Antigen Controls IL-7Rα Expression Levels on CD8 T Cells during Full Activation or Tolerance Induction¹

The IL-7R is expressed at high levels on resting naive and memory CD8⁺ T cells (1), and it plays an important role in promoting survival and homeostasis of the cells (2, 3, 4, 5, 6, 7, 8). The IL-7R is composed of the high-affinity IL-7Rα chain (CD127) and the common γ-chain (γ_c,³ CD132), which is shared with the receptors for IL-2, IL-4, IL-9, IL-15, and IL-21 (9). On naive T cells, IL-7Rα is regulated by its ligand, IL-7, in a manner that is unique for γ_c cytokines. Unlike most γ_c cytokines, which typically induce increased expression of their receptor, IL-7 transiently decreases expression of IL-7Rα (10). This may increase the probability that a greater number of cells will survive with a finite amount of IL-7 available in vivo, thus maintaining the size and diversity of the peripheral T cell pool and not selecting for a limited number of naive cells that have seen IL-7.

IL-7Rα expression levels are also regulated when CD8⁺ T cells respond to an antigenic challenge. Stimulated cells rapidly down-regulate IL-7Rα expression as they undergo clonal expansion and become effector cells (2, 11), whereas expression of receptors for IL-2 and IL-15 increases during this time. Considerable heterogeneity in IL-7Rα expression levels within the responding CD8 T cell populations are seen during the effector phase of responses to viruses or bacteria, including lymphocyte choriomeningitis virus (LCMV), EBV, CMV, HIV, and Listeria monocytogenes (11, 12, 13, 14, 15). Although most of the Ag-specific CD8 T cells are IL-7Rα^low at the peak of the response, a small fraction of the cells are IL-7Rα^high. The IL-7Rα^low cells typically produce high levels of IFN-γ and low levels of IL-2, whereas most IL-7Rα^high cells produce IL-2 but not IFN-γ (12, 13, 15). Kaech et al. (11) recently showed that it was the small population of IL-7Rα^high cells present at the peak of an LCMV response that was able to survive for the long term and become memory cells that can protect against secondary infection (11). These results suggested that high IL-7Rα expression marks the cells destined to become long-lived memory cells. Whether this small population represented cells that retained IL-7Rα at high levels throughout the response or that had down-regulated and subsequently re-expressed IL-7Rα was not clear. Subsequently, Lacombe et al. (16). found that a large fraction of the effector cells responding to peptide immunization expressed IL-7Rα at a high level at the peak of the response, but most did not survive to become memory cells. This suggested that high expression of IL-7Rα may be necessary, but not sufficient, to allow survival and conversion to memory. In contrast to these models where Ag is cleared, it has been shown that CD8 T cells responding to a chronic viral infection fail to develop a population of cells having high IL-7Rα expression (17, 18).

Definition of the signal(s) required for down-regulating IL-7Rα expression, and potentially for allowing for its re-expression, during an Ag-specific CD8 T cell response will be important to understand how this critical receptor contributes to memory formation. Previous work has suggested several possibilities, including IL-7 (10), IL-2 (19), Ag (2, 17, 18, 20), IFN-γ, and inflammation (21). Although, IL-7 can cause decreased IL-7Rα expression on resting cells (10), a recent study showed that it was not involved in regulation of IL-7Rα expression on responding CD8 T cells (22). Activation of naive CD8 T cells to clonally expand, develop effector functions, and form a memory population requires Ag, costimulation, and a third signal that can be provided by either IL-12 or type I IFN (23, 24, 25, 26, 27, 28, 29), making these signals likely candidates for involvement in IL-7Rα expression regulation. The results described herein demonstrate that as CD8 T cells respond to Ag, Ag-dependent signals are sufficient to regulate IL-7Rα expression on CD8 T cells stimulated to become effector and memory cells. This regulation is not unique to memory cells, however, as tolerant populations of CD8 T cells, generated by activation in the absence of adjuvant or signal 3, also re-express IL-7Rα equivalently and persist for the long term but fail to provide long-term protection. Thus, high IL-7Rα expression on long-lived CD8 T cells that have responded to Ag does not uniquely mark responsive memory cells.

OT-I mice (30) having a transgenic TCR specific for H-2K^b and OVA_257–264 were a gift from Dr. F. Carbone (University of Melbourne, Melbourne, Australia). OT-I mice were also crossed with Thy1-congenic B6.PL-Thy1a/Cy (Thy1.1) mice (The Jackson Laboratory, Bar Harbor, ME) and bred to homozygosity. C57BL/6NCr and B6-Ly5.2/Cr mice were obtained from the National Cancer Institute. CD45.1⁺ OT-I mice were a gift from Dr. E. Ingulli (University of Minnesota). K14-OVAp mice were a gift from Dr. K. Hogquist (University of Minnesota). All mice were maintained under specific pathogen-free conditions. Reagents used included biotinylated anti-mouse IL-15Rα Ab (R&D Systems), NeutrAvidin Cascade Blue (Invitrogen: Molecular Probes), anti-phospho-STAT5 Ab (BD Pharmingen), and anti-Bcl-2 Ab (BD Pharmingen). All other Abs were purchased from BD Pharmingen, eBioscience, or BioLegend. Neutralizing anti-mouse IL-2 Ab and anti-mouse IFN-γ Ab were obtained from eBioscience. Cells were cultured in RPMI 1640 medium with 10% FCS, 4 mM l-glutamine, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 U/ml penicillin and streptomycin, 10 mM HEPES, and 5 × 10⁻⁶ M 2-ME (RP-10).

CD8⁺CD44^low naive cells were enriched from lymph node (LN) cells by negative selection using MACS magnetic cell sorting (Miltenyi Biotec) (24) and were >95% CD8⁺ and <0.5% CD44^high. In some experiments, cells were labeled with CFSE (Molecular Probes). Cells were placed in microtiter wells coated with Ag (DimerX H-2K^b:Ig fusion protein loaded with OVA_257–264 peptide; BD Pharmingen) and, in some cases, recombinant B7-1/Fc chimeric protein (R&D Systems) as previously described (31). Purified CD8⁺ T cells (5 × 10⁴) were placed in the Ag- or Ag/B7-coated microtiter wells in 200 μl of RP-10. Where indicated, cultures were supplemented with 2 U/ml murine rIL-12 (Genetics Institute, Cambridge, MA), 2.5 U/ml human rIL-2 (TECIN; National Cancer Institute, Biological Resources Branch), 20 μg/ml neutralizing anti-IL-2 Ab (eBioscience), 20 μg/ml rat IgG1 isotype control Ab (eBioscience), 100 U/ml IFN-γ, 10 μg/ml neutralizing anti-IFN-γ Ab (eBioscience), or 10 μg/ml rat IgG1 isotype control Ab (eBioscience). Cells were harvested at the indicated times, and their numbers and phenotypes were determined by flow cytometric analysis.

CD8⁺CD44^low naive cells were purified from pooled LN cells as described above. CD8⁺ cells (1.5–3 × 10⁶) in 0.2 ml of HBSS were transferred via tail vein injection into age- and sex-matched 6- to 8-wk-old C57BL/6NCr or B6-Ly5.2/Cr mice. Recipient mice were then rested for 24 h before immunization. For immunization, 50 μg/mouse of OVA_257–264 peptide (SIINFEKL; ResGen) was injected via tail vein in 0.2 ml of HBSS. Where indicated, mice also received 1 μg of recombinant IL-12 (Genetics Institute) or 50 μg of LPS (Sigma-Aldrich) in the same injection. Mice were sacrificed for analysis at the indicated time after immunization, and numbers and phenotype of OT-I/PL cells were determined by flow cytometry. In experiments analyzing receptor expression 24 h after priming, LN and spleen were digested in collagenase D (Roche) for 25 min at 37°C to isolate activated T cells. In vivo killing assays (32) were performed as previously described in detail ((24) using CFSE-labeled C57BL/6 spleen cells pulsed with 0.2 μM of OVA_257–264 peptide as targets. LN and spleen cells from mice used for the in vivo killing assays were also stained with Abs to CD8 and Thy1.1 to determine the number of OT-I/PL cells in each tissue.

For some experiments, CD8⁺CD44^low OT-I/PL cells were cultured in vitro (as described) before transfer. At the end of the culture period, cells were harvested, washed twice in 37°C HBSS, and 1.5 × 10⁶ cells in 0.2 ml of HBSS were transferred via tail vein injection into age- and sex-matched C57BL/6NCr or K14-OVAp mice. Mice were sacrificed and perfused for analysis 48 h after in vivo transfer. Isolation of liver-infiltrating lymphocytes was performed as described previously by Calzascia et al. (33) using collagenase D digestion followed by a Percoll gradient to isolate lymphocytes.

To determine IL-7Rα expression on OT-I T cells during in vitro culture, cells were harvested at the indicated times, washed twice, and stained with PE-conjugated Ab to IL-7Rα (CD127) (eBioscience) for 30 min at 4°C. Stained cells were then were fixed in Cytofix buffer (BD Biosciences) for 15 min at 4°C before being analyzed on a FACSCalibur flow cytometer using CellQuest software (BD Biosciences). For analysis of OT-I cells in vivo, mice were killed at the indicated times after adoptive transfer and in vivo challenge. Spleen cells and LN cells (pooled from axillary, brachial, cervical, inguinal, and mesenteric nodes), as well as lung and liver infiltrating lymphocytes where indicated, were washed twice, stained with the Abs CD8 and Thy 1.1 or CD45.1 to detect the transferred OT-I/PL cells, and stained with PE-conjugated Ab to IL-7Rα (CD127) for 30 min at 4°C. Stained cells were then fixed and analyzed as described above.

Cells were cultured in vitro, harvested, washed twice with 37°C serum-free RMPI, resuspended in serum-free RMPI, and serum starved for 20 min at 37°C. Cells were then plated at 5 × 10⁴ cells/well and cultured in serum-free RMPI supplemented with 25 ng/ml IL-7 at 37°C for 30 min (phospho-STAT5) or 24 h (Bcl-2). Following restimulation with IL-7, cells were washed, fixed, permeabilized, and stained with anti-P-STAT5 (Invitrogen) according to the manufacturer’s instructions. Cells stained for Bcl-2 were fixed in Cytofix buffer for 15 min at 4°C and permeabilized in saponin-containing Perm/Wash buffer (BD Biosciences) for 15 min at 4°C before staining.

Recombinant Listeria monocytogenes expressing a secreted form of OVA (LM-OVA) (34) were a gift from Dr. S. Jameson (University of Minnesota). LM-OVA was grown in tryptic soy broth (BD Biosciences) supplemented with 50 μg/ml streptomycin sulfate to an OD of 0.1. 5 × 10⁴–1 × 10⁵ CFU in 0.3 ml of PBS were administered via tail vein injection into C57BL/6Ncr mice. The actual number of CFU injected was determined for each experiment by plate count. At the indicated times, mice were sacrificed and numbers of LM-OVA per tissue were determined. For quantitation of bacterial titers, spleen and liver were homogenized and lysed in 0.2% IGEPAL (Sigma-Aldrich) and dilutions were plated on tryptic soy broth agar supplemented with 50 μg/ml streptomycin sulfate and grown for 2 days at 37°C. Spleen and liver cells were also stained with Abs to CD8 and Thy1.1 to determine the number of OT-I/PL cells in each tissue, and some of the cells were permeabilized and stained for IFN-γ production in response to a 5-h in vitro pulse with 0.2 μM of OVA_257–264 peptide in the presence of GolgiStop (BD Biosciences).

IL-7Rα is expressed on resting CD8⁺ T cells but is down-regulated upon activation (2). In vitro stimulation of naive CD8⁺ T cells to clonally expand and develop effector functions requires three signals, Ag, costimulation, and a signal 3 cytokine (IL-12 or type I IFN) (23, 24, 25, 26, 27). To determine which of these signals contribute to down-regulation of IL-7Rα, naive OT-I CD8⁺ T cells (>95% CD8⁺, <0.5% CD44^high) specific for H-2K^b/OVA_257–264 (30) were stimulated in flat-bottom microtiter wells with various combinations of immobilized Ag, immobilized B7-1 ligand, and IL-12. Consistent with previous results, naive cells undergo multiple rounds of division by 72 h, as measured by CFSE dye dilution, in response to Ag alone, Ag with B7-1 (Ag/B7), or Ag/B7 + IL-12 (Fig. 1,A). However, despite extensive proliferation in all cases, strong clonal expansion only occurs when IL-12 is present to promote survival of the dividing cells (Fig. 1,B). When stimulated with Ag/B7 + IL-12, all of the cells rapidly down-regulate IL-7Rα expression, with a large decrease occurring within 6 h and minimal expression remaining by 12 h (Fig. 1,C). Comparable down-regulation was seen when cells were incubated with Ag alone, Ag/B7, or Ag/B7 + IL-12 (Fig. 1 D). In the absence of Ag, neither B7-1, IL-12, or their combination resulted in any IL-7Rα down-regulation (data not shown). Thus, IL-7Rα expression is uniformly down-regulated on virtually all naive CD8⁺ T cells, and this occurs very rapidly (before any significant cell division) and requires only signals that derive from recognition of Ag.

It has been recently been shown that IL-2 and IFN-γ can induce a slight down-regulation of IL-7Rα expression on naive resting T cells (10). It was therefore possible that the down-regulation of IL-7Rα during T cell activation was caused by IL-2 and/or IFN-γ produced during Ag-specific stimulation, even though Ag alone is a very suboptimal stimulus for production of these cytokines. This does not appear to be the case, however, since the addition of blocking anti-IL-2 and anti-IFN-γ Abs to in vitro cultures of OT-I cells stimulated with Ag did not inhibit down-regulation of IL-7Rα (data not shown), while clonal expansion at 72 h in response to Ag/B7 + IL-12 was severely compromised by addition of the blocking anti-IL-2 Ab (data not shown). Furthermore, the addition of IL-2 or IFN-γ to OT-I cells stimulated with Ag/B7 and IL-12 did not enhance IL-7Rα down-regulation, but it did result in a small increase in cell recovery at 72 h (data not shown). Thus, although there is the potential for these cytokines to be produced in response to Ag, they do not appear to influence the Ag-dependent down-regulation of IL-7Rα expression.

IL-7Rα down-regulation is one of several changes in cell-surface receptor expression that occurs rapidly upon Ag signaling. Within 24 h of contact with Ag, almost all of the naive cells have an IL-7Rα^low phenotype and up-regulated expression of CD69 and CD25 (data not shown). IL-15 can also contribute to T cell survival and homeostasis, and the expression level of IL-15Rα is increased on effector and memory cells in comparison to naive cells (35). When naive OT-I cells were stimulated with Ag/B7 and IL-12, the expression level of IL-15Rα increased within 12 h and was maximal by 24 h (data not shown). As was the case for IL-7Rα down-regulation, increased expression of IL-15Rα required only Ag-dependent signals, and expression levels did not increase further in the presence of B7-1 or IL-12 (data not shown).

At the peak of CD8⁺ T cell response to many viral pathogens, IL-7Rα expression on effector cells is heterogeneous such that there is a variable fraction of the cells that are IL-7Rα^high at the peak of the response (11, 12, 13, 14, 15). It is unclear whether these are cells that have down-regulated and then re-expressed the receptor or if they represent a subpopulation of cells that never down-regulate IL-7Rα, potentially being a distinct subpopulation of naive cells that are the memory precursors (11). The ability of Ag-dependent signals to cause rapid down-regulation of IL-7Rα on virtually all naive OT-I cells in vitro (Fig. 1) suggested that re-expression may occur and that cessation of Ag-dependent signaling might be sufficient to allow the re-expression.

The ability of cells to re-express IL-7Rα and the requirements for re-expression were examined by stimulating naive OT-I cells for varying times in vitro using Ag, Ag/B7, or Ag/B7 and IL-12. Cells were then harvested, washed, transferred to new wells that did or did not have immobilized Ag, and incubated, and IL-7Rα expression levels were determined. Cells stimulated with either Ag/B7 or Ag/B7 + IL-12 were found to be IL-7Rα^low within 12 h (Fig. 2,A), as expected. When these IL-7Rα^low cells were then removed from Ag, transferred into empty wells, and incubated for an additional 48 h, most celles re-expressed IL-7Rα, and levels were comparable regardless of whether IL-12 was present during the initial culture period (Fig. 2 A). When cells are exposed to Ag/B7 and IL-12 for only 12 h, subsequent clonal expansion is minimal (31) (and data not shown), thus ruling out selective expansion of a small population of IL-7Rα^high cells accounting for the re-expression.

To determine whether CD8⁺ T cells retain the ability to re-express IL-7Rα^high as differentiation and clonal expansion proceed, we examined the effects of progressively longer culture times on re-expression. Cells that were initially cultured for 24 or 48 h re-expressed IL-7Rα equally well when removed from Ag, and re-expression was the same in the presence or absence of IL-12 (data not shown). OT-I cells cultured with Ag/B7 and IL-12 for 72 h have developed effector functions, including cytolytic activity and the ability to produce IFN-γ upon restimulation. In initial experiments to examine whether IL-7Rα^high can be re-expressed on the effector cells, we found that some re-expression had occurred by 72 h, although its extent was variable (data not shown). In these cultures, however, strong clonal expansion had occurred (Fig. 1,A) and, owing to the large number of cells, many were no longer in contact with Ag on the bottom of the well. To avoid this, naive cells were cultured for 60 h in the presence of Ag/B7 and IL-12 and then transferred at a lower cell density per well to new wells having the same stimuli present. At 72 h the effector cells uniformly expressed a very low level of IL-7Rα and remained low when transferred into Ag-bearing wells (Fig. 2 B). Upon transfer into wells that lacked Ag, however, the effector cells re-expressed IL-7Rα within 24 h, similar to what was seen on cells during the earlier stages of differentiation. Additional experiments further demonstrated that re-expression was not affected by the addition of IL-2, IFN-γ, or anti-IL-2 and anti-IFN-γ Abs to the cultures (data not shown).

These results demonstrate that naive cells stimulated with Ag, B7-1, and IL-12 to become effector cells rapidly down-regulate IL-7Rα upon interacting with Ag, and that they rapidly re-express the receptor upon removal of Ag. Additionally, Ag-dependent down-regulation and re-expression occur when cells are stimulated with Ag and B7-1 in the absence of a third signal, where they fail to develop effector functions and, in vivo, are rendered tolerant for the long term (24, 36). To determine whether the IL-7R complex is functional on effector cells following re-expression in the absence of Ag, as well as on tolerant cells, naive OT-I cells were stimulated with Ag and B7 or with Ag/B7 and IL-12 for 72 h and then transferred to wells that did or did not have immobilized Ag. Twenty-four hours later the cells were then treated with IL-7 and analyzed for increased phospho-STAT5 and Bcl-2. As expected, IL-7Rα remained low if cells were re-cultured in the presence of Ag but was re-expressed in the absence of Ag (Fig. 3,A). Cells that were re-cultured in the absence of Ag and stimulated with IL-7 had increased levels of both phospho-STAT5 (Fig. 3,B) and Bcl-2 (Fig. 3 C). Moreover, the expression of phospho-STAT5 and Bcl-2 were equivalent regardless of whether IL-12 was present during the initial 72 h of culture. Thus, the IL-7Rα re-expressed on both effector and tolerant cells forms a functional signaling complex.

The results described above demonstrate that Ag-dependent signals alone control IL-7Rα expression levels when the minimal signals required for full activation of naive CD8⁺ T cells are examined in vitro. Whether this is also the case for in vivo responses where there is the potential for contributions from other receptor-ligand and/or cytokine interactions was examined using a peptide immunization model (37). OT-I T cells were adoptively transferred into normal C57BL/6 recipients, allowed to equilibrate for 1 day, and the mice were then challenged by i.v. injection of OVA_257–264 peptide alone or along with either IL-12 or LPS adjuvant. Twenty-four hours later, OT-I cells in the LN and spleen were analyzed for IL-7Rα expression. In mice that were not immunized, the transferred OT-I cells maintained an IL-7Rα^high phenotype (Fig. 4,A). In mice that received peptide alone or along with IL-12 or LPS, most OT-I cells had down-regulated IL-7Rα expression in both the LN (Fig. 4,A) and spleen (data not shown), and ∼70–80% of the cells were IL-7Rα^low. At this time the OT-I cells were blasting, as indicated by forward- and side-light scatter (data not shown), and they had up-regulated CD25 expression (Fig. 4 A). Thus, as predicted by the in vitro results, administration of Ag was sufficient to down-regulate IL-7Rα in vivo, and neither IL-12 nor LPS altered this response.

Similar results were obtained when OT-I cells from the immunized mice were examined on day 3, the peak of the response. The total numbers of OT-I cells in the LNs (Fig. 4,B) and spleens (data not shown) of primed mice demonstrated that IL-12 and LPS were acting effectively as adjuvants, as the OT-I cells expanded ∼30-fold in mice that received just peptide, whereas expansion when LPS or IL-12 was also administered was 70- and 120-fold, respectively. Nevertheless, most OT-I cells were IL-7Rα^low (Fig. 4 A) whether mice had received peptide Ag alone or along with IL-12 or LPS. In each case, however, ∼40% of the cells did express increased levels of IL-7Rα, consistent with the results that Lacombe et al. (16) obtained in experiments examining a similar model. In other experiments using CFSE dye-labeled OT-I cells for the adoptive transfer, we found that all of the cells had diluted CFSE by day 3 (data not shown). Thus, the cells expressing IL-7Rα on day 3 had responded to Ag by proliferating.

The increasing percentage of cells expressing IL-7Rα between days 1 and 3 suggested that the cells may have begun re-expressing the receptor, and therefore we examined this possibility during longer periods. Adoptively transferred OT-I cells undergo significant clonal expansion in response to peptide and IL-12, with the number of IL-7Rα^high cells peaking at day 3 and then declining to result in a small population by about day 13 that is stable and long lived (Fig. 5,A). By day 3, all of the OT-I cells are CD44^high and have completely diluted CFSE (data not shown), and ∼40% of the cells are IL-7Rα^high (Fig. 5,A). The percentage of IL-7Rα^high cells increased through day 17, when ∼90% of the remaining cells were IL-7Rα^high, yet the number of OT-I cells that are IL-7Rα^high declines dramatically during this time, demonstrating that not all of the cells that were IL-7Rα^high at the peak of the response survive for the long term. Nevertheless, it is possible that a fraction of IL-7Rα^high cells may preferentially migrate to peripheral tissues and become CD62L⁻CD127⁺ effector memory cells (38), which could result in decreased numbers of IL-7Rα⁺ cells in secondary lymphoid tissues. The presence of an increasing percentage of IL-7Rα^high cells between days 1 and 3 (Fig. 4,A) suggested that the level of Ag is declining and thereby allowing re-expression on the receptor. To estimate the kinetics of Ag clearance following priming, naive CD45.1⁺ OT-I T cells (indicator cells) were CFSE labeled and adoptively transferred at various times into primed mice that had originally received Thy1.1⁺ OT-I cells, recovered 3 days later, and CFSE levels were measured. With the amount of Ag used for priming in these experiments, all OT-I T cells (Thy1.1⁺ and CD45.1⁺) present at day 0 fully diluted their CFSE within 3 days (data not shown). When indicator cells were transferred on day 3, all had undergone some dilution of CFSE within 3 days, but it was not complete (Fig. 5 B), suggesting that Ag had declined significantly by this time. Indicator cells transferred after day 3 underwent successively less CFSE dilution with time, and by day 17 essentially all of the cells remained undivided, indicating that little or no Ag remained. Thus, the appearance of an increasing fraction of IL-7Rα^high cells correlates with clearance of Ag. This strongly suggests for early times (days 1–3) that loss of Ag signaling is allowing re-expression of the receptor. For longer periods, the continued increase in the fraction of IL7Rα^high cells likely reflects both continued decline in Ag signaling to allow re-expression, and selective survival of those cells that have re-expressed the receptor.

A large fraction of clonally expanded OT-I cells re-express IL-7Rα by day 6 postpriming, a time when the Ag load has substantially declined (Fig. 5). To determine whether IL-7Rα remains subject to Ag-dependent down-regulation at this time, and thus whether persisting Ag would keep the cells in an IL-7Rα^low state, we examined the effects of administering additional peptide Ag 6 days after the initial priming. OT-I cells were adoptively transferred and the mice were primed with peptide Ag alone, Ag and IL-12, or Ag and LPS. By day 6 postpriming the number of OT-I cells in the LN of mice primed with just Ag was ∼2-fold greater than in unprimed mice, while substantially greater numbers of OT-I cells were present when priming was with Ag along with either IL-12 or LPS, as expected. For all conditions, ∼70–80% of the OT-I cells were IL-7Rα^high at this time and were no longer blasting, as indicated by low forward scatter (Fig. 6,A). In contrast, when groups of mice received additional Ag on day 6 and were examined 24 h later, most cells had become IL-7Rα^low, and forward scatter indicated that they had begun to blast (Fig. 6 B). Thus, effector CTL resulting from priming with Ag and either IL-12 or LPS have re-expressed IL-7Rα at a high level by day 6, but again rapidly down-regulate the receptor upon reencountering Ag. Similarly, cells primed with Ag alone, that is, cells that fail to develop effector functions and are tolerant for the long term (36), re-express IL-7Rα by day 6 and down-regulate the receptor in response to Ag signals. Thus, persistence of Ag following the peak of the effector response can keep IL-7Rα expression low on the responding CD8 T cells.

Intravenous injection of peptide Ag results in Ag presentation to virtually all cells in the lymphoid system (37). During infections or tumor growth, however, Ag presentation may only occur in specific sites and could result in heterogeneity in IL-7Rα expression on the Ag-specific CD8 T cells at the various sites. To examine this possibility, we used K14/OVAp mice that express the OT-I OVA peptide epitope under the control of the human keratin 14 promoter that targets expression of the peptide to the thymus, as well as stratified epithelial cells of the skin, esophagus, and bile ducts of the liver (39, 40). Naive OT-I cells were activated in vitro with peptide Ag, B7.1 ligand, and IL-12 and cultured for 60 h. To ensure that all cells remained in contact with Ag and were thus IL-7Rα^low before transfer into recipient mice, the cells were then harvested, washed, and replated at lower cell density into new wells containing Ag, B7.1, and IL-12 and cultured for an additional 12 h. Cells were then washed in PBS and injected i.v. into CD57BL/6 or K14/OVAp mice. Forty-eight hours later, OT-I cells were collected from Ag-draining LNs (skin-draining: inguinal, axillary, brachial; liver-draining: hepatic), non-Ag-draining LNs (mesenteric), spleen, and liver, and the expression of IL-7Rα was evaluated by flow cytometry.

At the time of transfer, <5% of the OT-I cells were IL-7Rα^high (Fig. 7,A) and were uniformly CD25^high and CD69^high (data not shown). Forty-eight hours after transfer into C57BL/6 recipients, most (>80%) OT-I cells recovered at all sites had re-expressed the receptor and were IL-7Rα^high (Fig. 7, B and C). In contrast, a smaller fraction of the OT-I cells was IL-7Rα^high in the K14/OVAp mice at all sites. Most notably, only 40% of the OT-I cells in the livers (a site of Ag expression) were IL-7Rα^high, compared with 82% of OT-I cells in the livers of C57BL/6 mice. The fraction of IL-7Rα^high OT-I mice in the skin- and liver-draining LNs was also significantly lower in the K14/OVAp mice. Although Ag is not expressed directly in the LNs of the K14/OVAp mice, this difference could be explained by dendritic cells migrating from the site of Ag and presenting to the OT-I cells, as has been reported by Mayerova et al. (41). Differences between C57BL/6 and K14/OVAp mice were less dramatic in the nondraining LNs and spleens, but the IL-7Rα^high fractions were still significantly smaller in the K14/OVAp mice, possibly due to cells that have recently migrated from sites of Ag. Not only was a greater percentage of IL-7Rα^low cells found in the liver and draining LNs of K14/OVAp mice (Fig. 7,C), but OT-I cells were present in greater numbers in these same tissues (Fig. 7 D), possibly reflecting both sequestration of OT-I cells to the sites of Ag as well as increased survival in the presence of Ag. Taken together, these results suggest that in the presence of Ag, CD8 T cells do not re-express IL-7R, and thus cells responding to Ag remain predominantly IL-7Rα^low. However, when Ag is eliminated or cells migrate away from sites of Ag, IL-7Rα is re-expressed on CD8 T cells.

OT-I cells in mice primed with peptide Ag in the absence of IL-12 or adjuvant undergo some clonal expansion but fail to develop effector functions. Some of these cells survive for the long term and are tolerant, in that they do not make a productive response when rechallenged with a potent stimulus (24, 36). Cells that have responded to peptide challenge, in the presence and absence of adjuvant, re-express IL-7Rα at high levels within 6 days (Fig. 6), suggesting that the long-lived tolerant cells may be IL-7Rα^high, as are responsive memory cells (11). To examine this, naive OT-I T cells were adoptively transferred into recipient C57BL/6 mice, and groups of mice were either left unchallenged (naive) or immunized with peptide alone (tolerized) or peptide + IL-12 (full activation). Mice from each group were sacrificed at day 3, and the OT-I cells were examined. As expected, priming with peptide alone resulted in some clonal expansion, but the cells produced little IFN-γ and lacked in vivo cytolytic activity, whereas cells primed with peptide and IL-12 expanded more, produced IFN-γ, and had cytolytic activity (Fig. 8, A and B).

When OT-I cells from the mice were examined 45 days after priming, comparable numbers of cells were present in the LNs, spleens, and livers whether priming was with peptide alone or peptide and IL-12 (Fig. 8,C). Furthermore, the cells at all sites expressed high IL-7Rα levels, regardless of whether IL-12 was present during the priming (Fig. 8,D and data not shown). To confirm that the OT-I cells that resulted from priming with peptide alone were tolerant, groups of mice were challenged with LM-OVA, and the OT-I cells were examined 5 days later. Clonal expansion in the LNs, spleens, and livers was seen for naive and memory (peptide and IL-12) cells, but there was minimal expansion of OT-I cells from mice primed with just peptide (Fig. 8,E). In all cases, however, IL-7Rα expression was down-regulated on the OT-I cells in response to LM-OVA challenge (data not shown). Mice having cells primed with peptide alone (tolerant) also had higher numbers of LM-OVA CFU in both spleens and livers compared with mice having naive or memory (peptide- and IL-12-primed) OT-I cells (Fig. 8 F), suggesting that peptide-primed cells remained tolerant for the long term. Thus, resting long-lived tolerant cells express high IL-7Rα levels, comparable to the expression on resting memory cells. Furthermore, as on naive cells, the IL-7Rα on memory and tolerant cells is again down-regulated upon re-exposure to Ag.

IL-7 plays an important role in promoting survival of resting naive and memory T cells, as well as establishing memory cells following the effector phase of a response (2, 3, 4, 5, 6, 7, 8, 42). The high-affinity chain of the receptor, IL-7Rα, is uniformly expressed at high levels on naive CD8 T cells, but it is lost from a large fraction of stimulated cells as they undergo clonal expansion and become effector cells (2, 11). CD8 T cells responding to pathogens exhibit variable heterogeneity in IL-7Rα expression at the peak of the effector phase, but most of those that survive as long-term memory cells express high IL-7Rα (11, 12, 13, 14, 15). How IL-7Rα expression is regulated on CD8 T cells and why only a fraction of Ag-specific effector cells express IL-7Rα have been unclear. Dissecting the potential contributions of signals from Ag, costimulatory ligands, and cytokines is difficult during an in vivo response to pathogens that induce a plethora of cytokines and changes in surface ligand and receptor expression on APC. Indeed, IL-7 (10), IL-2 (19), IFN-γ, and inflammation (21) have all been implicated as having roles in regulating IL-7Rα expression.

Examination of highly purified naive CD8 T cells responding to immobilized Ag and costimulation in vitro allowed the clear demonstration that Ag-dependent signals are sufficient to rapidly down-regulate IL-7Rα on virtually all of the cells (Fig. 1). Furthermore, removal of cells from Ag following IL-7Rα down-regulation resulted in re-expression on most cells within 2 days (Fig. 2). This is in agreement with the results of Li et al. (20), who found that CD4 T cells uniformly re-expressed IL-7Rα when removed from Ag. Re-expression upon cessation of Ag-dependent signals also appears to be the case in vivo: the majority of Ag-specific CD8 T cells rapidly down-regulate IL-7Rα upon administration of Ag, but most re-express the receptor at the peak of the effector response and beyond (Fig. 4). Re-expression correlates with declining Ag levels, and down-regulation and re-expression are not affected by the presence of adjuvant to activate DC and induce inflammation (Figs. 4 and 5).

These results strongly suggest that TCR engagement is sufficient to cause IL-7Rα down-regulation, and that cessation of TCR signaling is sufficient to allow re-expression to occur. This conclusion is further supported by experiments demonstrating that effector cells that have re-expressed IL-7Rα again rapidly down-regulate the receptor when the cells reencounter Ag (Figs. 6 and 7). Thus, these results are consistent with the finding that IL-7Rα expression remains low on virus-specific CD8 T cells during a persistent viral infection (17, 18), where frequent encounter with persisting Ag would prevent re-expression. Our in vivo and in vitro results indicate that after a short exposure to Ag (72 h), most cells have the potential to re-express IL-7R. In contrast, Kaech et al. (11) found that when IL-7Rα^low cells from LCMV-infected mice were isolated between 8 and 15 days after infection and transferred into Ag-free mice, they failed to re-express the receptor. CD8 T cells that persist in the presence of chronic Ag are IL-7Rα^low and also fail to re-express the receptor when transferred into Ag-free hosts (17). These results suggest the possibility that the duration of Ag stimulation may affect the ability to re-express IL-7Rα, with prolonged exposure resulting in exhaustion of the ability to re-express the receptor. Further work examining the relationship between the duration of Ag exposure and IL-7Rα re-expression will be needed to determine whether this is the case.

Although our results suggest that Ag-dependent signals to the CD8 T cell control IL-7Rα expression levels, results from studies examining in vivo responses to pathogens have suggested that cytokines or other factors may also regulate expression of the receptor. It is not clear, however, whether these act directly by delivering signals to the CD8 T cells to lead to altered IL-7Rα expression, or indirectly by influencing the level and duration of Ag presentation. In two such studies, Takemoto et al. (43) and Badovinac et al. (21) demonstrated that in IL-12^−/− and IFN-γ^−/− mice, respectively, a higher fraction of effector cells express IL-7R compared with cells in wild-type mice. Although these results might suggest that IL-12 and IFN-γ can regulate IL-7Rα expression, we did not see effects of either IL-12 or IFN-γ on IL-7Rα down-regulation and re-expression under conditions where the cytokines could signal directly to the T cells. Thus, it may be the case that IL-12 and IFN-γ influence IL-7Rα indirectly by altering the presentation of pathogen-derived peptides, irrespective of the clearance of live virus or bacteria, such that in the absence of these cytokines a lower number of peptide/MHC I complexes are available to effector cells, thus permitting larger numbers of IL-7Rα-expressing cells. This would be consistent with the role of IFN-γ in enhancing multiple components of the MHC I pathway involving both MHC I expression and stability and Ag processing and loading (reviewed in Refs. (44) and (45)). Only detailed studies in specific pathogen infections will reveal whether signals to the CD8 T cell other than those from the TCR might also modulate IL-7Rα expression levels.

During the course of an LCMV infection, a small fraction of virus-specific CD8 T cells expresses high IL-7Rα at the peak of the response. It was shown that these were the cells capable of becoming functional memory cells, leading to the suggestion that expression of IL-7Rα on effector cells marks those cells that will become memory cells (11). However, subsequent work has shown that in some responses a large fraction of the effector cells can express high IL-7Rα, but the most of these cells do not survive the contraction phase to become memory cells. Lacombe et al. (16) demonstrated this in a peptide immunization model, and our results confirm their findings (Fig. 5). Similarly, Badovinac et al. (46) showed that ∼70% of CD8 effector cells at the peak of the response to peptide-DC immunization expressed high IL-7Rα, but most of these cells were lost during the contraction phase. In examining CD8 T cells responding to Vaccinia virus, we found that 30–60% of the cells expressed high IL-7Rα at the peak of expansion, and, again, most of these IL-7Rα^high cells did not survive the contraction phase (47). Thus, while expression of IL-7Rα may be required for progression of effector cells to memory, it is not sufficient for survival and progression.

CD8 T cells that respond to Ag in the absence of adjuvant undergo some clonal expansion but fail to develop effector functions, and those that survive for the long term are tolerant (24, 48, 49, 50, 51, 52). This is due, at least in part, to the absence of a required ‘third signal’ that can be provided by IL-12 or type I IFN (25, 26), and IL-12 can replace the need for adjuvant in supporting a productive response to peptide Ag in vivo (24, 36). Despite being tolerized upon administration of only peptide Ag (Fig. 8), the responding CD8 T cells still down-regulate and re-express IL-7Rα comparably to cells that have been fully activated by IL-12 or adjuvant (Figs. 4 and 6). In the absence of a third signal dependence on IL-7Rα expression, this would be expected, as the frequency of Ag encounters by tolerant and fully activated cells should decrease as soluble Ag is cleared. Moreover, the ability of tolerant and effector cells to comparably up-regulate Bcl-2 in response to IL-7 (Fig. 3) likely contributes to the long-term survival of the tolerant cells. Thus, high IL-7Rα expression on Ag-experienced cells cannot be interpreted as reliably marking responsive memory cells.

The results described herein suggest a model in which naive CD8⁺ T cells rapidly down-regulate IL-7Rα expression upon encountering Ag, undergo clonal expansion, and re-express the receptor when they are no longer receiving an Ag signal. When Ag clearance is rapid and a large fraction of the effector cells re-expresses IL-7Rα at the peak of clonal expansion, most of them die during the contraction phase, with only a small fraction remaining for the long term as IL-7Rα^high memory cells. Numerous signals are likely to be involved in determining long-term memory survival of effector CD8 T cells, and there is considerable evidence to suggest that IL-7 may provide one of these signals. Given the results described herein, it would appear that the IL-7-dependent component of signaling for memory may involve both instructive and stochastic processes, with cessation of TCR signals “instructing” the cells to re-express IL-7Rα at high levels. If large numbers of effectors re-express IL-7Rα at the peak of expansion, there may then be a “stochastic” process involving competition for the limited amounts of IL-7 present in vivo, so that only some of the cells receive an effective IL-7 signal. Because IL-7Rα expression is regulated by just Ag-dependent signals, cells that have responded to Ag in the absence of a third signal (IL-12, adjuvant) and become tolerant still re-express IL-7Rα, and this may also promote their long-term survival. Thus, high IL-7Rα expression on Ag-experienced cells cannot be interpreted as reliably marking responsive memory cells.

We thank Debra Lins for expert technical assistance and Dr. Marc Jenkins for critical reading of the manuscript.

The authors have no financial conflicts of interest.