Late Interleukin-6 Escalates T Follicular Helper Cell Responses and Controls a Chronic Viral Infection

See allHide authors and affiliations

Science  11 Nov 2011:
Vol. 334, Issue 6057, pp. 825-829
DOI: 10.1126/science.1208421


Multiple inhibitory molecules create a profoundly immunuosuppressive environment during chronic viral infections in humans and mice. Therefore, eliciting effective immunity in this context represents a challenge. Here, we report that during a murine chronic viral infection, interleukin-6 (IL-6) was produced by irradiation-resistant cells in a biphasic manner, with late IL-6 being absolutely essential for viral control. The underlying mechanism involved IL-6 signaling on virus-specific CD4 T cells that caused up-regulation of the transcription factor Bcl6 and enhanced T follicular helper cell responses at late, but not early, stages of chronic viral infection. This resulted in escalation of germinal center reactions and improved antibody responses. Our results uncover an antiviral strategy that helps to safely resolve a persistent infection in vivo.

Chronic viral infections, such as human immunodeficiency virus–type 1 (HIV-1) and hepatitis B and C viruses (HBV and HCV) in humans and lymphocytic choriomeningitis virus (LCMV) in rodents, create an altered immune environment in the infected host. This is characterized by deletion and functional exhaustion of T cell responses (1, 2), delayed and often dysfunctional appearance of antibodies (3, 4), and dysregulation of innate immunity (5, 6). These enable the virus to persist and make the host extremely susceptible to a range of secondary infections, inflammatory disorders, and cancers (7, 8). Despite this inhibitory environment, the remaining immune responses can often elicit partial (or even complete) control over persistent infections, but the molecules promoting such responses remain poorly understood. Classical antiviral mediators such as type I interferons (IFNs) are attenuated early and throughout the course of chronic viral infection (4, 5, 9), whereas CD4-derived interleukin-21 (IL-21) is critical for helping CD8 T cell responses and viral control during chronic LCMV and HIV-1 infections (1014). This suggests that the host immune system uses only select antiviral strategies to contain a pathogen once it has productively spread in vital tissues. A greater understanding of such strategies may lead to more effective, and safer, therapeutic approaches to alleviate chronic infections.

To gain insight into the molecules governing immunity during chronic viral infections, we infected mice with LCMV clone 13 (Cl 13), a persistent variant of LCMV (15), and analyzed cytokine production throughout infection. We determined the serum levels of over 30 different cytokines and chemokines between day 1 and 30 post infection (p.i.), with higher resolution between days 20 to 30 p.i., a time period that precedes the decline in viremia during LCMV Cl 13 infection. As we have previously reported (5), type I IFN levels rapidly increased on day 1 p.i., with little or no detectable IFN-α in the serum between day 5 p.i. and the end of the study at day 30 p.i. (fig. S1A). A similar pattern of acute secretion was observed with most cytokines studied during Cl 13 infection (fig. S1B). In contrast, a profile of two-wave inflammation was revealed by IL-6 and granulocyte colony-stimulating factor, with a strong initial peak on days 1 and 3 p.i., followed by a second significant peak around day 25 p.i. (Fig. 1A and fig. S1C). Acute infection with LCMV Armstrong 53b (ARM) resulted solely in the initial peak of both IFN-α and IL-6 (fig. S2). Remarkably, IL-6 was essential for clearing LCMV Cl 13 from blood and all tissues studied. IL-6 knockout (ko) mice (16) had between 105 to 107 plaque-forming units (PFU) of virus up to 450 days p.i., in stark contrast to wild-type (WT) mice that had eradicated the virus from most tissues, except kidneys, where low levels remained (Fig. 1B and fig. S3A). As previously reported, IL-6 ko mice showed normal viral clearance during acute ARM infection (16) (fig. S3B). These data revealed a biphasic inflammatory response that was specific for chronic LCMV infection and involved IL-6 production, which was vital for viral control.

Fig. 1

Biphasic IL-6 is produced by radiation-resistant cells and is essential for virus control during chronic LCMV infection. (A) C57BL/6 WT mice were infected with LCMV Cl 13, and serum IL-6 concentrations were determined by enzyme-linked immunosorbant assay (ELISA) throughout infection. (B) WT or IL-6 ko mice were infected with LCMV Cl 13, and virermia was determined by plaque assay. (C and D) WT or IL-6 ko mice were lethally irradiated and reconstituted with BM from either WT or IL-6 ko mice, and 8 weeks later mice were infected with LCMV Cl 13. Serum IL-6 levels (C) and viremia (D) were determined. (E) Il6 expression relative to glyceraldehyde phosphate dehydrogenase (gapdh) was determined in fluorescence-activated cell sorting (FACS)–isolated PICD45, PICD45FDCM1CD21/35, and PICD45FDCM1+CD21/35+ splenocytes from WT mice either naïve, at day 1, or day 25 after Cl 13 infection. (A) to (D) show the mean ± SEM of ≥four mice per group representative of more than two independent experiments, and statistical comparison were performed by two-way analysis of variance (ANOVA). *WT>WT versus WT>IL-6 ko or #WT>WT versus IL-6 ko>IL-6 ko. (E) data are pooled from more than two mice per group and representative of three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001.

Infection of fully reconstituted bone marrow (BM) chimeras of IL-6ko BM into lethally irradiated WT hosts (IL-6 ko>WT) resulted in serum IL-6 levels similar to those seen in the WT>WT mice (Fig. 1C and fig. S4). In contrast, WT>IL-6 ko mice showed only minor IL-6 production at day 1 p.i. and no detectable IL-6 for the remainder of infection. Viremia in these mice mirrored serum IL-6, with WT>WT and IL-6 ko>WT mice showing significantly reduced viral loads by day 60 p.i. and thereafter compared with either WT>IL-6ko or IL-6ko>IL-6ko mice (Fig. 1D). The spleen appeared to be an important source for IL-6 (fig. S5A), and this was consistent with up-regulated Il6 transcript in splenic leukocytes at day 1 p.i. (fig. S5B) and CD45 cells at day 1 and 25 p.i. (Fig. 1E). Notably, CD45 FDC-M1+ CD21/35+ cells, which showed size, granularity, and gene expression associated with follicular dendritic cells (FDCs) (fig. S6) (17, 18), exhibited the highest levels of Il6 RNA at day 25 (but not day 1) p.i. (Fig. 1E), suggesting that irradiation-resistant FDCs were an important source of late IL-6 during chronic LCMV infection.

IL-6 is a pleiotropic cytokine with described roles in cell survival, differentiation, proliferation, and inflammation (19). This includes induction of IL-21 in CD4+ T cells that could aid CD8 T cell responses (1113, 20, 21). We, however, observed no difference between Il21 RNA or IL-21 protein levels in WT versus IL-6 ko virus-specific CD4 T cells (fig. S7, A and B). Moreover, CD8 T cell responses in WT and IL-6 ko mice at day 30 p.i. (i.e., after the second wave of IL-6 but before viremia became different) were indistinguishable in the number of H2-Db NP396-404– or H2-Db GP33-41–specific CD8+ T cells, their surface expression of the T cell exhaustion marker PD-1 (22), and the degree of functional exhaustion (figs. S7, C to F, and S8). There was also no significant reduction in the numbers of I-Ab GP67-77–specific CD4+ T cells in IL-6 ko Cl 13–infected mice compared to WT mice at day 30 p.i. (fig. S9A). At this time, virus-specific IFN-γ, IL-2, and tumor necrosis factor α (TNF-α) production from CD4 T cells was also similar, implying IL-6 had no role in conventional CD4+ T helper type 1 (TH1) cell development (fig. S9B). IL-6 can also inhibit transforming growth factor–β (TGF-β)–dependent development of regulatory T cells (Tregs) while driving the differentiation of IL-17–secreting T helper (TH17) cells (23). We previously reported sustained TGF-β activity during chronic LCMV infection (24), but neither the FoxP3+ CD4+ T cell responses nor the RNA levels of the TH17 master regulator Rorc were affected by IL-6 deficiency during Cl 13 infection (fig. S9, C and D).

T follicular helper (TFH) cells are defined by a combination of cell-surface markers including antigen specificity, CXCR5, PD-1, CD200, and inducible T cell costimulator (ICOS) and the absence of the signaling lymphocytic activation marker (SLAM) in CD4+ T cells (25). The TFH transcription factor, Bcl6, has recently been identified as being required and sufficient for TFH differentiation (2628), but the signals that lead to Bcl6 up-regulation during viral infection in vivo remain unclear. During Cl 13 infection in WT mice, we and others (29) observed a significant increase in virus-specific TFH cells (defined as I-Ab GP67-77tetramer+CXCR5+CD200+ICOS+SLAMPD1+), with the majority of virus-specific CD4 T cells showing a TFH phenotype by day 30 p.i. (Fig. 2A and fig. S10A). The loss of IL-6 led to a significant reduction in percentage and number of LCMV-specific TFH cells at day 30 (but not day 9) p.i. (Fig. 2A and fig. S10B). Specifically, ICOS and CD200 expression on TFH cells were significantly reduced in the absence of IL-6 at day 30 p.i. (fig. S10C). Whereas Bcl6 transcript and protein levels in LCMV-specific CD4+ T cells normally increased from day 9 to day 30 p.i., this increase was absent in IL-6 ko mice (Fig. 2, B and C, and fig. S10, D and E), a result also seen when CXCR5+BCL6+ CD4 T cells were analyzed (fig. S11). Notably, as described for Rorc (fig. S9D), the expression of Tbx21 and Gata3 master transcriptional regulators for TH1 and TH2 subsets and the BCL6 antagonist, Prdm1, were mostly similar in WT and IL-6 ko LCMV-specific CD4+ T cells (fig. S12). A limitation in their inducing signals [e.g., IL-2 (30)] combined with repression by residual Bcl6 expression (Fig. 2C) (27) may explain the lack of up-regulation in the aforementioned transcriptional regulators. Again, we did not find differences in Il21 even when virus-specific TFH and non-TFH cells were separately analyzed in WT versus IL-6 ko mice in vivo (fig. S13).

Fig. 2

TFH cell and GC responses are increased in an IL-6–dependent fashion at late stages of chronic LCMV infection. WT or IL-6 ko mice were infected with LCMV Cl 13, and splenocytes were analyzed at day 30 p.i. (A) The number and percentages of CD4+ I-Ab GP67-77 tetramer+ CXCR5+ICOS+SLAMCD200+PD1+ TFHfh cells were determined by flow cytometry. (B) Sorted WT and IL-6 ko I-Ab GP67-77 tetramer+ and total CD4+ T cells were isolated, and bcl6 expression, relative to gapdh, was determined by quantitative polymerase chain reaction (qPCR). (C and D) BCL6 protein levels within WT and IL-6 ko I-Ab GP67-77 tetramer+ CD4+ T cells relative to levels in naive CD4+ T cells (C) and GC B cell (CD19+GL7+CD38) formation (D) were determined by flow cytometry. (E and F) LCMV-specific IgG1 and IgG2a (E) and Ig avidity (F) were determined by ELISA in serum from WT and IL-6 ko mice at day 30 p.i. Data are presented as individual mice or as mean ± SEM of at least four mice per group and representative of at least two experiments, with indicative FACS plots and % gated population shown where necessary. *P < 0.05, **P < 0.01, and ***P < 0.001.

TFH cells are central in the development of fully matured germinal center (GC) B cells and the production of high-affinity antibodies (25). Consistent with TFH kinetics, GC B cell responses increased over time in WT mice during Cl 13 infection (fig. S14, A and B). IL-6ko mice had significantly reduced GC B cells at day 30, but not day 9, after Cl 13 infection (Fig. 2D and fig. S14C). LCMV-specific immunoglobulin (Ig) was reduced in IL-6 ko Cl 13–infected mice, with a significant decrease in the LCMV-specific IgG1 subtype but minimal difference in LCMV-specific IgG2a antibodies (Abs) (Fig. 2E and fig. S14D). Antibody avidity was also reduced in IL-6 ko mice (Fig. 2F). IL-6 produced by irradiation-resistant cells was sufficient for TFH differentiation and reconstitution of the GC B cell response and anti-LCMV Ab levels (fig. S15). As previously shown (31), Bcl6 expression and GC responses were not affected by IL6 deficiency during acute LCMV infection (fig. S16). In conclusion, despite the pleiotropic functions ascribed to IL-6, we identified Bcl6 up-regulation in CD4 T cells and induction of TFH–B cell responses as the central effects of IL-6 during chronic viral infection.

We next investigated whether the second wave of IL-6 production was responsible for escalating TFH and B cell responses during chronic LCMV infection. Administration of IL-6 or IL-6R monoclonal (m) Abs into WT Cl 13–infected mice at days >20 p.i. resulted in a significant drop in the number and proportion of TFH cells and Bcl6 expression in LCMV-specific CD4 T cells compared with isotype control administration (Fig. 3, A and B, and fig. S17, A to C). GC B cells and LCMV-specific Abs were also reduced (Fig. 3C and fig. S17D). No changes were observed in virus-specific CD8 T cell number, their PD1 expression, or CD4 Treg numbers during late treatment with IL-6 or IL-6 receptor (IL-6R) mAbs when analyzed at day 30 p.i. (fig. S17, E to G). Additionally we could not observe any change in late TFH responses, GC reactions, or CD8 T cells responses when IL-6R mAb was administered early, on day –1, to day 5 p.i. (fig. S18, A to F). Late treatment of WT Cl 13–infected mice with anti–IL-6R or anti–IL-6 mAbs resulted in prolonged viremia, revealing that IL-6 signaling during this period was essential for optimal viral control (Fig. 3, D and E). These results indicated that late (rather than early) IL-6 was vital for maximizing TFH and GC responses and restraining viral replication in the face of the profound immunosuppressive environment that characterizes established chronic infections (1).

Fig. 3

Late blockade of IL-6 or IL-6R reduces TFH and B cell responses and delays viral clearance. (A to E) WT mice were infected with LCMV Cl 13. Mice received either 150 μg of IL-6R mAb intraperitoneally (ip) every 5 days between days 20 and 45 p.i. or an initial dose of 0.5 mg followed by 0.25 mg ip every 2 days of IL-6 mAb between days 20 and 35 p.i. Control groups were given the equivalent dose, isotype, and regime of Ab treatment. (A to C) At day 30 p.i., splenocytes were analyzed for TFH virus-specific CD4+ T cells by flow cytometry (A), Bcl6 expression in sorted I-Ab GP67-77 tetramer+ CD4+ T cells by qPCR (B), and GC B cell (CD19+GL7+CD38) formation by flow cytometry (C). (D and E) Viremia was determined in mAb-treated mice at indicated time points p.i. by plaque assay. Representative of two experiments with n = 5 mice per group each. *P < 0.05, **P < 0.01, and ***P < 0.001.

Lastly, we sought to elucidate whether CD4 and/or B cells were the direct IL-6 targets during Cl 13 infection. Ex vivo IL-6 stimulation of total or LCMV-specific CD4+ T cells, but not B cells, led to rapid phosphorylation of the main IL-6 transcription factor, signal transducer and activator of transcription 3 (STAT 3), regardless of infection status (Fig. 4A and fig. S19A). Ex vivo IL-6 stimulation of total or LCMV-specific CD4+ T cells isolated at day 0, 8, or 18 after Cl 13 infection resulted in similar increase in the IL-6 prototypical target genes Il6ra and socs3 (Fig. 4B and fig. S19B). Bcl6 and Il-21, however, were more rapidly and/or strongly induced in CD4+ T cells isolated at day 18 p.i. (Fig. 4B and fig. S19B). These data indicated that, despite comparable signaling, the outcome of IL-6 stimulation in virus-specific CD4+ T cells was dynamic and resulted in rapid Bcl6 induction only at late stages of chronic LCMV infection.

Fig. 4

Cell-intrinsic IL-6 signaling on virus-specific CD4 T cells up-regulates BCL-6 and TFH responses during chronic LCMV infection. (A and B) Adoptively transferred CD45.1+ SMARTA CD4 T cells (A) or LCMV-specific I-Ab GP67-77 tetramer+ CD4+ (B) were FACS-isolated from WT Cl 13–infected mice at day 8 and 18 p.i. and stimulated with recombinant mouse IL-6 or nontreated (NT) ex vivo. Phosphorylation of STAT3 was determined by flow cytometry 1 hour after stimulation (A). Levels of socs3, bcl6 (12 hours after stimulation), and Il21 (6 hours after stimulation) were determined by qPCR, and data are shown as -fold increase over unstimulated (B). (C to E) CD45.1 WT mice were lethally irradiated and reconstituted with matched BM from CD45.1 and IL6-R ko, and 8 weeks later mixed chimeras were infected with LCMV Cl 13 and analyzed at day 30 p.i. The proportion of splenic CD45.1 and CD45.2 IAb GP67-77 tetramer+ CD4 TFH cells was determined by flow cytometry (C). IAb GP67-77 tetramer+ CD4 T cells were FACS-isolated, and bcl6 expression relative to gapdh was determined by qPCR (D). Bcl6 protein levels were determined in IAb GP67-77 tetramer+ by flow cytometry (E). (F) About 2 × 106 CD4+SLAMCD62L (TFH-enriched) or CD4+SLAM+ (non-TFH) cells from day 30 after LCMV Cl 13–infected WT mice were transferred intravenously into infection-matched IL-6 ko mice (day 30 p.i.); control IL-6 ko mice received phosphate-buffered solution (PBS). Serum viremia was determined by plaque assay, and significance determined by two-way ANOVA. Changes in the CD4 T cell compartment were assessed by paired t tests between the two populations in individual mice. (A) and (B) are representative of two experiments with at least four mice per group. (C) to (E) are representative of two experiments of at least seven mice per group each. (F) represents one experiment with five mice per group. *P < 0.05, **P < 0.01, and ***P < 0.001.

To determine the cell-intrinsic effect of IL-6 signaling in vivo, we generated mixed chimeras of WT and IL-6R ko mice (32). WT and IL-6R ko cells showed successful T cell and B cell reconstitution before infection (fig. S20), and total as well as LCMV specific CD8+ and CD4+ T cells were similarly represented in WT versus IL-6R ko compartments at day 30 after Cl 13 infection (fig. S21, A and B). The proportions of LCMV-specific TFH cells (analyzed with two different sets of TFH markers) were, however, significantly biased toward WT with respect to IL-6R ko cells (Fig. 4C). WT LCMV-specific CD4 T cells also exhibited up-regulated bcl6 RNA and Bcl6 protein expression compared with their IL-6R ko counterparts at day 30 p.i. (Fig. 4, D and E). On the other hand, the proportion of total and GC B cells were comparable in WT versus IL-6R ko compartments (fig. S21C). These data demonstrated that IL-6R promoted virus-specific TFH responses in a cell-intrinsic fashion but did not directly control GC differentiation, suggesting that the decreased GC responses observed in IL-6 ko mice were secondary to TFH impairment. Accordingly, adoptive transfer of TFH-enriched cells from day 30 Cl 13–infected WT mice into infection-matched IL6 ko recipients resulted in improved GC and Ab responses and enhanced viral control contrasted with either untreated mice or mice that received non-TFH cells from the same WT donors (Fig. 4F and fig. S22). Conversely, SLAM associated protein (SAP) ko mice that showed impaired TFH responses during Cl 13 infection exhibited reduced GCs and failed to clear viremia, despite normal IL-6 levels and enhanced CD8 T cell responses (fig. S23) (33, 34). All together, these results support the idea that TFH are central to IL-6–mediated viral control.

The mechanisms of Bcl6 up-regulation and TFH cell generation remain unclear. CFA immunization requires IL-6 for TFH differentiation, but alum immunization or acute virus infection does not (21, 31, 35). A recent report described functional redundancy between IL-6 and IL-21 to induce TFH cells during acute LCMV infection (36), and this may explain the residual Bcl6 expression and TFH features in IL-6 ko chronically infected mice. However, IL-6 was absolutely essential to reach the optimal Bcl6 and TFH up-regulation during late chronic infection. Although differential location or cell source (i.e., FDCs) of IL-6 may play a role in determining the effect of late, versus early, IL-6 signaling, our data suggest that CD4 T cells are also intrinsically more prone to up-regulate Bcl6 in response to IL-6 at later stages of infection This may be determined by a combination of precise T cell receptor (TCR) affinity (37), sustained TCR stimulation (29, 38), low IL-2Rα signaling (30, 39), and possibly other signals (or lack thereof) that integrate with the IL-6 pathway at different times during infection. On the other hand, because IL-21 was unchanged in IL6 ko mice but ex vivo IL-6 stimulation is capable of driving IL-21 in noninfectious (21, 35) as well as chronically infectious conditions, it is conceivable that redundancy may occur in vivo to secure IL-21 induction during persistent viral infection.

FDCs produce IL-6 that supports GC reactions during immunization (40, 41) and are likely the biologically relevant IL-6 source during late chronic viral infection. Whether late IL-6 production and escalation of TFH cells occur in HIV-1, HCV, and/or other infections during which delayed emergence of GC responses and/or neutralizing Abs have been observed (4, 8, 42) is worthy of further investigation. Indeed, elevated IL-6 has been found in serum from HIV-, HCV-, and HBV-infected patients, but its immune functions in these contexts remain elusive (4345). Boosting IL-6 signaling in CD4 T cells and/or the downstream TFH responses could aid therapies to combat persistent viruses.

Supporting Online Material

Materials and Methods

Figs. S1 to S23

Table S1


References and Notes

  1. Acknowledgments: The data reported in this paper are tabulated in the supporting online material. The authors thank A. Drew (University of Cincinnati), P. Schwartzberg (NIH), M. David (University of California, San Diego), and S. Crotty (La Jolla Institute of Allergy and Immunology) for providing IL-6R ko, SAP ko, stat3 ko, and CD45.1+ Smarta mice, respectively; S. Crotty for insightful discussion; L.-Y. Liou for technical help with initial experiment; and A. Dolgoter for technical assistance throughout. This work was supported by grants from the NIH (AI072752 and AI081923 to E.I.Z. and AI09484). J.A.H. and E.I.Z. have a provisional patent (no. 61/475,511) relating in part to methods of treating chronic viral infections by administering compounds that boost IL-6 signaling and/or TFH responses.

Stay Connected to Science

Navigate This Article