidis et al., 1987), infection with several caspases culminating in DNA fragmentation and this virus provides a useful system with which to study in the rapid apoptotic death of the target cell (Darmon et the effector mechanisms of CD8 CTL in vivo. al., 1995; Duan et al., 1996; Gu et al., 1996; Quan et al., Two major pathways of CTL-mediated cytotoxicity 1996; Talanian et al., 1997). In addition, GzmB has been have been described. These are the secretory and non- reported to directly induce apoptosis independent of secretory pathways (Ka¨gi et al., 1994a,b; Lowin et al., caspase activation (Andrade et al., 1998; Talanian et al., 1994; Kojima et al., 1994). In the nonsecretory pathway, 1997). the interaction between Fas ligand (FasL) on the surface The primacy of the secretory pathway of CTL-medi- of the activated CTL and Fas on the surface of the target ated killing for the elimination of LCMV infection has cell results in apoptotic death of the target cell. This been conclusively demonstrated in studies using mice deficient in perforin (Ka¨gi et al., 1994a, 1995b; Walsh et al., 1994). Perforin-deficient mice do not have detectable 1 To whom correspondence and reprint requests should be ad- Fas-independent CTL and, consequently, fail to control dressed. Fax: (708) 216-9574. infection with LCMV. It is unclear, however, whether

mice (B6129SF2/J mice). We found no difference in the kinetics of clearance of LCMV from tissues of B6 mice and from tissues of B6129SF2/J mice. The control mice indicated in Fig. 1 represent three B6 mice and three B6129SF2/J mice at each time point. The levels of virus in both spleens and livers of GzmBϪ mice were similar to the levels in these tissues of control mice up to 4 days following infection. In contrast to control mice, however, peak viral titers are maintained in GzmBϪ mice until day 6 following intraperitoneal (ip) infection. By day 8 after infection splenic viral titers are substantially reduced in GzmBϪ mice, but the levels of virus are still significantly higher than the levels detected in the spleens of control mice at this time point. By contrast, although the viral titers in livers seem somewhat higher in GzmBϪ mice than B6 mice at this time point, this difference was not significant. By 10 days after infection, infectious virus is undetectable in spleens and livers of GzmBϪ mice and these animals remain virus-free for greater than 120 days following infection (data not shown).

Expansion of virus-specific T cells in GzmBϪ mice Since acute infection with LCMV is controlled by CTL, we reasoned that the delayed clearance of LCMV by FIG. 1. Kinetics of clearance of LCMV from spleens and livers of GzmBϪ CTL could be due to either of two possibilities. Ϫ ϫ 5 GzmB mice. Mice were infected with 2 10 PFU of LCMV. Spleens First, the effector function of GzmBϪ cells could be suf- (A) and livers (B) were removed at the indicated times following infection and viral titers were determined by plaque assay on Vero cell ficiently compromised such that they function less effi- Ϫ monolayers. ■, wild-type mice, n ϭ 6 at each time point (three B6 and ciently in vivo. Second, it is possible that the GzmB mice three B6129S2/J); Œ, GzmBϪ mice, n ϭ 5 at each time point; F, PfpϪ used in this study generated fewer numbers of virus- mice, n ϭ 3 at each time point. Data shown are the means Ϯ SEM of specific T cells following infection with LCMV. We first Ͻ Ͻ the viral titers in spleens at each time point. *P 0.01; **P 0.05. addressed the possibility that the delayed clearance of LCMV in GzmBϪ mice is due to the expansion of fewer virus-specific T cells compared with normal mice. We perforin-mediated loss of membrane integrity is suffi- enumerated T cells specific for the immunodominant cient for elimination of LCMV-infected cells in vivo or if viral CD4 and CD8 epitopes at various times following ip rapid induction of target cell apoptosis is also required. infection with LCMV using ELISPOT to detect IFN-␥- In this study we have investigated whether GzmB ex- secreting cells. There is a good correlation between the pression and, consequently, the rapid GzmB-dependent numbers of IFN-␥ producing CD8 cells and the total induction of DNA fragmentation in target cells are critical numbers of virus-specific cells (Murali-Krishna et al., for CTL-mediated control of LCMV infection of mice. 1998). As previously demonstrated (Butz and Bevan, 1998; Kamperschroer and Quinn, 1999; Murali-Krishna et RESULTS al., 1998), the numbers of virus-specific CD8 cells peak at

Ϫ 8 days postinfection with LCMV (Figs. 2A–2C). At this Kinetics of virus clearance in GzmB mice 7 time point there are, combined, approximately 5 ϫ 10 We initially examined whether GzmBϪ mice could con- cells in the spleen that are specific for the three domi- trol infection with the noncytopathic virus, LCMV. Infec- nant class I restricted epitopes of LCMV. The numbers of tion of normal mice with LCMV results in peak viral titers virus-specific cells subsequently decline with approxi- in spleens and livers at day 4 postinfection, following mately 2 ϫ 10 6 cells being maintained for greater than 90 which virus is controlled and eliminated by 8–10 days days following infection. As seen in Fig. 2, the numbers postinfection (Fig. 1). That this control is dependent on of virus-specific CD8 cells that are generated and main- CTL is illustrated by the persistent infection that occurs tained in GzmBϪ mice are indistinguishable from the in PfpϪ mice (Fig. 1) (Ka¨gi et al., 1994a; Walsh et al., 1994). numbers of these cells in B6 mice. Since CD4 cells are Because the GzmBϪ mice used in this study were on a important in promoting and maintaining the function of mixed B6x129 genetic background, we also assessed CD8 cells (Battegay et al., 1994; Matloubian et al., 1994; virus clearance from spleens and livers of control B6x129 von Herrath et al., 1996; Zajac et al., 1998), we examined LCMV INFECTION OF GzmB-DEFICIENT MICE 3

FIG. 2. Numbers of LCMV-specific T cells following infection of B6 and GzmBϪ mice. Spleen cells were prepared from B6 mice (F) and from GzmBϪ mice (E) at the indicated days after ip infection with LCMV. The numbers of spleen cells that secrete IFN-␥ in response to stimulation with the class I restricted LCMV epitopes GP33-41 (A), NP396-404 (B), or GP276-286 (C), and with the class II restricted epitopes GP61-80 (D) or NP309-328 (E) were determined by ELISPOT. Data shown are the means Ϯ SEM of the numbers of epitope-specific cell/spleen from a minimum of three mice at each time point. whether the defect in virus clearance that we observe in Together these data indicate that the impaired clear- Ϫ GzmB mice could be explained by inefficient expansion ance of LCMV from GzmBϪ mice is not due to expansion of virus-specific CD4 cells. Using ELISPOT, we enumer- of insufficient numbers of virus-specific T cells. We next ated T cells specific for the immunodominant I-Ab-re- examined whether the epitope-specific CD8 cells gener- stricted epitopes of LCMV in spleens of infected mice. ated during LCMV infection of GzmBϪ mice were capa- We found that the numbers of epitope-specific I-Ab-re- ble of killing virus-infected cells in vitro. stricted cells in spleens of GzmBϪ mice were similar to the numbers of these cells in spleens of B6 mice at all Virus-specific cytolytic activity by GzmBϪ CTL time points following infection with LCMV (Figs. 2D–2E). In addition, the percentages of CD4 cells and CD8 cells Since clearance of LCMV requires cytolytic T cell func- in spleens of GzmBϪ and B6 mice following infection tion (Ka¨gi et al., 1994a; Walsh et al., 1994), we examined were similar at all time points examined (data not whether GzmBϪ T cells could efficiently kill LCMV-in- shown). fected target cells in vitro. As shown in Figs. 3A and 3B, 4 ZAJAC, DYE, AND QUINN

Control of secondary infection in GzmBϪ mice The basis of protective immunity is that the memory cells generated following an initial exposure to a pathogen will allow for a more rapid response against that same pathogen upon secondary exposure. We wished to examine whether GzmBϪ CTL could eliminate virus suf- ficiently rapidly to protect the host from a normally lethal challenge with LCMV. Infection of naı¨ve B6 or GzmBϪ mice intracranially (i.c.) with LCMV results in death (Fig. 4). All mice that were depleted of CD8 cells prior to infection survived as expected, since lethal LCMV disease results from CTL-mediated destruction of infected cells in the brain (Baenziger et al., 1986; Leist et al., 1987; Moskophidis et al., 1987). Normal mice that have previously been exposed to LCMV are protected from lethal immunopathology following i.c. rechallenge with LCMV because they have developed protective immunity against this virus (Borrow and Oldstone, 1997; Buch- meier and Zajac, 1998). Similar to wild-type mice, we found that GzmBϪ animals also survived secondary i.c. infection with LCMV (Fig. 4). However, brain tissue taken from GzmBϪ mice 3 days after i.c. rechallenge with 2 ϫ 3 FIG. 3. Virus-specific cytolytic activity. B6 and GzmBϪ mice were 10 PFU LCMV contained 10 to 100 times more infectious infected ip with 2 ϫ 10 5 PFU of LCMV. Eight days later, spleen cells virus than brains from B6 mice at the same time point from these mice were tested for cytotoxic activity against LCMV in- (Fig. 5). These data indicate that GzmBϪ mice are not fected (solid lines) and noninfected (broken lines) MC57 cells using only compromised in their ability to control peripheral 51Cr-release assays (A, B) or JAM assays (C, D). The assays were incubated for either 4 h (A, C) or 8 h (B, D). F, B6; ■, GzmBϪ; Œ, PfpϪ. infection with LCMV, but are also less efficient at con- Representative results from four (A, B) and three (C, D) independent trolling LCMV infection within the CNS. The virus was, Ϫ experiments are shown. however, eliminated from the brains of GzmB mice by day 4 following rechallenge with no overt detrimental effect on the host. Similar results were obtained when spleen cells taken from B6 and GzmBϪ mice 8 days after animals were rechallenged i.c. with 2 ϫ 10 5 PFU LCMV infection with LCMV are equally potent at killing LCMV- (data not shown). infected target cells in vitro as measured using a 51Cr- Since LCMV is cleared from the CNS by CD8 T cells, release assay. Since the target cells used in this assay we next examined whether the delayed clearance of Ϫ do not express Fas, this assay is not measuring Fas- LCMV from the brains of GzmB mice was due to de- FasL-dependent cytotoxicity. Furthermore, the inability of effector cells from PfpϪ mice to kill the infected target cells indicates that this assay is measuring perforin- dependent killing activity. We next measured the ability of the effector cells to induce DNA fragmentation in the target cells using the JAM assay. In contrast to cells from B6 mice, the GzmBϪ cells are unable to induce rapid (within 4 h) DNA fragmentation in infected target cells (Fig. 3C). However, LCMV-infected target cells exposed to GzmBϪ CTL undergo DNA fragmentation by8hof culture (Fig. 3D). Overall, these data indicate that LCMV infection of GzmBϪ mice results in expansion of virus- FIG. 4. Protective immunity against secondary i.c. infection. Nonin- specific CTL, which kill infected cells in vitro using the fected mice (1°, broken lines) or mice that had been infected ip three secretory pathway of CTL-mediated lysis, but which fail months previously with LCMV (2°, solid lines) were infected i.c. with to induce rapid apoptosis of LCMV-infected cells. Our 2 ϫ 10 3 PFU of LCMV. One group of naive GzmBϪ mice was depleted of CD8 cells prior to i.c. infection with LCMV. The groups of mice are finding that clearance of LCMV is impaired in the ab- Ϫ sence of GzmB suggests that this inability of GzmBϪ indicated on the figure. All of the B6 and GzmB mice that had been previously infected ip with LCMV survived the i.c. rechallenge; conse- cells to induce rapid apoptosis of LCMV-infected cells is quently the lines representing these groups of mice are superimposed of consequence in vivo. on the figure. Each group consisted of either five or six mice. LCMV INFECTION OF GzmB-DEFICIENT MICE 5

expansion of virus-specific T cells in these animals. We found that GzmBϪ and B6 mice had equivalent numbers of virus-specific CD4 and CD8 T cells at all time points examined following infection. This indicates that the delay in clearance of virus from these animals cannot be explained by a failure to generate sufficient numbers of antiviral T cells. It has been conclusively demonstrated that clearance of acute infection with LCMV is dependent on CD8 T cells (Leist et al., 1987; Moskophidis et al., 1987). Since we found no difference in the numbers of FIG. 5. Viral titers in brains three days after i.c. rechallenge. Groups CD8 T cells that are generated following LCMV infection Ϫ Ϫ of three naive GzmB mice or of three GzmB and three B6 mice that in GzmBϪ mice compared with B6 mice, we reasoned had been infected ip 3 months previously with LCMV were infected i.c. 3 Ϫ that the impaired clearance of LCMV that we observed in with 2 ϫ 10 PFU LCMV. Another group of previously infected GzmB Ϫ mice (n ϭ 3) was depleted of CD8 cells prior to i.c. rechallenge with GzmB mice must be due to a difference in the function LCMV. Three days after i.c. infection, viral titers in brains were deter- of these T cells. We therefore examined the cytolytic mined by plaque assay. The horizontal bars indicate the mean value for activity of spleen cells from LCMV-infected GzmBϪ mice each group. *P Ͻ 0.01 compared with B6. in vitro. Our assays demonstrated that GzmBϪ spleen cells were competent to lyse LCMV-infected target cells in vitro, but that, as expected, these cells were incapable creased accumulation of virus-specific CD8 cells in this of inducing rapid DNA fragmentation in target cells. After tissue. Using ELISPOT, we enumerated virus-specific Ϫ Ϫ a longer incubation, GzmB cells did induce DNA frag- CD8 cells in brains of B6 and GzmB mice 3 days mentation in LCMV-infected target cells. This later effect following i.c. rechallenge with LCMV. As shown in Fig. 6, is most likely due to GzmA (Simon et al., 1997; Shresta et there were equivalent numbers of CD8 cells infiltrating Ϫ al., 1999). Our finding that clearance of LCMV is delayed the brains of GzmB and B6 mice that were specific for Ϫ in GzmB mice suggests that this GzmB-dependent early the immunodominant GP33-41 and NP396-404 class I pathway of apoptosis induction is relevant in vivo during restricted epitopes. This indicates that the elevated titers Ϫ CTL-mediated elimination of this virus. of LCMV detected in the brains of GzmB mice at this Several studies have addressed the role of GzmA and time point is not due to impaired infiltration of the CNS by GzmB in immune-mediated control of virus infections. virus-specific CD8 cells. Mu¨llbacher and colleagues (Mu¨llbacher et al., 1999) examined the response of GzmϪ mice to infection with the DISCUSSION Ϫ Ϫ poxvirus ectromelia. They found that GzmA and GzmB It has been argued that the major mechanism whereby mice were approximately 10 times and 100 times more the immune system rids the body of noncytopathic viral sensitive to infection with this virus, respectively, than infections is dependent on CTL (Ka¨gi et al., 1995a,b). The were wild-type mice. This same group also demon- requirement for perforin-dependent CD8 T cells for elimination of the classic noncytopathic virus, LCMV, has been clearly demonstrated (Ka¨gi et al., 1994a,b; Leist et al., 1987; Moskophidis et al., 1987; Walsh et al., 1994; Zajac et al., 1995). It is unclear, however, what role the other components of the secretory pathway of CTL killing play in clearance of this virus. To begin to address this question, we have studied the course of LCMV infection in mice that are genetically deficient in GzmB expression. Following ip infection, LCMV replicates to similar levels in GzmBϪ mice and control mice, and peak viral titers in spleens and livers of these animals are attained by day 4 following infection. However, peak viral titers in FIG. 6. Numbers of virus-specific CD8 cells in brains 3 days after i.c. these tissues are maintained for at least 2 days longer in Ϫ Ϫ rechallenge. Groups of four GzmB (open bars) and four B6 mice (solid GzmB animals than are the peak titers in tissues of bars) that had been infected ip 3 months previously with LCMV were control mice. Consequently, the virus is cleared some- infected i.c. with 2 ϫ 10 3 PFU LCMV. Three days after i.c. infection, mice what later from spleens of GzmBϪ mice than it is from were sacrificed and mononuclear cells were isolated from the brains. wild-type mice. These data indicate a role for GzmB in The numbers of GP33-41- and NP396-404-specific cells in the brains were determined by ELISPOT. The numbers of epitope-specific cells in efficient clearance of LCMV in vivo. brains of noninfected mice were below the level of detection of the We next determined whether the delayed clearance of assay. Bars represent the means Ϯ SEM of the numbers of cells/g of Ϫ LCMV by GzmB mice could be explained by impaired brain. 6 ZAJAC, DYE, AND QUINN strated that GzmAϪ and GzmBϪ mice were more suscep- GzmBϪ mice is not due to impaired trafficking of CTL into tible to murine cytomegalovirus (MCMV) infection (Riera the infected CNS. We found no difference in the numbers et al., 2000). In each of these reports the enhanced of LCMV-specific CD8 cells infiltrating the brains of in- susceptibility of GzmBϪ mice, compared with controls, to fected GzmBϪ mice and the numbers of these cells infection with the respective virus was substantially infiltrating the CNS of infected B6 mice. Since trafficking greater than what we observe with LCMV infection. This of GzmBϪ CD8 cells into the infected CNS is not im- might be due to the role played by natural killer (NK) cells paired, our data suggest that the delayed clearance of in each of these infections. Similar to CTL, NK cells LCMV from the brains of i.c.-infected GzmBϪ mice is due secrete perforin and granzymes and exert potent antiviral to an inability of GzmBϪ CTL to eliminate LCMV-infected activity early following infection with poxviruses (Delano cells in vivo. and Brownstein, 1995) and with MCMV (Bukowski and Several studies have shown that the predominant Welsh, 1985; Orange and Biron, 1996). The impairment of mechanism of CTL-mediated control of tumor cells is NK activity during infection with ectromelia and MCMV perforin-dependent (Davis et al., 2001; Smyth et al., 2000; might allow these viruses to replicate to high levels early van den Broek et al., 1996). However, the use of gene- following infection. By contrast, acute LCMV infection is targeted mice has revealed unsuspected complexities in controlled by CTL and is insensitive to the antiviral ac- the interaction between CTL and virus-infected cells in tions of NK cells (Orange and Biron, 1996). In light of this vivo. Whereas viruses such as ectromelia and MCMV it is not surprising that we only observe elevated titers of require both perforin and granzymes for effective control LCMV in GzmBϪ mice beginning at day 6 after infection, (Mu¨llbacher et al., 1996, 1999; Riera et al., 2000), control the time at which the virus is being brought under control of vesicular stomatitis virus can be effected by GzmB by CTL in normal mice. Another study also has ad- independent of perforin (Hommel-Berrey et al., 1997). We dressed the role of granzymes and perforin in control of have shown that clearance of LCMV, although exquisitely LCMV infection (Balkow et al., 2001). These authors re- dependent on perforin expression, is also facilitated by ported no difference in the kinetics of clearance of LCMV GzmB. It is likely that the disparate requirements for from livers of GzmBϪ mice and control mice. However, CTL-mediated control of different viruses is likely to be the earliest time point at which these authors measured determined by factors such as the mode of replication of viral titers was day 8 after infection. At this time point we the virus as well as the cell type that is infected in vivo. also find no difference between the viral titers in livers of Ϫ GzmB mice and control mice (Fig. 1B). Balkow et al. did MATERIALS AND METHODS not examine the kinetics of clearance of LCMV from spleens of GzmBϪ mice. Viruses and cell lines The studies alluded to above using ectromelia virus The Armstrong-3 strain of LCMV was obtained from (Mu¨llbacher et al., 1999) and MCMV (Riera et al., 2000), J. L. Whitton (The Scripps Research Institute, La Jolla, CA) as well as studies using herpes simplex virus (Pereira et and was propagated in BHK-21 cells (American Type al., 2000), also revealed a role for GzmA in control of Culture Collection, Rockville, MD), and infectious super- these pathogens. By contrast, the lack of GzmA does not natants were titrated on Vero cell monolayers as previ- appear to influence the outcome of infection with LCMV ously described (Ahmed et al., 1984). The fibroblastoid (Ebnet et al., 1995); however, without a thorough analysis b k MC57 (H-2 ) and L cell (H-2 ) lines have been described of the kinetics of virus elimination it is possible that a previously (Muller et al., 1992). MC57 cells were cultured slight delay in virus clearance could have gone undetec- in HEPES-buffered RPMI-1640 supplemented with 10% ted in that study. That GzmA cannot totally compensate FCS, and 50 ␮M 2-mercaptoethanol (complete medium) for the absence of GzmB would not be surprising, how- (Life Technologies Inc., Gaithersburg, MD) and L cells ever, as these enzymes have different substrate speci- were propagated in Dulbecco’s modified Eagle’s medium ficities and induce apoptosis via disparate pathways. (Life Technologies Inc.), which was also supplemented Although clearance of LCMV is somewhat delayed in Ϫ with 10% FCS and 50 ␮M 2-mercaptoethanol. GzmB mice, GzmB is not absolutely required for elimination of this virus. We reasoned, however, that the Ϫ Treatment of mice delayed kinetics of virus clearance in GzmB mice might compromise their ability to generate a protective immune C57BL/6J (B6), B6;129S-Gzmbtm1Ley (GzmBϪ) (Heusel et response against secondary infection with LCMV. Three al., 1994), C57BL/6-Pfptm1Sdz (PfpϪ)(Ka¨gi et al., 1994a), and days following i.c. rechallenge with LCMV, the levels of B6129SF2/J mice were obtained from Jackson Laborato- virus in the brains of GzmBϪ mice were elevated com- ries (Bar Harbor, ME) and were maintained in specific pared with viral titers in brains of B6 mice. As with pathogen-free, AAALAC-accredited facilities at Loyola peripheral infection, the virus was ultimately cleared University Medical Center, Maywood, IL. Female mice from the brains of GzmBϪ mice and the animals survived. were used at 7–20 weeks of age. Mice were infected via The delayed clearance of LCMV from the brains of the ip route with 2 ϫ 10 5 plaque forming units (PFU) of LCMV INFECTION OF GzmB-DEFICIENT MICE 7

LCMV in a volume of 0.4 ml. Spleens and livers were 3 h of the infection period the cells were labeled using 3 removed from mice at the indicated time points following ␮Ci/ml methyl-3H-thymidine. After washing with com- ip infection with LCMV for determination of viral titers. plete medium, the labeled target cells were incubated

The tissues were homogenized and viral titers in the with effector cells for 4 or8hat37°C in 6% CO2. supernatants were measured by plaque assay on Vero Following the incubation, the cells were harvested using cell monolayers as previously described (Ahmed et al., a Skatron Cell Harvester (Skatron Instruments) and the 1984). amount of DNA-associated radiolabel remaining was In experiments to assess the efficacy of secondary determined using a Packard ␤-Counter (Packard Instru- antiviral CTL responses, mice that had been infected ip ments). The percentage of 3H-DNA released was calcu- with 2 ϫ 10 5 PFU LCMV 3 months earlier were anesthe- lated using the formula: % 3H Ϫ DNA-release ϭ {[(c.p.m. tized using Metofane (Pittman-Moore, Mundelein, IL) and remaining without effectors) Ϫ (c.p.m. remaining with were rechallenged by i.c. injection of either 2 ϫ 10 3 or effectors)]/(c.p.m. remaining without effectors)} ϫ 100. 2 ϫ 10 5 PFU LCMV in a volume of 20 ␮l. Following i.c. infection, mice were monitored daily for morbidity and Isolation of mononuclear cells from brains mortality. At the indicated time points following i.c. infec- Mononuclear cells were isolated from brains essen- tion, mice were sacrificed and perfused with ice-cold tially as described previously (Hawke et al., 1998). Brains phosphate-buffered saline (PBS), and brains were har- were harvested from mice as outlined above and were vested. A portion of each brain was taken for determina- disrupted in complete medium by passage through a tion of viral titers as outlined above. Mononuclear cells sterile mesh (100-␮m pore size, Becton–Dickinson, San were isolated from the remaining portion of each brain Jose, CA) and by successive passages through 19- and as described below. In certain experiments mice were 21-gauge needles (Becton–Dickinson). Cell suspensions, depleted of CD8 cells prior to infection by ip injection of each derived from a single animal, were centrifuged at anti-CD8 antibody (clone 2.43) as previously described Ͼ 700 g (at 4°C) for 7 min and the pellets were resus- (Su et al., 1998). This treatment removed 95% of the pended in 3 ml of a Percoll solution composed of a 10:1 CD8 cells as assessed by flow cytometry. mixture of Percoll (Sigma Chemical Co., St. Louis, MO) and 10ϫ RPMI. Dilutions of 60, 40, and 0% of this Percoll Cytotoxicity assays solution were overlaid on top of the cell suspensions. Spleens were removed from mice 8 days following ip Following centrifugation for 15 min at 1000 g (at 4°C), infection with LCMV. The spleens were disrupted in com- mononuclear cells were harvested from the 40–60% in- plete medium and erythrocytes were removed by os- terface. Cells were washed three times with complete motic lysis. These effector cells were assayed for LCMV- medium prior to use in ELISPOT as described below. specific CTL activity using standard 51Cr-release assays as previously described (Muller et al., 1992). The target ELISPOT to enumerate cytokine secreting cells cell lines used in these assays were LCMV-infected and The numbers of cells reactive against the immuno- mock-infected MC57 and L cell lines. Infection of MC57 dominant major histocompatibility complex (MHC) class cells and L cells with LCMV was performed exactly as I and MHC class II restricted epitopes of LCMV in described previously (Muller et al., 1992). Effector cells spleens or brains of infected mice were determined were incubated with the target cells for 4 or8hat37°C using an interferon-␥ (IFN-␥) ELISPOT exactly as de- in 6% CO2. Supernatants were harvested using a Skatron scribed previously (Kamperschroer and Quinn, 1999). harvesting system (Skatron Instruments, Sterling, VA) Cells were prepared from infected mice as outlined and the amount of 51Cr released was quantified using a above and were serially diluted in complete medium to Packard ␥-counter (Packard Instruments, Sterling, VA). give a range of cell numbers per well. Each dilution of All assays were performed in triplicate and the percent- effector cells was assayed in duplicate or triplicate. age specific lysis was determined using the formula: % Wells also contained 50 units/ml of recombinant human specific lysis ϭ {[(c.p.m. released in the presence of IL-2 (Sigma Chemical Co.) and 1 ϫ 10 6 syngeneic stim- effectors) Ϫ (spontaneous release)]/[(total release) Ϫ ulator spleen cells together with the appropriate pep- (spontaneous release)]} ϫ 100, where total release is the tides. Peptides corresponding to amino acids 33–41, 61– c.p.m. released from target cells by addition of 5% (v/v) 80, and 276–286 of the LCMV glycoprotein (GP), and to Triton X-100 (final concentration), and spontaneous re- 309–328 and 396–404 of the LCMV nucleoprotein (NP), lease is the c.p.m. released from target cells in the were synthesized by the Loyola University Macromolec- absence of effectors. ular Synthesis Facility using a Synergy 432A automated To measure CTL-mediated target cell DNA degrada- synthesizer. The LCMV class I restricted peptides, GP33- tion, we used the JAM assay (Matzinger, 1991). MC57 41, GP276-286, and NP396-404 (Oldstone et al., 1988; cells and L cells were infected with LCMV as described Whitton et al., 1988a,b), were added at 1 ␮g/ml and the previously (Muller et al., 1992) except that during the final LCMV class II restricted peptides, GP61-80 and NP309- 8 ZAJAC, DYE, AND QUINN

328 (Oxenius et al., 1998), were added at 10 ␮g/ml. T. W., and Zinkernagel, R. M. (1994). Enhanced establishment of a ϩ Previous work in this laboratory has shown these to be virus carrier state in adult CD4 T-cell-deficient mice. J. Virol. 68, the optimal concentrations of the respective peptides for 4700–4704. Borrow, P., and Oldstone, M. B. A. (1997). Lymphocytic choriomeningitis use in ELISPOT. Control wells were treated as described virus. In “Viral Pathogenesis” (N. Nathanson, et al., Eds.), pp. 593–627. above except that no peptide was added. The ELISPOT Lippincott-Raven, Philadelphia. assays were developed and spots were counted as Buchmeier, M. J., and Zajac, A. J. (1998). Lymphocytic choriomeningitis described previously (Kamperschroer and Quinn, 1999). virus. In “Persistent Viral Infections” (R. Ahmed and I. Chen, Eds.), pp. Results were expressed as the total numbers of peptide- 575–606. John Wiley & Sons Ltd., New York. ␥ Bukowski, J. F., and Welsh, R. M. (1985). Adoptive transfer studies specific IFN- -secreting cells in the spleens of infected demonstrating the antiviral effect of NK cells in vivo. J. Exp. Med. 161, animals or as the numbers of peptide specific IFN-␥- 40–52. secreting cells per gram of brain. Butz, E. A., and Bevan, M. J. (1998). Massive expansion of antigen- specific CD8ϩ T cells during an acute virus infection. Immunity 8, Flow cytometry 167–175. Darmon, A. J., Nicholson, D. W., and Bleackley, R. C. (1995). Activation of Aliquots of cells in ice-cold PBS supplemented with 1% the apoptotic protease CPP32 by cytotoxic T-cell-derived granzyme (v/v) FCS and 15 mM sodium azide (FACS buffer) were B. Nature 377, 446–448. Davis, J. E., Smyth, M. J., and Trapani, J. A. (2001). Granzyme A and incubated with PE-conjugated anti CD4 (clone RM4-5; B-deficient killer lymphocytes are defective in eliciting DNA fragmen- BD-Pharmingen, San Diego, CA) and FITC-anti-CD8 tation but retain potent in vivo anti-tumor capacity. Eur. J. Immunol. 31, (clone 53-6.7; BD-Pharmingen) at a concentration of 1 39–47. ␮g/106 cells, for 30 min at 4°C. Samples were washed Delano, M. L., and Brownstein, D. G. (1995). Innate resistance to lethal with FACS buffer and analyzed using a FACSCalibur flow mousepox is genetically linked to the NK gene complex on chromo- some 6 and correlates with early restriction of virus replication by cytometer (Becton–Dickinson). Dead cells were ex- cells with an NK phenotype. J. Virol. 69, 5875–5877. cluded on the basis of forward and side scatter, and data Doherty, P. C., and Ahmed, R. (1997). Immune responses to viral infec- analysis was performed on 10,000 acquired events using tion. In “Viral Pathogenesis” (N. Nathanson, et al., Eds.), pp. 143–161. the computer program CellQuest (Becton–Dickinson). Lippincott-Raven, Philadelphia. Duan, H., Orth, K., Chinnaiyan, A. M., Poirier, G. G., Froelich, C. J., He, Statistical analyses W. W., and Dixit, V. M. (1996). ICE-LAP6, a novel member of the ICE/Ced-3 gene family, is activated by the cytotoxic T cell protease The Student’s t test was used to compare control with granzyme B. J. Biol. Chem. 271, 16720–16724. experimental groups. Values of P Ͼ 0.05 were consid- Ebnet, K., Hausmann, M., Lehmann-Grube, F., Mu¨llbacher, A., Kopf, M., Lamers, M., and Simon, M. M. (1995). Granzyme A-deficient mice ered not significant. retain potent cell-mediated cytotoxicity. EMBO J. 14, 4230–4239. Graubert, T. A., DiPersio, J. F., Russell, J. H., and Ley, T. J. (1997). ACKNOWLEDGMENTS Perforin/granzyme-dependent and independent mechanisms are both important for the development of graft-versus-host disease after This work was supported in part by a grant from the Crohn’s and murine bone marrow transplantation. J. Clin. Invest. 100, 904–911. Colitis Foundation of America, by a Schweppe Foundation Fellowship, Gu, Y., Sarnecki, C., Fleming, M. A., Lippke, J. A., Bleackley, R. C., and and by NIH Grant AI44861-03 (D.G.Q.) and by NIH training Grant T32 Su, M. S.-S. (1996). Processing and activation of CMH-1 by granzyme AI07508-01 (J.M.D.). The authors acknowledge the assistance of P. B. J. Biol. Chem. 271, 10816–10820. Simms for assistance with flow cytometry. Hawke, S., Stevenson, P. G., Freeman, S., and Bangham, C. R. (1998). Long-term persistence of activated cytotoxic T lymphocytes after REFERENCES viral infection of the central nervous system. J. Exp. Med. 187, 1575– 1582. Ahmed, R., Salmi, A., Butler, L. D., Chiller, J. M., and Oldstone, M. B. A. Heusel, J. W., Wesselschmidt, R. L., Shresta, S., Russell, J. H., and Ley, (1984). Selection of genetic variants of lymphocytic choriomeningitis T. J. (1994). Cytotoxic lymphocytes require granzyme B for the rapid virus in spleens of persistently infected mice. Role in suppression of induction of DNA fragmentation and apoptosis in allogeneic target cytotoxic T lymphocyte response and viral persistence. J. Exp. Med. cells. Cell 76, 977–987. 60, 521–540. Hommel-Berrey, G. A., Bochan, M. R., Montel, A., Goebel, W. S., Froelich, Andrade, F., Roy, S., Thornberry, N., Rosen, A., and Casciola-Rosen, L. C. J., and Brahmi, Z. (1997). Granzyme B independently of perforin (1998). Granzyme B directly and efficiently cleaves several down- mediates cytolytic intracellular inactivation of vesicular stomatitis stream caspase substrates: Implications for CTL-induced apoptosis. virus. Cell. Immunol. 180, 1–9. Immunity 8, 451–460. Ka¨gi, D., Ledermann, B., Burki, K., Seiler, P., Odermatt, B., Olsen, K. J., Baenziger, J., Hengartner, H., Zinkernagel, R. M., and Cole, G. A. (1986). Podack, E. R., Zinkernagel, R. M., and Hengartner, H. (1994a). Cyto- Induction or prevention of immunopathological disease by cloned toxicity mediated by T cells and natural killer cells is greatly impaired cytotoxic T cell lines specific for lymphocytic choriomeningitis virus. in perforin-deficient mice. Nature (Lond.) 369, 31–37. Eur. J. Immunol. 16, 387–393. Ka¨gi, D., Ledermann, B., Burki, K., Zinkernagel, R. M., and Hengartner, Balkow, S., Kersten, A., Tran, T. T., Stehle, T., Grosse, P., Museteanu, C., H. (1995a). Lymphocyte-mediated cytotoxicity in vitro and in vivo: Utermo¨hlen, O., Pircher, H., Von Weizsa¨cker, F., Wallich, R., Mu¨ll- Mechanisms and significance. Immunol. Rev. 146, 95–115. bacher, A., and Simon, M. M. (2001). Concerted action of the FasL/ Ka¨gi, D., Seiler, P., Palvovic, J., Ledermann, B., Burki, K., Zinkernagel, Fas and perforin/granzyme A and B pathways is mandatory for the R. M., and Hengartner, H. (1995b). The role of perforin- and Fas- development of early viral hepatitis but not for recovery from viral dependent cytotoxicity in protection against cytopathic and noncyto- infection. J. Virol. 75, 8781–8791. pathic viruses. Eur. J. Immunol. 25, 3256–3262. Battegay, M., Moskophidis, D., Rahemtulla, A., Hengartner, H., Mak, Ka¨gi, D., Vignaux, F., Ledermann, B., Burki, K., Depraetere, V., Nagata, S., LCMV INFECTION OF GzmB-DEFICIENT MICE 9

Hengartner, H., and Golstein, P. (1994b). Fas and perforin pathways noncytolytic component of CD8ϩ cell granules, restricts the spread of as major mechanisms of T cell-mediated cytotoxicity. Science (Wash. herpes simplex virus in the peripheral nervous systems of experi- DC) 265, 528–530. mentally infected mice. J. Virol. 74, 1029–1032. Kamperschroer, C., and Quinn, D. G. (1999). Quantification of epitope- Quan, L. T., Tewari, M., O’Rourke, K., Dixit, V., Snipas, S. J., Poirier, G. G., specific MHC class-II-restricted T cells following lymphocytic cho- Ray, C., Pickup, D. J., and Salvesen, G. S. (1996). Proteolytic activation riomeningitis virus infection. Cell. Immunol. 193, 134–146. of the cell death protease Yama/CPP32 by granzyme B. Proc. Natl. Kojima, H., Shimohara, N., Hanaoka, S., Someya-Shirota, Y., Takagaki, Acad. Sci. USA 93, 1972–1976. Y., Ohnoa, H., Saito, T., Katayama, T., Yagita, H., and Okumura, K. Riera, L., Gariglio, M., Valente, G., Mu¨llbacher, A., Musteanu, C., Land- (1994). Two distinct pathways of specific killing revealed by perforin olfo, S., and Simon, M. M. (2000). Murine cytomegalovirus replication mutant cytotoxic T lymphocytes. Immunity 1, 357–364. in salivary glands is controlled by both perforin and granzymes Leist, T. P., Cobbold, S. P., Waldmann, H., Aguet, M., and Zinkernagel, during acute infection. Eur. J. Immunol. 30, 1350–1355. R. M. (1987). Functional analysis of T lymphocyte subsets in antiviral Shresta, S., Graubert, T. A., Thomas, D. A., Raptis, S. Z., and Ley, T. J. host defense. J. Immunol. 138, 2278–2281. (1999). Granzyme A initiates an alternative pathway for granule- Lowin, B., Hahne, M., Mattmann, C., and Tschopp, J. (1994). Cytolytic mediated apoptosis. Immunity 10, 595–605. T-cell cytotoxicity is mediated through perforin and Fas lytic path- Simon, M. M., Hausmann, M., Tran, T., Ebnet, K., Tschopp, J., Tha Hla, ways. Nature (Lond.) 370, 650–652. R., and Mu¨llbacher, A. (1997). In vitro and ex-vivo-derived cytolytic Matloubian, M., Conception, R. J., and Ahmed, R. (1994). CD4ϩ T cells leucocytes from granzyme A ϫ B double knockout mice are defective are required to sustain CD8ϩ T cytotoxic T cell responses during in granule-mediated apoptosis but not lysis of target cells. J. Exp. chronic viral infection. J. Virol. 68, 8056–8063. Med. 186, 1781–1786. Matzinger, P. (1991). The JAM test: A simple assay for DNA fragmenta- Smyth, M. J., Thia, K. Y. T., Street, S. E. A., MacGregor, D., Godfrey, D. I., tion and cell death. J. Immunol. Methods 145, 185–192. and Trapani, J. A. (2000). Perforin-mediated cytotoxicity is critical for Moskophidis, D., Cobbold, S. P., Waldmann, H., and Lehmann-Grube, F. surveillance of spontaneous lymphoma. J. Exp. Med. 192, 755–760. (1987). Mechanism of recovery from acute virus infection: Treatment Su, H. C., Cousens, L. P., Fast, D., Slifka, M. K., Bungiro, R. D., Ahmed, of lymphocytic choriomeningitis virus infected mice with monoclonal R., and Biron, C. A. (1998). CD4ϩ and CD8ϩ T cell interactions in antibodies reveals that Lyt2ϩ T lymphocytes mediate clearance of IFN-␥ and IL-4 responses to viral infections: Requirements for IL-2. virus and regulate the antiviral antibody response. J. Virol. 61, 1867– J. Immunol. 160, 5007–5017. 1874. Talanian, R. V., Yang, X.-H., Turbov, J., Seth, P., Ghayur, T., Casiano, C. A., Motyka, B., Korbutt, G., Pinkoski, M. J., Heibein, J. A., Caputo, A., Orth, K., and Froelich, C. J. (1997). Granule-mediated killing: Pathways Hobman, M., Barry, M., Shostak, I., Sawchuk, T., Holmes, C. F. B., for granzyme B-initiated apoptosis. J. Exp. Med. 186, 1323–1331. Gauldie, J., and Bleackley, R. C. (2000). Mannose-6-phosphate/insu- van den Broek, M. F., Ka¨gi, D., Ossendorp, F., Toes, R., Vamvakas, S., lin-like growth factor II receptor is a death receptor for granzyme B Lutz, W. K., Melief, C. J., Zinkernagel, R. M., and Hengartner, H. (1996). during cytotoxic T cell-induced apoptosis. Cell 103, 491–500. Decreased tumor surveillance in perforin-deficient mice. J. Exp. Med. Mu¨llbacher, A., Ebnet, K., Blanden, R. V., Tha Hla, R., Stehle, T., Mus- 184, 1781–1790. teanu, C., and Simon, M. M. (1996). Granzyme A is critical for recov- von Herrath, M. G., Yokoyama, M., Dockter, J., Oldstone, M. B. A., and ery of mice from infection with the natural cytopathic viral pathogen, Whitton, J. L. (1996). CD4-deficient mice have reduced levels of ectromelia. Proc. Natl. Acad. Sci. USA 93, 5783–5787. memory cytotoxic T lymphocytes after immunization and show di- Mu¨llbacher, A., Waring, P., Tha Hla, R., Tran, T., Chin, S., Stehle, T., minished resistance to subsequent virus challenge. J. Virol. 70, Musteanu, C., and Simon, M. M. (1999). Granzymes are the essential 1072–1079. downstream effector molecules for the control of primary virus in- Walsh, C. M., Matloubian, M., Liu, C.-C., Ueda, R., Kurahara, C. G., fections by cytolytic leukocytes. Proc. Natl. Acad. Sci. USA 96, 13950– Christensen, J. L., Huang, M. T. F., Young, J. D.-E., Ahmed, R., and 13955. Clark, W. R. (1994). Immune function in mice lacking the perforin Muller, D., Koller, B. H., Whitton, J. L., LaPan, K. E., Brigman, K. K., and gene. Proc. Natl. Acad. Sci. USA 91, 10854–10858. Frelinger, J. A. (1992). LCMV-specific, class II-restricted cytotoxic T Whitton, J. L., Gebhard, J. R., Lewicki, H., Tishon, A., and Oldstone,

cells in ␤2-microglobulin-deficient mice. Science (Wash. DC) 255, M. B. A. (1988a). Molecular definition of a major cytotoxic T-lympho- 1576–1578. cyte epitope in the glycoprotein of lymphocytic choriomeningitis Murali-Krishna, K., Altman, J. D., Suresh, M., Sourdive, D. J., Zajac, A. J., virus. J. Virol. 62, 687–695. Miller, J. D., Slansky, J., and Ahmed, R. (1998). Counting antigen- Whitton, J. L., Southern, P. J., and Oldstone, M. B. A. (1988b). Analyses specific CD8 T cells: A reevaluation of bystander activation during of the cytotoxic T lymphocyte responses to glycoprotein and nucleo- viral infection. Immunity 8, 177–187. protein components of lymphocytic choriomeningitis virus. Virology Oldstone, M. B. A., Whitton, J. L., Lewicki, H., and Tishon, A. (1988). Fine 162, 321–327. dissection of a nine amino acid glycoprotein epitope, a major deter- Zajac, A. J., Blattman, J. N., Murali-Krishna, K., Sourdive, D. J., Suresh, minant recognized by lymphocytic choriomeningitis virus-specific M., Altman, J. D., and Ahmed, R. (1998). Viral immune evasion due to class I-restricted H-2Db cytotoxic T lymphocytes. J. Exp. Med. 168, persistence of activated T cells without effector function. J. Exp. Med. 559–570. 188, 2205–2213. Orange, J. S., and Biron, C. A. (1996). An absolute and restricted Zajac, A. J., Muller, D., Pederson, K., Frelinger, J. A., and Quinn, D. G. requirement for IL-12 in natural killer cell IFN-␥ production and (1995). Natural killer cell activity in lymphocytic choriomeningitis

antiviral defense. Studies of natural killer and T cell responses in virus-infected ␤2-microglobulin-deficient mice. Int. Immunol. 7, 1545– contrasting viral infections. J. Immunol. 156, 1138–1142. 1556. Oxenius, A., Bachmann, M. F., Zinkernagel, R. M., and Hengartner, H. Zajac, A. J., Quinn, D. G., Cohen, P. L., and Frelinger, J. A. (1996). (1998). Virus-specific MHC-class II-restricted TCR-transgenic mice: Fas-dependent CD4ϩ cytotoxic T-cell-mediated pathogenesis during Effects on humoral and cellular immune responses after viral infec- virus infection. Proc. Natl. Acad. Sci. USA 93, 14730–14735. tion. Eur. J. Immunol. 28, 390–400. Zinkernagel, R. M. (1996). Immunology taught by viruses. Science Pereira, R. A., Simon, M. M., and Simmons, A. (2000). Granzyme A, a (Wash. DC) 271, 173–178.