IFN-λ Exerts Opposing Effects on T Cell Responses Depending on the Chronicity of the Virus Infection

Interferons play a key role in limiting virus replication and stimulating adaptive immune responses against virus infections. The IFN-λs (also known as type III IFN; IL-28/29) are a new family of IFNs (1–3) that are found in many species, including humans, mice, bats, chickens, amphibians, and fish (4–7). There are three subtypes of IFN-λ in humans (λ1, λ2, λ3) and two in mice (λ2 and λ3; λ1 is a pseudogene). IFN-λ is highly conserved in human populations, implying strong evolutionary selection for these genes for protection against infections (8). Genetic polymorphisms in IFN-λ are associated with either enhanced clearance of hepatitis C virus (HCV) or poor outcomes (9–13). Although several models demonstrate that IFN-λ signals reduce virus replication in cell lines or in vivo, the role of type III IFNs in adaptive immune responses is less well understood.

IFN-λ are induced by many cell types, including plasmacytoid dendritic cells (pDCs), conventional dendritic cells, peritoneal macrophages, T cells, B cells, eosinophils, hepatocytes, neuronal cells, and epithelial cells, after virus infections or after activation of TLR3, TLR4, TLR7, TLR9, stimulation of RIG-I, or Ku70 (9, 14–25). IFN-λs are induced by either IFN regulatory factor 3 (IRF3), IRF7, or NF-κB pathways (1). The IFN-λs bind as monomers to the λR1 (IL-28Rα), which then pairs with IL-10Rβ to form the functional heterodimer receptor (2, 3). λR signals are transmitted through the JAK1/TyK2, STAT1, STAT2, STAT3, STAT5, and IRF9 pathways to induce transcription of IFN-stimulated genes via ISGF3 (1, 26–28). These signals result in the induction of 2′-5′oligoadenylate synthetase, serine/threonine protein kinase (PKR), ISG56, and IFN-λ2/3 (14, 28). By comparison with IFN-αβR signals, IFN-λR induces longer lived activated (tyrosine-phosphorylated) STAT1 and STAT2, and more strongly induces IFN responsive genes (MX-1, ISG15, TRAIL, SOCS1) (29).

IFN-λ blocks the replication of numerous viruses in vitro, including encephalomyocarditis virus (14), West Nile virus (30), vaccinia virus (22), vesicular stomatitis virus (31), foot and mouth disease (32), HSV 1 (33–35), influenza A virus (36–38), HIV (39), HCV, and hepatitis B virus (31, 40–43). The importance of this pathway in host defense is highlighted by the finding that several viruses have independently evolved mechanisms to block IFN-λ or its function. For example, the NS1 and NS2 proteins of pneumonia virus, a member of the Paramyxoviridae, blocks IFN-λ induction (44). The NS3/4A protease of HCV blocks IFN-λ production (45). A secreted glycoprotein from Yaba-like disease virus blocks IFN-λ signals (46). Vaccinia virus inhibits IFN-λR–mediated signals and gene expression, and blocks the IFN-λR antiviral response through a PKR-dependent pathway (47); moreover, a recombinant vaccinia virus that overexpresses IFN-λ shows reduced levels of replication in vivo, indicating that poxviruses are susceptible to this IFN and have evolved mechanisms to limit its activity (22). Promising results from clinical trials indicate that pegylated IFN-λ treatment diminishes HCV RNA in patients (9, 48).

The receptor for IFN-λ is expressed on dendritic cells (DCs), macrophages, and epithelial cells in the gastrointestinal and respiratory tracts, although evidence suggests that other cell types are responsive to IFN-λ (36, 49, 50). Although some evidence indicates that IFN-λR expression on peripheral leukocytes is not functional (51), other evidence shows clear antiviral innate defense in some of these cells, and IFN-λ signals stimulate monocytes and macrophages to produce IL-6, IL-8, and IL-10 (52). These findings suggest that IFN-λ has an important role in innate immunity as a first-line defense against invading pathogens through skin and mucosal surfaces but may also function during systemic infections. Mice that lack both IFN-αβR and IFN-λR show increased susceptibility against respiratory viruses compared with wild type (WT), IFN-αβR1–knock out (KO), or IL-28Rα-KO (36, 37). rIFN-λ given to mice protects against influenza infection because the IFN-λR is expressed on the epithelial cells targeted by the virus (37). At intestinal mucosal sites, IFN-λ shows antiviral functions that are independent of type 1 IFNs (50), and IFN-λR–deficient (λR-deficient) mice are more susceptible to rotavirus infection than IFN-λR–sufficient mice (50). Thus, the antiviral activities of IFN-λ largely depend on the type of virus and the route of administration of rIFN-λ (36, 37). IFN-λ may exert protective antiviral functions in multiple tissues. For example, pegylated IFN-λ treatment improves immunity to chronic HCV (9, 48), which indicates that it can improve immunity in the liver. Together, these data suggest that IFN-λ may function in immune defense against systemic infections, because numerous immune cell types respond to IFN-λ, IFN-λ improves T cell–based vaccination (53, 54), and IFN-λ reduces hepatropic infection (9, 48).

The effect of IFN-λ on adaptive immune responses is unclear because multiple groups working in different model systems have arrived at varying conclusions. For example, in a macaque DNA vaccination model, vaccine-induced CTL responses were improved when an additional IFN-λ–expressing plasmid was included to enhance granzyme-dependent killing (53). Similarly, an IL-28B–adjuvanted DNA vaccine strongly induced IFN-γ⁺ CD8⁺ T cells and long-lasting Th1-phenotype memory CD4⁺ T cells in macaques (54). However, IFN-λ may have immunoregulatory functions in addition to the demonstrated antiviral functions. The IFN-λR β-chain is shared with several additional cytokine receptors, including IL-10, IL-19, IL-20, and IL-22, which have well-defined immunomodulatory functions after infection. Consistent with this, IFN-λ–stimulated DCs induced the proliferation of regulatory T cells (Tregs) in vitro (55), although overexpression of IFN-λ in vivo resulted in fewer Tregs in a DNA vaccination model (56). IFN-λ signals inhibit the in vitro differentiation of Th2 cells but stimulate Th1 cells (57, 58). Respiratory syncytial virus–infected monocyte-derived dendritic cells (DCs) secrete IFN-λ that limits the in vitro proliferation of CD4⁺ T cells (59). Thus, a mixture of in vitro and in vivo data show that IFN-λ–mediated signals can exert positive or negative effects on T cells.

The overall influence of IFN-λ on innate and adaptive immune responses against systemic virus infections is not understood. In this study, we explored the role of IFN-λ using λR-deficient mice (24) that were given either acute Armstrong CA-1371 strain of lymphocytic choriomeningitis virus (LCMV-Arm) infection or the highly disseminating variant, LCMV-Clone13. We evaluated the effects of λR deficiency on IFN induction, NK cell frequencies, virus-specific B cell responses, and primary and memory T cell responses. We found that λR-deficient mice efficiently induced type 1 IFNs and eliminated acute infection with kinetics indistinguishable from those of WT mice. Virus-specific memory B cell responses and Ab also appeared normal without IFN-λ signals. However, λR-deficient mice showed a 3-fold increase in primary and memory T cell responses compared with WT mice. In contrast, λR-deficient mice were unable to sustain T cell responses when exposed to persistent virus infection. Thus, IFN-λR signals limit T cell responses during acute infection but support T cell responses during persisting virus infection.

BALB/c mice were purchased from Jackson Laboratory (Bar Harbor, ME) and were used as controls for the λR-deficient mice. In some experiments, BALB/cBy.PL-Thy1a/ScrJ mice from the Jackson Laboratory were used as recipients of BALB/c or λR-deficient cells. Mice deficient in IFN-λR1 (IL-28Rα^−/−; λR-deficient) on the BALB/c background were originally generated by ZymoGenetics (Seattle, WA). All animal experiments were performed in accordance with the Institutional Animal Care and Use Committee at the University of North Carolina at Chapel Hill. Adult mice (8–10 wk old) received an i.p. injection of 2 × 10⁵ PFU of LCMV-Arm. Some mice were given an i.v. injection of 2 × 10⁶ PFU of LCMV-Arm or LCMV-Clone13. Viral stocks of plaque-purified LCMV were prepared from infected BHK-21 monolayers. The virus titer in various organs was determined by plaque assay on Vero cell monolayers (60). Some mice were infected with 1 × 10³ CFU recombinant Listeria monocytogenes that expresses OVA (61, 62).

A reverse transcriptase reaction was performed to identify viral RNA (63). RNA was extracted from 5 mg spleen using RNeasy mini kit (Qiagen; http://www.qiagen.com), and cDNA was synthesized using SuperScript II with random primers (Promega; http://www.promega.com). The cDNA was at 37°C for 20 min followed by a denaturation step at 98°C for 5 min. PCR was performed using NP5-001 (5′-TCCATGAGTGCACAGTGCGGGGTGAT-3′) and NP3-001 (5′-GCATGGGAAAACACAACAATTGATC-3′) primers to amplify the NP region of LCMV S RNA. The PCR conditions were 95°C, 15 min (94°C for 30 s; 60°C for 90 s; 72°C for 90 s) × 35 cycles; 72°C for 10 min. Ten microliters of the PCR product were run on gel electrophoresis on a 1.5% agarose gel using a 100-bp ladder as a size reference.

Single-cell leukocyte suspensions were prepared from spleens, and erythrocytes were removed using ACK lysing buffer (Life Technologies-BRL, Grand Island, NY). Single-cell suspensions of splenocytes were surface stained with combinations of fluorescently labeled mAbs that were specific for CD4 (clone RM4-5), CD8 (53-6.7), CD44 (IM7), CD62L (MEL-14), CD19 (6D5), B220 (RA3-6B2), CD11a (M17/4), KLRG1 (2F1/KLRG1), PD1 (10F.9G2), CD127 (A7R34), CD49b (DX5), CD335 (NKp46; 29A1.4), Thy1.2 (30-H12). The intracellular cytokine staining (ICCS) assay was performed as described previously (64, 65). For ICCS for T cells, splenocytes were cultured with or without LCMV or recombinant L. monocytogenes peptide in the presence of brefeldin A. After 5.5 h of incubation, cells were stained for surface markers, washed, fixed with formaldehyde, then permeabilized and exposed to mAbs specific for IFN-γ (XMG1.2), IL-2 (JES6-5H4), TNF (MP6-XT22), or T-bet (4B10). Spleen cells were stimulated with NP_118–126 peptide to quantify virus-specific CD8⁺ T cells, or spleen cells were stimulated with one of several peptide epitopes that bind to I-A^d (GP_176–190; NP_116–130), I-E^d (NP_6–20), or both (Z_31–45) to identify cytokine-producing CD4⁺ T cells (66). Listeria p60_217–225 or LLO_91–99 were used to stimulate CD8⁺ T cells after recombinant L. monocytogenes infection (61, 62). Ab-stained cells were detected by a FACSCalibur cytometer (BD Biosciences), and the data were analyzed with FlowJo software (Tree Star). All mAbs listed earlier were purchased from Biolegend, except for CD11a (eBioscience).

An ELISPOT-based assay to identify LCMV-specific memory B cells capable of differentiating into Ab-secreting cells was performed as previously described (67, 68) with some modifications. In brief, single-cell suspensions of splenocytes at a final concentration of 2 × 10⁶ cells/ml were cultured in 96-well round-bottom culture plates for 5.5 d in the presence of protein A from Staphylococcus aureus (Cowan strain bacteria; Sigma-Aldrich), PWM (Sigma-Aldrich), and CpG oligonucleotide (ODN-2006; Invitrogen); at the end of the culture, the cells were replica plated onto ELISPOT plates (MAHA N4510; Millipore) that were coated with virus-infected BHK lysates, uninfected BHK lysates, or 1 μg/well donkey polyclonal anti-mouse IgG (H+L) (Jackson Immunoresearch). The ELISPOT plates were counted by a reader (Cellular Technology). We calculated the frequency of Ag-specific B cells among total IgG⁺ memory B cells.

Serum IL-10 was measured using the Quantikine Mouse IL-10 Immunoassay Kit (R&D Systems, Minneapolis, MN) according to manufacturer’s instructions. IFN-α was quantified by VeriKine Mouse IFN-α ELISA Kit (PBL, Piscataway, NJ) according to manufacturer’s instructions. LCMV-specific serum Ab was quantified by using ELISA plates (Greiner bio-one, Monroe, NC) that were coated with lysates from LCMV-infected BHK cells as described previously (69).

The graphs show mean data ± SEM. The statistical significance was determined by Student two-tailed t test with Prism 5 software (http://www.graphpad.com). Comparisons were considered significantly different when p < 0.05.

Type 1 IFNs are induced early after LCMV infection. pDCs are responsible for production of IFN-α by day 1, and other cell types make lower amounts of IFN-α later (70). Type 1 IFNs directly inhibit virus replication, activate NK cell responses against infected cells, and stimulate adaptive T cell responses. Given the parallel relationship between IFN-λ signals and type 1 IFN expression, we considered that type 1 IFN expression might be perturbed in the λR-deficient mice. However, λR deficiency did not impact type 1 IFN levels after virus infection, because the λR-deficient mice generated normal levels of IFN-α in the serum (Supplemental Fig. 1A).

Type 1 IFNs and IFN-λ act on NK cells to increase their antiviral functions (71). Cohorts of WT or λR-deficient mice were given LCMV-Arm infection to determine whether IFN-λ affects NK cell responses after infection. Before infection, ∼5% of cells in the blood were NK cells (DX5⁺NKp46⁺) in both groups of mice (Supplemental Fig. 1B). Three days after infection, 10–20% of blood cells were DX5⁺NKp46⁺ NK cells in both groups of Armstrong-infected mice, indicating that NK cell frequencies in the blood increased independently of IFN-λ signals. The number of NK cells in the spleen was similar in both groups of mice at day 3 postinfection (Supplemental Fig. 1C).

IFN-λs induce antiviral activity against a number of viruses, including influenza, poxvirus, herpesvirus, and HCV (14, 22, 35–37, 47, 50, 72). Early analyses showed that λR-deficient mice and WT mice replicate similar levels of virus in the spleen 3 d after LCMV-Arm infection (24), but the analyses did not extend to other tissues or later times to evaluate whether the mice eventually resolved the infection. Because early virus loads can influence T cell differentiation processes, we quantified the amount of infection at several times after infection with LCMV-Arm. At day 4, λR-deficient mice replicated virus in the liver, lung, and kidney to levels comparable with those found in WT mice (Fig. 1A), and viral RNA could be detected in the spleens of both groups by RT-PCR (Fig. 1B). By day 8, the λR-deficient mice and the WT mice reduced the infection to levels below detection by plaque assay, and the RT-PCR analyses showed no evidence of viral RNA at day 8 or later in the spleens (Fig. 1B). IFN-λs play a key role in limiting mucosal infections (36, 37, 50, 72, 73). We considered that the λR-deficient mice might show impaired immunity to LCMV when the virus was given through a mucosal route. However, the intranasal delivery of LCMV resulted in similar levels of infection in the spleen and lung in both groups of mice at day 4 (Supplemental Fig. 2A); by day 8, both groups of mice showed major reductions in the viral load, resulting in similar levels of infection. Tissue samples from long-term immune mice were also analyzed and found to have no infectious virus (data not shown), indicating that there is no virus recrudescence in these mice. Thus, λR signaling does not limit the early burden of virus and is not critical for the rapid resolution of LCMV-Arm when it is given parentally or mucosally.

The resolution of LCMV depends on the formation of large numbers of CD8⁺ CTL, so the earlier viral clearance data imply that λR signals are not needed to generate functional CTL. Type 1 and type 2 IFNs augment antiviral T cell responses (64, 74–77). These IFNs signal directly on responding T cells to increase their accumulation after acute infection. There is also evidence that these IFNs can act through indirect processes to support responding T cell responses. Because the signaling pathways between IFN-λR and IFN-αβR overlap, we investigated whether IFN-λ signals augment virus-specific T cells by comparing peak CD8⁺ T cell responses in WT and λR-deficient mice. The λR-deficient mice showed normal abundances of resting CD8⁺ T cells before infection (Supplemental Fig. 3A and data not shown), which implies that IFN-λ signals are not involved in naive T cell development or seeding of peripheral organs. Upon infection, there was a tremendous increase in the frequency of activated, CD44^hi, CD62L^lo, and CD11a^hi CD8⁺ T cells in the spleens of WT and λR-deficient mice (Supplemental Fig. 3A); however, the frequencies of these cells were higher in the λR-deficient mice compared with the WT mice and corresponded to ∼2-fold greater numbers of CD44^hi, CD62L^lo, and CD11a^hi CD8⁺ T cells in the λR-deficient mice (Fig. 2A). The overall cellularity of the spleens was similar in the two groups before and after infection (Supplemental Fig. 3B), indicating that there was a selective increase in activated CD8⁺ T cells in the λR-deficient mice.

All virus-specific T cells are contained within the emergent CD44^hi and CD11a^hi populations of cells (78–80), so the data in Fig. 2A and Supplemental Fig. 3A suggest that there are more virus-specific CD8⁺ T cells in the λR-deficient mice. Therefore, ICCS was used to quantify epitope-specific CD8⁺ T cells in WT and λR-deficient mice before and at the peak of the T cell response. There was a vigorous NP₁₁₈-specific CD8⁺ T cell response in the λR-deficient mice that was ∼4-fold greater than that seen in the WT mice (Fig. 2B). Combined with total spleen cells counts, this frequency corresponded to >10 × 10⁶ more NP₁₁₈-specific CD8⁺ T cells in the λR-deficient mice than in the λR-sufficient mice (Fig. 2B, bar graph).

Differentiating virus-specific T cells acquire the ability to make large amounts of TNF and IL-2 as they progress into effector and memory cells. Among NP₁₁₈-specific CD8⁺ T cells in both groups of mice, ∼60% of IFN-γ+ve cells also made TNF (Fig. 2C, left panel), and 10–15% of IFN-γ+ve cells also made IL-2 (data not shown). There was no significant difference in the amount of IFN-γ, TNF, or IL-2 produced by NP₁₁₈-specific CD8⁺ T cells at a per-cell level, as indicated by geometric mean fluorescence intensity (data not shown). However, the λR-deficient mice generated 3-fold greater numbers of IFN-γ⁺TNF⁺ and IFN-γ⁺IL-2⁺ CD8⁺ T cells at day 8 compared with λR-sufficient mice (Fig. 2C, right panel). Thus, λR signals do not impact the cytokine output of virus-specific CD8⁺ T cells but restrict T cell number. In contrast with acute LCMV infection, we observed no difference in the expansion of Listeria monocytogenes–specific CD8⁺ T cells in WT and λR-deficient mice (Supplemental Fig. 3C), and the mice cleared the infection in the liver as determined by colony counts on BHI agar (data not shown). Thus, the expression of IFN-λR leads to reduced T cell responses to some acute infections, perhaps correlating with the magnitude of the T cell expansion, which is far greater after LCMV than Listeria monocytogenes. Therefore, we focused our analyses on T cell responses induced by LCMV infection.

The increased T cell response in the λR-deficient mice suggests that IFN-λ is suppressive in WT mice. This effect might be mediated by direct signaling into T cells or could be an indirect consequence of IFN-λ signaling into other cell types. To determine whether T cells need to express IFN-λR, splenocytes from WT or λR-deficient mice were adoptively transferred to separate congenic recipient mice that were subsequently given acute LCMV infection (Fig. 3A). At day 9 postinfection, the NP₁₁₈-specific donor cells were identified by ICCS assay in the recipient mice. The overall number of WT and λR-deficient CD8⁺ T cells was similar in the WT recipient mice (Fig. 3B). These data imply that IFN-λ signaling into other cell types accounts for the difference in the virus-specific T cell number when comparing WT and λR-deficient mice.

Earlier studies using in vitro–stimulated T cell cultures indicated that IFN-λ stimulates Th1 cells and inhibits Th2 cells (57, 58, 81, 82). Therefore, we examined the effects of IFNλ on CD4⁺ T cell responses to infection. At day 8 postinfection, modest levels of CD4⁺ T cell activation were revealed by slight increases in the proportion of CD4⁺ T cells that were either CD44^hi or CD62L^lo (data not shown). Several H-2^d–restricted epitopes have been identified (66), which enables the quantitation of epitope-specific T cells in BALB/c mice using ICCS. In WT mice, ∼0.25% of CD4⁺ T cells were specific for GP₁₇₆ or NP₆, 0.4% for NP₁₁₆, and 0.13% for Z₃₁ (Fig. 4A). These frequencies corresponded to 3–10 × 10⁴ epitope-specific T cells per spleen (Fig. 4B). A similar analysis was performed for the λR-deficient mice, which showed moderately increased percentages of CD4⁺ T cells specific for each peptide. Overall, there was a 2-fold increase in the number of CD4⁺ T cells specific for GP₁₇₆, NP₆, NP₁₁₆, and no difference in responses to Z₃₁ (Fig. 4B). The NP_116–130 peptide contains NP_118–126, which efficiently stimulates H-2L^d–restricted CD8⁺ T cells that are the likely source of non-CD4⁺ T cells that made IFN-γ⁺ in the center dot plots. In both groups of mice, the IFN-γ⁺ CD4⁺ T cells specific for GP₁₇₆, NP₆, and Z₃₁ also produced IL-2. There was a >2-fold increase in the number of IFN-γ⁺IL-2⁺ CD4⁺ T cells specific for GP₁₇₆ and NP₆ (Fig. 4C). In total, these data indicate the CD4⁺ T cell response in WT H-2^d mice matures into small populations capable of making IFN-γ and IL-2, but those responses are only moderately increased in the absence of λR signals.

CD4⁺ T cell responses drive strong antiviral B cell responses after LCMV (69, 83). We examined whether the greater CD4⁺ T cell responses in the λR-deficient mice affected the humoral response after LCMV-Arm infection. WT and λR-deficient mice showed comparable numbers of CD19⁺ B cells before and several times postinfection (Supplemental Fig. 4A and data not shown). The percentage of cells with phenotypic markers of GC B cells (CD19⁺GL7⁺) at day 15 was higher in λR-deficient mice versus WT mice (20 ± 1 versus 15 ± 1%); however, when measured by an in vitro memory B cell assay (67, 68) at day 120, there was no statistically significant difference in memory B cell numbers between WT and λR-deficient mice (Supplemental Fig. 4B). Thus, the elevated GC-phenotype B cells seen at day 15 did not result in more LCMV-specific memory B cell responses. Correspondingly, LCMV-specific serum Ab levels were similar in both groups of mice at days 14 and 49 (Supplemental Fig. 4C). These data indicate that λR signals are not involved in differentiating memory B cell or plasma cells after acute infection.

The earlier data (Figs. 2, 4) show that IFN-λR deficiency leads to an increase in virus-specific CD8⁺ and CD4⁺ T cells. For CD8⁺ T cells, IFNs influence the formation of short-lived effector cells (SLECs) and long-lived memory precursor effector cells (MPECs) (84). We explored whether λR signals affect the development of SLECs (KLRG1^hiIL-7R^lo) or MPECs (KLRG1^loIL-7R^hi). Among activated CD11a⁺ T cells, the λR-deficient mice showed normal frequencies of SLEC but 2-fold higher frequencies of MPECs at day 8 (Supplemental Fig. 3D), which corresponded to a 1.7-fold increase in SLEC number and a 4-fold increase in MPEC number (Supplemental Fig. 3D). These data suggest that the expression of IFN-λR restricts the accumulation of cells with memory potential, which implies there could be long-term effects on T cell memory. Therefore, cohorts of mice were infected with LCMV-Arm, and the virus-specific CD8⁺ T cell number was quantified at various times after infection. The λR-deficient mice continued to have 2- to 3-fold greater numbers of NP₁₁₈-specific CD8⁺ T cells after the peak response (Fig. 5A, top panel). The proportions of memory T cells capable of making IFN-γ with TNF (Fig. 5A, middle panel) or IFN-γ with IL-2 (Fig. 5A, bottom panel) were also increased in λR-deficient mice compared with WT mice. At day 180, the λR-deficient mice continued to have nearly 3-fold more NP₁₁₈-specific CD8⁺ T cells and more CD8⁺CD11a⁺ T cells that expressed the IL-7R^hi MPEC phenotype (Fig. 5B) that is associated with memory.

No information is available about the number and epitope specificity of memory LCMV-reactive CD4⁺ T cells in BALB/c mice. In this study, we followed CD4⁺ T cell memory responses in WT (BALB/c) and λR-deficient mice. Virus-specific memory IFN-γ⁺ CD4⁺ T cells were quantified by ICCS at late times postinfection. At day 180, the NP₁₁₆-specific CD4⁺ T cells were the largest population of epitope-specific T cells in the WT mice; that population was 2-fold greater in the λR-deficient mice (Fig. 5C). Approximately 10⁴ CD4⁺ T cells per spleen were specific for GP₁₇₆, NP₆, and Z₃₁ in both groups of mice. Among the NP₁₁₆-reactive CD4⁺ T cells, a higher percentage made IL-2 in the λR-deficient mice compared with the WT mice (Fig. 5D). These data indicate that IFN-λR signals limit the abundance and cytokine content of immune-dominant memory CD4⁺ and CD8⁺ T cells.

IL-2 signals are involved in both the establishment of memory and in the recall response (85–89). CD8⁺ T cells in λR-deficient mice made slightly more IL-2 per cell than did CD8⁺ T cells from WT mice at days 40, 120, and 180 postinfection (data not shown). The effect was small but significant based on an unpaired Student t test p value of 0.01 to 0.03 at each time, with three to five mice per group. A similar pattern was observed for CD4⁺ T cells (Fig. 5D). This suggests that λR-deficient cells are qualitatively improved compared with cells in WT mice. Therefore, cohorts of WT and λR-deficient mice were immunized, and 4 mo later some were rechallenged with a higher dose of LCMV-Arm. As expected, WT mice mounted a robust recall response by 6 d, with ∼6% of CD8⁺ T cells specific for NP₁₁₈ and able to make IFN-γ (Fig. 5E); the λR-deficient mice generated 2-fold higher frequencies of these cells, and the same pattern was apparent when TNF and IL-2 production were quantified (Fig. 5E). These percentages corresponded to 5–10 × 10⁶ more NP₁₁₈-specific cytokine-producing CD8⁺ T cells per spleen in the rechallenged λR-deficient mice compared with the WT mice. The larger response in the λR-deficient mice upon rechallenge corresponded to the greater number of memory cells in these mice before challenge. Thus, there was no significant increase (Student t test, p = 0.07) for the λR-deficient mice compared with the WT mice when the ratio of effector cells after challenge was normalized to the number of memory cells before challenge. These data indicate that although λR-deficient mice establish more memory cells than WT mice, those memory cells are not inherently better at proliferative responses compared with memory cells in WT mice.

The recall response in both groups of immune mice led to efficient control of the infection (Fig. 5F, triangles). By comparison, both groups of naive mice given the challenge dose continued to show high levels of infection at day 6 (Fig. 5F, circles). Cumulatively, these data show that λR-dependent signals are dispensable for generating protective primary and memory T cell and B cell responses after acute infection and act to restrict the size of the overall response.

The earlier data indicate λR signals are dispensable for primary and memory T cell formation and protection against LCMV-Arm infection. The immunobiology of persisting virus infection is often very different from acute infection, and is largely impacted by the detrimental effects of sustained T cell stimulation, inhibitory molecule expression on T cells, and the presence of immune-suppressive cytokines. Consequently, T cells undergo exaggerated deletion or functional inactivation during these conditions. T cell exhaustion is observed in mice with persisting LCMV infection and in people who are persistently infected with HIV or HCV. To better understand the role of IFN-λ signals on the resolution of chronic virus infection, WT and λR-deficient mice were given LCMV-Clone13, which disseminates and persists. Both groups of mice showed an initial weight loss followed by partial recovery, although the λR-deficient mice endured greater weight loss than WT mice (Fig. 6A). Both groups of mice showed similar levels of virus across time (Fig. 6B). At days 9–20, there was 10⁵ PFU/ml in the serum and >10⁷ PFU/g in the liver, lung, and kidneys in both groups. By day 40, there was a reduction in viremia, and eventually both groups reduced the virus burden to the limits of detection, likely through a combination of T cell and neutralizing Ab-mediated mechanisms. Thus, there is no apparent effect of λR expression on peak viral titers or the longevity of the infection.

Compared with the WT mice, there was a significantly greater decline in the total spleen size in the λR-deficient mice by 8 d postinfection (Fig. 6C). The overall cellularity continued to decline across time to 6 × 10⁶ cells by day 40 in the λR-deficient mice, whereas the WT mice had ∼3-fold more splenocytes. Virus-specific T cell responses were analyzed by ICCS assay at days 8, 40, and 90 after LCMV-Clone13 infection. At day 8, the overall abundance of IFN-γ⁺ CD8⁺ T cells and CD4⁺ T cells was similar for the two groups (data not shown). In the λR-deficient mice, a smaller percentage of activated CD8⁺ T cells were KLRG1^hi, fewer expressed the inhibitory molecule PD-1, and more were CD127^hi (data not shown), which suggested that more T cells might survive to contribute to immune control in the λR-deficient mice. However, by day 40, there were 4-fold fewer virus-specific NP₁₁₈-specific CD8⁺ T cells (Fig. 6D) and 3-fold fewer GP₁₇₆-specific or NP₆-specific CD4⁺ T cells (Fig. 6E) in the λR-deficient mice compared with the WT mice. The reduced T cell responses were near or below the limits of detection, and this pattern was sustained to day 90 (data not shown). Much of this reduction in virus-specific T cell number was due to the lower number of spleen cells in the λR-deficient mice (Fig. 6C). In contrast with Clone13, LCMV-Arm led to 2- to 3-fold greater numbers of virus-specific CD8⁺ and CD4⁺ T cells in the λR-deficient mice compared with WT mice (Fig. 6D, 6E). These data show that IFN-λ functions vary with the chronicity of the infection. During persisting infection, IFN-λ signals protect against infection-induced weight loss and sustain IFN-γ⁺ T cell responses. Following acute infection, IFN-λ limits the size of the overall T cell response and memory.

Earlier analyses of IFN-λ focused on its role in limiting virus infection or replication, but much less is known in terms of how this pathway influences adaptive immunity. In this article, we examined primary and memory T cell responses after acute and persisting virus infection in λR-deficient mice. We found IFN-λR is not essential for mounting innate and adaptive antiviral defense, and appears to regulate peak effector T cell responses and memory cell number after acute infection. In contrast, T cell responses to disseminating infection were reduced in the λR-deficient mice, implying that IFN-λ signals maintain T cell responses during persisting infection. The λR-deficient mice also showed greater weight loss that was prolonged compared with WT mice.

Although the main effect of λR deficiency was on T cells, we observed a modest trend toward increased NK cell frequencies in the blood after infection in the λR-deficient mice (Supplemental Fig. 1B, 1C). The increase achieved significance after Clone13 infection but not after Armstrong infection. NK cells are induced by type 1 and type 2 IFNs, so the increased NK cell response in the λR-deficient mice suggests that IFN-λ signals may counterbalance the inducing effects of these other IFNs. Other investigators have shown that the antitumor activity of IFN-λ is partly mediated by positive effects on NK cell recruitment into the liver and increased tumor killing by NK cells (71). IFN-λ is being tested in clinical trials to treat HCV infection (9, 48, 90); in addition to directly limiting HCV replication in liver cells, IFN-λ may stimulate antiviral activity in hepatic NK cells to reduce HCV. NK cells restrain virus-specific T cell responses after LCMV-Clone13 infection and contribute to the formation of functionally exhausted T cells (91–94). Thus, the increased NK cell response in the Clone13-infected λR-deficient mice might contribute to greater NK cell activity and the eventual decline in virus-specific T cell responses during the chronic stage.

The levels of infectious virus early after LCMV-Arm infection were comparable in λR-deficient and WT mice (Fig. 1A, Supplemental Fig. 2), consistent with an earlier report (24) and another study showing that pretreating mice with 10 μg rIFN-λ does not reduce LCMV levels at day 2 postinfection (14). This implies that λR-deficient mice do not have new cellular targets of infection. IFN-λR signals do not limit the replication of the arenavirus, Lassa virus, in macrophages or DCs in vitro (95) and mice lacking IFN-λR did not show increased levels of Lassa virus replication compared with mice with the IFN-λR following intranasal infection (36). IFN-λR is important for the resolution of other infections at mucosal or liver sites; however, our analyses showed no difference in the lung or liver when LCMV-Arm was given intranasally (Supplemental Fig. 2). LCMV causes a systemic infection that induces robust levels of IFN-αβ and IFN-γ. In this context, direct IFN-αβR and IFN-γR signals may induce sufficient antiviral activity to impair early virus replication, thus overshadowing any antiviral effects mediated by IFN-λ (24).

We found that the overall CD8⁺ T cell response to acute infection was increased in the absence of IFN-λR signals. There are several potential explanations for our findings. First, pDCs express IFN-λR and are a primary source of IFNαβ immediately after LCMV infection. It is plausible that pDCs (55, 96) or a monocyte-derived lineage (52) were less active in the λR-deficient mice, leading to slightly higher Ag loads that stimulate T cells. Such slight differences in Ag load would not be revealed by the plaque assay or RT-PCR analysis but may be sufficient to induce more T cell accumulation. Second, IFN-λ may affect Treg activity. IFN-λ are increased during the chronic stage of HCV and act on DCs to stimulate CD4⁺FOXP3⁺ Tregs (55, 97), perhaps suppressing immune responses to HCV. A novel population of CD4⁺Foxp3⁺CD25⁻ T cells makes IFN-λ and induce tolerance in a mouse experimental autoimmune encephalomyelitis model (98); thus, potentially these tolerance-inducing populations of cells are induced in the infected WT mice but are diminished in the λR-deficient mice. Finally, it has been shown that IFN-λ signals can induce SOCS1 and SOCS3 (99) that suppress inflammatory processes. All of these potential mechanisms are likely to be indirect because it does not appear that IFN-λ signals act directly on T cells. In some models, activated T cells express λR (58, 100, 101); however, we could not detect IFN-λR expression on naive or activated T cells using flow cytometry, and mRNA levels for IFN-λR appeared to be very low based on RT-PCR (data not shown), consistent with other findings for human T cells (102). Moreover, λR-deficient T cells did not respond better than WT T cells when the cells were placed in acutely infected WT mice (Fig. 3). Thus, the effects we observe on antiviral T cell number likely occurred as a result of IFN-λ signaling in other cell types.

Interestingly, the expression of IFN-λR was associated with sustained T cell responses during persisting infection, which contrasts with the restraining effects of IFN-λ on T cells during acute infection. We do not know why antiviral T cell responses collapsed in the chronically infected λR-deficient mice, nor do we know why the effects of IFN-λ are opposite following acute and chronic infection. The effects of cytokines on T cell responses can differ between acute and chronic infections. For example, IL-10 and type 1 IFN also can stimulate or inhibit T cell responses to LCMV, depending on the duration of the infection. IL-10 signals during acute LCMV infection increase the formation of MPECs and long-term memory cells (103), yet limit T cell responses during persisting infection (104–108). Because IL-10 has been implicated in restricting T cell responses after LCMV infection (104, 107–109), we considered that IFN-λR1 deficiency might impact the amount of IL-10 that is present in the infected mice. Serum IL-10 levels rapidly increased on day 1 and declined in both WT and λR-deficient mice after Armstrong or Clone13 infection (Supplemental Fig. 1D). An interesting rebound of serum IL-10 levels on day 15 after LCMV-Clone13 infection was observed, consistent with an earlier report (104), but there was no significant difference in serum levels of IL-10 in the WT or λR-deficient mice at any time during LCMV-Clone13 infection. Type 1 IFNs contribute to the vigorous T cell responses after acute infection but diminish T cell responses during the chronic stage of LCMV-Clone13 infection (110–112) and can induce lymphopenia after virus infections (113, 114). We observed that the Clone13-infected, λR-deficient mice had somewhat smaller spleens than WT mice (Fig. 6C), and we considered that IFN-λ might affect type 1 IFN levels but found that serum levels of IFN-α were similar between WT and λR-deficient mice (Supplemental Fig. 1).

An alternative hypothesis is that IFN-λ inhibits T cell responses to both acute and chronic infections, but the differential effects observed on T cell number after LCMV-Arm and LCMV-Clone13 (Fig. 5) are linked to the tropism of these strains. Thus, LCMV-Clone13 more efficiently infects fibroblastic reticular cells (FRCs) (115, 116) and DCs and their progenitors (117, 118). Once these cells are infected, they become targets for CTL and are destroyed, leading to generalized immune suppression and sustained T cell loss during chronic infection (115–118). FRCs provide factors that support the retention and expansion of T cell responses to acute infection. However, LCMV-clone13 infection of FRCs leads to the disintegration of this important stromal cell network and a subsequent loss of cell populations (115, 116). Thus, a greater CTL response early on in the Clone13-infected λR-deficient mice may explain the exaggerated loss of spleen cell number in these mice (Fig. 6C). Alternatively, it may be that IFN-λ acts on FRCs to protect them from destruction by CTL in the WT mice. Analogous effects could occur for DCs: LCMV-Arm efficiently activates DCs to stimulate T cells and IFN-λ restrains T cell responses. In contrast, LCMV-Clone13 targets DCs and leads to their rapid destruction by CTL (117, 118), and IFN-λ acts to limit CTL-mediated destruction of these DCs. Thus, in the presence of IFN-λ, there could be lengthier DC-driven T cell responses during chronic infection. Finally, it is possible that IFN-λ supports T cell responses during chronic infection by limiting Clone13 infection of FRC or DCs without significantly changing the overall viral burden in the spleen.

Overall, our data indicate that IFN-λ improves T cell responses during chronic LCMV infection. Recent data from clinical trials show that rIFN-λ enhances protection against HCV (48). Although IFN-λ synergizes with IFN-β to reduce viral genomes in infected hepatocytes, our data in Fig. 6 suggest an alternative hypothesis: rIFN-λ improves antiviral T cell number or function during persisting infection to reduce virus levels.

We thank Sean Doyle of ZymoGenetics and Bristol-Myers Squibb for kindly supplying the λR-deficient mice and Jason Botten (University of Vermont), who provided helpful discussions concerning the RT-PCR analysis.

The authors have no financial conflicts of interest.