Dynamic Regulation of IFN- Signaling in Antigen-Specific CD8+ T Cells Responding to Infection1
http://www.100md.com
免疫学杂志 2005年第11期
Abstract
IFN- plays a critical role in the CD8+ T cell response to infection, but when and if this cytokine directly signals CD8+ T cells during an immune response is unknown. We show that naive Ag-specific CD8+ T cells receive IFN- signals within 12 h after in vivo infection with Listeria monocytogenes and then become unresponsive to IFN- throughout the ensuing Ag-driven expansion phase. Ag-specific CD8+ T cells regain partial IFN- responsiveness throughout the contraction phase, whereas the memory pool exhibits uniform, but reduced, responsiveness that is also modulated during the secondary response. The responsiveness of Ag-specific CD8+ T cells to IFN- correlated with modulation in the expression of IFN-R2, but not with IFN-R1 or suppressor of cytokine signaling-1. This dynamic regulation suggests that early IFN- signals participate in regulation of the primary CD8+ T cell response program, but that evading or minimizing IFN- signals during expansion and the memory phase may contribute to appropriate regulation of the CD8+ T cell response.
Introduction
Almost immediately after pathogenic insult, numerous cytokines and chemokines are released into tissues. One of the functions of this protein milieu is to assist in a rapid, direct counterattack mediated by the innate immune system against the invading organism. Another important role is to initiate and foster the propagation of the appropriate adaptive immune response. This interplay involves multiple signals delivered to immune cells via cytokines and cell-cell contacts. Optimally, an orchestrated response to these signals results in the appropriate immune response to clear the specific infection.
IFN- is one of the most well-studied proinflammatory cytokines produced by innate and adaptive immune cells. The importance of IFN- is evident in mice that have been made genetically deficient for the cytokine or its receptor components. These mice are profoundly susceptible to many intracellular pathogens, including Listeria monocytogenes, Leishmania major, and mycobacteria (1, 2, 3, 4, 5). IFN- is primarily made by NK cells and activated T cells (6, 7, 8) and functions as a noncovalently linked homodimer (9). The receptor for IFN- consists of two chains, IFN-R1, the ligand-binding portion, and IFN-R2, which is required for signaling. Binding of IFN- to IFN-R1 causes oligomerization of the receptor chains, which are constitutively associated with JAK1 and -2, but not normally preassociated with each other (10, 11). Once in proximity, the JAKs transphosphorylate and activate each other as well as phosphorylate the IFN-R1 chains on residue Tyr440. This creates a pair of docking sites for the latent cytosolic transcription factor STAT1, which is recruited to the receptor and phosphorylated on Tyr701 by the activated JAKs. Once phosphorylated, STAT1 subunits quickly dissociate from the receptor, translocate into the nucleus, and initiate transcription by binding to IFN--activated site DNA sequences (reviewed in Ref.10). Phosphorylation of STAT1 by protein kinase C on Ser727 before nuclear entry is important for optimal transcriptional activity (12, 13).
IFN- signaling is under tight control. Binding of IFN- to its receptor causes internalization and dissociation of the complex (14). In most cells, including murine macrophages, internalized IFN- is quickly degraded, and IFN-R1 is efficiently recycled back to the cell surface (15). Intracellular levels of phospho-STAT1 peak 15–30 min after in vitro IFN- stimulation, followed by a rapid decline to undetectable levels within 1–2 h (16). It has been demonstrated in HeLa cells that activated STAT1 homodimers are removed via ubiquitination and degradation by the proteosome (16). In addition to these control mechanisms, the Src homology 2 domain-containing protein suppressor of cytokine signaling-1 (SOCS-1)3functions as a negative feedback regulator of IFN- signaling (17). Expression of SOCS-1 in cell lines is under the control of several STATs, including STAT1, and SOCS-1 mRNA is quickly induced upon IFN- signaling. SOCS-1 binds to and inhibits the catalytic activity of JAKs, which attenuates IFN- signaling at a step proximal to STAT1 activation (18, 19, 20).
IFN- can regulate the expression of hundreds of genes, mostly, but not exclusively, via the JAK-STAT pathway (11, 21). Exposure to this cytokine can lead to a wide range of cellular responses in immune and nonimmune cells. Data concerning the effect(s) of IFN- specifically on T cells has been almost exclusively derived from in vitro studies. It has been shown that in vitro polarized Th1 CD4+ T cell lines are unresponsive to IFN- due to down-regulation of IFN-R2. This is thought to be a result of exposure to IFN-, because Th2 CD4+ T cells, which normally express IFN-R2, also down-regulate this receptor component when cultured with IFN-. In these experiments, surface protein expression of IFN-R1 was constitutive on both Th1 and Th2 CD4 T cells (22, 23). In contrast to murine CD4+ T cell clones, human T cells appear to regulate the expression of IFN-R2 in a ligand-independent manner (24). Resting and PHA-stimulated human CD3+ T cells maintain large cytoplasmic stores of IFN-R2, whereas surface expression remains low. The intracellular pools of IFN-R2 are the result of constitutive recycling of IFN-R2 between the cell surface and the cytoplasm. This process has been demonstrated to be ligand independent, because recycling of IFN-R2 still occurs in the absence of surface IFN-R1 or in the presence of neutralizing Abs for IFN- (24). Allospecific CD8+ T cell lines also maintain expression of IFN-R1, but are unresponsive to IFN- due to down-regulation of IFN-R2 at the mRNA level (25). In the absence of IFN- or STAT1, activated CD4+ T cells fail to up-regulate the expression of caspases 3, 6, and 8 despite the expression of normal levels of Fas and Fas ligand, suggesting that both IFN- and STAT1 are critical for activation-induced cell death of activated CD4+ T cells (26).
In addition to inducing apoptosis, IFN- can inhibit normal cell cycle progression in many cell types. When cultured with IFN-, bone marrow-derived macrophages arrest in G1/S phase (27). In a more detailed study using a murine macrophage cell line, it was shown that arrest at this cell cycle transition was due to an accumulation of the cyclin-dependent kinase inhibitor p27Kip1 (28). These studies suggest that active evasion of IFN- signals may allow T cells to escape the apoptotic and antiproliferative effects of this cytokine.
Despite intense investigation of the general cellular effects of IFN- and the aforementioned in vitro studies with T cell clones, we know little about how this cytokine directly affects T cells in vivo during an immune response. Mice that were engineered to constitutively express IFN-R2 (IFN-R2 transgenic (Tg)) mice were unable to mount productive Th1 immune responses to L. monocytogenes or Leishmania, and thus resembled IFN--deficient mice (29). Allospecific CD8+ T cell lines made from these mice had impaired cytotoxic capabilities in vitro despite being able to make IFN- and proliferate in response to Ag (25). In another study, T cells activated with the bacterial superantigen staphylococcal enterotoxin B exhibited a decreased ability to phosphorylate STAT1 in response to IFN- stimulation early after activation. This decreased responsiveness was interpreted to be independent of receptor down-regulation (30). These data combined with the aforementioned in vitro studies suggest that regulation of IFN- responsiveness may be required for normal T cell function.
IFN- mRNA is elevated very early after infection with L. monocytogenes (31), when T cells are first being activated, and recent studies suggest that this early IFN- production may control the eventual contraction program of CD8+ T cells (32). Consistent with this idea, activated CD4+ T cells persist in IFN- deficient B6 mice after infection with Mycobacterium bovis bacillus Calmette-Guérin or induction of experimental autoimmune encephalomyelitis (33, 34). Similarly, Ag-specific CD8+ T cells in IFN--deficient BALB/c mice undergo normal expansion, but exhibit a dramatically prolonged contraction phase after infection with L. monocytogenes or lymphocytic choriomeningitis virus (35). In addition, the hierarchy of epitope dominance was altered in the absence of IFN-. It is not known whether IFN- directly affected Ag-specific T cells during the immune response, or if its role was via indirect mechanisms (36).
In summary, these data suggest that IFN- may play a key role in regulating T cell responses in vivo. In the current study we investigated if, how, and when Ag-specific CD8+ T cells altered their responsiveness to IFN- in vivo during primary and secondary responses to infection with L. monocytogenes. The results reveal a dynamic pattern of IFN- responsiveness in Ag-specific CD8+ T cells during expansion and contraction that coincides with alterations in receptor component expression. This pattern of responsiveness is consistent with early exposure to the cytokine and an important role for IFN- in programming of the subsequent CD8+ T cell immune response.
Materials and Methods
PCR analyses
cDNA made from purified T cells and whole splenocytes was used as the template in probe-based, real-time PCR to assess the expression of IFN-R1 (data not shown), IFN-R2, and SOCS-1. Amplification of GAPDH was used as a control. Primer-probe sets for the targets of interest were designed using Primer Express software version 1.5 and were synthesized by IDT. These probes were labeled with the reporter dye FAM and the quencher dye TAMRA. GAPDH reagents were purchased from Applied Biosystems. The GAPDH probe was labeled with the reporter dye VIC and the quencher dye TAMRA. TaqMan universal PCR master mix (Applied Biosystems) was used for all reactions. All experiments were performed using an ABI PRISM 7700 sequence detection system (Applied Biosystems). The relative amplification of each unknown to GAPDH at different times p.i. was directly compared with the relative expression of that same target to GAPDH in purified naive or memory donor OT-1.PL Tg T cells. The data presented were obtained using the cycle thresholds to calculate 2–CT, which is the amount of target (unknown), normalized to the endogenous reference (GAPDH) and relative to a calibrator (naive OT-1.PL or memory donor cells).
Amplification of cDNA for IFN regulatory factor-2 (IRF-2) and TAP-2 from purified IFN-R1–/– and wild-type (wt) OT-1s was performed using RT-PCR. GAPDH was amplified together with each target gene and was used to normalize expression levels. Appropriate bands on ethidium bromide-stained gels were quantitated using ImageQuant 3.3.
Results
Response kinetics of Ag-specific CD8+ T cells after infection
An adoptive transfer system using Thy1.1 congenic OT-1 TCR Tg T cells (37) specific for OVA257 (herein called OT-1s) was established to investigate IFN- responsiveness in Ag-specific CD8+ T cells after bacterial infection. This system was instituted to allow detection of Ag-specific CD8+ T cells very early after infection and to facilitate purification of these cells without the use of MHC class I/peptide tetramers, which have the potential to affect intracellular signaling events (43). After transfer to C57BL/6 (Thy1.2) hosts, the OT-1s were activated by infection with L. monocytogenes that had been engineered to express OVA (LM-OVA) (38).
Preliminary experiments were performed to determine the number of OT-1s that could be adoptively transferred and still exhibit expansion and contraction after infection with kinetics similar to those of the endogenous CD8+ T cell response. CD8+-enriched splenocytes from OT-1 mice were transferred into recipient B6 mice 1 day before infection with 107 actA– LM-OVA. The results from these preliminary experiments (data not shown) indicated that after adoptive transfer of 5 x 104 or fewer OT-1s, the response kinetics most resembled the endogenous OVA257-specific CD8+ T cell response in B6 mice after actA– LM-OVA. With the addition of OT-1s, we achieved a 5-fold increase in the total number of memory OVA257-specific CD8+ T cells compared with nonadoptive transfer recipients (Fig. 1B). Attenuated actA– LM-OVA was used so that a sufficiently high dose of bacteria could be delivered to the mice to ensure activation of all adoptively transferred transgenic T cells, which was confirmed by CFSE dilution studies (data not shown). When OT-1s were analyzed at different times p.i. for functional activation, virtually all Thy1.1+ OT-1s responded to stimulation with cognate peptide (OVA257) by making IFN-, as detected by intracellular cytokine staining (ICS). An example of ICS for IFN- on day 10 p.i. is shown in Fig. 1A. Thus, the Thy1.1 marker alone is a suitable way to enumerate functionally activated OT-1s after infection.
Intracellular staining for phospho-STAT1 is a sensitive method for measuring IFN- responsiveness
If and how IFN- affects CD8+ T cells during an immune response are currently unknown. Because phosphorylation of STAT1 is a proximal step in the IFN- signaling pathway, the presence of this activated transcription factor in a cell after stimulation with IFN- indicates competence to receive a signal initiated via the binding of IFN- to its receptor. Typically, Western blotting is used to assess the phosphorylation status of STAT1 in cytokine-stimulated cells. However, Western blotting provides data on a whole population and cannot distinguish between global changes in IFN- responsiveness and the presence of subpopulations of responsive or unresponsive cells. In addition, Western blotting requires purification of a relatively large number of cells, which would not be possible very early after infection, when numbers of Ag-specific T cells are low. To overcome these limitations, we chose to use intracellular staining for phospho-STAT1 to assess the ability of Ag-specific CD8+ T cells to respond to a short in vitro stimulation with IFN-.
We performed preliminary experiments to determine the resolution provided by flow cytometric detection of intracellular phospho-STAT1 in IFN--stimulated T cells. CD8+-enriched splenocytes from naive B6 mice were stimulated with IFN- for 20 min, which was the peak of STAT1 phosphorylation as determined in time-course experiments (42) (data not shown). After stimulation, parallel samples of cells were immediately lysed for Western blotting or fixed, permeabilized, and stained for intracellular phospho-STAT1. Identically cultured, unstimulated splenocyte samples were lysed or stained as controls. To compare the sensitivities of the techniques, IFN--stimulated and unstimulated samples were mixed in specific ratios before blotting or staining. As shown in Fig. 2A, phospho-STAT1 was only detected in CD8+ T cells stimulated with IFN- (lanes 1 and 2) on a Western blot. We were readily able to detect phospho-STAT1 when the lysate contained 20% IFN--stimulated cells, but phospho-STAT1 was only variably detected when the lysate contained 10% stimulated IFN- cells (Fig. 2A, lanes 5 and 6).
In comparison, naive CD8+ T cells were highly responsive to IFN-, as detected by a uniform shift in staining with anti-phospho-STAT1 compared with unstimulated cells (Fig. 2B). Isotype control staining of stimulated and unstimulated cells was similar to anti-phospho-STAT1 staining in unstimulated cells and was omitted from the figure for clarity. As the frequency of IFN--stimulated cells in mixed samples decreased, the portion of the peak that was shifted compared with the unstimulated cell peak decreased (Fig. 2B). Importantly, it was still possible to observe positive signal when only a small fraction of the cells were stimulated with IFN-. These data validate intracellular staining for phospho-STAT1 as a viable technique to measure responsiveness to IFN- in T cells and highlight the unique ability of this method to resolve subpopulations of cells with different IFN- responsiveness.
Ag-specific CD8+ T cells become unresponsive to IFN- very early after primary infection and do not regain responsiveness until expansion is completed
As shown in Fig. 2B, naive CD8+ T cells from wt B6 mice are highly responsive to IFN-. Naive OT-1s exhibited the same high degree of responsiveness as polyclonal CD8+ T cells from naive B6 mice (Fig. 3C, first panel). In contrast, OT-1s obtained 3 days after actA– LM-OVA infection were unable to phosphorylate STAT1 after IFN- stimulation (Fig. 3B). OT-1s remained entirely unresponsive to IFN- until day 5 p.i., the time when the cells began to transition from the expansion to the contraction phase (Fig. 3, A and B). As cells entered the contraction phase, a fraction of Ag-specific CD8+ T cells reacquired the ability to respond to IFN-. On day 7 p.i., 5–7% of the expanded OT-1s stained positively for phospho-STAT1 after stimulation with IFN-. This population of IFN--responsive cells increased in proportion as the number of Ag-specific CD8+ T cells dropped. On day 8 p.i., 10–15% of the Tg T cells were positive for phospho-STAT1 after IFN- stimulation, and on day 9 p.i., this frequency increased substantially to 40–50% (Fig. 3B). Both early (day 30 p.i.) and late (day 101 p.i.) memory OT-1s retained the ability to respond to IFN-; however, the levels of phospho-STAT1 were reduced compared with those of naive OT-1s (Fig. 3C). These data indicate that despite being highly responsive to IFN- in their naive state, Ag-specific CD8+ T cells rapidly lose responsiveness to IFN- after infection with actA– LM-OVA and remain unresponsive throughout their expansion phase. During contraction, cells that exhibit some degree of responsiveness to IFN- accumulate to seed the memory population.
Before day 3 p.i., we were unable to detect OT-1s in the spleens of infected mice that had received adoptive transfer of 5 x 104 or fewer Tg T cells (data not shown). To study Ag-specific CD8+ T cells very early after infection, it was necessary to transfer greater numbers of OT-1s. In experiments performed through day 2 p.i., recipient B6 mice received 5 x 105 Tg T cells. Parallel CFSE dilution studies showed that all the OT-1s were recruited after infection with 107 actA– LM-OVA (data not shown). Strikingly, within 12 h after infection, OT-1s completely lost the ability to phosphorylate STAT1 in response to IFN- and remained unresponsive through day 2 p.i. (Fig. 3C). In combination, these data show that Ag-specific CD8+ T cells rapidly become unresponsive to IFN- within 12 h after infection, but regain responsiveness 5–7 days later. Thus, Ag-specific CD8+ T cells exhibit dynamic IFN- responsiveness during the primary immune response in vivo.
Expression of IFN-R1 does not correlate with IFN- responsiveness
IFN-R1 (CD119) is the ligand-binding portion of the IFN-R. Expression of IFN-R1 is constitutive on most cells (14). However, it has been demonstrated that IFN-R1 is transiently down-regulated in Tg CD4+ T cells after exposure to Ag or TCR ligation with Ab in vitro (44). To uncover the mechanism by which Ag-specific CD8+ T cells alter their responsiveness to IFN- in vivo, we investigated receptor expression on these cells at different times after infection.
As shown in Fig. 4A, naive OT-1s express high levels of surface IFN-R1, as do OT-1s adoptively transferred into mice that were subsequently left uninfected. Splenocytes from adoptive transfer recipients were harvested and analyzed directly ex vivo at different times p.i. for expression of CD8, Thy1.1, and IFN-R1. In preliminary experiments, at 12 and 24 h p.i. it appeared that most OT-1s had extensively down-regulated surface expression of IFN-R1 (Fig. 4A). However, the Ab used to detect IFN-R1 (clone GR20) can be blocked by bound IFN- (45, 46). IFN- bound to IFN-R1 can be released via stripping with a short incubation in very low pH medium (47). Acid stripping of OT-1s isolated from naive mice did not substantially alter the level of detectable surface IFN-R1 (pre-acid treatment mean fluorescence intensity (MFI), 672; after acid treatment MFI, 679; Fig. 4A). When splenocytes from OT-1 recipient mice 12 and 24 h p.i. were treated with acidic medium, followed immediately by staining for IFN-R1, we detected IFN-R1 on the surface of OT-1s, although surface expression was reduced compared with expression on naive OT-1 Tg T cells (naive MFI, 679; 12 h p.i. MFI, 517; 24 h p.i. MFI, 511; Fig. 4A). Similar results were obtained when OT-1s were stained with the anti-IFN-R1 Ab 2E2, which is not blocked by bound IFN- (23). Surface levels of IFN-R1 on OT-1s, as detected by 2E2, were decreased at 12 and 24 h p.i., but not at 48 h p.i. (data not shown). In addition, mRNA for IFN-R1 (quantitated by real-time PCR) was decreased in OT-1s at 12 h p.i. compared with cDNA samples made from naive OT-1s and cells purified from mice 48 h p.i. (data not shown).
On days 3 and 5 p.i., surface expression of IFN-R1 remained relatively high, even without acid stripping (Fig. 4B) and returned to the same level as that observed in naive T cells by day 7 p.i. At all time points analyzed after day 7 p.i., including days 8–10 and out to day 283 p.i., surface expression of IFN-R1 remained high (data not shown and Fig. 4B). These data indicate that OT-1s do not completely lose surface expression of IFN-R1 and express high levels of IFN-R1 during the expansion phase when they are unresponsive to IFN-. Thus, down-regulation of IFN-R1 is not the mechanism by which Ag-specific CD8+ T cells are rendered unresponsive to IFN- after L. monocytogenes infection.
The data presented in Fig. 4A demonstrate that IFN-R1 expressed on Ag-specific CD8+ T cells was almost completely bound by IFN- at 12 and 24 h p.i. Occupancy of the receptor may prevent these cells from responding to additional stimulation with IFN- in vitro before intracellular staining for phospho-STAT1 (Fig. 3C). To address this possibility, Ag-specific CD8+ T cells were subjected to acid stripping before in vitro IFN- stimulation and phospho-STAT1 staining. As shown in Fig. 4C, even after surface bound IFN- was removed, Ag-specific CD8+ T cells were still unable to respond to exogenously added IFN- by phosphorylating STAT1. The acid stripping treatment did not alter the ability of naive CD8+ T cells to respond to IFN- stimulation. It remains possible that as a consequence of exposure to IFN- in vivo, these cells have exhausted their ability to respond to this cytokine, perhaps through down-regulation of total levels of STAT1; however, the data support the conclusion that Ag-specific CD8+ T cells lose the ability to respond to IFN- by 12 h p.i.
The early blocking of surface IFN-R1 by IFN- on OT-1s was coincident with the expression of IFN- mRNA in the spleen (Fig. 4D). The expression of IFN- mRNA, detected by probe-based, real-time PCR, was dramatically increased at 12 and 24 h p.i. compared with that in naive mice (Fig. 4D). IFN- mRNA levels returned to baseline on day 2 p.i. and remained at the same levels as in naive mice throughout the remainder of the expansion and contraction phases of the CD8+ T cell response. Similarly, IFN- protein was readily detected in the serum of infected mice 12 and 24 h p.i., but was below the limit of detection by 48 h p.i. (Fig. 4E). No IFN- protein was detected in the serum of naive mice. These data suggest that the modest down-regulation of IFN-R1 observed 12 and 24 h p.i. was ligand induced. Although we cannot rule out that the acid stripping may not have removed all bound IFN- from the IFN-R1, similar results were obtained using an anti-IFN-R1 Ab (2E2) that was not blocked by the presence of IFN- (23), and mRNA for IFN-R1 was also decreased in OT-1s 12 h p.i. (data not shown). Internalization of a fraction of IFN--bound IFN-R1 could also be an explanation for the observed decrease in surface expression of IFN-R1; however, it has been demonstrated that Th1-polarized CD4+ T cell clones maintain detectable levels of surface IFN-R1 during long-term culture with IFN- (23). Thus, not all IFN-R1 is internalized during IFN- signaling. Alternatively, recycling of IFN-R1 back to the surface of cells after internalization is a very efficient process and could contribute to the amount of detectable surface receptor.
In summary, Ag-specific CD8+ T cells are exposed to IFN- early after infection with L. monocytogenes, and these cells undergo a transient, most likely ligand-induced, down-regulation of surface expression of IFN-R1. However, IFN-R1 was found to be uniformly expressed by these cells at times p.i. (days 5 and 7) when they remained largely unresponsive to IFN-. Therefore, there must be an additional mechanism(s) by which Ag-specific CD8+ T cells lose IFN- responsiveness during expansion.
Ag-specific CD8+ T cells regain responsiveness to IFN- upon re-expression of IFN-R2
Because altered expression of IFN-R1 by Ag-specific CD8+ T cells did not strictly correlate with their ability to respond to IFN-, we next investigated the expression of IFN-R2 at time points throughout the primary immune response. IFN-R2 is required for downstream signaling events to occur once IFN- binds to IFN-R1 (5, 22, 48). Decreased or absent IFN-R2 mRNA expression has been demonstrated in in vitro polarized Th1 CD4+ T cell lines, Th2 polarized cell lines cultured with IFN-, and CD8+ allospecific T cell lines via Northern blot (23, 25). In contrast with an earlier report (23), we were unable to detect surface expression of IFN-R2 on naive OT-1s via flow cytometry with the MOB-47 Ab despite the ability of these cells to respond to exogenous IFN-. We were also unable to verify the specificity of Abs that are suggested to identify IFN-R2 in Western blot analysis. Blotting with MOB-47 did not detect any specific protein in lysates from naive OT-1s. Another Ab (Q-20, rabbit polyclonal anti-IFN-R2) detected a single band of the appropriate size in lysates from wt T cells; however, the same sized band was also detected in lysates from IFN-R2–/– mice (5); therefore, the specificity of this polyclonal Ab could not be validated (data not shown). Thus, we used probe-based, real-time PCR to provide a quantitative measure of IFN-R2 mRNA expression.
OT-1s were purified from adoptive transfer recipients at different times p.i. via labeling with Thy1.1-PE, anti-PE-coated magnetic beads, and magnetic separation. At early time points only (days 0.5 and 2 p.i.), B6.RAG 1–/– mice were used as adoptive transfer recipients to increase the frequency of OT-1s in the spleen to improve purification. An example of a typical purification yielding >95% pure OT-1s from a starting population of 4% is shown in Fig. 5A.
We performed probe-based, real-time PCR for IFN-R2 mRNA using cDNA made from each purified cell population and compared expression in these cells to expression in naive OT-1s. As shown in Fig. 5, B and C, by 12 h p.i., IFN-R2 mRNA in OT-1s was already decreased compared with mRNA levels detected in naive OT-1s. IFN-R2 mRNA was maximally down-regulated on days 2 and 3 p.i., plummeting to 50- to 80-fold below expression detected in naive OT-1s. On day 5 p.i., mRNA for IFN-R2 increased to 25- to 30-fold below expression in naive cells. By day 7 p.i., when we observed a portion of the Ag-specific CD8+ T cells regaining responsiveness to IFN-, the expression of IFN-R2 mRNA in the population of OT-1s was 10-fold below the expression in naive OT-1s. IFN-R2 mRNA expression then remained constant throughout the contraction phase. The expression of IFN-R2 mRNA in memory OT-1s never again reached the level of expression in naive cells, even on day 101 p.i. (Fig. 5, B and C). Decreased IFN-R2 expression correlated with the reduced ability of memory OT-1s to phosphorylate STAT1 in response to IFN- compared with naive T cells (Fig. 3, B and C).
In addition to altered IFN-R expression, another potential mechanism by which Ag-specific CD8+ T cells could regulate IFN- responsiveness is through expression of the IFN- signaling inhibitor SOCS-1 (49). Although low levels of SOCS-1 transcripts are detectable in unstimulated cells, mRNA for SOCS-1 can be induced within 15–30 min after cytokine stimulation (18). The expression of this inhibitor is critical in controlling IFN- signaling in vivo, as evidenced by severe inflammatory disease and early death in SOCS-1–/– mice (50, 51). This phenotype is abrogated in the absence of lymphocytes or IFN- (17, 52).
SOCS-1 protein is markedly unstable, with a half-life reported to be <2 h (53). Due to this limitation and because we wanted to analyze endogenous SOCS-1 expression in nontransfected cells, we chose to perform real-time PCR using the same cDNA templates that were used to measure the expression of IFN-R2. As shown in Fig. 5, D and E, at 12 h p.i., SOCS-1 expression in OT-1s was 2- to 7-fold higher than that in naive T cells. This increase was similar to the SOCS-1 control template, which consisted of naive splenocytes cultured in vitro with IFN- for 4 h (Fig. 5, D and E) (54). These data suggest that OT-1s received an IFN- signal in vivo and induced SOCS-1 mRNA after only 12 h of infection with L. monocytogenes. Coincident with decreased mRNA for IFN-R2, SOCS-1 mRNA was decreased in OT-1s on day 2 p.i. (40-fold reduced) and was maximally down-regulated on day 3 p.i. (60-fold reduced) compared with SOCS-1 mRNA expression in naive OT-1s. The expression of SOCS-1 mRNA was increased on day 5 p.i. and remained constant during the contraction phase (days 7–10) when Ag-specific CD8+ T cells begin to recover responsiveness to IFN-.
This pattern of expression was exactly the opposite of what was predicted if SOCS-1 functioned as an IFN- signaling inhibitor in OT-1s during expansion. Instead, the observed pattern of SOCS-1 expression directly correlated with IFN-R2 expression and the ability of cells to phosphorylate STAT1 in response to IFN-. These data suggest that SOCS-1 was not responsible for the regulation of IFN- responsiveness in Ag-specific CD8+ T cells during expansion. Furthermore, although the promoter region of SOCS-1 has binding sites for STAT1, -3, and -6 (20, 55), and in vitro transcription of the SOCS-1 gene has been shown to be induced in response to multiple cytokines (18), it appears that in CD8+ T cells responding to an infection in vivo, SOCS-1 expression is limited to cells capable of productive IFN- signaling.
Although the up-regulation of mRNA for SOCS-1 in Ag-specific CD8+ T cells 12 h p.i. was an indication that these cells received a productive IFN- signal, which we hypothesized may result in the down-regulation of IFN-R2, we extended our PCR analysis to include other gene targets downstream of IFN-. The transcription factor IRF-2, which is a primary response gene up-regulated by many cells types in response to IFN- (11, 56), was expressed >3 times more in wt OT-1s than in IFN-R1–/– OT-1s 12 h p.i. (Fig. 5F). In addition, >2-fold more mRNA for TAP-2, another gene up-regulated by IFN- (11, 57), was detected in wt OT-1s compared with T cells lacking the IFN-R (Fig. 5G). These analyses provide additional evidence that Ag-specific CD8+ T cells receive and respond to signals delivered through the IFN-R early after L. monocytogenes infection.
These data demonstrate that Ag-specific CD8 T cells undergoing a primary immune response exhibit dynamic responsiveness to IFN- by regulating the expression of IFN- receptor components and suggest that the loss of IFN- responsiveness may be important during the expansion of these cells. The pattern of IFN- responsiveness was tightly associated with the expression of IFN-R2 mRNA, but not with the continued expression of the negative signaling inhibitor SOCS-1.
IFN- responsiveness during the secondary response of memory Ag-specific CD8+ T cells
Because memory CD8+ T cells never exhibit the same degree of IFN- responsiveness as naive CD8+ T cells, and because the secondary response of memory CD8+ T cells exhibits delayed contraction compared with the primary CD8+ T cell response (58, 59, 60, 61, 62), we wanted to determine whether Ag-specific CD8+ T cells responding to infection for a second time exhibited dynamic IFN- responsiveness. Memory OT-1s (1 x 104; donor mice were day 283 post-primary infection) were transferred into naive B6 mice before infection with 107 actA– LM-OVA. Fig. 6A shows the kinetics of the secondary response of memory OT-1s. We used intracellular phospho-STAT1 staining as a measure of IFN- responsiveness in OT-1s and Ab staining to detect surface expression of IFN-R1. The donor memory OT-1s exhibited a low degree of IFN- responsiveness, but expressed high levels of IFN-R1 (Fig. 6B). By 12 h. p.i., the population of memory OT-1s became entirely unresponsive to IFN-. Detectable surface expression of IFN-R1 was decreased on OT-1s at 12 and 24 h p.i. More surface IFN-R1 was detected after removal of bound IFN- via acid stripping, but IFN-R1 levels were still reduced compared with expression on donor memory cells before activation (donor cell MFI (MFI IFN-R1 – MFI control Ig), 505; 12 h p.i. MFI, 300; 24 h p.i. MFI, 356; Fig. 6B). The expression of mRNA for IFN-R1, as detected by real-time PCR, was decreased in responding memory OT-1s at 12 h. p.i. (data not shown). By day 3 p.i., the surface expression of IFN-R1 on responding memory OT-1s was again similar to IFN-R1 expression on donor memory OT-1s ( MFI, 523); however, the memory OT-1s undergoing a secondary response remained unresponsive to IFN- until day 7 p.i., which corresponded to the beginning of the protracted contraction phase (Fig. 6). A small fraction of OT-1s regained very low responsiveness to IFN- throughout the contraction phase and out to day 30 p.i. The expression of IFN-R1 remained consistently high during this interval (Fig. 6C). Cells that survived to seed the secondary memory pool of Ag-specific CD8+ T cells exhibited minimal responsiveness to IFN-. These data indicate that, similar to Ag-specific CD8+ T cells undergoing a primary response, memory Ag-specific CD8+ T cells undergoing a secondary response become unresponsive to IFN- during their expansion phase. However, the rate at which Ag-specific CD8+ T cells regained responsiveness to IFN- was slower during the secondary contraction phase than during the contraction phase following a primary response, perhaps reflecting the prolonged kinetics of the secondary contraction phase, slower cell turnover in the population, and the accumulation of T cells with very low IFN- responsiveness to seed the secondary memory pool. Surface expression of IFN-R1 was down-regulated to a lesser degree on Ag-specific CD8+ T cells early after secondary infection than after primary infection. This is most likely due either to differential exposure to IFN- at these times p.i., their initial decreased ability to respond to IFN-, or a difference in the programmed response of cells exposed to Ag for a second time.
These data demonstrated that decreased expression of IFN-R2 mRNA was even more striking in Ag-specific CD8+ T cells responding to infection for a second time compared with T cells responding to infection for the first time. However, the same overall pattern of expression in the two populations was observed, with re-expression of IFN-R2 mRNA being coincident with reacquisition of some degree of IFN- responsiveness. In addition, SOCS-1 did not appear to regulate IFN- responsiveness in Ag-specific CD8+ T cells undergoing a secondary immune response.
Discussion
T cells are bombarded with signals during every stage of their existence. Some signals keep them alive, some lead to their activation and differentiation into effector and memory cells, and still others result in their death. With the exception of positive and negative selection during thymic education, the signals received and processed by T cells after infection are perhaps the most complex and important determinants they will encounter. The ultimate functions of T cells are to rid the body of infection with minimal damage to the host and to provide long-lasting, specific memory. It is the ability to be ignorant of, evade, or productively respond to one or a combination of signals that determines how these tasks are accomplished.
Several lines of evidence suggested that the proinflammatory cytokine IFN- might directly affect CD8+ T cells in vivo during infection. IFN- mRNA and protein are highly expressed very early after infection when Ag-specific CD8+ T cells are initially activated (31) (Fig. 4, C and D). In addition, contraction of expanded CD8+ T cell populations was significantly prolonged after L. monocytogenes or lymphocytic choriomeningitis virus infection of BALB/c-IFN- deficient mice (35). In the present study we demonstrate that Ag-specific CD8+ T cells rapidly up-regulate SOCS-1 after infection, most likely as a result of an early IFN- signal, then quickly lose the ability to respond to IFN- by down-regulating, to different degrees, both components of the IFN-R. Ag-specific CD8+ T cells remained unable to phosphorylate STAT1 in response to stimulation with IFN- throughout the remainder of the expansion phase, as a consequence of loss of IFN-R2 expression. As discussed previously, IFN- has the potential to induce a wide range of cellular responses, including activation-induced cell death/apoptosis (11, 26). Given that CD8 T cells produce IFN- themselves during encounters with Ag (63) or cytokines such as IL-12/IL-18 (64, 65), it may be necessary to evade the potentially apoptotic effects of IFN- during the portion of the immune response when it is important for these cells to not only survive, but to expand to sufficient numbers to clear rapidly growing pathogens.
The loss of IFN- responsiveness by Ag-specific CD8+ T cells was not permanent. A small fraction of T cells regained IFN- responsiveness each day during the contraction phase, ultimately resulting in a memory population of Ag-specific CD8+ T cells that exhibited reduced, but uniform, responsiveness to IFN-. The ability to phosphorylate STAT1 in response to stimulation with IFN- was coincident with re-expression of IFN-R2. In parallel, recent studies show that Ag-specific CD8+ T cells down-regulate CD127 (the IL-7R -chain) after in vivo priming; however, a fraction of these cells appears to regain CD127 expression at the peak of the expansion phase. These CD127+ CD8+ T cells survive contraction and initiate the memory pool (66, 67). In addition, newly activated T cells up-regulate CD25, the high affinity IL-2R component, but subsequently lose expression of CD25 through the expansion phase (68, 69). Considered together, these studies suggest that regulation of cytokine receptor expression may be a critical method to control T cell survival by modulating responses to the complex mixture of cytokines present after infection.
Although a fraction of Ag-specific CD8+ T cells regained IFN- responsiveness during contraction, it appeared unlikely that IFN- was directly causing the death of Ag-specific CD8+ T cells during this phase. This conclusion is based on the fact that IFN- mRNA in the spleen and IFN- protein in the serum were not up-regulated at time points during the contraction phase (31) (Fig. 4, C and D). In addition, cells that survived contraction into the memory pool were able to respond to IFN-, albeit at a lower level than naive CD8+ T cells. Instead, we strongly favor the idea that IFN- signals delivered to CD8+ T cells very early after infection work in concert with signals delivered through the TCR and other receptors to initiate the expansion/contraction program. This hypothesis is supported by recent experiments performed in our laboratory suggesting that early inflammation and IFN- control the CD8+ T cell contraction program (32). This concept of a program executed in response to early signals has also recently been suggested by studies documenting the ability of CD8+ T cells to undergo extensive Ag-independent proliferation after only a short exposure to Ag (60, 70, 71, 72).
Except at 12 and 24 h p.i., the amount of detectable IFN- in the spleen or serum of L. monocytogenes-infected mice was not different from the amount detected in naive mice (Fig. 4, D and E), yet Ag-specific CD8+ T cells undergoing contraction and memory cells express decreased amounts of IFN-R2 mRNA compared with naive Ag-specific CD8+ T cells. This long-term down-regulation of IFN-R2 mRNA expression may be a consequence of IFN- exposure during primary activation. Alternatively, it has been shown that memory phenotype (CD44high) CD8+ T cells make IFN- in a non-Ag-specific manner when exposed to IL-12 and IL-18 (64, 65). Autocrine exposure to IFN- may serve to keep constitutive expression of IFN-R2 lower in previously activated cells compared with naive cells.
The selective accumulation of cells during the contraction phase with low responsiveness to IFN- is intriguing. The decreased capacity of memory cells to respond to IFN- compared with naive cells may be important for programming of the secondary response. In multiple instances it has been demonstrated that memory cells undergo protracted contraction compared with naive cells after a primary infection (58, 59, 60, 61, 62) (Figs. 1 and 6). Perhaps this difference in response kinetics is related to the strength of IFN- signal that memory cells receive early after secondary infection. Alternatively, memory cell responsiveness to IFN- may be important in the persistent turnover of memory cells (73, 74, 75) or in the global attrition of memory cells following unrelated infections (76).
During secondary infection, memory Ag-specific CD8+ T cells lose responsiveness to IFN- with similar kinetics as naive cells during primary infection. Memory CD8+ T cells responding to a secondary challenge with Ag remain unable to phosphorylate STAT1 in response to IFN- until they enter the contraction phase, during which populations of cells with very low responsiveness to IFN- begin to be detected. Secondary memory cells retain minimal responsiveness to IFN-. More moderate down-regulation of IFN-R1 was observed in responding memory Ag-specific CD8+ T cells compared with cells undergoing a primary response (Figs. 4 and 6), but down-regulation of IFN-R2 was greater in memory cells (Figs. 5 and 7). These differences most likely reflect disparate IFN- signals received by the two populations of Ag-specific CD8+ T cells, a hypothesis supported by the decreased expression of SOCS-1 mRNA in memory cells compared with Ag-specific CD8+ T cells undergoing a primary response (Figs. 5 and 7).
It has been clearly demonstrated in studies using knockout mice that SOCS-1 is essential for controlling IFN- signaling under basal conditions (17, 52). It has also been shown by in vitro experiments that SOCS-1 can be up-regulated by many cytokines, and it has been proposed that SOCS-1 induced by one cytokine may regulate signaling through other cytokine receptors by virtue of binding and inhibiting multiple members of the JAK family (18, 19, 50). Therefore, it was interesting that during in vivo infection, SOCS-1 expression was intimately linked to the ability of the CD8+ T cells to respond to IFN-. The expression of other cytokine receptors on Ag-specific CD8+ T cells was not evaluated; thus, we cannot say with absolute certainty that SOCS-1 expression was solely the result of competent IFN- signaling; however, these data do highlight a specific negative feedback relationship between IFN- and SOCS-1 in vivo.
Before this study, it was unknown whether or how IFN- directly affected T cells in vivo during an infection, which is the setting where T cells have the greatest exposure to this cytokine. We have demonstrated that naive CD8+ T cells are receptive to IFN- signals, productively respond to this cytokine very early after bacterial infection, and consequently undergo molecular changes to avoid signals delivered by IFN- until the Ag-driven expansion phase is complete. It is during this phase of the immune response when T cells would most be susceptible to autocrine effects of IFN-, which could potentially induce apoptosis in cells stimulated to make this cytokine upon exposure to Ag (26). Experiments are currently underway to determine whether early down-regulation of IFN-R1 and IFN-R2 are ligand-induced. Preliminary data indicate that IFN-R1–/– OT-1s express 3 times more mRNA for IFN-R2 than wt OT-1s on day 2 p.i., suggesting that the down-regulation of IFN-R2 is at least partially ligand dependent (data not shown). However, more work will be required to substantiate this conclusion.
In summary, these data reveal a dynamic pattern of IFN- responsiveness in Ag-specific CD8+ T cells after infection with L. monocytogenes. The results suggest that IFN- directly affects CD8+ T cells during an immune response and indicate that curtailing the ability to respond to IFN- early after activation may be important to achieve normal kinetics of the CD8+ T cell response.
Acknowledgments
We thank Dr. Kevin Knudtson, Lora Huang, Jessica Linton, and Jan Fuller at University of Iowa DNA facility for their excellent technical assistance with the real-time PCR experiments; Kate Rensberger and Rebecca Podyminogin for their laboratory technical assistance; and Drs. Vladimir Badovinac and Gail Bishop for critical review of this manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants AI42767, AI46653, and AI50073 (to J.T.H.) and T32AI0726-19 (to J.S.H.).
2 Address correspondence and reprint requests to Dr. John T. Harty, Department of Microbiology, 3-512 BSB, 51 Newton Road, Iowa City, IA 52242. E-mail address: john-harty{at}uiowa.edu
3 Abbreviations used in this paper: SOCS-1, suppressor of cytokine signaling-1; ICS, intracellular cytokine staining; IRF, IFN regulatory factor; MFI, mean fluorescence intensity; OT-1s, congenic Thy1.1+ OT-1 transgenic T cell; p.i., postinfection; Tg, transgenic; wt, wild type.
Received for publication November 9, 2004. Accepted for publication March 23, 2005.
References
Dalton, D. K., S. Pitts-Meek, S. Keshav, I. S. Figari, A. Bradley, T. A. Stewart. 1993. Multiple defects of immune cell function in mice with disrupted interferon- genes. Science 259: 1739-1742.[
Huang, S., W. Hendriks, A. Althage, S. Hemmi, H. Bluethmann, R. Kamijo, J. Vilcek, R. M. Zinkernagel, M. Aguet. 1993. Immune response in mice that lack the interferon- receptor. Science 259: 1742-1745.
Kamijo, R., J. Le, D. Shapiro, E. A. Havell, S. Huang, M. Aguet, M. Bosland, J. Vilcek. 1993. Mice that lack the interferon- receptor have profoundly altered responses to infection with bacillus Calmette-Guerin and subsequent challenge with lipopolysaccharide. J. Exp. Med. 178: 1435-1440.
Harty, J. T., M. J. Bevan. 1995. Specific immunity to Listeria monocytogenes in the absence of IFN-. Immunity 3: 109-117.
Lu, B., C. Ebensperger, Z. Dembic, Y. Wang, M. Kvatyuk, T. Lu, R. L. Coffman, S. Pestka, P. B. Rothman. 1998. Targeted disruption of the interferon- receptor 2 gene results in severe immune defects in mice. Proc. Natl. Acad. Sci. USA 95: 8233-8238.
Sad, S., R. Marcotte, T. R. Mosmann. 1995. Cytokine-induced differentiation of precursor mouse CD8+ T cells into cytotoxic CD8+ T cells secreting Th1 or Th2 cytokines. Immunity 2: 271-279.
Perussia, B.. 1991. Lymphokine-activated killer cells, natural killer cells and cytokines. Curr. Opin. Immunol. 3: 49-55.
Mosmann, T. R., R. L. Coffman. 1989. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu. Rev. Immunol. 7: 145-173(Jodie S. Haring*, Gail A.)
IFN- plays a critical role in the CD8+ T cell response to infection, but when and if this cytokine directly signals CD8+ T cells during an immune response is unknown. We show that naive Ag-specific CD8+ T cells receive IFN- signals within 12 h after in vivo infection with Listeria monocytogenes and then become unresponsive to IFN- throughout the ensuing Ag-driven expansion phase. Ag-specific CD8+ T cells regain partial IFN- responsiveness throughout the contraction phase, whereas the memory pool exhibits uniform, but reduced, responsiveness that is also modulated during the secondary response. The responsiveness of Ag-specific CD8+ T cells to IFN- correlated with modulation in the expression of IFN-R2, but not with IFN-R1 or suppressor of cytokine signaling-1. This dynamic regulation suggests that early IFN- signals participate in regulation of the primary CD8+ T cell response program, but that evading or minimizing IFN- signals during expansion and the memory phase may contribute to appropriate regulation of the CD8+ T cell response.
Introduction
Almost immediately after pathogenic insult, numerous cytokines and chemokines are released into tissues. One of the functions of this protein milieu is to assist in a rapid, direct counterattack mediated by the innate immune system against the invading organism. Another important role is to initiate and foster the propagation of the appropriate adaptive immune response. This interplay involves multiple signals delivered to immune cells via cytokines and cell-cell contacts. Optimally, an orchestrated response to these signals results in the appropriate immune response to clear the specific infection.
IFN- is one of the most well-studied proinflammatory cytokines produced by innate and adaptive immune cells. The importance of IFN- is evident in mice that have been made genetically deficient for the cytokine or its receptor components. These mice are profoundly susceptible to many intracellular pathogens, including Listeria monocytogenes, Leishmania major, and mycobacteria (1, 2, 3, 4, 5). IFN- is primarily made by NK cells and activated T cells (6, 7, 8) and functions as a noncovalently linked homodimer (9). The receptor for IFN- consists of two chains, IFN-R1, the ligand-binding portion, and IFN-R2, which is required for signaling. Binding of IFN- to IFN-R1 causes oligomerization of the receptor chains, which are constitutively associated with JAK1 and -2, but not normally preassociated with each other (10, 11). Once in proximity, the JAKs transphosphorylate and activate each other as well as phosphorylate the IFN-R1 chains on residue Tyr440. This creates a pair of docking sites for the latent cytosolic transcription factor STAT1, which is recruited to the receptor and phosphorylated on Tyr701 by the activated JAKs. Once phosphorylated, STAT1 subunits quickly dissociate from the receptor, translocate into the nucleus, and initiate transcription by binding to IFN--activated site DNA sequences (reviewed in Ref.10). Phosphorylation of STAT1 by protein kinase C on Ser727 before nuclear entry is important for optimal transcriptional activity (12, 13).
IFN- signaling is under tight control. Binding of IFN- to its receptor causes internalization and dissociation of the complex (14). In most cells, including murine macrophages, internalized IFN- is quickly degraded, and IFN-R1 is efficiently recycled back to the cell surface (15). Intracellular levels of phospho-STAT1 peak 15–30 min after in vitro IFN- stimulation, followed by a rapid decline to undetectable levels within 1–2 h (16). It has been demonstrated in HeLa cells that activated STAT1 homodimers are removed via ubiquitination and degradation by the proteosome (16). In addition to these control mechanisms, the Src homology 2 domain-containing protein suppressor of cytokine signaling-1 (SOCS-1)3functions as a negative feedback regulator of IFN- signaling (17). Expression of SOCS-1 in cell lines is under the control of several STATs, including STAT1, and SOCS-1 mRNA is quickly induced upon IFN- signaling. SOCS-1 binds to and inhibits the catalytic activity of JAKs, which attenuates IFN- signaling at a step proximal to STAT1 activation (18, 19, 20).
IFN- can regulate the expression of hundreds of genes, mostly, but not exclusively, via the JAK-STAT pathway (11, 21). Exposure to this cytokine can lead to a wide range of cellular responses in immune and nonimmune cells. Data concerning the effect(s) of IFN- specifically on T cells has been almost exclusively derived from in vitro studies. It has been shown that in vitro polarized Th1 CD4+ T cell lines are unresponsive to IFN- due to down-regulation of IFN-R2. This is thought to be a result of exposure to IFN-, because Th2 CD4+ T cells, which normally express IFN-R2, also down-regulate this receptor component when cultured with IFN-. In these experiments, surface protein expression of IFN-R1 was constitutive on both Th1 and Th2 CD4 T cells (22, 23). In contrast to murine CD4+ T cell clones, human T cells appear to regulate the expression of IFN-R2 in a ligand-independent manner (24). Resting and PHA-stimulated human CD3+ T cells maintain large cytoplasmic stores of IFN-R2, whereas surface expression remains low. The intracellular pools of IFN-R2 are the result of constitutive recycling of IFN-R2 between the cell surface and the cytoplasm. This process has been demonstrated to be ligand independent, because recycling of IFN-R2 still occurs in the absence of surface IFN-R1 or in the presence of neutralizing Abs for IFN- (24). Allospecific CD8+ T cell lines also maintain expression of IFN-R1, but are unresponsive to IFN- due to down-regulation of IFN-R2 at the mRNA level (25). In the absence of IFN- or STAT1, activated CD4+ T cells fail to up-regulate the expression of caspases 3, 6, and 8 despite the expression of normal levels of Fas and Fas ligand, suggesting that both IFN- and STAT1 are critical for activation-induced cell death of activated CD4+ T cells (26).
In addition to inducing apoptosis, IFN- can inhibit normal cell cycle progression in many cell types. When cultured with IFN-, bone marrow-derived macrophages arrest in G1/S phase (27). In a more detailed study using a murine macrophage cell line, it was shown that arrest at this cell cycle transition was due to an accumulation of the cyclin-dependent kinase inhibitor p27Kip1 (28). These studies suggest that active evasion of IFN- signals may allow T cells to escape the apoptotic and antiproliferative effects of this cytokine.
Despite intense investigation of the general cellular effects of IFN- and the aforementioned in vitro studies with T cell clones, we know little about how this cytokine directly affects T cells in vivo during an immune response. Mice that were engineered to constitutively express IFN-R2 (IFN-R2 transgenic (Tg)) mice were unable to mount productive Th1 immune responses to L. monocytogenes or Leishmania, and thus resembled IFN--deficient mice (29). Allospecific CD8+ T cell lines made from these mice had impaired cytotoxic capabilities in vitro despite being able to make IFN- and proliferate in response to Ag (25). In another study, T cells activated with the bacterial superantigen staphylococcal enterotoxin B exhibited a decreased ability to phosphorylate STAT1 in response to IFN- stimulation early after activation. This decreased responsiveness was interpreted to be independent of receptor down-regulation (30). These data combined with the aforementioned in vitro studies suggest that regulation of IFN- responsiveness may be required for normal T cell function.
IFN- mRNA is elevated very early after infection with L. monocytogenes (31), when T cells are first being activated, and recent studies suggest that this early IFN- production may control the eventual contraction program of CD8+ T cells (32). Consistent with this idea, activated CD4+ T cells persist in IFN- deficient B6 mice after infection with Mycobacterium bovis bacillus Calmette-Guérin or induction of experimental autoimmune encephalomyelitis (33, 34). Similarly, Ag-specific CD8+ T cells in IFN--deficient BALB/c mice undergo normal expansion, but exhibit a dramatically prolonged contraction phase after infection with L. monocytogenes or lymphocytic choriomeningitis virus (35). In addition, the hierarchy of epitope dominance was altered in the absence of IFN-. It is not known whether IFN- directly affected Ag-specific T cells during the immune response, or if its role was via indirect mechanisms (36).
In summary, these data suggest that IFN- may play a key role in regulating T cell responses in vivo. In the current study we investigated if, how, and when Ag-specific CD8+ T cells altered their responsiveness to IFN- in vivo during primary and secondary responses to infection with L. monocytogenes. The results reveal a dynamic pattern of IFN- responsiveness in Ag-specific CD8+ T cells during expansion and contraction that coincides with alterations in receptor component expression. This pattern of responsiveness is consistent with early exposure to the cytokine and an important role for IFN- in programming of the subsequent CD8+ T cell immune response.
Materials and Methods
PCR analyses
cDNA made from purified T cells and whole splenocytes was used as the template in probe-based, real-time PCR to assess the expression of IFN-R1 (data not shown), IFN-R2, and SOCS-1. Amplification of GAPDH was used as a control. Primer-probe sets for the targets of interest were designed using Primer Express software version 1.5 and were synthesized by IDT. These probes were labeled with the reporter dye FAM and the quencher dye TAMRA. GAPDH reagents were purchased from Applied Biosystems. The GAPDH probe was labeled with the reporter dye VIC and the quencher dye TAMRA. TaqMan universal PCR master mix (Applied Biosystems) was used for all reactions. All experiments were performed using an ABI PRISM 7700 sequence detection system (Applied Biosystems). The relative amplification of each unknown to GAPDH at different times p.i. was directly compared with the relative expression of that same target to GAPDH in purified naive or memory donor OT-1.PL Tg T cells. The data presented were obtained using the cycle thresholds to calculate 2–CT, which is the amount of target (unknown), normalized to the endogenous reference (GAPDH) and relative to a calibrator (naive OT-1.PL or memory donor cells).
Amplification of cDNA for IFN regulatory factor-2 (IRF-2) and TAP-2 from purified IFN-R1–/– and wild-type (wt) OT-1s was performed using RT-PCR. GAPDH was amplified together with each target gene and was used to normalize expression levels. Appropriate bands on ethidium bromide-stained gels were quantitated using ImageQuant 3.3.
Results
Response kinetics of Ag-specific CD8+ T cells after infection
An adoptive transfer system using Thy1.1 congenic OT-1 TCR Tg T cells (37) specific for OVA257 (herein called OT-1s) was established to investigate IFN- responsiveness in Ag-specific CD8+ T cells after bacterial infection. This system was instituted to allow detection of Ag-specific CD8+ T cells very early after infection and to facilitate purification of these cells without the use of MHC class I/peptide tetramers, which have the potential to affect intracellular signaling events (43). After transfer to C57BL/6 (Thy1.2) hosts, the OT-1s were activated by infection with L. monocytogenes that had been engineered to express OVA (LM-OVA) (38).
Preliminary experiments were performed to determine the number of OT-1s that could be adoptively transferred and still exhibit expansion and contraction after infection with kinetics similar to those of the endogenous CD8+ T cell response. CD8+-enriched splenocytes from OT-1 mice were transferred into recipient B6 mice 1 day before infection with 107 actA– LM-OVA. The results from these preliminary experiments (data not shown) indicated that after adoptive transfer of 5 x 104 or fewer OT-1s, the response kinetics most resembled the endogenous OVA257-specific CD8+ T cell response in B6 mice after actA– LM-OVA. With the addition of OT-1s, we achieved a 5-fold increase in the total number of memory OVA257-specific CD8+ T cells compared with nonadoptive transfer recipients (Fig. 1B). Attenuated actA– LM-OVA was used so that a sufficiently high dose of bacteria could be delivered to the mice to ensure activation of all adoptively transferred transgenic T cells, which was confirmed by CFSE dilution studies (data not shown). When OT-1s were analyzed at different times p.i. for functional activation, virtually all Thy1.1+ OT-1s responded to stimulation with cognate peptide (OVA257) by making IFN-, as detected by intracellular cytokine staining (ICS). An example of ICS for IFN- on day 10 p.i. is shown in Fig. 1A. Thus, the Thy1.1 marker alone is a suitable way to enumerate functionally activated OT-1s after infection.
Intracellular staining for phospho-STAT1 is a sensitive method for measuring IFN- responsiveness
If and how IFN- affects CD8+ T cells during an immune response are currently unknown. Because phosphorylation of STAT1 is a proximal step in the IFN- signaling pathway, the presence of this activated transcription factor in a cell after stimulation with IFN- indicates competence to receive a signal initiated via the binding of IFN- to its receptor. Typically, Western blotting is used to assess the phosphorylation status of STAT1 in cytokine-stimulated cells. However, Western blotting provides data on a whole population and cannot distinguish between global changes in IFN- responsiveness and the presence of subpopulations of responsive or unresponsive cells. In addition, Western blotting requires purification of a relatively large number of cells, which would not be possible very early after infection, when numbers of Ag-specific T cells are low. To overcome these limitations, we chose to use intracellular staining for phospho-STAT1 to assess the ability of Ag-specific CD8+ T cells to respond to a short in vitro stimulation with IFN-.
We performed preliminary experiments to determine the resolution provided by flow cytometric detection of intracellular phospho-STAT1 in IFN--stimulated T cells. CD8+-enriched splenocytes from naive B6 mice were stimulated with IFN- for 20 min, which was the peak of STAT1 phosphorylation as determined in time-course experiments (42) (data not shown). After stimulation, parallel samples of cells were immediately lysed for Western blotting or fixed, permeabilized, and stained for intracellular phospho-STAT1. Identically cultured, unstimulated splenocyte samples were lysed or stained as controls. To compare the sensitivities of the techniques, IFN--stimulated and unstimulated samples were mixed in specific ratios before blotting or staining. As shown in Fig. 2A, phospho-STAT1 was only detected in CD8+ T cells stimulated with IFN- (lanes 1 and 2) on a Western blot. We were readily able to detect phospho-STAT1 when the lysate contained 20% IFN--stimulated cells, but phospho-STAT1 was only variably detected when the lysate contained 10% stimulated IFN- cells (Fig. 2A, lanes 5 and 6).
In comparison, naive CD8+ T cells were highly responsive to IFN-, as detected by a uniform shift in staining with anti-phospho-STAT1 compared with unstimulated cells (Fig. 2B). Isotype control staining of stimulated and unstimulated cells was similar to anti-phospho-STAT1 staining in unstimulated cells and was omitted from the figure for clarity. As the frequency of IFN--stimulated cells in mixed samples decreased, the portion of the peak that was shifted compared with the unstimulated cell peak decreased (Fig. 2B). Importantly, it was still possible to observe positive signal when only a small fraction of the cells were stimulated with IFN-. These data validate intracellular staining for phospho-STAT1 as a viable technique to measure responsiveness to IFN- in T cells and highlight the unique ability of this method to resolve subpopulations of cells with different IFN- responsiveness.
Ag-specific CD8+ T cells become unresponsive to IFN- very early after primary infection and do not regain responsiveness until expansion is completed
As shown in Fig. 2B, naive CD8+ T cells from wt B6 mice are highly responsive to IFN-. Naive OT-1s exhibited the same high degree of responsiveness as polyclonal CD8+ T cells from naive B6 mice (Fig. 3C, first panel). In contrast, OT-1s obtained 3 days after actA– LM-OVA infection were unable to phosphorylate STAT1 after IFN- stimulation (Fig. 3B). OT-1s remained entirely unresponsive to IFN- until day 5 p.i., the time when the cells began to transition from the expansion to the contraction phase (Fig. 3, A and B). As cells entered the contraction phase, a fraction of Ag-specific CD8+ T cells reacquired the ability to respond to IFN-. On day 7 p.i., 5–7% of the expanded OT-1s stained positively for phospho-STAT1 after stimulation with IFN-. This population of IFN--responsive cells increased in proportion as the number of Ag-specific CD8+ T cells dropped. On day 8 p.i., 10–15% of the Tg T cells were positive for phospho-STAT1 after IFN- stimulation, and on day 9 p.i., this frequency increased substantially to 40–50% (Fig. 3B). Both early (day 30 p.i.) and late (day 101 p.i.) memory OT-1s retained the ability to respond to IFN-; however, the levels of phospho-STAT1 were reduced compared with those of naive OT-1s (Fig. 3C). These data indicate that despite being highly responsive to IFN- in their naive state, Ag-specific CD8+ T cells rapidly lose responsiveness to IFN- after infection with actA– LM-OVA and remain unresponsive throughout their expansion phase. During contraction, cells that exhibit some degree of responsiveness to IFN- accumulate to seed the memory population.
Before day 3 p.i., we were unable to detect OT-1s in the spleens of infected mice that had received adoptive transfer of 5 x 104 or fewer Tg T cells (data not shown). To study Ag-specific CD8+ T cells very early after infection, it was necessary to transfer greater numbers of OT-1s. In experiments performed through day 2 p.i., recipient B6 mice received 5 x 105 Tg T cells. Parallel CFSE dilution studies showed that all the OT-1s were recruited after infection with 107 actA– LM-OVA (data not shown). Strikingly, within 12 h after infection, OT-1s completely lost the ability to phosphorylate STAT1 in response to IFN- and remained unresponsive through day 2 p.i. (Fig. 3C). In combination, these data show that Ag-specific CD8+ T cells rapidly become unresponsive to IFN- within 12 h after infection, but regain responsiveness 5–7 days later. Thus, Ag-specific CD8+ T cells exhibit dynamic IFN- responsiveness during the primary immune response in vivo.
Expression of IFN-R1 does not correlate with IFN- responsiveness
IFN-R1 (CD119) is the ligand-binding portion of the IFN-R. Expression of IFN-R1 is constitutive on most cells (14). However, it has been demonstrated that IFN-R1 is transiently down-regulated in Tg CD4+ T cells after exposure to Ag or TCR ligation with Ab in vitro (44). To uncover the mechanism by which Ag-specific CD8+ T cells alter their responsiveness to IFN- in vivo, we investigated receptor expression on these cells at different times after infection.
As shown in Fig. 4A, naive OT-1s express high levels of surface IFN-R1, as do OT-1s adoptively transferred into mice that were subsequently left uninfected. Splenocytes from adoptive transfer recipients were harvested and analyzed directly ex vivo at different times p.i. for expression of CD8, Thy1.1, and IFN-R1. In preliminary experiments, at 12 and 24 h p.i. it appeared that most OT-1s had extensively down-regulated surface expression of IFN-R1 (Fig. 4A). However, the Ab used to detect IFN-R1 (clone GR20) can be blocked by bound IFN- (45, 46). IFN- bound to IFN-R1 can be released via stripping with a short incubation in very low pH medium (47). Acid stripping of OT-1s isolated from naive mice did not substantially alter the level of detectable surface IFN-R1 (pre-acid treatment mean fluorescence intensity (MFI), 672; after acid treatment MFI, 679; Fig. 4A). When splenocytes from OT-1 recipient mice 12 and 24 h p.i. were treated with acidic medium, followed immediately by staining for IFN-R1, we detected IFN-R1 on the surface of OT-1s, although surface expression was reduced compared with expression on naive OT-1 Tg T cells (naive MFI, 679; 12 h p.i. MFI, 517; 24 h p.i. MFI, 511; Fig. 4A). Similar results were obtained when OT-1s were stained with the anti-IFN-R1 Ab 2E2, which is not blocked by bound IFN- (23). Surface levels of IFN-R1 on OT-1s, as detected by 2E2, were decreased at 12 and 24 h p.i., but not at 48 h p.i. (data not shown). In addition, mRNA for IFN-R1 (quantitated by real-time PCR) was decreased in OT-1s at 12 h p.i. compared with cDNA samples made from naive OT-1s and cells purified from mice 48 h p.i. (data not shown).
On days 3 and 5 p.i., surface expression of IFN-R1 remained relatively high, even without acid stripping (Fig. 4B) and returned to the same level as that observed in naive T cells by day 7 p.i. At all time points analyzed after day 7 p.i., including days 8–10 and out to day 283 p.i., surface expression of IFN-R1 remained high (data not shown and Fig. 4B). These data indicate that OT-1s do not completely lose surface expression of IFN-R1 and express high levels of IFN-R1 during the expansion phase when they are unresponsive to IFN-. Thus, down-regulation of IFN-R1 is not the mechanism by which Ag-specific CD8+ T cells are rendered unresponsive to IFN- after L. monocytogenes infection.
The data presented in Fig. 4A demonstrate that IFN-R1 expressed on Ag-specific CD8+ T cells was almost completely bound by IFN- at 12 and 24 h p.i. Occupancy of the receptor may prevent these cells from responding to additional stimulation with IFN- in vitro before intracellular staining for phospho-STAT1 (Fig. 3C). To address this possibility, Ag-specific CD8+ T cells were subjected to acid stripping before in vitro IFN- stimulation and phospho-STAT1 staining. As shown in Fig. 4C, even after surface bound IFN- was removed, Ag-specific CD8+ T cells were still unable to respond to exogenously added IFN- by phosphorylating STAT1. The acid stripping treatment did not alter the ability of naive CD8+ T cells to respond to IFN- stimulation. It remains possible that as a consequence of exposure to IFN- in vivo, these cells have exhausted their ability to respond to this cytokine, perhaps through down-regulation of total levels of STAT1; however, the data support the conclusion that Ag-specific CD8+ T cells lose the ability to respond to IFN- by 12 h p.i.
The early blocking of surface IFN-R1 by IFN- on OT-1s was coincident with the expression of IFN- mRNA in the spleen (Fig. 4D). The expression of IFN- mRNA, detected by probe-based, real-time PCR, was dramatically increased at 12 and 24 h p.i. compared with that in naive mice (Fig. 4D). IFN- mRNA levels returned to baseline on day 2 p.i. and remained at the same levels as in naive mice throughout the remainder of the expansion and contraction phases of the CD8+ T cell response. Similarly, IFN- protein was readily detected in the serum of infected mice 12 and 24 h p.i., but was below the limit of detection by 48 h p.i. (Fig. 4E). No IFN- protein was detected in the serum of naive mice. These data suggest that the modest down-regulation of IFN-R1 observed 12 and 24 h p.i. was ligand induced. Although we cannot rule out that the acid stripping may not have removed all bound IFN- from the IFN-R1, similar results were obtained using an anti-IFN-R1 Ab (2E2) that was not blocked by the presence of IFN- (23), and mRNA for IFN-R1 was also decreased in OT-1s 12 h p.i. (data not shown). Internalization of a fraction of IFN--bound IFN-R1 could also be an explanation for the observed decrease in surface expression of IFN-R1; however, it has been demonstrated that Th1-polarized CD4+ T cell clones maintain detectable levels of surface IFN-R1 during long-term culture with IFN- (23). Thus, not all IFN-R1 is internalized during IFN- signaling. Alternatively, recycling of IFN-R1 back to the surface of cells after internalization is a very efficient process and could contribute to the amount of detectable surface receptor.
In summary, Ag-specific CD8+ T cells are exposed to IFN- early after infection with L. monocytogenes, and these cells undergo a transient, most likely ligand-induced, down-regulation of surface expression of IFN-R1. However, IFN-R1 was found to be uniformly expressed by these cells at times p.i. (days 5 and 7) when they remained largely unresponsive to IFN-. Therefore, there must be an additional mechanism(s) by which Ag-specific CD8+ T cells lose IFN- responsiveness during expansion.
Ag-specific CD8+ T cells regain responsiveness to IFN- upon re-expression of IFN-R2
Because altered expression of IFN-R1 by Ag-specific CD8+ T cells did not strictly correlate with their ability to respond to IFN-, we next investigated the expression of IFN-R2 at time points throughout the primary immune response. IFN-R2 is required for downstream signaling events to occur once IFN- binds to IFN-R1 (5, 22, 48). Decreased or absent IFN-R2 mRNA expression has been demonstrated in in vitro polarized Th1 CD4+ T cell lines, Th2 polarized cell lines cultured with IFN-, and CD8+ allospecific T cell lines via Northern blot (23, 25). In contrast with an earlier report (23), we were unable to detect surface expression of IFN-R2 on naive OT-1s via flow cytometry with the MOB-47 Ab despite the ability of these cells to respond to exogenous IFN-. We were also unable to verify the specificity of Abs that are suggested to identify IFN-R2 in Western blot analysis. Blotting with MOB-47 did not detect any specific protein in lysates from naive OT-1s. Another Ab (Q-20, rabbit polyclonal anti-IFN-R2) detected a single band of the appropriate size in lysates from wt T cells; however, the same sized band was also detected in lysates from IFN-R2–/– mice (5); therefore, the specificity of this polyclonal Ab could not be validated (data not shown). Thus, we used probe-based, real-time PCR to provide a quantitative measure of IFN-R2 mRNA expression.
OT-1s were purified from adoptive transfer recipients at different times p.i. via labeling with Thy1.1-PE, anti-PE-coated magnetic beads, and magnetic separation. At early time points only (days 0.5 and 2 p.i.), B6.RAG 1–/– mice were used as adoptive transfer recipients to increase the frequency of OT-1s in the spleen to improve purification. An example of a typical purification yielding >95% pure OT-1s from a starting population of 4% is shown in Fig. 5A.
We performed probe-based, real-time PCR for IFN-R2 mRNA using cDNA made from each purified cell population and compared expression in these cells to expression in naive OT-1s. As shown in Fig. 5, B and C, by 12 h p.i., IFN-R2 mRNA in OT-1s was already decreased compared with mRNA levels detected in naive OT-1s. IFN-R2 mRNA was maximally down-regulated on days 2 and 3 p.i., plummeting to 50- to 80-fold below expression detected in naive OT-1s. On day 5 p.i., mRNA for IFN-R2 increased to 25- to 30-fold below expression in naive cells. By day 7 p.i., when we observed a portion of the Ag-specific CD8+ T cells regaining responsiveness to IFN-, the expression of IFN-R2 mRNA in the population of OT-1s was 10-fold below the expression in naive OT-1s. IFN-R2 mRNA expression then remained constant throughout the contraction phase. The expression of IFN-R2 mRNA in memory OT-1s never again reached the level of expression in naive cells, even on day 101 p.i. (Fig. 5, B and C). Decreased IFN-R2 expression correlated with the reduced ability of memory OT-1s to phosphorylate STAT1 in response to IFN- compared with naive T cells (Fig. 3, B and C).
In addition to altered IFN-R expression, another potential mechanism by which Ag-specific CD8+ T cells could regulate IFN- responsiveness is through expression of the IFN- signaling inhibitor SOCS-1 (49). Although low levels of SOCS-1 transcripts are detectable in unstimulated cells, mRNA for SOCS-1 can be induced within 15–30 min after cytokine stimulation (18). The expression of this inhibitor is critical in controlling IFN- signaling in vivo, as evidenced by severe inflammatory disease and early death in SOCS-1–/– mice (50, 51). This phenotype is abrogated in the absence of lymphocytes or IFN- (17, 52).
SOCS-1 protein is markedly unstable, with a half-life reported to be <2 h (53). Due to this limitation and because we wanted to analyze endogenous SOCS-1 expression in nontransfected cells, we chose to perform real-time PCR using the same cDNA templates that were used to measure the expression of IFN-R2. As shown in Fig. 5, D and E, at 12 h p.i., SOCS-1 expression in OT-1s was 2- to 7-fold higher than that in naive T cells. This increase was similar to the SOCS-1 control template, which consisted of naive splenocytes cultured in vitro with IFN- for 4 h (Fig. 5, D and E) (54). These data suggest that OT-1s received an IFN- signal in vivo and induced SOCS-1 mRNA after only 12 h of infection with L. monocytogenes. Coincident with decreased mRNA for IFN-R2, SOCS-1 mRNA was decreased in OT-1s on day 2 p.i. (40-fold reduced) and was maximally down-regulated on day 3 p.i. (60-fold reduced) compared with SOCS-1 mRNA expression in naive OT-1s. The expression of SOCS-1 mRNA was increased on day 5 p.i. and remained constant during the contraction phase (days 7–10) when Ag-specific CD8+ T cells begin to recover responsiveness to IFN-.
This pattern of expression was exactly the opposite of what was predicted if SOCS-1 functioned as an IFN- signaling inhibitor in OT-1s during expansion. Instead, the observed pattern of SOCS-1 expression directly correlated with IFN-R2 expression and the ability of cells to phosphorylate STAT1 in response to IFN-. These data suggest that SOCS-1 was not responsible for the regulation of IFN- responsiveness in Ag-specific CD8+ T cells during expansion. Furthermore, although the promoter region of SOCS-1 has binding sites for STAT1, -3, and -6 (20, 55), and in vitro transcription of the SOCS-1 gene has been shown to be induced in response to multiple cytokines (18), it appears that in CD8+ T cells responding to an infection in vivo, SOCS-1 expression is limited to cells capable of productive IFN- signaling.
Although the up-regulation of mRNA for SOCS-1 in Ag-specific CD8+ T cells 12 h p.i. was an indication that these cells received a productive IFN- signal, which we hypothesized may result in the down-regulation of IFN-R2, we extended our PCR analysis to include other gene targets downstream of IFN-. The transcription factor IRF-2, which is a primary response gene up-regulated by many cells types in response to IFN- (11, 56), was expressed >3 times more in wt OT-1s than in IFN-R1–/– OT-1s 12 h p.i. (Fig. 5F). In addition, >2-fold more mRNA for TAP-2, another gene up-regulated by IFN- (11, 57), was detected in wt OT-1s compared with T cells lacking the IFN-R (Fig. 5G). These analyses provide additional evidence that Ag-specific CD8+ T cells receive and respond to signals delivered through the IFN-R early after L. monocytogenes infection.
These data demonstrate that Ag-specific CD8 T cells undergoing a primary immune response exhibit dynamic responsiveness to IFN- by regulating the expression of IFN- receptor components and suggest that the loss of IFN- responsiveness may be important during the expansion of these cells. The pattern of IFN- responsiveness was tightly associated with the expression of IFN-R2 mRNA, but not with the continued expression of the negative signaling inhibitor SOCS-1.
IFN- responsiveness during the secondary response of memory Ag-specific CD8+ T cells
Because memory CD8+ T cells never exhibit the same degree of IFN- responsiveness as naive CD8+ T cells, and because the secondary response of memory CD8+ T cells exhibits delayed contraction compared with the primary CD8+ T cell response (58, 59, 60, 61, 62), we wanted to determine whether Ag-specific CD8+ T cells responding to infection for a second time exhibited dynamic IFN- responsiveness. Memory OT-1s (1 x 104; donor mice were day 283 post-primary infection) were transferred into naive B6 mice before infection with 107 actA– LM-OVA. Fig. 6A shows the kinetics of the secondary response of memory OT-1s. We used intracellular phospho-STAT1 staining as a measure of IFN- responsiveness in OT-1s and Ab staining to detect surface expression of IFN-R1. The donor memory OT-1s exhibited a low degree of IFN- responsiveness, but expressed high levels of IFN-R1 (Fig. 6B). By 12 h. p.i., the population of memory OT-1s became entirely unresponsive to IFN-. Detectable surface expression of IFN-R1 was decreased on OT-1s at 12 and 24 h p.i. More surface IFN-R1 was detected after removal of bound IFN- via acid stripping, but IFN-R1 levels were still reduced compared with expression on donor memory cells before activation (donor cell MFI (MFI IFN-R1 – MFI control Ig), 505; 12 h p.i. MFI, 300; 24 h p.i. MFI, 356; Fig. 6B). The expression of mRNA for IFN-R1, as detected by real-time PCR, was decreased in responding memory OT-1s at 12 h. p.i. (data not shown). By day 3 p.i., the surface expression of IFN-R1 on responding memory OT-1s was again similar to IFN-R1 expression on donor memory OT-1s ( MFI, 523); however, the memory OT-1s undergoing a secondary response remained unresponsive to IFN- until day 7 p.i., which corresponded to the beginning of the protracted contraction phase (Fig. 6). A small fraction of OT-1s regained very low responsiveness to IFN- throughout the contraction phase and out to day 30 p.i. The expression of IFN-R1 remained consistently high during this interval (Fig. 6C). Cells that survived to seed the secondary memory pool of Ag-specific CD8+ T cells exhibited minimal responsiveness to IFN-. These data indicate that, similar to Ag-specific CD8+ T cells undergoing a primary response, memory Ag-specific CD8+ T cells undergoing a secondary response become unresponsive to IFN- during their expansion phase. However, the rate at which Ag-specific CD8+ T cells regained responsiveness to IFN- was slower during the secondary contraction phase than during the contraction phase following a primary response, perhaps reflecting the prolonged kinetics of the secondary contraction phase, slower cell turnover in the population, and the accumulation of T cells with very low IFN- responsiveness to seed the secondary memory pool. Surface expression of IFN-R1 was down-regulated to a lesser degree on Ag-specific CD8+ T cells early after secondary infection than after primary infection. This is most likely due either to differential exposure to IFN- at these times p.i., their initial decreased ability to respond to IFN-, or a difference in the programmed response of cells exposed to Ag for a second time.
These data demonstrated that decreased expression of IFN-R2 mRNA was even more striking in Ag-specific CD8+ T cells responding to infection for a second time compared with T cells responding to infection for the first time. However, the same overall pattern of expression in the two populations was observed, with re-expression of IFN-R2 mRNA being coincident with reacquisition of some degree of IFN- responsiveness. In addition, SOCS-1 did not appear to regulate IFN- responsiveness in Ag-specific CD8+ T cells undergoing a secondary immune response.
Discussion
T cells are bombarded with signals during every stage of their existence. Some signals keep them alive, some lead to their activation and differentiation into effector and memory cells, and still others result in their death. With the exception of positive and negative selection during thymic education, the signals received and processed by T cells after infection are perhaps the most complex and important determinants they will encounter. The ultimate functions of T cells are to rid the body of infection with minimal damage to the host and to provide long-lasting, specific memory. It is the ability to be ignorant of, evade, or productively respond to one or a combination of signals that determines how these tasks are accomplished.
Several lines of evidence suggested that the proinflammatory cytokine IFN- might directly affect CD8+ T cells in vivo during infection. IFN- mRNA and protein are highly expressed very early after infection when Ag-specific CD8+ T cells are initially activated (31) (Fig. 4, C and D). In addition, contraction of expanded CD8+ T cell populations was significantly prolonged after L. monocytogenes or lymphocytic choriomeningitis virus infection of BALB/c-IFN- deficient mice (35). In the present study we demonstrate that Ag-specific CD8+ T cells rapidly up-regulate SOCS-1 after infection, most likely as a result of an early IFN- signal, then quickly lose the ability to respond to IFN- by down-regulating, to different degrees, both components of the IFN-R. Ag-specific CD8+ T cells remained unable to phosphorylate STAT1 in response to stimulation with IFN- throughout the remainder of the expansion phase, as a consequence of loss of IFN-R2 expression. As discussed previously, IFN- has the potential to induce a wide range of cellular responses, including activation-induced cell death/apoptosis (11, 26). Given that CD8 T cells produce IFN- themselves during encounters with Ag (63) or cytokines such as IL-12/IL-18 (64, 65), it may be necessary to evade the potentially apoptotic effects of IFN- during the portion of the immune response when it is important for these cells to not only survive, but to expand to sufficient numbers to clear rapidly growing pathogens.
The loss of IFN- responsiveness by Ag-specific CD8+ T cells was not permanent. A small fraction of T cells regained IFN- responsiveness each day during the contraction phase, ultimately resulting in a memory population of Ag-specific CD8+ T cells that exhibited reduced, but uniform, responsiveness to IFN-. The ability to phosphorylate STAT1 in response to stimulation with IFN- was coincident with re-expression of IFN-R2. In parallel, recent studies show that Ag-specific CD8+ T cells down-regulate CD127 (the IL-7R -chain) after in vivo priming; however, a fraction of these cells appears to regain CD127 expression at the peak of the expansion phase. These CD127+ CD8+ T cells survive contraction and initiate the memory pool (66, 67). In addition, newly activated T cells up-regulate CD25, the high affinity IL-2R component, but subsequently lose expression of CD25 through the expansion phase (68, 69). Considered together, these studies suggest that regulation of cytokine receptor expression may be a critical method to control T cell survival by modulating responses to the complex mixture of cytokines present after infection.
Although a fraction of Ag-specific CD8+ T cells regained IFN- responsiveness during contraction, it appeared unlikely that IFN- was directly causing the death of Ag-specific CD8+ T cells during this phase. This conclusion is based on the fact that IFN- mRNA in the spleen and IFN- protein in the serum were not up-regulated at time points during the contraction phase (31) (Fig. 4, C and D). In addition, cells that survived contraction into the memory pool were able to respond to IFN-, albeit at a lower level than naive CD8+ T cells. Instead, we strongly favor the idea that IFN- signals delivered to CD8+ T cells very early after infection work in concert with signals delivered through the TCR and other receptors to initiate the expansion/contraction program. This hypothesis is supported by recent experiments performed in our laboratory suggesting that early inflammation and IFN- control the CD8+ T cell contraction program (32). This concept of a program executed in response to early signals has also recently been suggested by studies documenting the ability of CD8+ T cells to undergo extensive Ag-independent proliferation after only a short exposure to Ag (60, 70, 71, 72).
Except at 12 and 24 h p.i., the amount of detectable IFN- in the spleen or serum of L. monocytogenes-infected mice was not different from the amount detected in naive mice (Fig. 4, D and E), yet Ag-specific CD8+ T cells undergoing contraction and memory cells express decreased amounts of IFN-R2 mRNA compared with naive Ag-specific CD8+ T cells. This long-term down-regulation of IFN-R2 mRNA expression may be a consequence of IFN- exposure during primary activation. Alternatively, it has been shown that memory phenotype (CD44high) CD8+ T cells make IFN- in a non-Ag-specific manner when exposed to IL-12 and IL-18 (64, 65). Autocrine exposure to IFN- may serve to keep constitutive expression of IFN-R2 lower in previously activated cells compared with naive cells.
The selective accumulation of cells during the contraction phase with low responsiveness to IFN- is intriguing. The decreased capacity of memory cells to respond to IFN- compared with naive cells may be important for programming of the secondary response. In multiple instances it has been demonstrated that memory cells undergo protracted contraction compared with naive cells after a primary infection (58, 59, 60, 61, 62) (Figs. 1 and 6). Perhaps this difference in response kinetics is related to the strength of IFN- signal that memory cells receive early after secondary infection. Alternatively, memory cell responsiveness to IFN- may be important in the persistent turnover of memory cells (73, 74, 75) or in the global attrition of memory cells following unrelated infections (76).
During secondary infection, memory Ag-specific CD8+ T cells lose responsiveness to IFN- with similar kinetics as naive cells during primary infection. Memory CD8+ T cells responding to a secondary challenge with Ag remain unable to phosphorylate STAT1 in response to IFN- until they enter the contraction phase, during which populations of cells with very low responsiveness to IFN- begin to be detected. Secondary memory cells retain minimal responsiveness to IFN-. More moderate down-regulation of IFN-R1 was observed in responding memory Ag-specific CD8+ T cells compared with cells undergoing a primary response (Figs. 4 and 6), but down-regulation of IFN-R2 was greater in memory cells (Figs. 5 and 7). These differences most likely reflect disparate IFN- signals received by the two populations of Ag-specific CD8+ T cells, a hypothesis supported by the decreased expression of SOCS-1 mRNA in memory cells compared with Ag-specific CD8+ T cells undergoing a primary response (Figs. 5 and 7).
It has been clearly demonstrated in studies using knockout mice that SOCS-1 is essential for controlling IFN- signaling under basal conditions (17, 52). It has also been shown by in vitro experiments that SOCS-1 can be up-regulated by many cytokines, and it has been proposed that SOCS-1 induced by one cytokine may regulate signaling through other cytokine receptors by virtue of binding and inhibiting multiple members of the JAK family (18, 19, 50). Therefore, it was interesting that during in vivo infection, SOCS-1 expression was intimately linked to the ability of the CD8+ T cells to respond to IFN-. The expression of other cytokine receptors on Ag-specific CD8+ T cells was not evaluated; thus, we cannot say with absolute certainty that SOCS-1 expression was solely the result of competent IFN- signaling; however, these data do highlight a specific negative feedback relationship between IFN- and SOCS-1 in vivo.
Before this study, it was unknown whether or how IFN- directly affected T cells in vivo during an infection, which is the setting where T cells have the greatest exposure to this cytokine. We have demonstrated that naive CD8+ T cells are receptive to IFN- signals, productively respond to this cytokine very early after bacterial infection, and consequently undergo molecular changes to avoid signals delivered by IFN- until the Ag-driven expansion phase is complete. It is during this phase of the immune response when T cells would most be susceptible to autocrine effects of IFN-, which could potentially induce apoptosis in cells stimulated to make this cytokine upon exposure to Ag (26). Experiments are currently underway to determine whether early down-regulation of IFN-R1 and IFN-R2 are ligand-induced. Preliminary data indicate that IFN-R1–/– OT-1s express 3 times more mRNA for IFN-R2 than wt OT-1s on day 2 p.i., suggesting that the down-regulation of IFN-R2 is at least partially ligand dependent (data not shown). However, more work will be required to substantiate this conclusion.
In summary, these data reveal a dynamic pattern of IFN- responsiveness in Ag-specific CD8+ T cells after infection with L. monocytogenes. The results suggest that IFN- directly affects CD8+ T cells during an immune response and indicate that curtailing the ability to respond to IFN- early after activation may be important to achieve normal kinetics of the CD8+ T cell response.
Acknowledgments
We thank Dr. Kevin Knudtson, Lora Huang, Jessica Linton, and Jan Fuller at University of Iowa DNA facility for their excellent technical assistance with the real-time PCR experiments; Kate Rensberger and Rebecca Podyminogin for their laboratory technical assistance; and Drs. Vladimir Badovinac and Gail Bishop for critical review of this manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants AI42767, AI46653, and AI50073 (to J.T.H.) and T32AI0726-19 (to J.S.H.).
2 Address correspondence and reprint requests to Dr. John T. Harty, Department of Microbiology, 3-512 BSB, 51 Newton Road, Iowa City, IA 52242. E-mail address: john-harty{at}uiowa.edu
3 Abbreviations used in this paper: SOCS-1, suppressor of cytokine signaling-1; ICS, intracellular cytokine staining; IRF, IFN regulatory factor; MFI, mean fluorescence intensity; OT-1s, congenic Thy1.1+ OT-1 transgenic T cell; p.i., postinfection; Tg, transgenic; wt, wild type.
Received for publication November 9, 2004. Accepted for publication March 23, 2005.
References
Dalton, D. K., S. Pitts-Meek, S. Keshav, I. S. Figari, A. Bradley, T. A. Stewart. 1993. Multiple defects of immune cell function in mice with disrupted interferon- genes. Science 259: 1739-1742.[
Huang, S., W. Hendriks, A. Althage, S. Hemmi, H. Bluethmann, R. Kamijo, J. Vilcek, R. M. Zinkernagel, M. Aguet. 1993. Immune response in mice that lack the interferon- receptor. Science 259: 1742-1745.
Kamijo, R., J. Le, D. Shapiro, E. A. Havell, S. Huang, M. Aguet, M. Bosland, J. Vilcek. 1993. Mice that lack the interferon- receptor have profoundly altered responses to infection with bacillus Calmette-Guerin and subsequent challenge with lipopolysaccharide. J. Exp. Med. 178: 1435-1440.
Harty, J. T., M. J. Bevan. 1995. Specific immunity to Listeria monocytogenes in the absence of IFN-. Immunity 3: 109-117.
Lu, B., C. Ebensperger, Z. Dembic, Y. Wang, M. Kvatyuk, T. Lu, R. L. Coffman, S. Pestka, P. B. Rothman. 1998. Targeted disruption of the interferon- receptor 2 gene results in severe immune defects in mice. Proc. Natl. Acad. Sci. USA 95: 8233-8238.
Sad, S., R. Marcotte, T. R. Mosmann. 1995. Cytokine-induced differentiation of precursor mouse CD8+ T cells into cytotoxic CD8+ T cells secreting Th1 or Th2 cytokines. Immunity 2: 271-279.
Perussia, B.. 1991. Lymphokine-activated killer cells, natural killer cells and cytokines. Curr. Opin. Immunol. 3: 49-55.
Mosmann, T. R., R. L. Coffman. 1989. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu. Rev. Immunol. 7: 145-173(Jodie S. Haring*, Gail A.)