Cross-Presentation of the Long-Lived Lymphocytic Choriomeningitis Virus Nucleoprotein Does Not Require Neosynthesis and Is Enhanced via Heat
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
Many viral proteins that contain MHC class I-restricted peptides are long-lived, and it is elusive how they can give rise to class I epitopes. Recently, we showed that direct presentation of an epitope of the long-lived lymphocytic choriomeningitis virus nucleoprotein (LCMV-NP) required neosynthesis in accordance with the defective ribosomal products hypothesis. In this study, we report that LCMV-NP can be cross-primed in mice using either LCMV-NP-transfected human HEK293 or BALB/c-derived B8 cells as Ag donor cells. In addition, we establish that contrary to direct presentation, cross-presentation required accumulation of the mature LCMV-NP and could not be sustained by the newly synthesized LCMV-NP protein, intermediate proteasomal degradation products, or the minimal NP396 epitope. Nevertheless, NP cross-presentation was enhanced by heat shock and was blunted by inhibitors of heat shock protein 90 and gp96. We propose that cross-presentation has evolved to sustain the presentation of stable viral proteins when their neosynthesis has ceased in infected donor cells.
Introduction
Peptides presented on MHC class I molecules are normally generated through fragmentation of proteins that have been synthesized within the cell. It is therefore remarkable that several structural proteins of viruses are quite stable in cells and nevertheless give rise to sufficient quantities of peptide to induce a CTL response (1, 2, 3). A hypothesis which reconciles this apparent contradiction is the defective ribosomal products (DRiPs)4 hypothesis that states that CTL epitopes are not derived from properly folded proteins but from newly synthesized polypeptides that are degraded within a few minutes after their translation (4). The lymphocytic choriomeningitis virus nucleoprotein (LCMV-NP) is an example of such proteins since it is stable over 3–4 days but nevertheless contains immunodominant epitopes like NP396. Epitopes generated from LCMV-NP shortly after neosynthesis were the sole factor for its direct presentation on MHC class I (5). However, the mechanisms involved in the presentation of DRiPs via the MHC class I pathway are only just surfacing (6, 7, 8).
Professional APCs (pAPCs) can generate class I peptides by two different mechanisms. Either from internally synthesized proteins (9) via "direct-presentation" or by acquiring exogenous Ags through "cross-presentation." Cross-presentation represents a vital pathway for the initiation of immune responses against tumors and viruses that do not have access to, or interfere with, the classical pathway for MHC class I presentation (10, 11, 12, 13, 14, 15). Dendritic cells and macrophages are pAPCs that have the unique capability of generating and presenting peptides on MHC class I molecules from exogenous Ags (16, 17, 18). Activation of naive TCD8+ cells by such pAPCs that have cross-processed Ags into the MHC class I pathway is known as cross-priming (13).
Although the DRiP hypothesis explicates how direct presentation of stable proteins in general can occur, it remains unknown whether stable viral proteins such as LCMV-NP can be cross-presented by pAPCs. This is particularly important since in vivo evidence questioned the ability of LCMV-NP to access the cross-priming pathway (19), yet, it was recently shown that proteins are the mediators of in vivo cross-priming (20, 21, 22). Moreover, immediate proteasomal products were reported to be the main source of Ags transferred to the pAPC in another virus system (23), while others substantiate that Ags could be transferred as peptides bound to heat shock proteins (HSPs) (24). It is possible that virus-specific differences, depending on the nature of virus proteins and their cellular localization, could exist during Ag processing. We therefore started our study by exploring the following avenues; are stable proteins, such as LCMV-NP, predisposed to access the cross-presentation pathway and, if so, what are the antigenic elements responsible for that (20, 21, 22, 23, 24, 25)?
In this study, we report on the cross-presentation of LCMV-NP in vitro and in vivo. We show that accumulated mature proteins probably aided by molecular chaperones are the source of Ags fueling the cross-presentation pathway. Interestingly, neither proteasomal products nor nascent proteins nor DRiPs per se were capable of providing Ags to the cross-presentation-pathway. Our data highlight the significant role of the cross-presentation machinery in degrading stable proteins that are otherwise resistant to processing via the proteasome in virus-infected cells.
Intracellular detection of LCMV-NP
For the detection of LCMV-NP, cells (1 x 105) were harvested and resuspended in 200 μl of PBS, followed by fixation with 4% Formalin for 20 min at room temperature. After washing with 1x PBS, the cells were permeabilized with 1% Triton X-100 for 20 min at room temperature, followed by incubation overnight at 4°C with rat anti-LCMV-NP Ab (clone VL4) (28) in PBS/2% FCS. After two further washing steps, FITC-conjugated polyclonal mouse anti-rat Ab (diluted 1/150 in PBS; BD Biosciences) was left with the cells for at least 1 h at room temperature. Data were acquired with a FACScan flow cytometer (BD Biosciences) using CellQuest (BD Biosciences) and analyzed with FlowJo software (TreeStar).
Intracellular cytokine staining (ICS)
In vitro To detect TCD8+ cell activation with ICS, effector cells were coincubated in round-bottom 96-well plates with (0.1 μM) peptide-loaded APCs (3 x 105/well). Either DMSO or irrelevant peptide (gp33) were used as controls for background activation in all assays. The coincubation ratio of 1:1 APCs (BMC-2) to responders was chosen when T cells were expanded for 5–6 days in vitro. When TCD8+ cell lines were used, using the DC2.4 cell line as stimulator cells, brefeldin A (BFA) was added directly and the cells were left together for 3 h before staining.
In vivo Ex vivo responders were incubated with the specific peptides for 2 h at 37°C and then left in the presence of BFA (Sigma-Aldrich) at 10 μg/ml for 3 h. The coincubation ratio of 1:10 APCs (BMC-2) to responders was chosen in the case of direct analysis ex vivo. In general, cells were then stained with PE-Cy5 anti-CD8 rat IgG Ab clone 53-6.7 (BD Biosciences) on ice for 20 min, washed, and fixed with 1% paraformaldehyde in PBS at room temperature for 20 min. After washing, we labeled the cells with FITC-conjugated rat anti-IFN- Ab clone XGM1 (BD Pharmingen) in PBS/0.1% saponin at 4°C overnight. Stained cells were analyzed with a live gate on the CD8+ cells after 80,000 gated cells were acquired.
Tetrameric MHC class I peptide analysis
Tetrameric complexes containing biotinylated H-2Db 2-microglobulin, the NP396 peptide, and extravidin-PE were generated as described previously (26, 29). Splenocytes (5 x 105 cells) were stained with 0.5–1 μg of tetramer in 50 μl of FACS buffer (PBS, 2% FCS, 0.01% NaN3, and 20 mM EDTA) at room temperature for 30 min, followed by addition of PE-Cy5 anti-CD8 rat IgG Ab clone 53-6.7 (BD Biosciences) on ice for 30 min. Cells were washed twice and data were acquired using CellQuest (BD Biosciences) as before by gating on CD8+ lymphocytes.
TCD8+ cell priming and induction of T cell lines
For in vivo priming, 8- to 10-wk-old female C57BL/6J mice were injected i.p. with 1 x 107 cells in PBS. In the case of primary responses, ICS measurements were performed directly 8 days after injection. In some experiments, spleens were harvested and restimulated with peptide-loaded APCs. On day 5 or 6 of the in vitro expansion, T cells were tested for IFN- production by ICS as before.
For generating NP396-specific T cell lines, C57BL/6 mice were injected i.v. with 200 μl of LCMV-WE (1000 PFU/ml). The spleens were harvested 4 wk after injection and the splenocytes were purified by density centrifugation (Ficoll-Paque). The splenocytes were cultured in RPMI 1640 containing mouse IL-2 (40 U/ml) along with gamma-irradiated peptide-loaded APCs. An additional density centrifugation step was conducted 5–6 days later when they were used for in vitro Ag presentation experiments. In all assays, we used the T cell lines at this time point and their specificity corresponded to at least 85% when tested with peptide (50 nM). For the detection of the epitope NP396, we resuspended the APCs in RPMI 1640 (1 x 106/ml), added them to NP396-specific TCD8+ cells, and stained for CD8 and IFN- as described above.
Preparation of Ag donor or presenting cells (ADCs and APCs)
ADCs were detached with trypsin-EDTA, washed with PBS, and resuspended in RPMI 1640 (1 x 106 cells/ml) and treated with different inhibitors where indicated. For interfering with HSPs, the HSP90-specific inhibitors geldanamycin (5 μM) and herbimycin A (5 μM) were used (Sigma-Aldrich). Cycloheximide (250 μg/ml) was obtained from Calbiochem. For testing different conditions, the cells were either further heated (14 min at 45°C), or UV irradiated (8 min), or snap frozen in liquid nitrogen (–180°C) and UV irradiated for 8 min (lysed and UV treated (LyUV)). UV-triggered apoptosis was induced using the Stratalinker 2400 UV cross-linker that provides a radiation intensity of 4 mW/cm2 per s. Interestingly, one round of lysis resulted in intact cells incorporating trypan blue, whereas with LyUV cells only 40% incorporated trypan blue. If the cells were subjected to repeated cycles of freezing and thawing, the outcome was mainly cell fragments.
As APCs, we used the macrophage cells line BMC-2 that was recently shown to posses competent phagosomes as organelles for Ag cross-presentation (30). Where indicated, the following inhibitors were used: BFA (10 μg/ml), chloroquine (100 μM), leupeptin (100 μM), pepstatin A (50 μg/ml), fucoidin (250 μg/ml; Sigma-Aldrich), lactacystin (LC, 50 μM; Biomol). During the cross-presentation assays, coincubation of ADCs with BMCs was conducted in round-bottom 96-well plates (greiner bio-one; Cellstar) at 37°C at a ratio of 1:1 for 16 h.
Results
Expression of LCMV-NP in transfected cells
To study LCMV-NP cross-presentation, we used human HEK293 cells as ADCs since they cannot directly present the H-2Db-restricted NP396 epitope to specific T cell lines that we used to monitor cross-presentation. We generated stable transfectants expressing either the full-length LCMV-NP (designated HEK-NP) or a ubiquitin-LCMV-NP fusion protein (HEK-Ub-NP). The N-terminal ubiquitin moiety is a G76A mutant that cannot be cleaved off by ubiquitin-specific proteases and targets the fusion protein for rapid degradation by the proteasome (31). Western blot analyses of LCMV-NP in either LCMV-infected HEK293 cells (HEKi) or HEK-NP cells demonstrated that LCMV-NP was abundantly expressed (Fig. 1A). LCMV-NP expression was very low in untreated HEK-Ub-NP cells but proteasome inhibitors markedly enhanced the expression level, thus confirming rapid proteasomal degradation of the fusion protein (Fig. 1A, HEK-Ub-NP + LLnL) as previously demonstrated in transiently transfected cells (31). The relative NP expression levels as assessed by Western blot analysis were further confirmed by pulse-chase analysis as well as intracellular NP staining and flow cytometry (data not shown and Fig. 1B).
To confirm these findings independently and to assess the role of peptide epitopes for cross-presentation, we infected either HEK293 cells or J774 cells (H-2d) with rVV encoding either the full-length LCMV-NP protein or the NP396 minigene. In this situation, donor cells were subjected to UV irradiation to prevent secondary reinfection of the APCs. As shown in Fig. 3B, only ADCs that were expressing the full-length protein served as an efficient source of Ag for the activation of NP396-specific TCD8+ cells. In marked contrast, the rVV encoding the NP396 minigene failed to induce activation via cross-presentation, suggesting that the minimal peptide Ags cannot be efficiently cross-presented.
Heat shock or lysis and UV irradiation increases cross-presentation of LCMV-NP but not of LCMV-Ub-NP
To test different factors that may affect cross-presentation, we subjected HEK-NP and HEK-Ub-NP cells to either heat shock or to lysis by freezing/thawing followed by UV irradiation (LyUV). The heat shock treatment (14 min at 45°C) did not significantly affect the viability of HEK cells as determined by propidium iodide exclusion staining (11 ± 4% dead cells in both control or heated cells) and trypan blue (data not shown). As shown in Fig. 3C, both protocols enhanced cross-presentation of the ADCs with LyUV showing a superior outcome. Remarkably, none of the two ADC treatment protocols were able to enhance cross-presentation of HEK-Ub-NP. However, cross-presentation of HEK-Ub-NP could be recovered when the ADCs were treated with LC (Fig. 3D). Incubation with LC for 8 h did not significantly affect cell viability (propidium iodide) exclusion staining (14 ± 6% dead cells for LC-treated control or heated cells). This suggests that HSPs can enhance cross-presentation that is biased toward proteins but not peptides as an Ag source in this pathway.
The HSP90 family enhances cross-presentation of LCMV-NP
Because initial experiments indicated that treatment of the ADCs leading to increased expression of HSPs enhanced cross-presentation, we examined the outcome of interfering with a prominent class of the HSP families. HEK-NP cells were treated with the very well-characterized HSP90 inhibitors geldanamycin and herbimycina before heat shock or LyUV treatment. As shown before, when ADCs were either heat shocked or LyUV treated, there was an increase in the signals (Fig. 3C). When we used the inhibitors that are specific for both HSP90 and gp96, the signals were almost at background levels (Fig. 4A, heated or LyUV). The effect of these inhibitors on intact cells was not apparent, as no significant differences were detected (Fig. 4A, intact). The possibility that the effect of HSP90 inhibitors was due to a direct inhibition of the responder T cell lines was ruled out as peptide-loaded APCs incubated with either untreated or inhibitor-treated ADCs induced similar activation of the T cell lines (activation was almost 85%, data not shown). In addition, we ruled out possible negative effects of these inhibitors on the NP levels in ADCs by intracellular NP staining and flow cytometry (Fig. 4B).
A fundamental issue regarding the presentation of LCMV-NP through the cross-presentation pathway centers on whether DRiPs per se are a source of Ag for cross-presentation. To address this point, we tried two different approaches: the first one was based on inhibition of de novo protein synthesis before injecting the ADCs in vivo. As shown in Fig. 6A, we could not detect any significant differences if ADCs were treated with cycloheximide, suggesting that the mature folded protein is responsible for the cross-presentation activity. In vitro, when de novo protein synthesis in ADCs was inhibited by cycloheximide, an inhibitory effect on cross-presentation in vitro was observed (data not shown). Unfortunately, using this approach is inadequate because translation inhibitors can affect a variety of factors in the cell.
To resolve these issues, we used the tight tetracycline (tet)-repressor system where systemic effects of inhibiting protein neosynthesis are ruled out. This system relies on the tet-repressible expression of LCMV-NP in the fibroblast transfectant B8tNP64. In these cells, NP neosynthesis can be switched on and off without using inhibitors, which has been instrumental in identifying LCMV-NP as a DRiP substrate (5). First, we checked whether LCMV-NP induced for 24–72 h in B8tNP64 cells was cross-presented in vivo. This occurred to a similar degree as we have observed for HEK-NP cells (Fig. 2 and data not shown). We have thus chosen the 24-h time point for further in vitro analyses. To assess the control of LCMV-NP expression and the specificity of cross-presentation in vitro, we used either B8 recipient cells (B8-wt) or B8tNP64 cells (NP64) that were tet suppressed and confirmed that cross-presentation was undetectable when LCMV-NP was not expressed (Fig. 6B). To examine the effect of neosynthesis, we removed tet for 5 h (Fig. 6B, –5 h) so that we could analyze the role of DRiPs as an Ag source, but we failed to detect NP396 presentation after exogenous processing, unlike direct presentation (data not shown). We could recover the activity if tet had been removed for a period of 24 h, thus allowing LCMV-NP accumulation (Fig. 6B, fourth column, tet–), strongly indicating that the accumulated mature protein rather than DRiPs or nascent proteins is responsible for cross-presentation. The LCMV-NP expression levels detected by FACS and immunoblot analyses in either B8tNP64 or HEK-NP cells support this proposition (Fig. 6C and data not shown). The measurements obtained by flow cytometry reflect the nature of the fully folded protein since this Ab does not detect native forms of NP in Western blots. Importantly, as indicated before with the HEK-NP cells, when we awaited NP expression to build up for 24 h and then applied the chaperone inhibitors, no significant negative effects on NP expressions were found (Fig. 6C, histogram overlays).
The ability to cross-present LCMV-NP was significantly reduced when chaperones of the HSP90 family were inhibited in B8tNP64 cells with geldanamycin and herbimycina that were added to cells after tet removal for 24 h (Fig. 6B, columns 5 and 6). To exclude that HSP90 inhibitors negatively affected NP396 presentation by APCs, we used peptide-pulsed APCs as a control (Fig. 6B, columns 8–10). To further scrutinize the requirement for neosynthesis in cross-presentation, we allowed protein synthesis to proceed for 24 h and then added tet for 5 h to block any further neosynthesis (Fig. 6B, ±5 h). If DRiPs were playing a role, then the signal would be drastically reduced. This experiment revealed that, when only the mature LCMV-NP protein was present but no DRiPs, cross-presentation was as efficient as if LCMV-NP synthesis was permitted over 24 h. Taken together, it is evident from our results that NP neosynthesis is not required as an Ag source in LCMV-NP cross-presentation.
Discussion
Better understanding of factors contributing to the cross-presentation pathway is an important issue at present in the field of Ag presentation and immune surveillance. Recently, a novel phagosomal pathway involved in cross-presentation, which merits further assessment, has been discovered (30, 32, 33). Furthermore, the involvement of DRiPs as a source of Ags is a new paradigm in direct presentation (4). These new concepts have prompted us to examine whether cross-presentation of the long-lived LCMV-NP (5) is attainable in APCs after its uptake, since it is not readily degraded in ADCs. Consequently, we set our goals to scrutinize the parameters involved in its degradation in APCs and investigated the relative contribution of proteins vs DRiPs or peptides as Ag sources and the role of HSPs therein.
We tested the ability of this long-lived protein and its degradable products to cross-present in vitro and in vivo to investigate whether it conforms with the recent findings regarding the role of proteins in cross-priming (20, 21, 22) or its stability does not allow it to access the cross-presentation pathway (19). Although the in vitro assays proved to be a more sensitive readout system compared with in vivo, the overall conclusion reached was similar in both systems. Accordingly, and contrary to previously reported data (19), we found that NP transfectants (xeno- or alloresponses) can prime TCD8+ cells in vivo via the exogenous pathway. Perhaps, our system was more sensitive to detect such activity because our transfectants expressed sufficient amounts of NP. Cross-priming and presentation of this long-lived protein was considerably reduced when it was targeted for rapid degradation by the proteasome. This conclusion was confirmed by different means; T cell frequencies were estimated using tetramer analysis and their activation was examined directly ex vivo or after expansion in vitro. Considering all of the data combined, it is clear that less efficient T cell activation, at least four times less, occurred when the NP was targeted for degradation. We hypothesized at this stage that the long-lived form of NP is a better source for cross-priming than the unstable ubiquitin (Ub)-NP fusion protein. This finding concurs with and further extends the concept that proteasomal substrates in general rather than proteasomal products are the main source of cross-priming in vivo (20, 21, 22).
Although the Ub-NP transfectants generate more peptides or polypeptides than full-length protein in the donor cells, they nevertheless were poor inducers of TCD8+ cells via cross-priming. We confirmed these findings with rVV encoding either the full-length protein or the minigene, where the full-length protein was superior in its ability to cross-present compared with the cytosolic peptide. These conclusions are supported by earlier reports showing that cytosolic peptides, unless bound to chromatin, are unstable with a very short half-life due to cytosolic proteases (34). Furthermore, cross-priming of T cell responses in vivo does not require antigenic peptides in the ER (35). According to our results, degraded proteins are unlikely to play a major role in cross-priming of T cell responses in agreement with recently published data (20, 21, 22). Yet, HSPs bound to antigenic peptides were suggested to be a potent source of Ag for cross-presentation (24, 36). In contrast, the circumstances under which chaperone-peptide interactions play a specific role in immunity has been put in question recently and need to be thoroughly investigated (36, 37). Our findings reconcile the proposed role of HSPs in cross-presentation on the one hand and the inability of peptides to cross-prime on the other hand. Based on our results, the ability of HSPs to enhance cross-presentation is evidently biased toward proteins interactions and not proteasomal products.
To examine conditions where possible danger signals may affect the activation of pAPCs (36, 37, 38), we compared intact HEK-NP cells with other stress settings where induction or release of chaperones is envisaged. Either LyUV or heat shock protocols led to a marked enhancement of cross-presentation. This was not the case if we just UV treated intact ADCs (data not shown), indicating that the differential cell treatment can influence immunogenicity (39, 40). In our system, the use of HSP90 and gp96 inhibitors impeded cross-presentation markedly, thus indicating the involvement of one or both of these chaperones after the induction of stress. A very large body of strong biochemical and pharmacological evidence has established geldanamycin and herbimycin A as specific inhibitors of HSP90 family chaperones (41, 42). The chaperone function of HSP90 relies on ATP hydrolysis (43), which is blocked by geldanamycin and herbimycin A because they bind to the ATP binding site of HSP90 according to x-ray crystallographic analysis (44). In addition, a very recent finding reported on reduced cross-priming in mice that were defective in heat shock factor 1 (Hsf1). These mice have a decreased expression of several HSPs including HSP90 and HSP70 (45).
There are several reasons, currently under investigation by us, why these chaperones could have led to increased presentation. It is possible that the effect is indirect due to stimulation of APCs via cell surface signaling receptors like TLRs or increased uptake of ADCs via HSP receptors (36, 37, 38). The chaperones may be required to maintain the proper folding or function of other unknown interacting factors that are generally needed for the cross-presentation pathway. Alternatively, if HSPs are tightly bound to NP, especially under stress conditions, then one could envisage a direct role of chaperones as Ag carrier molecules. Therefore, a possible model could be based on protein-bound HSPs rather than peptide-bound HSPs in enhancing cross-presentation. We also examined certain aspects of the uptake mechanism and the downstream events that follow. We looked into the role of SR-A since it has an array of ligands (46), including certain molecular chaperones (47). A strong reduction of the cross-presentation signal was observed when fucoidin was incubated with APCs. Since fucoidin, a polysaccharide that competitively binds to the ligand-binding domain of SR-A (48), interfered with the presentation signal, it is likely that SR-A is playing an important role in Ag uptake. However, in this instance we could not single out whether the fucoidin was competing out stress chaperones (47) or other undefined SR-A ligands. Recent studies have revealed that stressed/apoptotic or heated cells display HSPs on their surface (36, 40) and thus their recognition could be a factor in the internalization of ADCs by pAPCs.
Hitherto, three pathways have been described for MHC class I presentation of exogenous Ags by pAPCs. In addition to the phagosome-cytosol-ER pathway, and the vacuolar TAP-independent pathway, a new ER-phagosome fusion pathway has recently been discovered (7). It was therefore interesting for us to examine which pathway may be involved in LCMV-NP cross-presentation. NP cross-presentation was markedly reduced in LC- or BFA-treated APCs, indicating that the classical phagosome-cytosol-ER pathway is the major route for the presentation of this Ag. Furthermore, we made use of several inhibitors to examine the proteolytic pathways involved in NP396 cross-presentation (Fig. 7). This was an important point in our study, since this long-lived protein is not degraded in ADCs.
Accordingly, we examined what enzymes or organelles during cross-presentation are facilitating the processing before proteasomal degradation. Cross-presentation was affected by lowering the endo/lysosomal pH and cysteine and serine proteases, but not aspartate proteases. It appears that acid-dependent endolysosomal Ag preprocessing is required, and it may very well make LCMV-NP amenable to processing by the proteasome in the cytosol. These inhibition studies were controlled as described in Results, where peptide-pulsed cells (with limiting amounts of the peptide) gave similar responses to the control. In addition, viability tests clearly showed no significant differences to untreated cells. Ultimately, we addressed the role of de novo protein synthesis in ADCs in cross-presentation. Using the tight tet-repressor system where protein synthesis is controlled without causing other general inhibitory effects, we unequivocally proved that neosynthesis is not responsible for the generation of NP396 epitopes during cross-presentation of LCMV-NP.
In this study, we have addressed key questions regarding the processing pathways of virus proteins in APCs as shown in Fig. 7. We have scrutinized the potential for proteins in various cytosolic steps (newly synthesized, mature, proteasome-degraded products, and minimal epitopes) as a source of Ag in cross-presentation. Recently published data have uncovered some very interesting findings about the cross-presentation pathway, but at the same time exposed discrepancies regarding the role of proteasomal intermediates and DRiPs as a source of Ag. Besides, a role for HSP in cross-presentation was either not addressed or indirectly looked at (20, 21, 22, 23, 49), while others have reported that LCMV-NP does not access the cross-priming pathway (19). Of a particular interest is the data reported by Janda et al. (49) examining secreted bacterial proteins with a half-life of 90 min in ADCs (49) and reporting, in contrast to our findings, that unstable bacterial translation products were more efficient than mature proteins as an antigenic source for cross-presentation. In this system, the proteins examined for cross-presentation were synthesized in the bacterium before being secreted in the cytosol of infected cells. Consequently, it is difficult to directly compare their source of the cross-presenting Ag to our system wherein newly synthesized Ags are made by the cell’s own translation machinery. It is also possible that their findings reflect an alternative mechanism when fully folded newly synthesized proteins can directly access the cytosol in bacterially infected cells.
In this study, we highlight the extent of DRiPs’ contributions to the presentation of a stable virus protein via the exogenous Ag-processing pathway. Our study clearly shows that although proteasomal substrates are important in cross-presentation as previously reported (20, 21, 22), neither DRiPs per se nor de novo synthesized proteins can contribute to this pathway. It was the accumulated LCMV-NP protein possibly aided by a specific class of molecular chaperones, which was capable of providing Ags.
Since newly synthesized DRiPs constitute a large fraction of the proteasome’s substrates (50), it makes sense that DRiPs are crucial for the immediate direct NP presentation while cross-presentation awaits the build up of stable proteins that are likely to reflect heavily infected cells. Thus, cross-presentation can make long-lived viral proteins accessible for T cell priming, even if protein neosynthesis is curtailed in infected cells after the onset of the IFN response. Our data provide further insights into significant aspects regarding generation of viral immunity (10, 11, 12, 13, 14, 15) and highlight how the direct and cross-presentation pathways complement each other during viral infection (Fig. 7).
Acknowledgments
We thank U. Beck for excellent technical assistance, K. Tschannen for tetramer production, and Drs. L. Whitton, M. Buchmeier, and K. Rock for providing plasmids, Abs, and cell lines.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by Marie Curie Fellowship Grant MCFI-2002-01548from the European Commission (to S.B.).
2 S.B. and R.S. contributed equally.
3 Address correspondence and reprint requests to his current address: Department of Microbiology and Immunology, Queen’s University, Kingston K7L 3N6, Canada. E-mail address: bastas@post.queensu.ca
4 Abbreviations used in this paper: DRiP, defective ribosomal product; LCMV, lymphocytic choriomeningitis virus; NP, nucleoprotein; pAPC, professional APC; HSP, heat shock protein; ADC, Ag donor cell; rVV, recombinant vaccinia virus; LLnL, N-acetyl-L-leucinyl-L-leuciyl-norleucinal; ICS, intracellular staining; LC, lactacystin; BFA, brefeldin A; ER, endoplasmic reticulum; SR-A, scavenger receptor A; tet, tetracycline; LyUV, lysed and UV treated; Ub, ubiquitin.
Received for publication December 2, 2004. Accepted for publication May 16, 2005.
References
Yewdell, J. W., J. R. Bennink, G. L. Smith, B. Moss. 1985. Influenza A virus nucleoprotein is a major target antigen for cross-reactive anti-influenza A virus cytotoxic T lymphocytes. Proc. Natl. Acad. Sci. USA 82: 1785-1789.
Wills, M. R., A. J. Carmichael, K. Mynard, X. Jin, M. P. Weekes, B. Plachter, J. G. Sissons. 1996. The human cytotoxic T-lymphocyte (CTL) response to cytomegalovirus is dominated by structural protein pp65: frequency, specificity, and T-cell receptor usage of pp65-specific CTL. J. Virol. 70: 7569-7579.
Vanlandschoot, P., T. Cao, G. Leroux-Roels. 2003. The nucleocapsid of the hepatitis B virus: a remarkable immunogenic structure. Antiviral Res. 60: 67-74.
Yewdell, J. W., L. C. Anton, J. R. Bennink. 1996. Defective ribosomal products (DRiPs): a major source of antigenic peptides for MHC class I molecules. J. Immunol. 157: 1823-1826.
Khan, S., R. de Giuli, G. Schmidtke, M. Bruns, M. Buchmeier, M. van den Broek, M. Groettrup. 2001. Cutting edge: neosynthesis is required for the presentation of a T cell epitope from a long-lived viral protein. J. Immunol. 167: 4801-4804.
Munz, C.. 2004. Epstein-Barr virus nuclear antigen 1: from immunologically invisible to a promising T cell target. J. Exp. Med. 199: 1301-1304(Sameh Basta2,3,*, Ricarda)
Many viral proteins that contain MHC class I-restricted peptides are long-lived, and it is elusive how they can give rise to class I epitopes. Recently, we showed that direct presentation of an epitope of the long-lived lymphocytic choriomeningitis virus nucleoprotein (LCMV-NP) required neosynthesis in accordance with the defective ribosomal products hypothesis. In this study, we report that LCMV-NP can be cross-primed in mice using either LCMV-NP-transfected human HEK293 or BALB/c-derived B8 cells as Ag donor cells. In addition, we establish that contrary to direct presentation, cross-presentation required accumulation of the mature LCMV-NP and could not be sustained by the newly synthesized LCMV-NP protein, intermediate proteasomal degradation products, or the minimal NP396 epitope. Nevertheless, NP cross-presentation was enhanced by heat shock and was blunted by inhibitors of heat shock protein 90 and gp96. We propose that cross-presentation has evolved to sustain the presentation of stable viral proteins when their neosynthesis has ceased in infected donor cells.
Introduction
Peptides presented on MHC class I molecules are normally generated through fragmentation of proteins that have been synthesized within the cell. It is therefore remarkable that several structural proteins of viruses are quite stable in cells and nevertheless give rise to sufficient quantities of peptide to induce a CTL response (1, 2, 3). A hypothesis which reconciles this apparent contradiction is the defective ribosomal products (DRiPs)4 hypothesis that states that CTL epitopes are not derived from properly folded proteins but from newly synthesized polypeptides that are degraded within a few minutes after their translation (4). The lymphocytic choriomeningitis virus nucleoprotein (LCMV-NP) is an example of such proteins since it is stable over 3–4 days but nevertheless contains immunodominant epitopes like NP396. Epitopes generated from LCMV-NP shortly after neosynthesis were the sole factor for its direct presentation on MHC class I (5). However, the mechanisms involved in the presentation of DRiPs via the MHC class I pathway are only just surfacing (6, 7, 8).
Professional APCs (pAPCs) can generate class I peptides by two different mechanisms. Either from internally synthesized proteins (9) via "direct-presentation" or by acquiring exogenous Ags through "cross-presentation." Cross-presentation represents a vital pathway for the initiation of immune responses against tumors and viruses that do not have access to, or interfere with, the classical pathway for MHC class I presentation (10, 11, 12, 13, 14, 15). Dendritic cells and macrophages are pAPCs that have the unique capability of generating and presenting peptides on MHC class I molecules from exogenous Ags (16, 17, 18). Activation of naive TCD8+ cells by such pAPCs that have cross-processed Ags into the MHC class I pathway is known as cross-priming (13).
Although the DRiP hypothesis explicates how direct presentation of stable proteins in general can occur, it remains unknown whether stable viral proteins such as LCMV-NP can be cross-presented by pAPCs. This is particularly important since in vivo evidence questioned the ability of LCMV-NP to access the cross-priming pathway (19), yet, it was recently shown that proteins are the mediators of in vivo cross-priming (20, 21, 22). Moreover, immediate proteasomal products were reported to be the main source of Ags transferred to the pAPC in another virus system (23), while others substantiate that Ags could be transferred as peptides bound to heat shock proteins (HSPs) (24). It is possible that virus-specific differences, depending on the nature of virus proteins and their cellular localization, could exist during Ag processing. We therefore started our study by exploring the following avenues; are stable proteins, such as LCMV-NP, predisposed to access the cross-presentation pathway and, if so, what are the antigenic elements responsible for that (20, 21, 22, 23, 24, 25)?
In this study, we report on the cross-presentation of LCMV-NP in vitro and in vivo. We show that accumulated mature proteins probably aided by molecular chaperones are the source of Ags fueling the cross-presentation pathway. Interestingly, neither proteasomal products nor nascent proteins nor DRiPs per se were capable of providing Ags to the cross-presentation-pathway. Our data highlight the significant role of the cross-presentation machinery in degrading stable proteins that are otherwise resistant to processing via the proteasome in virus-infected cells.
Intracellular detection of LCMV-NP
For the detection of LCMV-NP, cells (1 x 105) were harvested and resuspended in 200 μl of PBS, followed by fixation with 4% Formalin for 20 min at room temperature. After washing with 1x PBS, the cells were permeabilized with 1% Triton X-100 for 20 min at room temperature, followed by incubation overnight at 4°C with rat anti-LCMV-NP Ab (clone VL4) (28) in PBS/2% FCS. After two further washing steps, FITC-conjugated polyclonal mouse anti-rat Ab (diluted 1/150 in PBS; BD Biosciences) was left with the cells for at least 1 h at room temperature. Data were acquired with a FACScan flow cytometer (BD Biosciences) using CellQuest (BD Biosciences) and analyzed with FlowJo software (TreeStar).
Intracellular cytokine staining (ICS)
In vitro To detect TCD8+ cell activation with ICS, effector cells were coincubated in round-bottom 96-well plates with (0.1 μM) peptide-loaded APCs (3 x 105/well). Either DMSO or irrelevant peptide (gp33) were used as controls for background activation in all assays. The coincubation ratio of 1:1 APCs (BMC-2) to responders was chosen when T cells were expanded for 5–6 days in vitro. When TCD8+ cell lines were used, using the DC2.4 cell line as stimulator cells, brefeldin A (BFA) was added directly and the cells were left together for 3 h before staining.
In vivo Ex vivo responders were incubated with the specific peptides for 2 h at 37°C and then left in the presence of BFA (Sigma-Aldrich) at 10 μg/ml for 3 h. The coincubation ratio of 1:10 APCs (BMC-2) to responders was chosen in the case of direct analysis ex vivo. In general, cells were then stained with PE-Cy5 anti-CD8 rat IgG Ab clone 53-6.7 (BD Biosciences) on ice for 20 min, washed, and fixed with 1% paraformaldehyde in PBS at room temperature for 20 min. After washing, we labeled the cells with FITC-conjugated rat anti-IFN- Ab clone XGM1 (BD Pharmingen) in PBS/0.1% saponin at 4°C overnight. Stained cells were analyzed with a live gate on the CD8+ cells after 80,000 gated cells were acquired.
Tetrameric MHC class I peptide analysis
Tetrameric complexes containing biotinylated H-2Db 2-microglobulin, the NP396 peptide, and extravidin-PE were generated as described previously (26, 29). Splenocytes (5 x 105 cells) were stained with 0.5–1 μg of tetramer in 50 μl of FACS buffer (PBS, 2% FCS, 0.01% NaN3, and 20 mM EDTA) at room temperature for 30 min, followed by addition of PE-Cy5 anti-CD8 rat IgG Ab clone 53-6.7 (BD Biosciences) on ice for 30 min. Cells were washed twice and data were acquired using CellQuest (BD Biosciences) as before by gating on CD8+ lymphocytes.
TCD8+ cell priming and induction of T cell lines
For in vivo priming, 8- to 10-wk-old female C57BL/6J mice were injected i.p. with 1 x 107 cells in PBS. In the case of primary responses, ICS measurements were performed directly 8 days after injection. In some experiments, spleens were harvested and restimulated with peptide-loaded APCs. On day 5 or 6 of the in vitro expansion, T cells were tested for IFN- production by ICS as before.
For generating NP396-specific T cell lines, C57BL/6 mice were injected i.v. with 200 μl of LCMV-WE (1000 PFU/ml). The spleens were harvested 4 wk after injection and the splenocytes were purified by density centrifugation (Ficoll-Paque). The splenocytes were cultured in RPMI 1640 containing mouse IL-2 (40 U/ml) along with gamma-irradiated peptide-loaded APCs. An additional density centrifugation step was conducted 5–6 days later when they were used for in vitro Ag presentation experiments. In all assays, we used the T cell lines at this time point and their specificity corresponded to at least 85% when tested with peptide (50 nM). For the detection of the epitope NP396, we resuspended the APCs in RPMI 1640 (1 x 106/ml), added them to NP396-specific TCD8+ cells, and stained for CD8 and IFN- as described above.
Preparation of Ag donor or presenting cells (ADCs and APCs)
ADCs were detached with trypsin-EDTA, washed with PBS, and resuspended in RPMI 1640 (1 x 106 cells/ml) and treated with different inhibitors where indicated. For interfering with HSPs, the HSP90-specific inhibitors geldanamycin (5 μM) and herbimycin A (5 μM) were used (Sigma-Aldrich). Cycloheximide (250 μg/ml) was obtained from Calbiochem. For testing different conditions, the cells were either further heated (14 min at 45°C), or UV irradiated (8 min), or snap frozen in liquid nitrogen (–180°C) and UV irradiated for 8 min (lysed and UV treated (LyUV)). UV-triggered apoptosis was induced using the Stratalinker 2400 UV cross-linker that provides a radiation intensity of 4 mW/cm2 per s. Interestingly, one round of lysis resulted in intact cells incorporating trypan blue, whereas with LyUV cells only 40% incorporated trypan blue. If the cells were subjected to repeated cycles of freezing and thawing, the outcome was mainly cell fragments.
As APCs, we used the macrophage cells line BMC-2 that was recently shown to posses competent phagosomes as organelles for Ag cross-presentation (30). Where indicated, the following inhibitors were used: BFA (10 μg/ml), chloroquine (100 μM), leupeptin (100 μM), pepstatin A (50 μg/ml), fucoidin (250 μg/ml; Sigma-Aldrich), lactacystin (LC, 50 μM; Biomol). During the cross-presentation assays, coincubation of ADCs with BMCs was conducted in round-bottom 96-well plates (greiner bio-one; Cellstar) at 37°C at a ratio of 1:1 for 16 h.
Results
Expression of LCMV-NP in transfected cells
To study LCMV-NP cross-presentation, we used human HEK293 cells as ADCs since they cannot directly present the H-2Db-restricted NP396 epitope to specific T cell lines that we used to monitor cross-presentation. We generated stable transfectants expressing either the full-length LCMV-NP (designated HEK-NP) or a ubiquitin-LCMV-NP fusion protein (HEK-Ub-NP). The N-terminal ubiquitin moiety is a G76A mutant that cannot be cleaved off by ubiquitin-specific proteases and targets the fusion protein for rapid degradation by the proteasome (31). Western blot analyses of LCMV-NP in either LCMV-infected HEK293 cells (HEKi) or HEK-NP cells demonstrated that LCMV-NP was abundantly expressed (Fig. 1A). LCMV-NP expression was very low in untreated HEK-Ub-NP cells but proteasome inhibitors markedly enhanced the expression level, thus confirming rapid proteasomal degradation of the fusion protein (Fig. 1A, HEK-Ub-NP + LLnL) as previously demonstrated in transiently transfected cells (31). The relative NP expression levels as assessed by Western blot analysis were further confirmed by pulse-chase analysis as well as intracellular NP staining and flow cytometry (data not shown and Fig. 1B).
To confirm these findings independently and to assess the role of peptide epitopes for cross-presentation, we infected either HEK293 cells or J774 cells (H-2d) with rVV encoding either the full-length LCMV-NP protein or the NP396 minigene. In this situation, donor cells were subjected to UV irradiation to prevent secondary reinfection of the APCs. As shown in Fig. 3B, only ADCs that were expressing the full-length protein served as an efficient source of Ag for the activation of NP396-specific TCD8+ cells. In marked contrast, the rVV encoding the NP396 minigene failed to induce activation via cross-presentation, suggesting that the minimal peptide Ags cannot be efficiently cross-presented.
Heat shock or lysis and UV irradiation increases cross-presentation of LCMV-NP but not of LCMV-Ub-NP
To test different factors that may affect cross-presentation, we subjected HEK-NP and HEK-Ub-NP cells to either heat shock or to lysis by freezing/thawing followed by UV irradiation (LyUV). The heat shock treatment (14 min at 45°C) did not significantly affect the viability of HEK cells as determined by propidium iodide exclusion staining (11 ± 4% dead cells in both control or heated cells) and trypan blue (data not shown). As shown in Fig. 3C, both protocols enhanced cross-presentation of the ADCs with LyUV showing a superior outcome. Remarkably, none of the two ADC treatment protocols were able to enhance cross-presentation of HEK-Ub-NP. However, cross-presentation of HEK-Ub-NP could be recovered when the ADCs were treated with LC (Fig. 3D). Incubation with LC for 8 h did not significantly affect cell viability (propidium iodide) exclusion staining (14 ± 6% dead cells for LC-treated control or heated cells). This suggests that HSPs can enhance cross-presentation that is biased toward proteins but not peptides as an Ag source in this pathway.
The HSP90 family enhances cross-presentation of LCMV-NP
Because initial experiments indicated that treatment of the ADCs leading to increased expression of HSPs enhanced cross-presentation, we examined the outcome of interfering with a prominent class of the HSP families. HEK-NP cells were treated with the very well-characterized HSP90 inhibitors geldanamycin and herbimycina before heat shock or LyUV treatment. As shown before, when ADCs were either heat shocked or LyUV treated, there was an increase in the signals (Fig. 3C). When we used the inhibitors that are specific for both HSP90 and gp96, the signals were almost at background levels (Fig. 4A, heated or LyUV). The effect of these inhibitors on intact cells was not apparent, as no significant differences were detected (Fig. 4A, intact). The possibility that the effect of HSP90 inhibitors was due to a direct inhibition of the responder T cell lines was ruled out as peptide-loaded APCs incubated with either untreated or inhibitor-treated ADCs induced similar activation of the T cell lines (activation was almost 85%, data not shown). In addition, we ruled out possible negative effects of these inhibitors on the NP levels in ADCs by intracellular NP staining and flow cytometry (Fig. 4B).
A fundamental issue regarding the presentation of LCMV-NP through the cross-presentation pathway centers on whether DRiPs per se are a source of Ag for cross-presentation. To address this point, we tried two different approaches: the first one was based on inhibition of de novo protein synthesis before injecting the ADCs in vivo. As shown in Fig. 6A, we could not detect any significant differences if ADCs were treated with cycloheximide, suggesting that the mature folded protein is responsible for the cross-presentation activity. In vitro, when de novo protein synthesis in ADCs was inhibited by cycloheximide, an inhibitory effect on cross-presentation in vitro was observed (data not shown). Unfortunately, using this approach is inadequate because translation inhibitors can affect a variety of factors in the cell.
To resolve these issues, we used the tight tetracycline (tet)-repressor system where systemic effects of inhibiting protein neosynthesis are ruled out. This system relies on the tet-repressible expression of LCMV-NP in the fibroblast transfectant B8tNP64. In these cells, NP neosynthesis can be switched on and off without using inhibitors, which has been instrumental in identifying LCMV-NP as a DRiP substrate (5). First, we checked whether LCMV-NP induced for 24–72 h in B8tNP64 cells was cross-presented in vivo. This occurred to a similar degree as we have observed for HEK-NP cells (Fig. 2 and data not shown). We have thus chosen the 24-h time point for further in vitro analyses. To assess the control of LCMV-NP expression and the specificity of cross-presentation in vitro, we used either B8 recipient cells (B8-wt) or B8tNP64 cells (NP64) that were tet suppressed and confirmed that cross-presentation was undetectable when LCMV-NP was not expressed (Fig. 6B). To examine the effect of neosynthesis, we removed tet for 5 h (Fig. 6B, –5 h) so that we could analyze the role of DRiPs as an Ag source, but we failed to detect NP396 presentation after exogenous processing, unlike direct presentation (data not shown). We could recover the activity if tet had been removed for a period of 24 h, thus allowing LCMV-NP accumulation (Fig. 6B, fourth column, tet–), strongly indicating that the accumulated mature protein rather than DRiPs or nascent proteins is responsible for cross-presentation. The LCMV-NP expression levels detected by FACS and immunoblot analyses in either B8tNP64 or HEK-NP cells support this proposition (Fig. 6C and data not shown). The measurements obtained by flow cytometry reflect the nature of the fully folded protein since this Ab does not detect native forms of NP in Western blots. Importantly, as indicated before with the HEK-NP cells, when we awaited NP expression to build up for 24 h and then applied the chaperone inhibitors, no significant negative effects on NP expressions were found (Fig. 6C, histogram overlays).
The ability to cross-present LCMV-NP was significantly reduced when chaperones of the HSP90 family were inhibited in B8tNP64 cells with geldanamycin and herbimycina that were added to cells after tet removal for 24 h (Fig. 6B, columns 5 and 6). To exclude that HSP90 inhibitors negatively affected NP396 presentation by APCs, we used peptide-pulsed APCs as a control (Fig. 6B, columns 8–10). To further scrutinize the requirement for neosynthesis in cross-presentation, we allowed protein synthesis to proceed for 24 h and then added tet for 5 h to block any further neosynthesis (Fig. 6B, ±5 h). If DRiPs were playing a role, then the signal would be drastically reduced. This experiment revealed that, when only the mature LCMV-NP protein was present but no DRiPs, cross-presentation was as efficient as if LCMV-NP synthesis was permitted over 24 h. Taken together, it is evident from our results that NP neosynthesis is not required as an Ag source in LCMV-NP cross-presentation.
Discussion
Better understanding of factors contributing to the cross-presentation pathway is an important issue at present in the field of Ag presentation and immune surveillance. Recently, a novel phagosomal pathway involved in cross-presentation, which merits further assessment, has been discovered (30, 32, 33). Furthermore, the involvement of DRiPs as a source of Ags is a new paradigm in direct presentation (4). These new concepts have prompted us to examine whether cross-presentation of the long-lived LCMV-NP (5) is attainable in APCs after its uptake, since it is not readily degraded in ADCs. Consequently, we set our goals to scrutinize the parameters involved in its degradation in APCs and investigated the relative contribution of proteins vs DRiPs or peptides as Ag sources and the role of HSPs therein.
We tested the ability of this long-lived protein and its degradable products to cross-present in vitro and in vivo to investigate whether it conforms with the recent findings regarding the role of proteins in cross-priming (20, 21, 22) or its stability does not allow it to access the cross-presentation pathway (19). Although the in vitro assays proved to be a more sensitive readout system compared with in vivo, the overall conclusion reached was similar in both systems. Accordingly, and contrary to previously reported data (19), we found that NP transfectants (xeno- or alloresponses) can prime TCD8+ cells in vivo via the exogenous pathway. Perhaps, our system was more sensitive to detect such activity because our transfectants expressed sufficient amounts of NP. Cross-priming and presentation of this long-lived protein was considerably reduced when it was targeted for rapid degradation by the proteasome. This conclusion was confirmed by different means; T cell frequencies were estimated using tetramer analysis and their activation was examined directly ex vivo or after expansion in vitro. Considering all of the data combined, it is clear that less efficient T cell activation, at least four times less, occurred when the NP was targeted for degradation. We hypothesized at this stage that the long-lived form of NP is a better source for cross-priming than the unstable ubiquitin (Ub)-NP fusion protein. This finding concurs with and further extends the concept that proteasomal substrates in general rather than proteasomal products are the main source of cross-priming in vivo (20, 21, 22).
Although the Ub-NP transfectants generate more peptides or polypeptides than full-length protein in the donor cells, they nevertheless were poor inducers of TCD8+ cells via cross-priming. We confirmed these findings with rVV encoding either the full-length protein or the minigene, where the full-length protein was superior in its ability to cross-present compared with the cytosolic peptide. These conclusions are supported by earlier reports showing that cytosolic peptides, unless bound to chromatin, are unstable with a very short half-life due to cytosolic proteases (34). Furthermore, cross-priming of T cell responses in vivo does not require antigenic peptides in the ER (35). According to our results, degraded proteins are unlikely to play a major role in cross-priming of T cell responses in agreement with recently published data (20, 21, 22). Yet, HSPs bound to antigenic peptides were suggested to be a potent source of Ag for cross-presentation (24, 36). In contrast, the circumstances under which chaperone-peptide interactions play a specific role in immunity has been put in question recently and need to be thoroughly investigated (36, 37). Our findings reconcile the proposed role of HSPs in cross-presentation on the one hand and the inability of peptides to cross-prime on the other hand. Based on our results, the ability of HSPs to enhance cross-presentation is evidently biased toward proteins interactions and not proteasomal products.
To examine conditions where possible danger signals may affect the activation of pAPCs (36, 37, 38), we compared intact HEK-NP cells with other stress settings where induction or release of chaperones is envisaged. Either LyUV or heat shock protocols led to a marked enhancement of cross-presentation. This was not the case if we just UV treated intact ADCs (data not shown), indicating that the differential cell treatment can influence immunogenicity (39, 40). In our system, the use of HSP90 and gp96 inhibitors impeded cross-presentation markedly, thus indicating the involvement of one or both of these chaperones after the induction of stress. A very large body of strong biochemical and pharmacological evidence has established geldanamycin and herbimycin A as specific inhibitors of HSP90 family chaperones (41, 42). The chaperone function of HSP90 relies on ATP hydrolysis (43), which is blocked by geldanamycin and herbimycin A because they bind to the ATP binding site of HSP90 according to x-ray crystallographic analysis (44). In addition, a very recent finding reported on reduced cross-priming in mice that were defective in heat shock factor 1 (Hsf1). These mice have a decreased expression of several HSPs including HSP90 and HSP70 (45).
There are several reasons, currently under investigation by us, why these chaperones could have led to increased presentation. It is possible that the effect is indirect due to stimulation of APCs via cell surface signaling receptors like TLRs or increased uptake of ADCs via HSP receptors (36, 37, 38). The chaperones may be required to maintain the proper folding or function of other unknown interacting factors that are generally needed for the cross-presentation pathway. Alternatively, if HSPs are tightly bound to NP, especially under stress conditions, then one could envisage a direct role of chaperones as Ag carrier molecules. Therefore, a possible model could be based on protein-bound HSPs rather than peptide-bound HSPs in enhancing cross-presentation. We also examined certain aspects of the uptake mechanism and the downstream events that follow. We looked into the role of SR-A since it has an array of ligands (46), including certain molecular chaperones (47). A strong reduction of the cross-presentation signal was observed when fucoidin was incubated with APCs. Since fucoidin, a polysaccharide that competitively binds to the ligand-binding domain of SR-A (48), interfered with the presentation signal, it is likely that SR-A is playing an important role in Ag uptake. However, in this instance we could not single out whether the fucoidin was competing out stress chaperones (47) or other undefined SR-A ligands. Recent studies have revealed that stressed/apoptotic or heated cells display HSPs on their surface (36, 40) and thus their recognition could be a factor in the internalization of ADCs by pAPCs.
Hitherto, three pathways have been described for MHC class I presentation of exogenous Ags by pAPCs. In addition to the phagosome-cytosol-ER pathway, and the vacuolar TAP-independent pathway, a new ER-phagosome fusion pathway has recently been discovered (7). It was therefore interesting for us to examine which pathway may be involved in LCMV-NP cross-presentation. NP cross-presentation was markedly reduced in LC- or BFA-treated APCs, indicating that the classical phagosome-cytosol-ER pathway is the major route for the presentation of this Ag. Furthermore, we made use of several inhibitors to examine the proteolytic pathways involved in NP396 cross-presentation (Fig. 7). This was an important point in our study, since this long-lived protein is not degraded in ADCs.
Accordingly, we examined what enzymes or organelles during cross-presentation are facilitating the processing before proteasomal degradation. Cross-presentation was affected by lowering the endo/lysosomal pH and cysteine and serine proteases, but not aspartate proteases. It appears that acid-dependent endolysosomal Ag preprocessing is required, and it may very well make LCMV-NP amenable to processing by the proteasome in the cytosol. These inhibition studies were controlled as described in Results, where peptide-pulsed cells (with limiting amounts of the peptide) gave similar responses to the control. In addition, viability tests clearly showed no significant differences to untreated cells. Ultimately, we addressed the role of de novo protein synthesis in ADCs in cross-presentation. Using the tight tet-repressor system where protein synthesis is controlled without causing other general inhibitory effects, we unequivocally proved that neosynthesis is not responsible for the generation of NP396 epitopes during cross-presentation of LCMV-NP.
In this study, we have addressed key questions regarding the processing pathways of virus proteins in APCs as shown in Fig. 7. We have scrutinized the potential for proteins in various cytosolic steps (newly synthesized, mature, proteasome-degraded products, and minimal epitopes) as a source of Ag in cross-presentation. Recently published data have uncovered some very interesting findings about the cross-presentation pathway, but at the same time exposed discrepancies regarding the role of proteasomal intermediates and DRiPs as a source of Ag. Besides, a role for HSP in cross-presentation was either not addressed or indirectly looked at (20, 21, 22, 23, 49), while others have reported that LCMV-NP does not access the cross-priming pathway (19). Of a particular interest is the data reported by Janda et al. (49) examining secreted bacterial proteins with a half-life of 90 min in ADCs (49) and reporting, in contrast to our findings, that unstable bacterial translation products were more efficient than mature proteins as an antigenic source for cross-presentation. In this system, the proteins examined for cross-presentation were synthesized in the bacterium before being secreted in the cytosol of infected cells. Consequently, it is difficult to directly compare their source of the cross-presenting Ag to our system wherein newly synthesized Ags are made by the cell’s own translation machinery. It is also possible that their findings reflect an alternative mechanism when fully folded newly synthesized proteins can directly access the cytosol in bacterially infected cells.
In this study, we highlight the extent of DRiPs’ contributions to the presentation of a stable virus protein via the exogenous Ag-processing pathway. Our study clearly shows that although proteasomal substrates are important in cross-presentation as previously reported (20, 21, 22), neither DRiPs per se nor de novo synthesized proteins can contribute to this pathway. It was the accumulated LCMV-NP protein possibly aided by a specific class of molecular chaperones, which was capable of providing Ags.
Since newly synthesized DRiPs constitute a large fraction of the proteasome’s substrates (50), it makes sense that DRiPs are crucial for the immediate direct NP presentation while cross-presentation awaits the build up of stable proteins that are likely to reflect heavily infected cells. Thus, cross-presentation can make long-lived viral proteins accessible for T cell priming, even if protein neosynthesis is curtailed in infected cells after the onset of the IFN response. Our data provide further insights into significant aspects regarding generation of viral immunity (10, 11, 12, 13, 14, 15) and highlight how the direct and cross-presentation pathways complement each other during viral infection (Fig. 7).
Acknowledgments
We thank U. Beck for excellent technical assistance, K. Tschannen for tetramer production, and Drs. L. Whitton, M. Buchmeier, and K. Rock for providing plasmids, Abs, and cell lines.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by Marie Curie Fellowship Grant MCFI-2002-01548from the European Commission (to S.B.).
2 S.B. and R.S. contributed equally.
3 Address correspondence and reprint requests to his current address: Department of Microbiology and Immunology, Queen’s University, Kingston K7L 3N6, Canada. E-mail address: bastas@post.queensu.ca
4 Abbreviations used in this paper: DRiP, defective ribosomal product; LCMV, lymphocytic choriomeningitis virus; NP, nucleoprotein; pAPC, professional APC; HSP, heat shock protein; ADC, Ag donor cell; rVV, recombinant vaccinia virus; LLnL, N-acetyl-L-leucinyl-L-leuciyl-norleucinal; ICS, intracellular staining; LC, lactacystin; BFA, brefeldin A; ER, endoplasmic reticulum; SR-A, scavenger receptor A; tet, tetracycline; LyUV, lysed and UV treated; Ub, ubiquitin.
Received for publication December 2, 2004. Accepted for publication May 16, 2005.
References
Yewdell, J. W., J. R. Bennink, G. L. Smith, B. Moss. 1985. Influenza A virus nucleoprotein is a major target antigen for cross-reactive anti-influenza A virus cytotoxic T lymphocytes. Proc. Natl. Acad. Sci. USA 82: 1785-1789.
Wills, M. R., A. J. Carmichael, K. Mynard, X. Jin, M. P. Weekes, B. Plachter, J. G. Sissons. 1996. The human cytotoxic T-lymphocyte (CTL) response to cytomegalovirus is dominated by structural protein pp65: frequency, specificity, and T-cell receptor usage of pp65-specific CTL. J. Virol. 70: 7569-7579.
Vanlandschoot, P., T. Cao, G. Leroux-Roels. 2003. The nucleocapsid of the hepatitis B virus: a remarkable immunogenic structure. Antiviral Res. 60: 67-74.
Yewdell, J. W., L. C. Anton, J. R. Bennink. 1996. Defective ribosomal products (DRiPs): a major source of antigenic peptides for MHC class I molecules. J. Immunol. 157: 1823-1826.
Khan, S., R. de Giuli, G. Schmidtke, M. Bruns, M. Buchmeier, M. van den Broek, M. Groettrup. 2001. Cutting edge: neosynthesis is required for the presentation of a T cell epitope from a long-lived viral protein. J. Immunol. 167: 4801-4804.
Munz, C.. 2004. Epstein-Barr virus nuclear antigen 1: from immunologically invisible to a promising T cell target. J. Exp. Med. 199: 1301-1304(Sameh Basta2,3,*, Ricarda)