Abstract
Chronic Ag exposure during persistent viral infection erodes virus-specific CD8 T cell numbers and effector function, with a concomitant loss of pathogen control. Less clear are the respective contributions of Ag-specific and Ag-nonspecific (bystander) events on the quantity, quality, and maintenance of antiviral CD8 T cells responding to persistent virus infection. In this study, we show that low-dose inoculation with mouse polyomavirus (PyV) elicits a delayed, but numerically equivalent, antiviral CD8 T cell response compared with high-dose inoculation. Low-dose infection generated virus-specific CD8 T cells endowed with multicytokine functionality and a superior per cell capacity to produce IFN-γ. PyV-specific CD8 T cells primed by low-dose inoculation also expressed higher levels of IL-7Rα and bcl-2 and possessed enhanced Ag-independent survival. Importantly, the quantity and quality of the antiviral CD8 T cell response elicited by dendritic cell-mediated immunization were mitigated by infection with a mutant PyV lacking the dominant CD8 T cell viral epitope. These findings suggest that the fitness of the CD8 T cell response to persistent virus infection is programmed in large part by early virus-associated Ag-nonspecific factors, and imply that limiting bystander inflammation at the time of inoculation, independent of Ag load, may optimize adaptive immunity to persistent viral infection.
Elucidation of the variables that influence the generation and maintenance of functional pathogen-specific T cells is essential for the development of effective vaccines (1). A substantial body of literature supports the concept that early events after Ag stimulation have long-lasting consequences on T cell responses. A single brief encounter with Ag is sufficient to program naive T cells to proliferate, differentiate into antipathogen effectors, and give rise to a reservoir of memory cells poised to rapidly contain repeat infection (2, 3, 4). T cell differentiation programs, however, can be modified. For example, naive T cell priming is fine-tuned by TCR signal strength and duration, such that extended stimulation is required for optimal in vivo proliferation and acquisition of effector capabilities (5, 6, 7, 8). A number of Ag-nonspecific factors can also influence the quantity and quality of responding T cells. These include cytokine and costimulatory receptor engagement, availability of T cell help, and the particular subset of Ag-presenting dendritic cell (DC)3 that engages the naive T cell (9, 10, 11, 12, 13, 14, 15). Thus, a complex array of variables must be considered in vaccine development to elicit optimal adaptive immune responses. Of these, initial inoculation dose has received little attention as an independent variable on the magnitude, function, and maintenance of the antiviral CD8 T cell response.
Whether Ag is entirely eliminated after resolution of acute infection or persists has dramatic effects on the quality of the memory CD8 T cell response. Acute memory CD8 T cells, which emerge after complete viral clearance, are maintained long-term, independent of Ag and MHC class I molecules, by IL-7- and IL-15-mediated survival and proliferation signals, respectively (16, 17, 18, 19). In contrast, chronic memory antiviral CD8 T cells (i.e., those present during the persistent phase of infection) are defective in homeostatic proliferation, express low levels of IL-7Rs and IL-15Rs, and have poor recall potential (20). In sharp contrast to acute memory CD8 T cells, maintenance of chronic memory CD8 T cells is Ag dependent (20, 21). It is well documented that persistent viral infections are associated with CD8 T cell functional deterioration and deletion, a phenomenon that negates viral clearance by the host (22, 23, 24, 25). Several recent reports have shown that antiviral therapy or blockade of the PD-1 inhibitory receptor or IL-10R during persistent infection restores T cell function and results in viral clearance (26, 27, 28, 29, 30). T cell responses to chronic viral infections have largely been studied using viruses that establish high-level persistent infection.
Polyomavirus (PyV), however, establishes a low-level systemic persistent infection in the mouse, its natural host (31). Like its murine counterpart, nearly all humans are infected lifelong by the JC and BK PyVs, which are asymptomatic in immunocompetent hosts; but, in the setting of immune suppression, JC and BK viruses can cause a fatal CNS demyelinating disease or kidney allograft dysfunction and failure, respectively (32, 33). In the present study, we used the PyV-mouse model to investigate whether inoculation dose impacts the antiviral CD8 T cell response. Inoculation dose was found to affect the onset of expansion, but, unexpectedly, not the peak magnitude of the virus-specific CD8 T cell response, and was inversely related to level of functional competence. Superior effector function was evident as early as the acute infection phase following low-dose inoculation and did not erode during persistent PyV infection. Using peptide-coated DC immunization and infection with a mutant PyV lacking the dominant CD8 T cell viral epitope, we found that Ag-nonspecific events associated with infection mitigate the magnitude and function of the PyV-specific CD8 T cell response. These results indicate that early virus-induced bystander factors shape the quality of the antiviral CD8 T cell response to persistent infection.
Materials and Methods
Mice
C57BL/6 (B6) female mice were purchased from the National Cancer Institute. C57BL/6-gld (Gld) mice (The Jackson Laboratory) and B6.SJL (Taconic Farms) were bred at the Emory University mouse facility. All animal protocols were conducted according to guidelines established by the Institutional Animal Care and Use Committee and the Department of Animal Resources. All mice were between 6 and 12 wk of age at the time of infection.
Viruses
Stocks of PyV A2 strain were prepared in baby mouse kidney cells, as previously described (34). Mice were injected in both hind footpads with 100 μl of virus suspension. Virus stocks were diluted in PBS for low-dose infections. The large T Ag (LT)359–368 epitope (SAVKNYCSKL) (31)-null PyV (nPyV) was generated using PCR-based site-directed mutagenesis (Stratagene) to change the codon for aa 363 in LT from arginine to threonine to remove the primary Db-peptide anchoring residue (35). An analog LT359.N363T peptide did not stabilize surface Db by TAP-deficient RMA/S cells (data not shown). Recombinant vesicular stomatitis virus (VSV) (rVSV-LT359) was generated by inserting primers containing the PyV LT359–368 sequence into a pVSV.XN2 plasmid containing the entire VSV genome (36). DNA sequence analysis confirmed the presence of the insert.
Synthetic peptides
The LT359–368Abu (SAVKNY[Abu]SKL) peptide, in which the cysteine residue at position 7 was replaced with α-aminobutyric acid, a thiolless cysteine analog residue, was synthesized by the solid-phase method using F-moc chemistries (Emory University Microchemical Core Facility) and HPLC purified to >90% purity. For simplicity, the LT359–368Abu peptide is referred to as LT359 peptide.
TaqMan real-time PCR to quantitate PyV DNA
DNA isolation and Taqman PCR were performed, as described (31). PyV DNA quantity is expressed in genome copies/mg tissue and is calculated based on a standard curve of known PyV genome copy number vs threshold cycle of detection. The detection limit with this assay is 10 copies of genomic viral DNA.
Tissue processing
Lymph nodes were ground between glass slides, and spleens were processed by mechanical disruption on wire screens, followed by RBC lysis with Tris-buffered ammonium chloride (Sigma-Aldrich).
Flow cytometry
One million cells per sample were stained in PBS containing 2% FBS and 0.1% sodium azide. Ab staining was performed at 4°C for 30 min; allophycocyanin-conjugated Db LT359 tetramer (31) staining was performed at room temperature for 30 min. Cells were surface stained with a FITC-conjugated mAb specific for CD8α (BD Biosciences) and PE-conjugated mAbs specific for CD25, CD44, CD62L, CD69, CD127, PD-1, and rat IgG isotype controls. All PE-conjugated Abs were obtained from Caltag Laboratories except for anti-CD127, anti-PD-1, and their isotype controls (eBioscience). TCR Vβ domain usage was determined using a FITC-conjugated panel of mAbs (BD Pharmingen). Fas ligand (FasL) expression on DCs was determined by staining with FITC-conjugated anti-CD11c and biotin-conjugated anti-FasL, followed by streptavidin-PerCP. Samples were acquired on a FACSCalibur (BD Biosciences), and data were analyzed using FlowJo software (Tree Star).
Intracellular staining
Cells were stimulated directly ex vivo with the indicated concentration of LT359 peptide, then stained for surface CD8α and intracellular cytokines, as described elsewhere (31). For bcl-2, spleen cells from persistently infected mice were surface stained directly ex vivo with anti-CD8α and Db LT359 tetramer, then permeabilized using the Cytofix/Cytoperm kit (BD Pharmingen), and then stained with anti-bcl-2 PE or the appropriate isotype control (BD Pharmingen).
rVSV-LT359 challenge infection
Mice persistently infected with PyV received 1 × 106 PFU of rVSV-LT359 via the tail vein. Four days later, lymphocytes from freshly explanted spleens were analyzed for LT359-specific CD8 T cells by flow cytometry.
In vivo cytotoxicity assays
B6 mice infected with high- or low-dose PyV 6 wk previously were tested for in vivo killing of LT359–368Abu peptide-pulsed naive B6 spleens cells over a 4-h period, as previously described (37).
Adoptive transfer
B6 and B6.SJL mice received high-dose (2 × 106 PFU) or low-dose (5 × 103 PFU) inoculations of wild-type PyV or nPyV, respectively. Six weeks after infection, B6 spleen cells were incubated for 1 h at 37°C in flasks coated with goat anti-mouse IgG (H + L) (Jackson ImmunoResearch Laboratories). Nonadherent cells were labeled with 4.0 μM CFSE (Invitrogen Life Technologies) for 10 min at 37°C. A total of 40 × 106 cells was injected i.v. into infection-matched B6.SJL hosts (e.g., high-dose PyV donors into high-dose nPyV recipients). CD45.1−CD8+ Db LT359 tetramer+ cells in the blood were tracked over time for CFSE fluorescence intensity and as a frequency of total CD8+ cells.
DC immunization
B6 bone marrow was cultured in IMDM (Invitrogen Life Technologies) containing 5% FBS, 100 ng/ml GM-CSF (PeproTech), and 1.0 μg/ml gentamicin sulfate (Mediatech). DCs were matured on the sixth day with 100 ng/ml LPS (Sigma-Aldrich). Cultures were harvested on day 7, CD11c expression was determined by flow cytometry, and cells were incubated at 37°C for 1.5 h with LT359 peptide. B6 mice received i.v. injections of 2.5 × 105 CD11c+ cells with or without s.c. injections of high-dose (2 × 106 PFU) or low-dose (5 × 103 PFU) nPyV on day 0 or 2 of DC injection. At the indicated days after immunization, spleen cells were isolated and LT359-specific CD8 T cells were analyzed by flow cytometry.
Statistics
Statistical significance was determined by unpaired Student’s t test, assuming unequal variances. A p value of ≤0.05 was considered statistically significant.
Results
PyV inoculation dose alters the kinetics, but not magnitude of CD8 T cell expansion
The magnitude of CD8 T cell expansion in response to many acutely resolved infections positively correlates with the initial pathogen dose (38, 39, 40). Reduced pathogen-derived epitope presentation resulting from lower dose inoculation is generally thought to limit recruitment of Ag-specific CD8 T cells. Using PyV as a model persistent viral infection, we re-examined the association between initial inoculum dose and the magnitude of the antiviral CD8 T cell response during acute and persistent phases of infection. B6 mice received 2 × 106 (high-dose) or 5 × 103 (low-dose) PFU of PyV. The peak frequencies and numbers of CD8 T cells specific for the dominant Db-restricted viral epitope, LT359–368, were similar after high- or low-dose infection (Fig. 1,A). However, low-dose PyV infection shifted the peak of CD8 T cell expansion to day 10, a delay of 2 days compared with high-dose infection. Even a 200,000-fold reduction in PyV dose (10 PFU) recruited comparable numbers of LT359-specific CD8 T cells, but resulted in peak expansion at day 14 postinfection (p.i.) (data not shown). Additionally, the initial dose of PyV did not alter CD8 T cell responses to the two defined subdominant viral epitopes (31) in terms of hierarchy and numbers at the peaks of acute infection (data not shown). Of note, the rate of LT359-specific CD8 T cell expansion was similar irrespective of initial inoculation dose. The 2-day delay in maximal Db LT359 tetramer+ CD8 T cell expansion between low- and high-dose inoculations parallels the delay in the former dose reaching peak viral DNA load (Fig. 1 B). Moreover, because splenic PyV DNA levels differed at least 10-fold between these initial inoculum doses throughout infection, viral load is dissociated from the magnitude of PyV-specific CD8 T cell response. The delayed, but kinetically equivalent expansion and peak magnitude of LT359-specific CD8 T cells with lower dose inoculation suggests that there is a threshold Ag density that, once reached, triggers maximal recruitment of anti-PyV CD8 T cells.
Legge and Braciale (41) recently demonstrated that high-dose influenza virus infection induced FasL up-regulation on lymph node-resident dendritic cells (DCs) that, in turn, triggered apoptosis of virus-specific CD8 T cells. We asked whether high-dose PyV infection similarly interferes with anti-PyV CD8 T cell expansion. Confirming this observation, we found that high-dose, but not low-dose, inoculation strongly induced FasL expression on draining lymph node CD11c+ cells by 2 days p.i. (Fig. 2,A, upper panel). Low-dose PyV inoculation also appeared to be less inflammatory, as this dose resulted in only weak cell surface up-regulation of CD80/CD86 on lymph node DCs (Fig. 2,A, lower panel), and draining lymph nodes contained 6-fold (day 2 p.i.) and 2-fold (day 4 p.i.) fewer mononuclear cells than after high-dose inoculation (data not shown). As shown in Fig. 2 B, wild-type and FasL-deficient Gld mice mounted equivalent LT359-specific CD8 T cell responses irrespective of initial virus inoculation dose. Thus, Fas-FasL-induced apoptosis does not appear to control the magnitude of PyV-specific CD8 T cell expansion.
Phenotypic analysis of CD8 T cells after high- or low-dose PyV inoculation
We next examined the effect of inoculation dose on the expression of surface molecules associated with T cell activation and memory differentiation. Throughout the course of PyV infection with either low- or high-dose inoculation, LT359-specific CD8 T cells retained the CD62Llow expression profile characteristic of CD8 T cells that continually encounter Ag (data not shown) (42, 43, 44). Because of the prominent role of IL-7Rs in memory CD8 T cell differentiation and survival (16, 45), we monitored IL-7R α-chain (CD127) expression on LT359-specific CD8 T cells under these different inoculation conditions. During the initial stages of PyV infection, CD127 levels on splenic LT359-specific CD8 T cells were similar regardless of the viral dose (Fig. 3,A). As the acute phase of infection resolved, however, CD127 expression patterns began to diverge with a larger fraction of the low-dose CD8 T cells becoming CD127+. We also assessed the effect of PyV inoculation dose on expression of the inhibitory receptor PD-1. At the peaks of the LT359-specific CD8 T cell response (days 8, 10, and 14 p.i. for 2 × 106, 5 × 103, and 10 PFU inoculation doses, respectively), a direct correlation was observed between initial PyV dose and PD-1 expression (Fig. 3,B, left panel). These patterns of CD127 and PD-1 expression were maintained 6 wk after infection (Fig. 3 B, right panel). As the IL-7Rhigh PD-1low expression pattern appears to be linked to CD8 T cell functional competence (26, 45, 46, 47), these phenotypic data suggest that antiviral CD8 T cells elicited by low-dose inoculation may be better equipped to survive and mediate immune surveillance during persistent infection.
Effect of low-dose PyV inoculation on CD8 T cell function
Because manipulation of the initial PyV inoculum dose elicited antiviral CD8 T cells with distinct cell surface phenotypes, we asked whether viral dose affected CD8 T cell function as well. At the peak expansion of high- and low-dose PyV inoculation (days 8 and 10 p.i., respectively), >90% of LT359-specific CD8 T cells produced IFN-γ after 5-h ex vivo peptide stimulation (data not shown). However, as early as 2 days before their peak expansion, low-dose CD8 T cells produced significantly more IFN-γ on a per cell basis as shown by higher IFN-γ mean fluorescence intensity (MFI) (Fig. 4,A). No statistically significant differences were seen in the frequencies of IL-2- and TNF-α-producing CD8 T cells (Fig. 4,B). In vivo cytotoxicity assays performed at the peaks of infection also showed that both inoculation groups were equally efficient in eliminating LT359 peptide-pulsed spleen cells (data not shown). Functional avidity, as determined by the frequency of IFN-γ+ CD8 T cells over an LT359 peptide titration curve, was equivalent for both PyV inoculation doses (Fig. 5,A), as was TCR Vβ usage (Fig. 5 B), suggesting that altering PyV dose does not affect the repertoire of Ag-specific CD8 T cells recruited into the response.
We next asked whether the difference in PyV-specific CD8 T cell function after high-dose and low-dose inoculations seen during acute infection held through persistent infection. For each inoculation dose, similar numbers of Db LT359 tetramer+ CD8 T cells were present at 6 wk p.i. in the spleen (Fig. 1,A). Although the numbers of IFN-γ-producing LT359-specific CD8 T cells in the spleen corresponded closely to those enumerated by Db LT359 tetramer staining irrespective of inoculation dose (data not shown), those elicited by low-dose inoculation exhibited higher IFN-γ MFIs and a broader cytokine profile (Fig. 6, A and B). As seen during the acute phase, in vivo LT359-specific cytotoxicity during persistent infection was similar for both groups (data not shown). Thus, the increased functional competence of low-dose inoculation antiviral CD8 T cells was not only maintained from acute to persistent infection, but was enhanced, because a larger fraction of these cells coproduced IL-2 and TNF-α. Low-dose inoculation also endowed PyV-specific CD8 T cells with greater recall potential. Four days after challenge with a rVSV encoding an LT359–368 minigene, mice receiving a low-dose inoculation mounted twice the LT359-specific CD8 T cell recall response of the high-dose inoculated mice (Fig. 6 C). These results reveal the long-term positive effect of low-dose inoculation on the anti-PyV CD8 T cell response.
CD8 T cell homeostasis during persistent PyV infection
Persistent viral infection can result in defective antiviral CD8 T cell homeostasis and a subsequent decline in pathogen-specific CD8 T cell numbers (20, 21). Because low-dose PyV inoculation produced antiviral CD8 T cells with improved function, we next asked whether PyV dose affected CD8 T cell homeostasis during persistent infection. Pathogen-specific CD8 T cells that survive contraction are maintained long-term by IL-7R- and IL-15R-mediated survival and proliferation signals (16, 17, 48). At 6 wk p.i., the IL-2R/IL-15Rβ chain (CD122) was expressed at similar levels by all Db LT359 tetramer+ CD8 T cells at both inoculation doses (data not shown). However, a greater fraction of low-dose LT359-specific CD8 T cells (64%) expressed the IL-7Rα chain (CD127) than did high-dose CD8 T cells (40%) (Fig. 3,A). The observation that IL-7R signaling is associated with elevated bcl-2 levels (49) fits with the 50% increase in MFI for bcl-2 intracellular staining of anti-PyV CD8 T cells in mice given a low virus inoculum (Fig. 7 A).
We recently demonstrated that PyV-specific CD8 T cells from persistently infected mice are defective in homeostatic proliferation and progressively decline in number upon transfer to wild-type PyV infection-matched mice (21). Given the difference in IL-7Rα and bcl-2 expression by PyV-specific CD8 T cells as a function of inoculation dose, we next asked whether inoculation dose affected the survival of antiviral CD8 T cells. Spleen cells from persistently infected high-dose and low-dose PyV-inoculated B6 (CD45.2) mice were CFSE labeled and transferred to persistently infected B6.SJL (CD45.1) recipient mice that had been inoculated with an equivalent dose of a mutant PyV lacking the LT359–368 epitope (designated nPyV). The use of nPyV serves two purposes. First, by eliciting a virus-neutralizing humoral response, infection with nPyV eliminates virus carryover with the donor T cells, as confirmed by the absence of endogenous Db LT359 tetramer+ CD8 T cells in the recipient mice (data not shown). Secondly, this mutant virus allowed us to further ask whether lack of maintenance of persistent infection-phase anti-PyV CD8 T cells (21) depends on cognate viral Ag stimulation. As shown in Fig. 7,B, donor tetramer+ cells steadily declined in number over time and did not divide, as shown by an absence of CFSE dilution (Fig. 7 C). Of note, the t1/2 of the low-dose tetramer+ cells was 10 days longer, indicating a survival advantage over the high-dose cells. Neither exogenous IL-7 nor IL-15 corrected the defect in proliferative capacity of these chronic memory CD8 T cells in vitro (data not shown), suggesting that increased CD127 expression by PyV-specific CD8 T cells elicited by low-dose inoculation is not sufficient to confer their survival advantage.
Bystander factors attenuate PyV-specific CD8 T cell expansion and function
We have demonstrated that a lower-dose PyV inoculum favors induction of a more potent antiviral CD8 T cell response during persistent infection. Delayed CD8 T cell expansion (Fig. 1) and reduced DC maturation (Fig. 2) after low-dose PyV infection suggest that early priming events occur efficiently even under conditions of reduced PyV Ag presentation and inflammation. The increased viral burden in mice receiving high-dose PyV infection (Fig. 1,B) precludes a conclusion on whether functional differences during persistence are the product of early inflammatory events or reflect repetitive encounter with cognate Ag. We therefore sought to isolate the effects of Ag stimulation from bystander factors in induction of PyV-specific CD8 T cells. We first performed DC immunizations to evaluate the effects of Ag density on the extent of CD8 T cell expansion and acquisition of effector function. Bone marrow-derived DCs were pulsed with different concentrations of LT359 peptide, 2.5 × 105 CD11c+ cells were injected i.v. into naive B6 mice, and Ag-specific responses were analyzed 7 days later. Mice receiving DCs pulsed with lower amounts of peptide exhibited slightly decreased, but not statistically significant, numbers of splenic LT359-specific CD8 T cells (Fig. 8,A, left panel). Interestingly, varying peptide dose also did not affect IFN-γ MFI (Fig. 8 A, right panel), CD127 expression, or the fraction of cells coproducing IL-2 and TNF-α (data not shown). Thus, Ag density during priming had little phenotypic or functional impact on the anti-PyV CD8 T cell response.
To analyze the effects of virus-induced inflammation on CD8 T cell responses without the confounding variable of persistent Ag, mice were immunized as above, or were immunized and received 2 × 106 or 5 × 103 PFU nPyV s.c. To determine whether postpriming bystander infection affected the quality or quantity of the CD8 T cell response, one group received 2 × 106 PFU nPyV at day 2 postimmunization, when transferred DCs are undetectable (10). Although low-dose nPyV infection had little effect on peak Ag-specific CD8 T cell numbers at day 7 postimmunization, high-dose nPyV attenuated the response >2-fold (Fig. 8,B, bar graph). High-dose nPyV infection also dampened LT359 peptide-stimulated IFN-γ (Fig. 8,B, right panels) and IL-2 production (Fig. 8,C, left panel), but had less effect on the frequency of cells coproducing IFN-γ and TNF-α (Fig. 8,C, right panel). Administering high-dose nPyV after CD8 T cell priming (day 2 postimmunization) also resulted in an inferior CD8 T cell response in terms of total numbers, IFN-γ MFI, and cytokine coproduction (Fig. 8, B and C). These results suggest that the induction of a superior anti-PyV CD8 T cell response is less dependent on Ag density during priming, but is negatively affected by bystander infection-associated events during priming and subsequent expansion.
Using the above DC immunization model, we performed a kinetic analysis of LT359-specific CD8 T cell responses to determine whether differences in CD8 T cell fitness incurred by nPyV infection were maintained long-term in the absence of cognate Ag. With an increasing dose of nPyV, LT359-specific CD8 T cells attained a lower peak and a smaller pool of memory cells (Fig. 8,D). There was also an accelerated loss of IFN-γ production on a per-cell basis that occurred in a nPyV dose-dependent manner (Fig. 8 E). These data demonstrate that bystander factors associated with PyV infection negatively impact memory antiviral CD8 T cell numbers and function.
Discussion
In this study, we show that viral inoculation dose is a critical variable guiding differentiation of the CD8 T cell response to a persistent viral infection. Low-dose inoculation with PyV, a persistent natural mouse pathogen, is associated with generation of virus-specific CD8 T cells having multicytokine effector capability, improved per-cell cytokine production, enhanced expression of molecules controlling T cell survival, and a more robust recall response. Reduced inoculation doses affected the kinetics, but not the magnitude of the anti-PyV CD8 T cell response. As early as the acute infection phase, virus-specific CD8 T cells elicited by lower dose inoculation demonstrated improved IFN-γ production. Using peptide-DC immunization and a dominant CD8 T cell epitope-null mutant PyV, we further demonstrate that the bystander infection environment during priming and expansion modulates CD8 T cell functional competence. These data indicate that Ag-independent factors during early phases of infection control the quality of antiviral CD8 T cell response to persistent virus infection.
Equivalent CD8 T cell expansion in response to varying doses of PyV fits with the autopilot scenario in which brief Ag encounter by naive T cells is sufficient to initiate a full program of expansion and differentiation (2, 3, 4). The equivalent rate of antiviral CD8 T cell expansion despite marked differences in inoculation dose is consistent with the interpretation that viral epitope density needs to achieve a minimum threshold to initiate priming of Ag-specific T cells. Our findings contrast with those of other groups showing that the duration or density of Ag stimulation is directly proportional to the magnitude of CD8 T cell expansion, although not necessarily the acquisition of effector function or recall capacity (50, 51). An explanation for this discrepancy may lie in the possibility that, in these studies, Ag is cleared before complete recruitment of naive T cells. In contrast, the inability of mice to clear PyV even at very low doses may permit recruitment of all naive virus-specific CD8 T cells.
Several recent reports provide evidence that the timing of recruitment of naive CD8 T cells can affect T cell phenotype and function, a possible reflection of unique early and late priming microenvironments (21, 52, 53). Our recent studies show that naive PyV-specific CD8 T cells are continuously recruited in persistently infected mice and that late-primed cells are phenotypically distinct from those primed earlier (21, 31). By extension, the confluence of viral Ag and inflammatory mediators during acute-phase low-dose PyV infection may mimic the environment found months after high-dose inoculation. In support of this possibility, IL-2 production by high-dose inoculum-elicited virus-specific CD8 T cells takes nearly 1 year to reach the levels seen by low-dose antiviral CD8 T cells at 6 wk p.i. (31). Antiviral therapy during high-level chronic infections has been shown to improve the function of pre-existing pathogen-specific CD8 T cells (54, 55). The data presented in this study, together with our recent studies (21, 31), suggest that these therapies may have beneficial consequences for the quality of late-primed antiviral CD8 T cells as well.
Persistent viral infections often result in deterioration of CD8 T cell function due to repetitive encounter with cognate Ag (22, 23, 24, 25). In these models, loss of CD8 T cell function is gradual and does not manifest until months after infection. Our results demonstrate that CD8 T cell functional competence can diverge in response to differing PyV inoculation doses as early as the peak of CD8 T cell expansion. This finding is consistent with a recent report demonstrating that virus-specific CD8 T cells generated in response to infection with the aggressive lymphocytic choriomeningitis virus clone 13 variant display defects in cytokine-producing capability as early as day 9 of infection (27). The continued presence of PyV during persistent infection also did not degrade the superior functional quality of low-dose CD8 T cells seen during the acute phase of infection. The diminished PyV-specific CD8 T cell responses after DC immunization with superimposed nPyV infection are consistent with the interpretation that CD8 T cell fitness is programmed during the initial priming and expansion phase by virus-elicited Ag-nonspecific factors (Fig. 8). In line with this conclusion, these detrimental early bystander effects carried forward to reduced memory CD8 T cell numbers and function. These defects appear to be cell intrinsic; Ag-specific CD8 T cells from high-dose donors underwent faster attrition than those from low-dose donors upon relocation to cognate Ag-free, persistently infected recipients (Fig. 7 B). Taken together, these data suggest that the enhanced fitness of anti-PyV CD8 T cells during persistent low-dose infection bears the imprint of early priming events.
Because maintenance of stable numbers of CD8 T cells during persistent infection is dependent on antigenic stimulation (20, 21), a potential consequence of successful antiviral therapy is a decline in virus-specific CD8 T cells (56, 57). Interventions that improve the ability of such chronic memory CD8 T cells to respond to IL-7R- and/or IL-15R-mediated signaling may reduce this Ag dependence (20). In this connection, van Leeuwen et al. (45) recently provided evidence that maintenance of CD8 T cells specific for human CMV requires both TCR stimulation and helper factors. Several studies in both mice and humans have shown an association between persistent viral infection and reduced bcl-2, IL-7Rα, and IL-2 expression by virus-specific CD8 T cells (58, 59, 60, 61). Our results demonstrate that this phenotype is reversed by reduction of the initial viral inoculum. This infection regimen had functional consequences, as a lower initial dose of PyV promoted survival of virus-specific CD8 T cells (Fig. 7,B). In addition, the defect in chronic memory CD8 T cell maintenance we previously described (21) was not reversed in the absence of cognate Ag. Although in vitro assays did not demonstrate a role for IL-7 or IL-15 in this process, culture conditions may not replicate the in vivo cytokine environment. Alternatively, higher bcl-2 levels may be responsible for the longer t1/2 of low-dose inoculation-elicited PyV-specific CD8 T cells (Fig. 7 A). These data suggest that administration of antiviral agents soon after infection may improve chronic memory CD8 T cell survival.
Finally, our finding that bystander inflammation negatively affects the anti-PyV CD8 T cell response (Fig. 8, B–E) raises a cautionary note in the use of adjuvants to bolster pathogen-specific memory CD8 T cell responses. In line with this, Badovinac et al. (10) recently showed that CpG-mediated inflammation impairs the rate of differentiation of memory CD8 T cells induced by DC vaccination. With respect to PyV, particular TLRs have been implicated in inducing and shaping cytokine responses to infection with this virus (57), raising the possibility that the level of TLR engagement by PyV itself might influence the quality of the virus-specific CD8 T cell response. Our data also suggest that the efficacy of vaccine/adjuvant combinations may be assessed early after immunization based on differences in IFN-γ MFI and IL-7R expression during and soon after the peak of CD8 T cell expansion. From a vaccination perspective, our findings further suggest that manipulation of vaccine dose may enhance the quality of adaptive pathogen-specific immunity.
Acknowledgment
We thank Joshua Loomis for constructing the rVSV-LT359.
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.
This work was supported by National Institutes of Health Grants RO1CA71971 and RO1CA100644 (to A.E.L.), RO1CA095924 (to G.N.B.), and F32CA106090 (to V.V.).
Abbreviations used in this paper: DC, dendritic cell; FasL, Fas ligand; LT, large T Ag; MFI, mean fluorescence intensity; nPyV, LT359–368 epitope-null polyomavirus; p.i., postinfection; PyV, polyomavirus; VSV, vesicular stomatitis virus.