Proliferation of T cells is important for the expansion of specific T cell clones during immune responses. In addition, for the establishment of protective immunity against viruses, bacteria, and tumors, the expanded T cells must differentiate into effector T cells. Here we show that effector T cell generation is driven by activation of APCs and duration of antigenic stimulation. Adoptively transferred TCR-transgenic T cells extensively proliferated upon immunization. However, these T cells failed to differentiate into effector cells and died within 1 wk after immunization unless antigenic peptides persisted for >1 day or were presented by activated APCs. The induction of protective immunity in a nontransgenic system was more stringent, since activation of APCs or prolonged Ag persistence alone was not sufficient to drive immunity. In contrast, Ag had to be presented for several days by activated APCs to trigger protective T cell responses. Thus, activation of APCs and duration of Ag presentation together regulate the induction of protective T cell responses.
Induction of T cell responses in vivo is a complex process that depends on the interaction of a multitude of molecules expressed on T cells and APCs (1, 2, 3). For optimal T cell responses, Ag-specific stimulation of the TCRs together with costimulatory molecules is usually required. Under conditions of suboptimal stimulation, T cell activation does not occur as an all-or-none event. Rather, certain T cell responses may be induced in the absence of others (2, 4, 5, 6). For example, T cells may express activation markers without proliferating or may proliferate without producing relevant cytokines, a phenomenon summarized as partial T cell activation (7, 8, 9, 10). Altered peptide ligands (APLs),2 in particular, partial T cell agonists, have been useful in dissecting partial T cell activation (4, 5). Nevertheless, there is presently little evidence for a functional correlate of partial T cell activation during the course of regular immune responses in vivo, in particular if responses against pathogens are studied. Chronically stimulated lymphocytic choriomeningitis virus (LCMV)-specific T cells may be an interesting exception. Under conditions where the virus cannot be fully controlled, CD8+ T cells may proliferate but fail to secrete cytokines and gain effector function (11). Moreover, it has recently been suggested that two populations of memory T cells exist, a resting population, also called central memory T cells, and an activated population, also referred to as effector memory T cells (12). Importantly, only the effector memory T cells seem to be responsible for antiviral, antibacterial, or antitumor protection (13, 14, 15), while central memory T cells appear to proliferate without necessarily gaining effector function. Thus, for vaccination purposes it is critical to understand the balance between the proliferation of specific T cells (possibly linked to central T cell memory) and the efficient induction of effector T cells (possibly linked to effector T cell memory).
To address this question, we established an adoptive transfer system using CFSE-labeled naive TCR-transgenic T cells specific for peptide p33 derived from LCMV (16). Three different Ags were used for immunization of recipient mice. Live LCMV, which is known to induce strong effector CTL responses, free peptide p33, which does not usually induce a measurable T cell response in normal mice, and virus-like particles (VLPs) consisting of the hepatitis B core Ag fused to peptide p33 (p33-VLPs) (17, 18). Although not infectious, these p33-VLPs efficiently reach the MHC class I pathway in both a TAP-dependent and a TAP-independent fashion (17). Moreover, p33-VLPs induce weak CTL responses in normal mice (18). Using these three types of immunogens, we set out to identify parameters that regulate T cell proliferation vs effector T cell generation. Immunization with peptide alone induced the proliferation of transferred T cells, but failed to trigger the induction of effector T cells. Activation of APCs during vaccination or Ag persistence for several days was sufficient to drive effector cell induction and accumulation of specific T cells. However, to establish protective immunity in a nontransgenic system, activation of APCs together with prolonged Ag presentation were required.
Materials and Methods
C57BL/6 mice were purchased from Harlan Netherlands (Horst, The Netherlands) at the age of 8–12 wk. Transgenic mice expressing a TCR specific for peptide p33 in association with H-2Db have been described (19). To allow for discrimination between endogenous and transgenic CD8+ T cells in the adoptive transfer experiments, the transgenic animals were additionally crossed with Ly5.1+ mice. C57BL/10ScNCr, Toll-like receptor 4-deficient mice (20) were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were bred and kept in a specific pathogen-free facility at Cytos Biotechnology.
Viruses, p33-VLPs, peptides, and oligonucleotides
LCMV isolate WE was originally obtained from Dr. R. M. Zinkernagel (Institute of Experimental Immunology, University Hospital, Zurich, Switzerland) and propagated on L929 cells. Virus titers were determined using a focus-forming assay on MC57 fibroblasts (21). Vacc-GP, a recombinant vaccinia virus expressing the LCMV glycoprotein, was described previously (22). Vacc-GP was grown and plaqued on BSC40 cells (23). The production and purification of recombinant p33-VLPs were previously described in detail (18). LCMV glycoprotein peptide p33–41 (KAVYNFATM) was synthesized by a solid phase method and purchased from Neosystem Laboratoire (Strasbourg, France). Phosphorothioate-modified CpG-ODN was synthesized by Microsynth (Balgach, Switzerland). The following oligonucleotide sequence was used: 1668pt, 5′-TCC ATG ACG TTC CTG AAT AAT-3′ (18).
CFSE was purchased from Molecular Probes (Eugene, OR). Erythrocytes were removed from spleen and lymph node (LN) cell suspensions by Lympholyte-M Gradient (Cedarlane, Ontario, Canada). Transgenic CD8+ T cells were obtained by positive MACS MicroBeads isolation according to the manufacturer’s instructions (Miltenyi Biotec, Bergisch Gladbach, Germany) with a purity of at least 90%. Cells were labeled by diluting the 0.5 mM CFSE stock 1000-fold into the cell suspension (final concentration, 0.5 μM) and incubating them for 10 min at 37°C. After labeling, FCS was added to a final concentration of 10%, and cells were subsequently washed with PBS at 4°C.
Adoptive transfer and in vivo activation of transgenic T cells
Ly5.1+ TCR-transgenic T cells were labeled with CFSE as described above. Labeled T cells (5 × 106) were resuspended in 250 μl of PBS and injected into the tail vein of sex-matched C57BL/6 recipients. After 16 h, recipients were s.c. vaccinated in the neck region with p33 peptide (with or without CpG oligonucleotides) or p33-VLPs at the indicated doses. Alternatively, mice were i.v. infected with LCMV strain WE. For the Ag presentation kinetic experiments, mice were first immunized with peptide or virus-like particles and subsequently i.v. transferred with the transgenic T cells at the indicated time points.
To analyze cell proliferation and activation marker expression of adoptively transferred transgenic T cells, single-cell suspensions were prepared from draining cervical LN of treated C57BL/6 mice. Cells were then incubated with PE-labeled anti-CD25, anti-CD44, or anti-CD69 Abs in combination with CyChrome-labeled anti-CD8 abs (all from BD Biosciences, Mountain View, CA). Discrimination of transgenic T cells was achieved by incubating the cells with biotin-coupled anti-Ly5.1 Abs, followed by streptavidin-allophycocyanin staining (BD Biosciences).
To analyze the intracellular IFN-γ expression of the transferred p33-specific T cells, LN cells were resuspended in IMDM complemented with 10% FCS and restimulated in vitro with 2 × 10−6 M p33 peptide for 6 h. The cultures were supplemented with 10 μg/ml brefeldin A (Sigma-Aldrich, Buchs, Switzerland) during the last 4 h of incubation. Restimulated cells were then stained with CyChrome-labeled anti-CD8 and PE-coupled anti-Ly5.1 Abs for 30 min on ice. Cells were fixed in PBS/4% formaldehyde for 15 min and permeabilized in PBS/0.5% saponin (Sigma-Aldrich) for 30 min at room temperature. During permeabilization cells were incubated with allophycocyanin-labeled anti-IFN-γ Abs. Ly5.1+ T cells (1 × 104) were acquired in a FACSCalibur device and analyzed using CellQuest software (BD Biosciences).
Assessment of dendritic cell (DC) maturation in vivo
C57BL/6 mice were immunized s.c. in the neck region with 20 μg of p33-VLPs or, alternatively, with 20 nmol of CpGs and DCs from draining LN (cervical LN) were isolated 1, 2, and 3 days after priming. Briefly, LN from three mice per group were pooled and digested twice for 30 min each time at 37°C in IMDM supplemented with 5% FCS and 100 μg/ml collagenase D (Roche, Mannheim, Germany). Released cells were resuspended, and DC were purified by CD11c+ MACS beads isolation (Miltenyi Biotec). To determine the phenotype of DCs obtained from LN of immunized mice, CD11c+ cells were stained with anti-CD80, anti-CD86, anti-CD40, or anti-I-Ab MHC class II Abs (all from BD PharMingen, San Diego, CA) for 30 min on ice. Their expression pattern was then compared with that one of DCs obtained from LN of untreated mice. Where indicated, myeloid, lymphoid, and plasmacytoid DC subtypes were differentiated by additional anti-CD8 and anti-Gr1 Ab staining (also BD PharMingen).
Assessment of antiviral immunity in vivo
To examine antiviral immunity, vaccinated female C57BL/6 mice were infected i.p. with 1.5 × 106 PFU Vacc-GP. Five days later ovaries were collected, and the vaccinia titers were determined on BSC 40 cells as previously described (23).
Efficient T cell proliferation in the absence of effector T cell induction
T cells specific for peptide p33 were isolated from spleens and LN of Ly5.1+ TCR-transgenic mice. After labeling with CFSE, 5 × 106 cells were transferred into normal congenic Ly5.2 C57BL/6 recipient mice, which were subsequently immunized with LCMV (strain WE, 1200 or 1.2 × 106 PFU), free peptide p33 or p33-VLPs (both 20 μg). Additional mice were immunized with peptide plus CpGs to assess the role of APC activation. Three or 5 days later, cells from draining LN were isolated and stained for the expression of CD8 and Ly5.1, allowing us to specifically follow the transferred T cells (Fig. 1⇓). LCMV given at a low dose did not induce measurable proliferation within 3 days, but T cells had undergone extensive cell cycling by day 5. Moreover, free peptide clearly induced proliferation within 3 days, while on day 5 few specific T cells could be detected in draining LN. Addition of CpGs slightly enhanced proliferation up to day 3, but resulted in dramatically increased accumulation of specific T cells by day 5. The proliferation pattern induced by p33-VLPs appeared similar to that induced by peptide plus CpGs and also resulted in accumulation of specific T cells by day 5. Induction of effector cells was assessed in the same experiment by intracellular IFN-γ staining after in vitro restimulation with specific peptide for 6 h (Fig. 2⇓). Despite the induction of extensive proliferation by vaccination with peptide, no significant effector cell induction could be observed. However, the picture was reversed if APCs were activated by CpGs during vaccination, and a large population of effector T cells was induced under these conditions. Since the same APCs most likely presented peptide p33 in the presence or the absence of CpGs, it was the activation status, rather than the type of APCs, that regulated effector cell generation (see also below). Thus, immunization with p33 peptide alone induced T cell proliferation in the absence of effector cell induction. In addition, immunization with p33-VLPs or LCMV induced sizeable effector cell populations. It should also be pointed out that while p33-VLP-derived peptide is presented mainly by professional APCs (DCs and, to a minor extent, macrophage) (17, 18), free p33 peptide will probably be presented by all cells expressing MHC class I molecules, because it binds directly to surface-expressed MHC class I molecules and does not need to be further processed. Nevertheless, the coadministration of CpGs with peptide could significantly induce effector T cell generation.
Cytokine secretion depends on the number of cell cycles that T cells have gone through (24). It was therefore possible that the difference in effector cell induction may be due to less extensive cell cycling in the absence of CpGs. This was, however, not the case, since two doses of peptide induced proliferation as extensive as that observed after immunization with peptide plus CpGs. Nevertheless, no effector T cells were induced (not shown). Interestingly, both LCMV and, to a lesser degree, p33-VLPs were able to induce cytokine production in the absence of proliferation (Fig. 2⇑, left panel). This observation is reminiscent of earlier findings for CD4+ T cells, where cytokine production could be detected in the absence of T cell proliferation (25). Whether this finding is physiologically important or is due to bystander T cell activation in TCR-transgenic systems needs to be further investigated.
Up-regulation of activation markers despite the absence of effector T cells
Induction of activation markers on transferred cells was assessed next. Surprisingly, despite the clear-cut difference in effector cell induction in the presence or the absence of CpGs, no substantial difference in the expression of CD25, CD44, or CD69 was observed if free peptide p33 was used for immunization (Fig. 3⇓). Vaccination with peptide alone therefore resulted in fully activated T cells if the expression of activation markers (in particular CD44) or proliferation was assessed. Nevertheless, no effector cells were induced by this regimen. Interestingly, LCMV induced more pronounced expression of CD25, most likely because of Ag persistence due to viral replication, peaking between days 3 and 5 after infection (26, 27).
Interestingly, specific T cells could be identified that had not undergone cell division in mice primed with p33-VLPs 3 days earlier. Nevertheless, these cells expressed up-regulated levels of CD69. This observation indicated that these cells had been activated very recently, suggesting that p33 derived from p33-VLPs may persist for several days on APCs, leading to prolonged Ag presentation (see also below). Note that the injection of 20 nmol of CpGs did not influence basal CD25, CD44, and CD69 expression on endogenous Ly5.2+ T cells (not shown).
Up-regulation of costimulatory molecules or extended Ag presentation may result in the generation of effector cells
These results indicated that LCMV, p33-VLPs, or free peptide presented on activated APCs were able to drive generation of effector cells. Activation of APCs may result in increased expression of costimulatory molecules and cytokines or may lead to enhanced survival of DCs, resulting in prolonged Ag presentation. To distinguish between these possibilities, mice were vaccinated with peptide alone, peptide plus CpGs, or p33-VLPs, and CFSE-labeled TCR-transgenic T cells were transferred into the recipient mice 1, 2, or 3 days later. Under these conditions, the induction of T cell proliferation in recipient mice reflects the presence of MHC class I-associated Ag at the point of cell transfer. Peptide was still presented by APCs 1 day after immunization with peptide alone or peptide plus CpGs, but could not be detected with this sensitive in vivo assay at later time points. Thus, activation of APCs did not prolong Ag presentation. In contrast, vaccination with p33-VLPs resulted in prolonged Ag presentation, which could be detected up to 3 days after vaccination (Fig. 4⇓A). Note that at a molar ratio, 20 times more free peptide (20 μg) was injected than peptide fused to VLPs (20 μg), excluding the possibility that extended Ag persistence after vaccination with p33-VLPs was due to higher Ag load. Intracellular IFN-γ staining was then performed to analyze under which conditions effector cells are generated in this particular experimental setup. Interestingly, when mice were treated with p33-VLPs 1 day before T cell transfer, ∼50% of proliferating cells expressed IFN-γ. The number of IFN-γ+ T cells dropped to 26 and 12%, for days 2 and 3, respectively (Fig. 4⇓B). These results show that the shorter the Ag presentation is (e.g., days 2 and 3), the more reduced the IFN-γ production becomes (despite a significant T cell proliferation), providing thereby further evidence that the persistence of Ag is required for the induction of effector function.
Expression of costimulatory molecules on DCs isolated from draining LN at different time points after vaccination was assessed next. Maximal DC activation was observed early after CpG treatment (day 1); in fact, CpGs induced massive up-regulation of CD80, CD86, and CD40 (Fig. 5⇓). Furthermore, high CD80 expression was observed over the whole time window, while CD86 and CD40 expression dropped sharply after day 1 and reached normal levels on day 3. The increased CD80 expression until day 3 in mice receiving CpGs indicates that at least a part of the DCs may be present at late time points, even if they will probably deliver weaker costimulation signals to T cells. Although it remains possible that CpGs persist for 3 days in the host and induce up-regulation of costimulatory molecules on newly generated DCs, this seems unlikely, since only CD80 remained up-regulated for 3 days. VLPs induced only marginal, but reproducible, up-regulation of the activation markers; free peptide, by contrast, did not show any significant effect (not shown).
The activation pattern of different DC subtypes was investigated 1 day after CpG or p33-VLP treatment (maximal DC activation). Therefore, myeloid (CD8−), lymphoid (CD8+), and plasmacytoid (Gr1+) DCs were stained for CD80, CD86, CD40, and MHC class II expression. The analysis showed that even if the three different DC subtypes express relatively different levels of the activation markers in the nonstimulated state, CpG treatment induced significant up-regulation of all four molecules (at least by a factor of 3) in all DC subsets (Table I⇓). Thus, because CpGs do not prolong presentation of peptide p33, they seem to enhance T cell responses due to increased expression of costimulatory molecules on DCs. In contrast, enhanced induction of effector cells by p33-VLPs correlated with prolonged Ag exposure rather than superior expression of costimulatory molecules.
To exclude that a possible LPS contamination of p33-VLPs was responsible for the observed effector cell induction, the experiments were repeated in TLR-4-deficient LPS nonresponder mice. Comparable up-regulation of activation markers on DC, proliferation and effector function in T cells were observed after vaccination with p33-VLPs (not shown). Thus, LPS was not responsible for the difference between free peptide and p33-VLPs. Taken together, these results indicate that either increased expression of costimulatory molecules or prolonged exposure to Ag is sufficient to drive effector cell differentiation in this system. Nevertheless, it should be pointed out that p33-VLPs have the ability to trigger a low level activation of DCs (Fig. 5⇑) (17), which may facilitate the generation of effector cells.
Both up-regulation of costimulatory molecules and extended Ag presentation are required for protective immunity in a nontransgenic situation
CTLs are able to produce factors inducing the maturation of DCs and therefore may create a costimulatory milieu on their own, in particular if they are present at high frequency (28, 29, 30). We therefore tested requirements for protective immunity in a nontransgenic mouse model. Naive C57BL/6 mice were vaccinated with p33 plus CpGs, p33-VLPs alone, or p33-VLPs together with CpGs. The frequencies of specific T cells were measured 8 days later in the blood (Fig. 6⇓A). As reported previously (18), p33-VLPs alone did not induce high frequencies of specific CTLs. Moreover, p33 plus CpGs were not able to induce significant numbers of specific CTLs. Only vaccination with p33-VLPs together with CpGs resulted in a sizeable population of specific T cells, with an average of 2.0 ± 0.7% p33-specific CD8+ T cells in the total CD8+ T cell population. To determine whether vaccination with peptide plus CpGs results in significant antiviral protection, mice were challenged with recombinant vaccinia virus expressing LCMV-GP. Vaccination with p33-VLPs plus CpGs resulted in protection, confirming earlier results (18), while p33 plus CpGs failed to induce measurable protection from viral infection (Fig. 6⇓B). Thus, effector T cells may be induced in the TCR-transgenic T cell model, where unphysiologically high frequencies of naive T cells are present, by either prolonged Ag presentation or enhanced costimulatory molecule expression by DCs. However, for priming of a protective CTL response in a nontransgenic situation, both enhanced expression of costimulatory molecules and extended Ag presentation were required.
Protective immunity critically depends on the presence of effector T cells. We show here that the proliferation of specific T cells and the expression of activation markers are uncoupled from effector T cell generation. Either presentation of peptide on activated APCs or prolonged exposure to Ag facilitated generation of effector T cells. However, for the induction of protective T cell-mediated immunity under physiological conditions, peptide had to be presented for extended time periods by activated professional APCs.
It has previously been shown that a single, brief antigenic stimulation may be sufficient to induce extensive proliferation of T cells (31, 32). This was confirmed here, since injection of peptide triggered massive proliferation of specific T cells. In fact, virtually all transferred T cells underwent extensive cell cycling after peptide injection, and increased expression of CD44 and other activation markers could be observed. Despite proliferation, T cells stimulated by vaccination with peptide in saline failed to differentiate into effector cells. In fact, peptide-activated T cells divided up to four to five times, but eventually died without gaining effector function. Only if the peptide was presented by activated APCs or was present in the host for several days would TCR-transgenic T cells differentiate sufficiently to produce IFN-γ within 6 h of in vitro stimulation. Thus, activation of T cells by peptide alone results in partial T cell activation in vivo, since T cells efficiently proliferate and express activation markers, but fail to secrete cytokines. It is interesting to note that the survival and accumulation of specific T cells correlated with the presence of effector cells. Specifically, T cells stimulated by free peptide not only failed to produce cytokines, but also died or at least disappeared from draining lymphoid organs within a few days after activation. These results suggest that similar factors drive effector cell generation and T cell survival and are compatible with the hypothesis that memory T cells originate from effector T cells (33).
It should be pointed out that free peptide may be presented by different types of APCs than virus-like particle-derived Ag. In fact, VLPs require cellular internalization and processing for p33 presentation, and this is a restricted capacity of professional APCs, such as DCs or macrophages; B and T cells fail to present p33-VLPs for MHC class I presentation in vitro (17). Therefore, the duration of Ag presentation may not be the only difference between p33-VLPs and free p33; the quality of Ag presentation could also be important for the difference between the two antigenic forms. Nevertheless, CpG-mediated activation of professional APCs could overcome transient activation, followed by deletion induced by free peptide, and a sizable population of IFN-γ producing CD8+ T cells could be detected. In addition, p33-VLPs presented by professional APCs for a short time period failed to induce effector T cells, demonstrating that, indeed, the duration of Ag presentation drives effector T cell generation (Fig. 4⇑).
The ability of T cells to rapidly gain effector function is key for their protective capacity (34, 35, 36). In fact, both naive T cells and resting (central) memory T cells are inefficient at mediating antiviral (13, 14) or antibacterial protection (15). As an extreme example, naive mice expressing a transgenic TCR specific for peptide p33 are not protected from intracerebral infection with LCMV. Thus, despite a precursor frequency of almost 1:1, naive T cells cannot protect from such a type of infection (14). Similar observations were made for resting (central) memory T cells (13). Moreover, elevated frequencies of tumor-specific CTLs may not be sufficient to rid a tumor. Even the induction of effector cells is often not enough for protection, but a population of effector T cells has to be maintained for extended time periods for effective tumor therapy (37, 38). Interestingly, activation of DCs during vaccination has been shown to be effective at eliminating transplanted tumors (39, 40) and inducing autoimmunity (41). These observations are compatible with the data presented here showing that CpGs dramatically increase effector T cell induction while leaving T cell proliferation largely unaffected. Thus, the ability to induce effector T cells rather than the induction of T cell proliferation appears to be key for the induction of protective effector T cell responses.
TCR-transgenic T cells are ideal tools to study the differentiation of naive T cells into effector and memory T cells. Nevertheless, the presence of unphysiological high numbers of naive T cells may facilitate the induction of effector cells, leading to an underestimation of the in vivo thresholds that have to be overcome for successful vaccination. We and others have shown earlier that CD8+ T cells may produce factors that activate DC, leading to Th cell-independent CTL responses (28, 29). Very high precursor frequencies of specific CTLs may therefore facilitate the induction of effector CTL responses (30). Using adoptively transferred T cells, we found here that either activation of APCs during vaccination or prolonged Ag exposure was sufficient to drive effector cell function. However, these vaccination regimens failed to induce protective immunity in normal mice. To induce a protective CTL response under “real-life” conditions, APCs had to be activated, and Ag presentation had to occur for several days. If only one of the conditions was fulfilled, protective immune response could not be detected. Thus, the duration of Ag presentation and the activation of APCs together drive the induction of effector CTL responses.
The observation that peptide-stimulated T cells proliferated and expressed activation markers before cell death is reminiscent of the phenotype of T cells undergoing peripheral deletion (42, 43, 44). In fact, it has been shown earlier that high doses of peptide may induce tolerance, rather than immunity (45, 46, 47). This may suggest that antigenic stimulation in the absence of APC activation by inflammatory or pathogen-derived factors induces abortive immune responses, resulting in T cell tolerance. In marked contrast, if APCs presenting the peptide are activated, a protective effector T cell response may be mounted. APCs can be activated by T cell-specific factors, such as CD40 ligand (48, 49), RANKL/Trance (50), and others (28, 29). Often and perhaps even generally during pathogen-specific immune responses, APC-activating stimuli are generated by cells of the innate immune system (e.g., TNF or α/β-IFNs, etc). In addition, DCs and macrophages express receptors, such as Toll-like receptors, directly recognizing patterns associated with pathogens. Prominent examples are the CpG-containing DNA oligonucleotides used here, which are recognized by Toll-like receptor 9 (51), or LPS, which is recognized by Toll-like receptor 4 (52). Thus, the innate immune system is usually required for activation of APCs, while the adaptive immune system delivers the specific T cells. Only if both systems collaborate may protective effector T cell responses be mounted, while in the absence of innate responses, T cell tolerance is established.
We thank Manfred Kopf and Pamela Ohashi for helpful discussions, and Pia Wildhaber for excellent secretarial assistance.
↵1 Address correspondence and reprint requests to Dr. Martin F. Bachmann, Cytos Biotechnology AG, Wagistrasse 25, 8952 Schlieren-Zurich, Switzerland. E-mail address:
↵2 Abbreviations used in this paper: APL, altered peptide ligand; DC, dendritic cell; LCMV, lymphocytic choriomeningitis virus; LN, lymph node; VLP, virus-like particle.
- Received October 23, 2002.
- Accepted May 27, 2003.
- Copyright © 2003 by The American Association of Immunologists