CD4 T cells are principal producers of IL-21 and are often required for optimal CD8 T cell responses. Therefore, we investigated the importance of IL-21 in determining the phenotypic attributes, functional quality, and maintenance of antiviral CD8 T cells following acute infection with the prototypic mouse pathogen lymphocytic choriomeningitis virus. Previous reports have documented an obligatory role for IL-21 in sustaining CD8 T cell responses during chronic infections. Here we show that the requirements for IL-21 are less stringent following acute infections; however, in the absence of IL-21, the capacity of CD8 T cells to attain the polyfunctional trait of IL-2 production is consistently reduced during both the effector and memory phases. This is further supported by in vitro studies showing that the addition of IL-21 promotes the differentiation of IL-2–producing CD8 T cells. Although the generation of memory CD8 T cells, which are capable of mounting protective recall responses, proceeds independently of IL-21, we demonstrate that IL-21 does function to support secondary responses, especially under competitive conditions. Collectively, these studies highlight the potential roles of IL-21 in determining the quality of CD8 T cell responses postinfection.
The quality and effectiveness of pathogen-specific CD8 T cells is shaped by many factors including the presence of CD4 T cell help and the availability of cytokines. Ideally, effector CD8 T cell responses elaborate the necessary arrays of effector functions, such as cytokine production and cytotoxicity, required to clear infected cells. Typically, as homeostasis is restored following the control of the pathogen, a population of memory CD8 T cells emerges that is stably maintained over time (1–3). These memory T cells serve as sentinels and respond to protect the host if the original infection is re-encountered.
CD8 T cell responses are diverse and exhibit a spectrum of functional activities. The ability of effector CD8 T cells to produce IL-2 has been partially associated with their ability to form the memory pool (4, 5). In addition, CD8 T cells that produce IL-2 are generally considered to be highly polyfunctional and can often simultaneously elaborate multiple effector activities, including the coproduction of IFN-γ and TNF-α (6–8). The development of T cells that can functionally multitask is notable, as this has been correlated with a greater ability to control infections (7–12).
It is now well established that optimal CD8 T cell responses can be disrupted if the levels of CD4 T cell help is insufficient. Under these “helpless” conditions, reduced primary CD8 T cell responses may be elicited and atypical memory responses are often observed (13–15). Without CD4 T cell help, the pseudomemory CD8 T cells may not be maintained at steady-state levels, exhibit altered responsiveness to homeostatic cytokines, display a reduced portfolio of effector functions including diminished IL-2 production, and fail to mount prominent secondary immune responses. During chronic infections, the requirements for CD4 T cell help are even more stringent, as under these conditions, marked exhaustion of the CD8 T cell response can occur, which is characterized by the functional inactivation and subsequent deletion of the responding T cells (7, 8, 16–18).
IL-21 is a member of the common γ-chain family of cytokines, is mainly produced by CD4 T cells, and has a broad impact on the immune response (19, 20). IL-21 has been shown to limit the terminal differentiation of effector CD8 T cells, favoring the development of immature effector cells that are more effective at suppressing tumor growth (21). IL-21 has also been shown to suppress the development of T cell senescence, which is characterized by the loss of CD28 expression (22–24). In vitro studies indicate that IL-21 can act as a third signal, distinct from IFN-α and IL-12, to promote CD8 T cell effector functions (25). Recently, IL-21 has been implicated as a vital helper factor that supports CD8 T cell activities during chronic viral infections. In studies of HIV-1–infected individuals, the levels of IL-21 have been shown to positively correlate with CD4 T cell counts (26, 27). The highest amounts of IL-21 were detected in uninfected subjects and in long-term nonprogressors. In the case of chronic lymphocytic choriomeningitis virus (LCMV) infections, IL-21 was shown to act directly on CD8 T cells to sustain their physical presence and functional capabilities (28–31). The absence of IL-21 resulted in severe CD8 T cell exhaustion and viral persistence, outcomes that parallel those observed in chronically infected CD4−/− mice.
Given that CD4 T cells are a principal source of IL-21 and that CD4 T cells are necessary for optimal antiviral CD8 T cell responses, in this report, we set out to determine how IL-21 influences the functional quality and memory attributes of CD8 T cells following an acute viral infection. We demonstrate that following acute LCMV infection, effector and memory CD8 T cell responses develop independently of IL-21, but that IL-21 regulates the generation of IL-2–producing CD8 T cells. Although IL-21 deficiency has minimal effects on the formation of the memory pool, in the absence of IL-21 signaling, secondary responses are compromised under competitive conditions.
Materials and Methods
Mice and infections
C57BL/6J (B6), B6.PL Thy1a/CyJ, B6.129S2-Cd4tm1Mak/J (CD4−/−), B6.SJL-PtprcaPepcb/BoyJ (CD45.1), and B6.129S7-Rag1tm1Mom/J (RAG-1−/−) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). P14 TCR transgenic mice (P14) crossed onto a B6 background were originally kindly provided by Dr. R. Ahmed (Emory University, Atlanta, GA). B6;129S5-Il21tm1Lex (IL-21−/−) and B6;129S5-Il21rtm1Lex (IL-21R−/−) mice were obtained from the Mutant Mice Regional Resource Center (Davis, CA). IL-21−/− mice were crossed an additional 6, 10, or 12 generations onto the B6 background. IL-21R−/− mice (crossed an additional four generations onto the B6 background) were mated with P14 mice to generate IL-21R+/− and IL-21R−/− P14 TCR transgenics. The IL-21R−/− mice used for the bone marrow chimera studies were crossed eight generations onto the B6 background. All mice were bred and maintained in fully accredited facilities at the University of Alabama at Birmingham. For acute infections, mice were infected by i.p. injection with 2 × 105 PFU LCMV-Armstrong and with 2 × 106 PFU LCMV-Armstrong for rechallenge studies. In certain experiments, LCMV primed mice were challenged by i.v. inoculation with 5 × 104 to 1 × 105 CFU recombinant Listeria monocytogenes expressing the LCMV GP33/GP34 epitopes (rLM33), kindly provided by Dr. H. Shen (University of Pennsylvania, Philadelphia, PA) (32).
Splenocytes were prepared as previously described (6). Blood samples were collected into 4% (w/v) sodium citrate. Livers and lungs were isolated from mice that had been perfused with 20 ml PBS. Livers were minced and incubated for 30 min with 2 mg/ml Collagenase D (Roche, Indianapolis, IN) and 0.3 mg/ml DNase I (Sigma-Aldrich, St. Louis, MO) in HBSS. Liver samples were then further disrupted by passage through a 70-μm cell strainer. Lungs were disrupted into single-cell suspension using fine wire mesh. Lymphocytes from the blood, liver, and lungs were subsequently isolated by centrifugation over a layer of Histopaque-1083 (Sigma-Aldrich).
Cellular analyses and flow cytometry
MHC tetramer staining and intracellular cytokine analyses were performed essentially as previously described (6). For the detection of IL-21, a two-step intracellular cytokine staining was performed by initially using recombinant IL-21R/Fc fusion proteins (0.1 μg; R&D Systems, Minneapolis, MN), together with anti–IFN-γ FITC (XMG1.2) and anti–IL-2 allophycocyanin (JES6-5H4) (eBioscience, San Diego, CA). After incubation, cells were washed with Perm-Wash buffer (BD Biosciences, San Jose, CA) and stained with F(ab′)2 PE-conjugated goat anti-human Fcγ (Jackson ImmunoResearch Laboratories, West Grove, PA) (33). Samples were acquired using FACSCalibur or LSR II flow cytometers (BD Biosciences), and data were analyzed using FlowJo software (Tree Star, Ashland, OR).
In vitro cultures
Splenocytes from naive P14 TCR transgenic mice were labeled with CFSE (Molecular Probes; Eugene, OR) and cultured in the presence of 1 μg/ml GP33 (KAVYNFATM) peptide and combinations of recombinant IL-21 (50 ng/ml; R&D Systems) and IL-15 (50 ng/ml; Invitrogen, Carlsbad, CA). In certain experiments, naive (CD44lo) CD8+ T cells from IL-21R+/− and IL-21R−/− P14 TCR transgenic mice were isolated by cell sorting using a FACSAria instrument (BD Biosciences). These sorted naive CD8 T cells were labeled with CFSE and cultured together with irradiated (3000 radiation-absorbed dose) splenic feeder cells from B6.PL Thy1a/CyJ mice at a responder-to-feeder cell ratio of 1:5. The GP33 peptide was added to activate the responder cells, with and without the addition of IL-21, as described earlier. After ∼5 d, cultured splenocytes were restimulated with the GP33 peptide before intracellular cytokine staining.
Adoptive transfer of normalized numbers of memory CD8 T cells
Splenic CD8 T cells from LCMV primed IL-21+/+ and IL-21−/− mice (>150 d postinfection) were isolated to >96% purity using mouse CD8 FlowComp Dynabeads (Invitrogen). The frequencies of LCMV Db(GP33)-specific CD8 T cells in the purified cell pools were then immediately determined by MHC tetramer staining. Subsequently, purified cell populations containing 5 × 104 Db(GP33)-specific CD8 T cells were transferred by i.p. injection into allelically marked (CD45.1) recipients. One day later, recipient mice were challenged with rLM33, and the recall responses of the donor cells were analyzed after 6 d.
Mixed bone marrow chimeras
Bone marrow was prepared from the tibias and femurs of control CD45.1 mice, and CD45.2 IL-21R+/+ and IL-21R−/− littermates (31). Bone marrow suspensions were depleted of T cells using anti-CD5 (Ly-1) microbeads (Miltenyi Biotec, Auburn, CA). Recipient RAG-1−/− mice were irradiated twice, 3–4 h apart, to give a total exposure of ∼1000 radiation-absorbed doses. These mice were then reconstituted by i.v. injection of T cell-depleted bone marrow containing 5–7 × 106 IL-21R+/+ (CD45.1) cells mixed with equal numbers of IL-21R+/+ or IL-21R−/− (CD45.2) cells. Mice were maintained on chlorinated acidified water containing sulfamethoxazole, trimethoprim, and neomycin for 6 wk. Chimeras were infected with LCMV-Armstrong ∼8 wk following reconstitution and subsequently rechallenged with LCMV >60 d after primary challenge.
Two-tailed t tests were used to determine statistical significance between two groups, and a one-way ANOVA test was used to compare three groups. The p values were calculated using Prism software (Graph Pad, La Jolla, CA).
Reduced IL-2–producing primary CD8 T cell responses in the absence of IL-21
To evaluate the impact of IL-21 on the induction of antiviral effector CD8 T cell responses during an acute viral infection, wild-type (IL-21+/+), IL-21 heterozygous (IL-21+/−), and IL-21–deficient (IL-21−/−) mice were infected with LCMV-Armstrong. As expected, by 8 d postinfection, IL-21+/+ mice mounted a robust multiepitope-specific CD8 T cell response (Fig. 1). IFN-γ production by virus-specific CD8 T cells was readily detected following a brief (5 h) period of reactivation, and a subset of the responding cells displayed polyfunctional attributes as assessed by the coproduction of IL-2 (Fig. 1A). By comparison with IL-21+/+ mice, the frequencies of IFN-γ+ CD8 T cells capable of producing IL-2 were reduced in IL-21+/− hosts and were even lower in IL-21−/− mice (Fig. 1A–C). This decreased fraction of IL-2–competent effector CD8 T cells was also paralleled by a diminution in the mean fluorescence intensity (MFI) of IL-2 staining, indicating that the levels of IL-2 production are lower, on a per cell basis, in CD8 T cells derived from IL-21–deficient hosts (Fig. 1C). Evaluation of the overall size of the CD8 T cell pool and the numbers of MHC-tetramer binding virus-specific CD8 T cells showed no significant differences between IL-21+/+, IL-21+/−, and IL-21−/− mice (Fig. 1D). Enumeration of IL-2– and IFN-γ–producing epitope-specific CD8 T cells revealed that the reduced percentage of IL-2 competent effector CD8 T cells was attributable to both a slight decrease in the absolute numbers of IL-2+ cells and also a modest increase in the numbers of IFN-γ+ CD8 T cells (Fig. 1E). Thus, qualitative and quantitative differences in the functionality of the virus-specific CD8 T cell response manifest in the absence of IL-21.
We investigated whether the functional shifts that occur in the absence of IL-21 were mirrored by other phenotypic alterations within the effector CD8 T cell pool. Analyses of the expression of a panel of molecules associated with the generation of effector CD8 T cells, including CD27, CD43, CD62L, CD122, CD127, PD-1, granzyme B, and T-bet, showed no marked differences between epitope-specific CD8 T cells induced in IL-21+/+, IL-21+/−, or IL-21−/− mice (Fig. 2). Thus, even though IL-21 selectively impacted the proportion of IL-2–producing CD8 T cells, other phenotypic attributes of the response were unaltered.
Generation of memory CD8 T cells in the absence of IL-21
Because the loss of IL-21 negatively impacted the quality of the antiviral CD8 T cell response during the effector phase of the response, we next checked whether differences in functional quality were maintained into the memory phase. By day 120 following acute LCMV infection, memory CD8 T cells capable of coproducing IFN-γ and IL-2 were readily detectable. Although by comparison with the effector phase, the proportion of IL-2–producing CD8 T cells increased in IL-21+/+, IL-21+/−, and IL-21−/− mice, the fraction of IL-2–producing CD8 T cells was lower in the IL-21−/− mice compared with the IL-21+/+ mice (Fig. 3A, 3B). In addition, a significant reduction in the MFI of IL-2 production by CD8 T cells in the IL-21−/− cohort remained discernible at memory time points, paralleling the reductions observed during the effector phase (cf. Figs. 1C, 3C). Despite this reduction in functional quality, the numbers of virus-specific memory CD8 cells were similar in all of the groups, demonstrating that IL-21 is not required for the maintenance of these cells (Fig. 3D, 3E). Further analyses of the phenotypic attributes of the virus-specific memory CD8 T cells revealed no marked differences between the IL-21+/+, IL-21+/−, and IL-21−/− cohorts (Fig. 4). Collectively, these data show that IL-21 deficiency does not inhibit memory CD8 T cell formation, and a distinct population of IL-2+ CD8 T cells does form; nevertheless, a reproducible decrease in the fraction of IL-2+ CD8 T cells in IL-21−/− mice is apparent during the memory phase of the response.
CD8 T cells in the liver, lungs, and circulation
It has been proposed that the inability to produce IL-2 affects the accumulation of virus-specific CD8 T cells in extralymphoid organs (34). Therefore, we compared the frequencies of IFN-γ and IL-2–producing virus-specific CD8 T cells in the liver, lungs, and circulation at days 9 (Fig. 5A, 5B) and 32 (Fig. 5C, 5D) following infection. At 9 d after infection, comparable frequencies of IFN-γ+ CD8 T cells were detectable in IL-21+/+ and IL-21−/− mice following stimulation with the GP33 peptide (Fig. 5A). Consistent with the splenic analyses (Fig. 1), the fraction of the IFN-γ+ CD8 T cells capable of producing IL-2 were reduced in all of the compartments analyzed in IL-21−/− mice (Fig. 5A, 5B). By day 32 postinfection, the proportion of IL-2+ CD8 T cells had increased compared with the effector time point. Nevertheless, by comparison with IL-21+/+ mice, the IL-21−/− mice continued to exhibited a decrease in the frequency of IL-2–producing CD8 T cells (Fig. 5C, 5D). Thus, IL-21 deficiency impacts the ability of CD8 T cells to produce IL-2 in both lymphoid and nonlymphoid organs.
CD4 T cell responses in the absence of IL-21
CD4 T cells assist in the generation of optimal CD8 T cell responses following acute infections (13–15). Therefore, we next inspected whether the functional defects apparent in the virus-specific CD8 T cells in IL-21−/− mice were associated with altered CD4 T cell responses. Analysis at effector (day 8; Fig. 6A–C) and memory (day 120 or longer; Fig. 6D–F) time points revealed no marked differences in the overall virus-specific CD4 T cell response in the absence of IL-21. Thus, the altered CD8 T cell response observed in IL-21−/− mice is unlikely to be due to the failure to induce an antiviral CD4 T cell response.
Comparative analyses of CD8 T cells consistently shows the most pronounced IL-2+ response in IL-21+/+ mice, an intermediate response in IL-21+/− mice, and the lowest response in IL-21−/− mice (Figs. 1, 3). We predicted that these changes in the quality of CD8 T cell responses would be paralleled by decreases in the levels of IL-21 production by the virus-specific CD4 T cells. Therefore, we measured the ability of antiviral CD4 T cells from each group of mice to produce IL-21 following stimulation with the LCMV GP61-80 peptide epitope. IL-21 production was clearly discernible by CD4 T cells from IL-21+/+ mice, with intermediate levels produced by IL-21+/− CD4 T cells and, as expected, was not synthesized by IL-21−/− CD4 T cells (Fig. 6G, 6H).
IL-21 promotes the generation of IL-2+ CD8 T cells
The results showing that the generation of IL-2–producing antiviral CD8 T cells was reduced without IL-21 prompted us to determine whether the addition of IL-21 enhanced the functional quality of CD8 T cells. In a series of in vitro studies, CD8 T cells from P14 TCR transgenic mice, which are specific for the LCMV GP33 epitope, were activated by addition of the cognate GP33 peptide and cocultured with combinations of IL-15 and IL-21. To determine whether the functional attributes of the responses were altered by the addition of IL-21, intracellular cytokine analyses for IL-2 (Fig. 7A) and IFN-γ (Fig. 7B) were performed. After 18 h of culture, no pronounced differences in the fraction of IL-2–producing CD8 T cells were observed (data not shown). By ∼5 d of culture, the responding cells had undergone six to eight divisions, and notably, initial activation in the presence of exogenous IL-21 increased the fraction of IL-2–producing CD8 T cells (11.8 ± 4.9%) compared with cells cultured without cytokine (4.6 ± 3.1%; p = 0.0245; Fig. 7A). Examination of the impact of IL-21 on IFN-γ production showed no significant differences between the IL-21 cultured cells (21.9 ± 10.5%) and cells cultured without the addition of IL-21 (18.1 ± 7.2; p = 0.5293; Fig. 7B).
We next examined whether IL-21 was acting directly on the responding CD8 T cells to configure their functional traits (Fig. 7C–F). Naive (CD44lo) CD8 T cells sorted from P14 TCR transgenic IL-21R–deficient (IL-21R−/−) and IL-21R+/− mice were similarly activated in vitro using the GP33 peptide, with and without the addition of IL-21. After ∼5 d, culture in the presence of IL-21 increased the proportion of IL-2–producing IL-21R+/− CD8 T cells but did not augment the responses of IL-21R−/− CD8 T cells (Fig. 7C, 7D). Because IL-21 has been proposed to influence the induction of IL-17–producing T cells, we also checked whether Ag-activated CD8 T cells exposed to IL-21 gained the ability to produce IL-17 (Fig. 7E, 7F) (33, 35). No significant fraction of IL-17+ CD8 T cells were detectable following the culture period. These in vitro studies indicate that IL-21 acts directly on CD8 T cells to influence their differentiation and acquisition of functional traits.
CD8 T cells primed in the absence of IL-21 mount recall responses
Because functional discrepancies were detected in antiviral CD8 T cells generated in the presence or absence of IL-21, we next evaluated whether these “IL-21–less” memory CD8 T cells could elicit secondary recall responses. In an initial series of studies, IL-21+/+, IL-21+/−, IL-21−/−, and CD4−/− mice, which had been infected with LCMV-Armstrong at least 150 d earlier, were rechallenged with a high dose of LCMV (Fig. 8A). As shown previously, CD8 T cells primed in CD4−/− mice mount severely blunted recall responses (13–15), as demonstrated by the minimal increase in the numbers of total CD8 T cells present in the spleen at 7 d postrechallenge (Fig 8B). By contrast, the absence of IL-21 had no effect on the overall magnitude of the CD8 T cell response, as similar secondary responses were detected in IL-21+/+, IL-21+/−, and IL-21−/− mice (Fig. 8B). Further evaluation of viral-epitope–specific CD8 T cells confirmed that secondary CD8 T cell responses are impaired in CD4−/− hosts but are unaffected by the absence of IL-21 (Fig. 8C). Thus, the requirements for CD4 T cell help in generating memory CD8 T cells that can mount robust recall responses are far more stringent than the specific requirement for IL-21.
To further evaluate recall responses, we performed a series of adoptive transfer studies. Normalized numbers (5 × 104) of memory LCMV GP33 epitope-specific CD8 T cells from IL-21+/+ and IL-21−/− mice were transferred into congenic, allelically marked, naive CD45.1 recipient mice (Fig. 9A). Subsequently, the recall potential of these donor populations was compared following challenge of the recipients with rLM33 (32). Both IL-21+/+ and IL-21−/− donor CD8 T cells were readily detectable by 6 d following rLM33 challenge (Fig. 9B). Similar numbers of IFN-γ–producing GP33/GP34-specific donor cells were present in both groups of recipients; however, the number of IL-2–producing IL-21−/− donor cells was lower, which is consistent with the impaired functional CD8 T cells responses detected at acute and memory time points (Fig. 9C; cf. Figs. 1, 3). Analysis of the endogenous listeriolysin O-specific CD4 T cell responses in the recipient mice demonstrated the presence of IL-21–producing CD4 T cells (Fig. 9E). Nevertheless, this production of IL-21 did not enable the memory CD8 T cells generated in the absence of IL-21 to recover their ability to produce IL-2 and match the functional responsiveness of IL-21+/+ memory CD8 T cells.
IL-21 helps recall CD8 T cell responses under competitive conditions
To provide a more stringent readout of the potential roles of IL-21 in helping recall CD8 T cell responses, we directly compared the secondary responses of IL-21R+/+ and IL-21R−/− memory CD8 T cells in the same host. To accomplish this, we generated a series of chimeras by reconstituting lethally irradiated RAG-1−/− mice with a mixture of allelically distinguishable bone marrow from IL-21R+/+ and IL-21R−/− mice. In this way, the recipient mice are reconstituted with populations of CD8 T cells that could (IL-21R+/+) and could not (IL-21R−/−) perceive IL-21 signals. For control purposes, we also generated mixed chimeras using two populations of allelically marked IL-21R+/+ cells. The reconstitution efficiency was checked 6 wk following transfer, and the chimeric mice were primed by LCMV-Armstrong infection at 8 wk posttransfer. To determine whether IL-21 is required for optimal recall responses, the mixed chimeras were rechallenged with a high dose of LCMV-Armstrong after at least 60 d following the primary acute infection (Fig. 10A). Collectively, in the control chimeras, the median percentage of IL-21R+/+ (CD45.2) CD8 T cells before rechallenge was 44.3 with a range of 29.2–65. By comparison, in the experimental cohorts, the median percentage of IL-21R−/− (CD45.2) CD8 T cells was 41.45 with a range of 24–49.4.
Analysis of the blood on days 4, 7, 15, and 21 following secondary challenge showed similar overall CD8 T cell levels in the control and experimental chimeras (Fig. 10B). Comparison of the IL-21R+/+ and IL-21R−/− CD8 T cells did, however, reveal a selective defect in the recall responses of IL-21R−/− CD8 T cells (Fig. 10C). The recoveries of total and viral epitope-specific CD45.2 IL-21R+/+ CD8 T cells in the control cohorts (Fig. 10C, squares), as well as the fraction of total and viral epitope-specific CD45.2 IL-21R−/− CD8 T cells in the experimental cohorts (Fig. 10C, triangles) were normalized to 100% at day 4 following rechallenge. By day 7 after challenge, the proportion of the IL-21R−/− CD8 T cells clearly declined in the experimental cohort. This impediment of the IL-21R−/− CD8 T cells is most dramatic for the NP396-specific response, which typically dominates following a secondary challenge with LCMV (36, 37). Enumeration of the virus-specific CD8 T cells in the spleen at day 21 following a secondary challenge shows similar magnitudes of CD45.1 IL-21R+/+ and CD45.2 IL-21R+/+ CD8 T cells in the control chimeras (Fig. 10D). In the experimental groups, the numbers of GP33-, NP396-, and GP276-specific IL-21R−/− CD8 T cells were significantly lower than their IL-21R+/+ counterparts, mirroring the findings from the blood analyses (cf. Fig. 10B, 10D). Thus, the IL-21R−/− virus-specific memory CD8 T cells mount a less vigorous recall response and are outcompeted by their IL-21R+/+ counterparts in this mixed environment.
CD4 T cells have been proposed to help CD8 T cell responses in several ways, including by the licensing of dendritic cells, the production of chemokines and cytokines that promote T cell migration to the sites of priming and areas of infection, and by the secretion of cytokines such as IL-2 and IL-21 (28, 29, 31, 38–43). Previous studies have emphasized the critical role of IL-21 in supporting CD8 T cell responses in the context of chronic viral infections (28, 29, 31). In this report, we demonstrate that IL-21 has a subtle but significant role in structuring the functional attributes of antiviral CD8 T cells in acutely infected hosts, as well as a potential impact on recall responses, especially under competitive conditions.
As naive CD8 T cells become initially activated, a burst of IL-2 synthesis is induced that subsides as the primed cells proliferate and differentiate into effector T cells (44, 45). During the effector phase, only a fraction of the Ag-specific CD8 T cells are capable of producing IL-2 (7, 8) (Figs. 1, 3). The ability of the effector CD8 T cells to produce IL-2 tends to be associated with a preferential ability to form the memory CD8 T cell pool; however, clearly not all memory T cells are competent at producing IL-2 following antigenic activation. The emergence of effector and memory CD8 T cells following acute infections has been reported to occur independently of IL-21, which is generally consistent with our observations (28, 29). Nevertheless, we now show that IL-21 appears to shape the functional quality of effector and memory CD8 T cells that do develop, as our analyses of IL-21−/− mice demonstrate a selective impairment in the ability of antiviral CD8 T cells, primed by acute LCMV infection, to produce IL-2. In addition, in vitro priming studies further show that increasing the levels of IL-21 favors the development of IL-2–producing CD8 T cells, and these effects are predominately mediated by the direct actions of IL-21 on the responding T cells.
IL-21 has the potential to regulate IL-2 synthesis in CD8 T cells at several levels including by epigenetic changes, by direct effects on master transcriptional regulators, and by an overall skewing of T cell differentiation. IL-2 and IL-21 are closely genetically linked, and in mice, these genes are encoded 98 kb apart on chromosome 3 (46). This tight association may influence expression because chromatin remodeling at one locus may have regional effects on gene accessibility and activation. The transcription factor Blimp-1 (Prdm1) is a negative regulator of IL-2 production. Blimp-1 deficiency during viral infections results in increased levels of IL-2–producing CD8 T cells and favors the differentiation of KLRG-1loIL-7hi memory precursor/memory T cells (47–51). It is well documented that IL-21 induces Blimp-1 expression in B cells, and more recent results show that IL-21 also induces the expression of this transcription factor in Ag-activated CD8 T cells (52, 53). Nevertheless, Blimp-1 induction in T cells is associated with downregulation of IL-2. Consequently, the lower IL-2+ CD8 T cell response in IL-21−/− mice cannot be caused by IL-21–induced increases in Blimp-1 levels. In B cells and CD4 T cells, IL-21 has also been shown to drive the expression of Bcl-6, which can act antagonistically on Blimp-1 (54–57). Thus, it is plausible that one way in which IL-21 dictates the expression of IL-2 by CD8 T cells is, in part, by shifting the balance of Blimp-1 and Bcl-6 expression.
Several studies have now implicated exogenous IL-2 in inducing CD25 expression on CD8 T cells and driving the development of activated effector cells that are less competent at producing their own IL-2 and forming the memory compartment (44, 58, 59). IL-21 has also been shown to skew the differentiation of naive CD8 T cells, but the outcomes are distinct from those observed following IL-2 exposure. Activation of CD8 T cells in the presence of IL-21 allows the responding cells to attain an immature effector-like phenotype (21), and as illustrated in Fig. 7, IL-21 favors the development of IL-2 competent CD8 T cells. Thus, IL-21 may arrest the development and terminal differentiation of potent effector CD8 T cells, and instead may favor the generation of T cell subsets that are more likely to be maintained over time. The potential role of IL-21 in promoting CD8 T cell responses that are sustainable is further supported by studies showing that CD8 T cells activated in the presence of IL-21 are more effective at eradicating tumors in vivo, and that IL-21 deficiency can result in severe T cell exhaustion and loss of antiviral CD8 T cells in chronically infected hosts (21, 28, 29, 31).
Our findings indicate that IL-21 can influence the developmental fate of responding CD8 T cells, and one signature of this may be enhanced IL-2 expression. Notably, in the absence of IL-21, virus-specific CD8 T cells from all compartments examined, including the spleen, liver, lungs, and circulation, show a decreased intrinsic ability to produce IL-2. This reduction in IL-2 production is more pronounced during the effector phase than at memory time points. Thus, IL-21 deficiency does not preclude memory CD8 T cell development, and a prominent, albeit reduced, population of CD8 T cells capable of producing IL-2 develops under these conditions. It has been previously shown that CD8 T cell frequencies are altered in extralymphoid organs if the T cells cannot produce autonomous IL-2 (34). Thus, the level of IL-2 production by “IL-21–less” CD8 T cells may be sufficient to account for the comparable frequencies observed in nonlymphoid organs of IL-21+/+ and IL-21−/− mice (Fig. 5).
In this study, we show that autocrine production of IL-2 by CD8 T cells is impacted by IL-21. The ability of memory CD8 T cells to synthesize IL-2 is generally associated with polyfunctional attributes, which is characterized by the simultaneous elaboration of multiple effector functions by an individual cell (6, 7, 9, 60, 61). Preferential IL-2 production is also a signature of central memory CD8 T cells and can correlate with greater proliferative potential (62, 63). Thus, detection of IL-2 serves as a useful surrogate marker of phenotypically and functionally distinct CD8 T cell subsets. The functional significance of IL-2 production by CD8 T cells is intriguing. Memory CD8 T cell development and maintenance has been shown to be independent of autocrine IL-2 (34, 42). The requirements for intrinsic IL-2 production may also vary depending on the anatomical location of the memory cells, where it may dictate both expansion and apoptosis of the CD8 T cell population (64). It is also possible that the ability to produce IL-2 preferentially influences the fate of the CD8 T cell population during distinct phases of the immune response.
Although the direct effects of IL-21 on CD8 T cells may shape these responses, IL-21 impacts many immunologic functions. Consequently, the indirect actions of IL-21 on other cells of the immune system may also influence the outcome of the CD8 T cell response. IL-21 restricts the maturation of dendritic cells by reducing the expression of MHC class II complexes, CCR7, CD80, CD86, and cytokines including IL-6, IL-12, IL-1α, and TNF-α (65, 66). Therefore, in the absence of IL-21 altered Ag presentation, costimulation and the production of innate cytokines may affect CD8 T cell activities. IL-21 also has a profound impact on B cell responses by prolonging germinal center reactions and promoting Ig class switching (52, 56, 57, 67–69). Although marked changes in Ab responses have not been detected at early time points following acute LCMV infections, lower Ab titers are present during the memory phase (28, 31). Thus, it is possible that underlying changes in B cell activation and Ab levels could indirectly feed back on the CD8 T cell response.
One of the most striking features of CD8 T cells primed in the absence of CD4 T cells is often severely reduced secondary responses (13–15). Therefore, it was important to carefully evaluate whether IL-21 similarly influenced the development of memory CD8 T cell responses, which are capable of mounting vigorous recall responses. As shown in Fig. 8, IL-21–deficient mice are not phenocopies of CD4 T cell deficient hosts. Thus, CD4 T cells must use distinct mechanisms in addition to IL-21 production to support secondary CD8 T cell responses. Nevertheless, although not as pronounced as CD4 T cell deficiency, IL-21 can have at least some influence on recall responses. The analyses of mixed bone marrow chimeras revealed a “fitness” difference in the secondary responses of IL-21R−/− CD8 T cells. The kinetics of secondary CD8 T cell responses to LCMV are well characterized, and the responding cells proliferate rapidly between days 4 and 7 following rechallenge (36, 37). In the rechallenged mixed chimeras, the IL-21R−/− cells are outcompeted during this proliferative phase. Therefore, the ability to perceive IL-21 can play some role in shaping recall CD8 T cell responses even though other CD4 T cell–dependent functions must be required. Interestingly, during chronic LCMV infections, the direct actions of IL-21 are obligatory for sustaining CD8 T cell responses (28, 29, 31). These discordant outcomes in acute and chronically infected IL-21−/− mice indicate that the precise roles and requirements for IL-21 likely vary with Ag load and the differentiation state of the responding cells. Important next steps will be to dissect how IL-21 integrates with other activatory and suppressive cytokines to influence the development of CD8 T cells capable of producing IL-2 and sustaining responses under conditions of more repetitive or continuous antigenic stimulation.
We thank Laurie Harrington and all of the members of Dr. Zajac’s laboratory for advice and critical reading of this manuscript. We also thank the University of Alabama at Birmingham Center for AIDS Research flow cytometry core facility for services.
Disclosures The authors have no financial conflicts of interest.
This work was supported by Grants R01 AI049360, R01 AI067993, and U01 AI082966 (to A.J.Z.) and T32 AI007051 (to J.S.Y.) from the National Institutes of Health.
Abbreviations used in this paper:
lymphocytic choriomeningitis virus
mean fluorescence intensity
recombinant Listeria monocytogenes expressing the LCMV GP33/GP34 epitopes.