Cell-intrinsic gp130 Signaling on CD4⁺ T cells Shapes Long-lasting Antiviral Immunity

Abstract

The interleukin-6 (IL-6) cytokine family utilizes the common signal transduction molecule gp130, which can mediate a diverse range of outcomes. To clarify the role of gp130 signaling in vivo during acute viral infection we infected Cd4-cre Il6st^fl/fl mice, in which gp130 is conditionally ablated in T cells, with acute lymphocytic choriomeningitis virus (LCMV). We found that by day 12, but not at day 8, post infection the number of virus specific CD4⁺ T cells was reduced in the absence of gp130, and this was sustained for up to 2 months post infection. Additionally gp130 deficient T_FH had lower expression of Maf, IL-21 and ICOS and this was accompanied by a reduction in the proportion of germinal center B cells and plasmablasts. Remarkably, two months post-infection the proportion of IgG_2a/c⁺ memory B cells and the systemic levels of LCMV-specific IgG₂ Abs were dramatically decreased, while there was a corresponding increase in IgG₁⁺ memory B cells and virus-specific IgG₁ Abs. In the same animals Gp130 deficient virus specific CD8⁺ T cells showed a reduced proportion of memory cells, which expressed lower levels of Tcf7, and displayed diminished recall responses on secondary infection. Mixed bone marrow chimeras revealed that the aforementioned gp130 effects on CD4⁺ T cells were cell-intrinsic. Overall our data show that gp130 signaling in T cells influences the quantity and quality of long lasting CD4⁺ T cell responses as well as CD8⁺ T cell and antibody mediated immunity after acute viral infection.

Introduction

The signal transduction molecule gp130 mediates downstream signaling of all members of the IL-6 family of cytokines, which includes IL-6, IL-11, IL-27, leukemia inhibitory factor (LIF), and oncostatin M (1). In T cells, signaling through gp130 can have overlapping, unique and even contrasting effects depending on the cytokine milieu. IL-6, in combination with TGF-β, causes the differentiation of CD4⁺ T cells into T helper (Th)-17 cells at the expense of T regulatory (Treg) cells (2). Conversely LIF enhances the differentiation of Tregs, counteracting IL-6 (3, 4). IL-27 has the capacity to limit overall Th17 responses but can also promote the differentiation and function of both Tregs and Th17 cells (5–8). Additionally IL-27 can promote Th1 responses through induction of T-bet, and on both CD8⁺ T cell and natural killer cells increases cytotoxic capacity (9–11), while IL-11 can induce Th2 polarization and T cell dependent antibody responses in humans (12, 13).

Recently IL-6 and IL-27 have been implicated in the regulation of a specific subset of CD4⁺ T cells, T follicular helper cells (T_FH), which are critical in the promotion and regulation of humoral immunity (14, 15). In acute infection with lymphocytic choriomeningitis (LCMV) Armstrong 53b (ARM) early IL-6 signaling, via STAT-1, can promote the up-regulation of the transcription factor Bcl-6 and thus early stages of T_FH development (16). IL-6 is not, however, necessary for optimal T_FH or germinal center numbers at the peak of their response during acute LCMV infection or subsequent humoral immune responses (17, 18). In contrast, we found that during chronic infection with LCMV Clone 13 (Cl13) IL-6 is partly responsible for the escalation of T_FH and their optimal expression of ICOS and CD200 at late stages of infection, a process that appears vital for improving LCMV-specific antibodies and eventual viral containment (18, 19).

IL-27 is also important in the development of T_FH and subsequent humoral immunity after TNP-OVA vaccination (15). During chronic viral infection we recently found that while IL-27 signaling on CD4⁺ T cells is not necessary for differentiation of LCMV GP_66–77 specific T_FH it is critical for the survival of T_FH and non-T_FH virus specific CD4⁺ T cells (20). In addition we and others have shown that both IL-6 and IL-27 are capable of driving production of the key CD4⁺ T cell-derived cytokine IL-21, which promotes the proliferation and survival of germinal center B cells (7, 21), the differentiation of plasmablasts (22, 23), as well as survival and/or maintenance of CD8⁺ T cells during chronic viral infection (24–26). In LCMV Cl13 infection in vivo IL-21 production relies on the presence of gp130 on T cells, however neither IL-6 nor IL-27 are essential for its production suggesting redundancy amongst these cytokines and/or other gp130 signaling cytokines involved (18, 20). The role of T cell specific gp130 signaling during acute infection has, however, not yet been explored.

To clarify the role of signaling through gp130 on T cells, and eliminate any redundancy within the IL-6 family of cytokines, we infected mice with conditional ablation of gp130 in T cells with the prototypic acute arenavirus LCMV Armstrong 53b. It has previously been shown that T cell specific deletion of gp130 during infection with gastrointestinal nematode Trichuris muris strongly polarizes the immune responses away from pathogenic Th1/Th17 responses towards protective Th2 responses (27). In the strongly Th1 environment of LCMV ARM infection we found little evidence of increased Th2 cell differentiation in the absence of gp130. We did however find that the number of virus specific CD4⁺ T cells was compromised at day 12 and long after infection. Additionally gp130 deficient T_FH had lower expression of Maf, IL-21 and ICOS. These CD4⁺ T cell defects were cell intrinsic and accompanied by reductions in the proportion of germinal center B cells and plasmablasts. A significant decrease in IgG_2a/c⁺ memory B cells and virus specific IgG₂ levels was also observed in concert with increased IgG₁ two months after infection. At the same time post-infection, Gp130 deficient virus specific CD8⁺ T cells exhibited decreased proportion of memory cells, which exhibit reduced Tcf7 expression, and displayed a diminished recall response on secondary infection. Overall our data show that gp130 signaling in T cells is vital for optimal virus specific CD8⁺ and CD4⁺ T cell responses long after acute infection and that disrupting this pathway has significant effects on lasting humoral immunity and recall responses.

Materials and Methods

Mice and viral stocks

Cd4-cre Il6st^fl/fl mice (on a C57BL/6 background) were kindly provided by Dr. Werner Mueller (University of Manchester, U.K.). CD45.1⁺ (B6.SJL-Ptprc^aPrpc^b/BoyJ) and inbred C57BL/6 mice were obtained from The Jackson Laboratory. All mice were bred and maintained in a closed breeding facility and mouse handling conformed to the requirements of the National Institutes of Health and the Institutional Animal Care and Use Guidelines of UCSD. Unless otherwise stated, 6–8 week old mice were infected intravenously (i.v.) with 2 × 10⁶ PFU of LCMV Armstrong 53b or with 2 × 10⁶ LCMV Cl13. Viruses were grown, identified and quantified as described in (28, 29). Viral quantification was carried out by six well plate plaque assay on vero cells (ATCC). For mixed bone marrow chimeras recipient CD45.1⁺ mice were exposed to 1000 Rads. The following day they were injected with 5 × 10⁶ bone marrow cells i.v. (An equal mix of CD45.1⁺ or Cd4-cre Il6st^fl/fl donor cells). Mice were maintained on oral antibiotics for 2 weeks and reconstitution was continued for an additional 6 weeks prior to infection (18).

LCMV specific antibody ELISAs

LCMV specific ELISAs were done as we and others have previously described using antigen prepared by purifying LCMV on a renograffin gradient (18, 30).

Flow cytometry

Flow cytometry was done as previously described (31). The following fluorochrome labeled antibodies purchased from Biolegend, eBioscience or BD Biosciences were used to stain blood or spleen cells: anti-CD8-pacific blue, -CD4-APC-Cy7, -CD19-PE, -B220-PE-CF594, -CD38-Alexafluor700, -CD38-PE-Cy7, -GL7-FITC, -GL7-efluor660, -CD138-PE, -IgM-APC-Cy7, -IgD-PB, -PD1-PE-Cy7, -PD1-BV605, -ICOS-PE, -ICOS-PE-Cy7, -CD11a-FITC, -CD49d-PerCP-Cy5.5, -KLRG1-FITC, -CD127-PerCP-Cy5.5, -CD8-efluor450, -CD45.1-PE-CF594, -CD45.2-BV650, -IgG₁-FITC, -IgG_2a^b-biotin followed by Strepavidin-BV650, -IFN-γ-APC, -TNF-α-FITC, -IL-2-PE, -CXCR5-BV421. Two-step CXCR5 (BD), Foxp3 (eBioscience), T-bet (Santa Cruz Biotech, Dallas, Tx) and Bcl6 (BD, K112–91) staining were done as previously described (18). Biotinylated D^b GP_33–41 and D^b NP_396–404 monomers along with APC-I-A^b GP_67–77 tetramers were kindly provided by the NIH tetramer core facility (Atlanta, Ga). Monomers were folded using SA-PE or SA-APC (Molecular Probes, Life Technologies). Class I tetramers were stained 1:100 for 1h 15 mins on ice, Class II tetramers were stained 1:100 for 3 hrs at 37°C, followed by normal antibody staining. Cells were acquired using the Digital LSR II flow cytometer (Becton Dickinson, San Jose, CA). Staining for fluorescence assisted cell sorting was conducted in an identical fashion and cells were isolated on a FACSAria (Becton Dickinson, San Jose, CA) to >95% purify. Flow cytometric data were analyzed with FlowJo software (TreeStar, CA).

Ex-vivo T cell stimulation

For MHC class-I-restricted GP_33–41 peptide (2 µg/ml) or MHC class-II restricted GP_67–77 (5 µg/ml) stimulation and staining were carried out as we have previously described (31). For polyclonal stimulation we used PMA (10 ng/ml) and ionomicyn (0.5 µg/ml) in place of peptide. For intracellular IL-21 staining, cells were permeabilized with saponin and incubated with 1:25 dilution of mouse IL-21R-human Fc (R&D Systems) for 30 minutes at 4°C, washed twice and stained with 1:200 anti-human Fc-PE (BD Pharmingen).

Real-time RT-PCR

Total RNA was extracted from splenocytes using RNeasy kits (Qiagen), and reverse transcribed into cDNA using superscript III RT (Invitrogen). cDNA quantification was performed using SYBR Green PCR kits (Applied Biosystems) and a Real-Time PCR Detection System (ABI). Primers for the genes assessed are described in (18), as well as Tcf7-Forward, CAATCTGCTCATGCCCTACC; Tcf7-Reverse, CTTGCTTCTGGCTGATGTCC; Maf-Forward, AAATACGAGAAGCTGGTGAGCAA; Maf-Reverse, CGGGAGAGGAAGGGTTGTC..

Statistical analysis

Non-parametric Mann Whitney tests or ANOVA tests were performed using the InStat 3.0 software (GraphPad, CA) * P < 0.05, ** P < 0.01, *** P < 0.001 in all data shown.

Results

gp130 signaling sustains virus specific CD4⁺ T cell numbers

In order to determine the role of the IL-6 cytokine family on T cell responses during acute viral infection we infected Cd4-cre Il6st^fl/fl (T cell specific gp130 deficient) mice and littermate control, Il6st^fl/fl (cre-negative, herein referred to as WT), mice with LCMV Armstrong 53b (ARM).

During chronic LCMV infection, T cell specific deletion of gp130 significantly reduces the survival of virus specific CD4⁺ T cells at later stages of infection. After acute LCMV ARM infection the polyclonal virus specific CD4⁺ T cells response, as marked by high expression of both CD11a and CD49d (32), in the blood were similar in Il6st^fl/fl and Cd4-cre Il6st^fl/fl mice (Figure 1a). We did, however, find that by day 12 p.i. there was a significant reduction in the proportion and number of I-A^b GP_67–77 specific CD4⁺ T cells in the spleen in the absence of gp130, despite similar numbers being present at day 8 p.i. (Figure 1b). Reduced virus specific CD4⁺ T cell numbers remained observable out to day 60 p.i.. Supporting this observation the number of IFN-γ⁺ CD4⁺ T cells present in the spleen after ex vivo GP_67–77 peptide stimulation at day 12 p.i., but not day 8 p.i., was significantly reduced in the absence gp130 (Figure 1c). Production of IL-21 by virus specific CD4⁺ T cells was decreased in LCMV Cl13 infected animals that lack gp130 signaling in T cells (20). In LCMV ARM infection there also appeared to be a selective, but mild, alteration in cytokine production by virus specific IFN-γ⁺ CD4⁺ T cells by day 12 p.i. when stimulated ex vivo with GP_67–77 peptide (Figure 1d). Specifically TNF-α production was similar between WT and gp130 deficient animals while IL-21 producing CD4⁺ T cells were slightly yet significantly reduced, and IL-2 producing CD4⁺ T cells were increased. Overall, these data indicate that gp130 signaling influences both virus specific CD4⁺ T cell numbers, and cytokine production after LCMV ARM infection.

Figure 1. Gp130 signaling regulates virus specific CD4⁺ T cell numbers and cytokine production.

Il6st^fl/fl or Cd4-cre Il6st^fl/fl mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. (A) The number of CD11a^highCD49d⁺ CD4⁺ T cells in the blood was determined at indicated days p.i.. (B) I-A^b GP_67–77 tetramer⁺ CD4⁺ T cells were monitored in the spleen at indicated days p.i. (C) The proportion and number of IFN-γ⁺ CD4⁺ T cells after GP_67–77 peptide stimulation was determined at days 8 and 12 p.i.. (D) The frequency of TNF-α, IL-21 and IL-2 producing cells within IFN-γ⁺ CD4⁺ T cells after GP_67–77 peptide stimulation was quantified at day 12 p.i. Representative FACS plots are gated on CD4⁺ cells (A–C) or CD4⁺IFNγ⁺ cells (D). Data representative of n = 3 experiments with >4 mice per group. *<0.05, **<0.005.

Gp130 regulates T_FH function

Gp130 signaling is also known to affect a number of other CD4⁺ T cell differentiation fates (1). Furthermore, IL-21 can be thought of as a canonical cytokine produced by T_FH and IL-2, via STAT-5 signaling, has recently been identified as inhibiting T_FH cell fate (33–35), raising the possibility that gp130 induction of IL-21 during LCMV ARM infection (Figure 1d) would result from regulation of T_FH differentiation. We therefore evaluated CD4⁺ T cell differentiation during acute LCMV infection in the absence of gp130 signaling in T cells at day 12 p.i., the earliest time point at which altered CD4⁺ T cell responses were seen. We observed that the number of splenic Foxp3⁺ Treg cells was identical in WT and gp130 deficient animals (Figure 2a). In addition despite the reported role of IL-27 in regulating the canonical Th1 transciption factor T-bet (10), the expression of T-bet was similar in WT and gp130 deficient I-A^b GP_67–77 tetramer⁺ CD4⁺ T cells (Figure 2b). Despite the IL-21/IL-2 skew seen previously (Figure 1c), the proportion of virus specific CD4⁺ T cells that were T_FH (Bcl6⁺CXCR5⁺) was similar, although there was a trend towards decreased numbers of virus specific T_FH in the absence of gp130, which reached significance in 1 out of 3 independent experiments. This was most likely due to the reduced number of total virus specific CD4⁺ T cells in these mice (Figure 1b). Within the T_FH population Bcl6 expression was unchanged, however gp130 deficient T_FH did express less ICOS (Figure 2d and 2e).

Figure 2. Gp130 signaling promotes ICOS, Il21 and Maf upregulation but does not affect T_FH differentiation.

Il6st^fl/fl (filled squares/bars) or Cd4-cre Il6st^fl/fl (empty squares/bars) mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. Spleens were isolated at day 12 p.i. to determine (A) the number of Foxp3⁺ CD4⁺ T cells, (B) the MFI of T-bet in I-A^b GP_67–77⁺ CD4⁺ T cells, (C) the number and proportion of BCL6⁺ CXCR5⁺ T_FH within I-A^b GP_67–77⁺ CD4⁺ T cells, (D) the MFI of BCL6 and (E) the MFI of ICOS in I-A^b GP_67–77⁺ CXCR5⁺ PD1⁺ CD4 T cells by flow cytometry. Average MFI of Bcl6 on naive CD4⁺ T cells from uninfected mice is depicted by the broken line. (F) T_EFF (CD44⁺ PD1^+/− CXCR5⁻ CD4⁺ CD19⁻), T_FH (CD44⁺ PD1⁺ CXCR5⁺ CD4⁺ CD19⁻) and I-A^b GP_67–77 tetramer⁺ CD4⁺ CD19⁻ T cells were FACS isolated. The expression of Tbx21, Bcl6, Il21 and Maf relative to Gapdh was then determined by qPCR. Expression levels in naive (CD62L⁺ CD44⁻) CD4⁺ T cells from uninfected mice are depicted by the broken lines, except Il21 where transcript was undetectable. Data is representative of n = 3 experiments. *<0.05, **<0.01, ***<0.001.

Protein expression levels of T-bet and Bcl6 were confirmed by gene expression in FACS isolated CD4⁺ T effector cells (CD19⁻CD4⁺CD44⁺PD1^+/−CXCR5⁻), T_FH (CD19⁻CD4⁺CD44⁺PD1⁺CXCR5⁺) and virus specific CD4⁺ T cells (CD19⁻CD4⁺I-A^bGP_67–77⁺) from the spleen from day 12 p.i., with both Tbx21 and Bcl6 expression being similar between WT and Cd4-cre Il6st^fl/fl mice in all three cell populations (Figure 2f). In contrast the expression of Il21 was significantly reduced in isolated T_FH and virus specific CD4⁺ T cells, but not T_EFF. This was associated with a significant decrease in the expression of Maf in the T_FH and virus specific CD4⁺ T cells, a transcription factor known to play a role in IL-21 production in T cells (7) (Figure 2f). Taken together these data indicated that signaling through gp130 did not directly affect T_FH differentiation but controlled specific T_FH features such as IL-21 production as well as ICOS and Maf expression after acute LCMV infection.

T cell specific gp130 deficiency alters B cell responses

CD4⁺ T cell help is vital for subsequent humoral immune responses to infections and vaccination. We therefore wished to determine whether the differences observed in T_FH responses in the absence of gp130 signaling in T cells affected this process.

Analysis of B cell responses at day 12 p.i., showed that the number of germinal center (GC) B cells (IgD⁻IgM⁻CD38⁻GL7⁺) was decreased (Figure 3a). There was an even greater reduction in the number of CD138⁺ cells amongst isotype-switched IgD⁻IgM⁻ B cells indicating a strong effect of T cell-gp130 signaling on plasmablast formation (Figure 3b). The low levels LCMV specific IgG and the IgG₁ and IgG₂ subtypes detected in the serum at day 12 p.i. were similar in WT and gp130 deficient mice (data not shown). However, flow cytometric analysis of IgG_2a/c expression on B cells revealed a profound reduction in the proportion of isotype-switched IgD⁻IgM⁻ B cells that were IgG_2a/c⁺ in the absence of T cell specific gp130 at day 12 after ARM infection (Figure 3c). Skewing of antibody responses from IgG_2a toward IgG₁ has commonly been associated with increased Th2 immunity, and gp130 signaling has previously been reported to enhance Th2 to immune responses (27). However, we found that after LCMV ARM infection the expression of Gata3, the canonical transcription factor associated with Th2 differentiation, was not significantly altered in FACS isolated T_EFF, T_FH and virus specific (I-A^bGP_67–77⁺) CD4⁺ T cells in the absence of gp130 (Figure 3d). Indeed Gata3 expression in both WT and gp130 deficient cells remained similar to expression seen in naïve CD4⁺ T cells, indicating an absence of Th2 immune cells (Figure 3d).

Figure 3. Gp130 signaling on T cells regulates early germinal center and plasmablast responses and control the IgG₁ versus IgG2 balance.

Il6st^fl/fl or Cd4-cre Il6st^fl/fl mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. Spleens were isolated at day 12 to quantify (A) germinal center B cells (CD19⁺GL7⁺CD38⁻) and (B) isotype-switched plasmablasts (CD19⁺IgD⁻IgM⁻CD138⁺) by flow cytometry. The number of GC B cells or plasmablasts from uninfected mice are depicted by the broken lines. (C) At day 12 p.i. isotype-switched B cells (CD19⁺IgD⁻IgM) were analyzed for surface expression of IgG_2a/c. The proportion of IgG_2a/c⁺ B cells from uninfected mice is depicted by the broken line. (D) CD4⁺ T cells were isolated as in Figure 2F and Gata3 expression relative to Gapdh was determined by qPCR. Expression levels in naive (CD62L⁺ CD44) CD4⁺ T cells from uninfected mice are depicted by the broken lines. Data representative of 3 experimental repeats with n ≥ 4 mice per group except for (C) which is representative of 2 experimental repeats. *<0.05, **<0.01, ***<0.001.

Together, these data indicate that gp130 signaling in T cells is essential for mounting optimal GC and plasmablast responses as well as IgG_2a/c switched memory B cells after acute LCMV infection.

Gp130 signaling in T cells influences long-lasting T_FH and B cell responses

We next investigated if the aforementioned changes observed in T_FH and B cells from gp130 deficient mice at day 12 p.i. were either transient or long-lasting. For that, we analyzed CD4⁺ T cell and B cell responses in Il6st^fl/fl versus Cd4-cre Il6st^fl/fl mice at day 60 after LCMV ARM infection. We observed that even at this late time point, the frequency of LCMV-specific CD4⁺ T cells that were Bcl6⁺ CXCR5⁺ T_FH was similar between WT and gp130 deficient mice (Figure 4a). As at day 12 p.i., there was a non-significant trend to reduced numbers of virus specific T_FH in the spleen in the absence of gp130, as a result of the significantly reduced number of virus specific CD4⁺ T cells in Cd4-cre Il6st^fl/fl mice (Figure 4a and Figure 1b). Another phenotype maintained since day 12 p.i. was the reduced ICOS (but not Bcl6) expression in gp130-deficient compared to WT virus specific (I-A^bGP_67–77⁺) or polyclonal T_FH cells (Figure 4b and Supplemental Figure 1). Activated CD44⁺ CD4⁺ T cells were FACS isolated as a surrogate for virus specific CD4⁺ T cells given the low number of I-A^bGP_67–77 tetramer⁺ CD4⁺ T cells at this late timepoint. Consistent with an Il21 defect at day 12 p.i., expression of Il21 was significantly decreased in CD44⁺ CD4⁺ T cells from Cd4-cre Il6st^fl/fl compared to WT mice at day 60 p.i. (Figure 4c).

Figure 4. T cell-gp130 controls long lasting B cell responses.

Il6st^fl/fl or Cd4-cre Il6st^fl/fl mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. Spleens were isolated at day 60 p.i. to quantify (A) the proportion and number of I-A^b GP_67–77⁺ CXCR5⁺ Bcl6⁺ CD4⁺ T cells, and (B) their BCL6 and ICOS expression by flow cytometry. MFI of BCL6 and ICOS on naive CD4⁺ T cells from uninfected mice are depicted by the broken lines. (C) The expression of Il21 was determined in FACS isolated CD44⁺ CD4⁺ T cells by qPCR. Expression levels in naive (CD62L⁺ CD44) CD4⁺ T cells from uninfected mice are depicted by the broken lines. (D–E) The number of germinal center (GL7⁺CD38⁻CD138⁻), plasmablast (CD138⁺) and memory (CD38⁺GL7⁻CD138⁻) B cells amongst switched (IgM⁻IgD⁻CD19⁺B220⁺) B cells was determined by flow cytometry, as was the proportion of switched memory CD38⁺ B cells expressing IgG₁ and IgG_2a/c. The numbers and proportion of each B cell compartment from uninfected mice are depicted by the broken lines. Data represents 2 experimental repeats with n ≥ 3 mice per group. *<0.05, **<0.01, ***<0.001.

By day 60 p.i. there were far fewer GC B cells or CD138⁺ plasmablasts in LCMV ARM infected mice than at day 12 p.i. (Figure 3a versus Figure 4d); these numbers, however, were for the most part still above the levels detected in age-matched WT uninfected mice. Despite this contraction, T cell specific deletion of gp130 continued the trend toward reduced numbers of GC B cells and significantly fewer plasmablasts compared to WT mice at this later timepoint (Figure 4d). By day 60 p.i. the number of splenic memory B cells (IgD⁻IgM⁻CD138⁻GL7⁻CD38⁺) was only marginally elevated above uninfected animals and this was not dependent on T cell-gp130 signaling (Figure 4d). Importantly, while the number of memory B cells was unchanged, the loss of T cell-gp130 signaling resulted in a dramatic bias of memory B cells towards IgG₁ expression with a profound reduction in IgG_2a/c⁺ memory B cells (Figure 4e). This B cell phenotype was consistent with a significant shift in LCMV specific antibody isotypes in the serum: there was decreased IgG_2a/c and IgG_2b and increased IgG₁ in Cd4-cre Il6st^fl/fl mice versus WT littermates at both day 30 and 60 after ARM infection (Figure 5). Decreased total Ig levels were also observed in T cell-gp130 deficient mice in 2 of 4 experiments by day 60 p.i., correlating with fewer plasmablasts in these animals.

Figure 5. T cell-gp130 signaling promotes increased circulating IgG_2a/c over IgG₁.

Il6st^fl/fl or Cd4-cre Il6st^fl/fl mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. Anti-LCMV specific antibodies in the serum at day 30 and 60 p.i. were determined by ELISA. Data are representative of n = 4 experiments with n ≥ 3 mice per group. *<0.05, **<0.01, ***<0.001.

To confirm that the changes in B cell responses in Cd4cre Il6st^fl/fl mice were most likely via loss of gp130 on T cells, rather than deletion occurring on another Cd4 expressing cell, we analyzed gp130 expression in the spleen at day 60 p.i. Gp130 is constitutively expressed by all immune cells, however gp130 expression by B cells was considerably lower than that observed on the other cell types analyzed (Supplemental Figure 2). Loss of gp130 in Cd4-cre Il6st^fl/fl mice, even 60 days after LCMV ARM infection, appeared to be predominantly restricted to T cells with no decreased gp130 expression seen in any other cell type analyzed, even among CD4⁺ populations of DCs.

Altogether, these data suggest that when T cells do not receive signaling via gp130 there is significant alteration in long lasting T_FH and B cells, both in the scale and quality of the responses.

Gp130 signaling promotes formation of virus specific memory CD8⁺ T cells

As Cd4-cre Il6st^fl/fl mice also lack gp130 on their CD8⁺ T cells, we next analyzed the formation of virus specific CD8⁺ T cells after LCMV ARM infection. The absence of gp130 signaling did not affect the generation of immunodominant virus specific D^b NP_396–404 and D^b GP_33–41 CD8⁺ T cells, with similar numbers observed in the blood at day 9 post infection (p.i.) although there was a trend for slightly more gp130 deficient virus specific CD8⁺ T cells upon contraction of the response at days 15 and 30 p.i. while the numbers at day 60 p.i. were similar (Supplemental Figure 3a and 3b). Similarly in the spleen, absence of gp130 signaling in T cells had little effect on the proportion or number of D^b NP_396–404 or D^b GP_33–41 specific CD8⁺ T cells at days 8, 12 and 60 p.i. (Figure 6a).

Figure 6. Gp130 signaling on T cells promotes memory CD8⁺ T cell responses.

Il6st^fl/fl or Cd4-cre Il6st^fl/fl mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. (A) D^b NP_396–404 tetramer⁺ and D^b GP_33–41 tetramer⁺ CD8⁺ T cells were monitored in the spleen at day 8, 12 and 60 p.i. (B) The proportion of circulating D^b NP_396–404 tetramer⁺ CD8⁺ T cells expressing KLRG1, CD127 or both was examined throughout infection and (C) the number of splenic D^b NP_396–404 tetramer⁺ memory (KLRG1⁻CD127⁺) 60 days p.i. with LCMV ARM was calculated. (D) D^b NP_396–404 tetramer⁺ CD8⁺ T cells in the spleen at day 5 post re-challenge with 2 × 10⁶ pfu of LCMV Cl13 were determined. (E) The expression of Tcf7 was determined in FACS isolated KLRG1⁻CD127⁺ splenic D^b NP_396–404⁺ CD8⁺ T cells at day 60 post LCMV ARM infection. Data are representative of n ≥ 2 experiments with at least 3 mice per group. *<0.05, **<0.01, ***<0.001.

Previous reports have suggested that IL-27 signaling may regulate CD8⁺ T cell function through the transcription factor T-bet (10, 36), however we saw no significant difference in T-bet levels in D^b NP_396–404 specific CD8⁺ T cells in the absence of gp130 (Supplemental Figure 3c). In addition there was no observable difference in the number of NP_396–404 specific CD8⁺ T cells secreting the cytokines IFN-γ, TNF and IL-2 (Supplemental Figure 3d). Together these data indicated that while gp130 signaling affected the number of virus specific CD4⁺ T cells, it did not appear to influence the kinetics of virus specific CD8⁺ T cells.

We next analyzed the frequency of terminal effector (KLRG1⁺ CD127⁻) and memory (KLRG1⁻ CD127⁺) CD8⁺ T cells in the blood of LCMV infected mice (Figure 6b). At late/memory timepoints after acute infection (days 30 and 60 p.i.) gp130 deficient animals showed a significantly higher frequency of terminal effector (KLRG1⁺CD127⁻) NP_396–404 specific CD8⁺ T cells compared to WT mice. Concomitantly there was a significant reduction in the frequency of memory (KLRG1⁻CD127⁺) CD8⁺ T cells at the same timepoints. This observation was confirmed in the spleen where there was a significant reduction in the number and proportion of D^b NP_396–404⁺ memory CD8⁺ T cells at day 60 p.i. (Figure 6c). To determine if this affected CD8⁺ T cell recall responses, we then re-challenged these mice between day 60 and 70 with the more virulent LCMV Cl13 strain. Despite similar quantities of total D^b NP_396–404⁺ CD8⁺ T cells immediately prior to re-challenge (Figure 6a), the number of D^bNP_396–404⁺ CD8⁺ T cells in the spleen was significantly reduced in the absence of gp130 upon re-challenge (Figure 6d), consistent with a reduced presence of memory CD8⁺ T cells seen in (Figure 6c). These results indicate that gp130 signaling in T cells influences the formation of memory CD8⁺ T cells and can affect the virus specific CD8⁺ T cell re-call response on re-challenge. Tcf1, encoded by Tcf7, is a transcription factor known to be important in maintaining long lasting KLRG1⁻CD127⁺, but not KLRG1⁺CD127⁻, CD8⁺ T cells and Tcf7^−/− CD8⁺ T cells fail to expand on secondary challenge (37). In the absence of gp130, Tcf7 expression was significantly reduced in KLRG1⁻CD127⁺ CD8⁺ T cells, which may help explain both the reduced frequency of these cells, and their limited expansion, on re-challenge (Figure 6e).

There was also a reduced number of both polyclonal (CD11a⁺CD49d⁺) and virus specific (I-A^bGP_67–77⁺) specific CD4⁺ T cells at day 5 post challenge (Supplemental Figure 4). This could be indicative of a common failure of gp130 deficient memory cells in expansion on re-challenge, however unlike virus specific CD8⁺ T cells (Figure 6a), there were also fewer virus specific CD4⁺ T cells prior to challenge (Figure 1b).

Intrinsic Gp130 signaling is critical for CD4⁺ T cell numbers and function

T-cell specific deletion of gp130 resulted in long lasting alterations to both the CD4⁺ and CD8⁺ T cell compartments. CD4⁺ T cell help can be essential in maintenance of CD8⁺ T cell memory. Indeed IL-21, which is predominantly produced by CD4⁺ T cells in vivo and reduced in the absence of gp130, has been shown to potently upregulate Tcf1 in CD8⁺ T cells (38). To determine whether gp130 signaling on CD4⁺ T cells, CD8⁺ T cells or both was responsible for the T cell defects observed in Cd4cre Il6st^fl/fl mice we generated WT:Cd4cre Il6st^fl/fl mixed bone marrow chimeras and infected them with LCMV ARM. At day 12 p.i. the proportion of gp130 deficient virus specific (I-A^bGP_67–77⁺) CD4⁺ T cells was significantly lower than that observed in WT mice (Figure 7a). In addition while the per cell production of IFN-γ by each virus specific CD4⁺ T cell appeared similar, virus specific CD4⁺ T cells had reduced levels of TNF-α and IL-21 when stimulated ex vivo (Figure 7b). Consistently Il21 expression in I-A^b GP_67–77 tetramer⁺ CD4⁺ T cells was also lower in gp130 deficient cells than in WT cells directly ex vivo (Figure 7c) and was associated with lower expression of Maf in T_FH cells (Figure 7d). We also observed that significantly fewer gp130 deficient CD4⁺ T cells were T_FH (CXCR5⁺Bcl6⁺ or alternatively gated as CXCR5⁺PD-1⁺) than WT CD4⁺ T cells (Figure 7e), possibly due to a reduced proportion of virus specific CD4⁺ T cells in the gp130 deficient compartment (Figure 7a). Importantly, gated T_FH cells exhibited significantly reduced levels of ICOS (Figure 7f and 7g), demonstrating the intrinsic nature of ICOS downregulation described above in the non-chimeric setting (Figure 2e and Figure 4b). Of note, in contrast to the non-chimeric mice, BCL-6 expression was also reduced in gp130 deficient versus WT gated T_FH cells when analyzed in the competitive chimeric setting (Figure 7g). There was however no observable defect in the proportion of D^b NP_396–404⁺ CD8⁺ T cells nor the frequency of those cells that were CD127⁺ KLRG1⁻ in the absence of gp130 (Supplemental Figure 4c & d). This data supports a direct role for gp130 signaling in regulating the quantity and quality of CD4⁺ T cell responses after an acute viral infection.

Figure 7. CD4⁺ T cell intrinsic gp130 signaling regulates virus specific CD4⁺ T cell accumulation and T_FH functionality.

Mixed bone marrow chimeras were generated using a 1:1 ratio of bone marrow from WT C57BL/6 (CD45.1⁺) and Cd4cre Il6st^fl/fl (CD45.2⁺) donor cells. 8 weeks after reconstitution chimeras were infected with 2 × 10⁶ pfu of LCMV ARM and the spleen analyzed at day 12 p.i.. (A) The frequency of I-A^b_67–77 tetramer⁺ CD4⁺ T cells, (B) the proportion of I-A^b_67–77⁺ CD4⁺ T cells that were IFN-γ⁺ following GP_67–77 peptide stimulation and the proportions of IFN-γ⁺ cells that were TNF-α⁺ or IL-21⁺ in CD45.1⁺ and CD45.2⁺ compartments were determined. (C) Il21 expression in I-A^b_67–77⁺ CD4⁺ T cells and (D) Maf expression in T_FH cells within the CD45.1⁺ and CD45.2⁺ compartments, as isolated by FACS, were determined by qPCR. Expression levels in naive (CD62L⁺ CD44⁻) CD4⁺ T cells from uninfected mice are depicted by the broken lines. (E) Frequencies of polyclonal T_FH cells (gated either CXCR5⁺Bcl6⁺ or CXCR5⁺PD1⁺), (F) ICOS expression on CXCR5⁺Bcl6⁺ T_FH cells, and (G) BCL6 expression on CXCR5⁺PD1⁺ T_FH cells within the CD45.1⁺ and CD45.2⁺ compartments were determined by flow cytometry. Data represents 2 independent experiments of n = 5 mice per experiment. *<0.05, **<0.01, ***<0.001.

T cell gp130 signaling is not essential for viral control during acute LCMV infection

LCMV Cl13 infected Cd4-cre Il6st^fl/fl mice fail to control virus, leading to prolonged persistence (20). In contrast, viremia after LCMV ARM infection was similar in WT and Cd4-cre Il6st^fl/fl mice, and no virus was detectable by day 9 p.i. (Figure 8a). In addition, at day 5 p.i. tissue viral loads were also similar between WT and gp130 deficient animals (Figure 8b). As gp130 deficiency resulted in reduced memory CD8⁺ T cells and altered antibody immune response we next assessed whether these effects could result in reduced immune protection. We re-challenged WT or Cd4-cre Il6st^fl/fl mice with the more aggressive LCMV Cl13 virus 60 to 70 days after primary infection with LCMV ARM. By day 5 post challenge, both WT and gp130 deficient animals had no detectable virus in the serum (Figure 8c). These results suggested that while gp130 signaling on T cells regulates both long lasting humoral and CD8⁺ T cell memory it is not essential for viral control during either primary or secondary viral infection with LCMV.

Figure 8. T cell-gp130 signaling does not affect viral control during acute primary LCMV infection or secondary chronic LCMV infection.

Il6st^fl/fl or Cd4-cre Il6st^fl/fl mice were infected i.v. with 2 × 10⁶ pfu of LCMV ARM. (A–B) Viral loads were determined by plaque assay in serum at days 5 and 9 p.i. (A) or indicated tissues at day 5 p.i. (B). (C) At 2 months p.i., mice were re-challenged with 2 × 10⁶ pfu of LCMV Cl13 i.v. Viral loads were determined in the serum at 5 days post challenge and compared to WT mice receiving LCMV Cl13 that had not previously been infected with LCMV ARM (1° WT). Data represents 2 independent experiments. n.d., not detected.

Discussion

The IL-6 cytokine family can influence a wide diversity of T cell functions and fates in spite of common use of the gp130 signal transduction protein. Therefore to determine the direct role of signaling through gp130 we utilized T cell specific gp130 deficient mice. Upon infection with acute LCMV, gp130 signaling was critical for the maintenance of splenic virus specific CD4⁺ T cells, but not CD8⁺ T cells, numbers at later stages of infection. This signaling pathway intrinsically modulated expression of key molecules and gene associated with T follicular helper cells (T_FH) including Maf, ICOS and IL-21, consequently influencing humoral immune responses. Notably, deficiency in gp130 signaling appeared to not dramatically affect other T helper subsets. We also found that gp130 signaling in T cells regulated the ratio of effector to memory CD8⁺ T cells and impacted the recall of these cells during re-infection, which again appeared to be mediated through gp130 signaling events in virus specific CD4⁺ T cells.

CD4⁺ T cell help is critical in both cytotoxic CD8⁺ T cell-mediated and humoral B cell-mediated immunity. Here we found that signaling through the gp130 pathway on CD4⁺ T cells is vital for the maintenance of higher virus specific CD4⁺ T cell numbers during the contraction and memory phases of the T cell response. IL-6 can act in an anti-apoptotic capacity in resting and primed CD4⁺ T cells, as well as CD4⁺ T cells during recall responses (39–41). Likewise, IL-27 signaling was found to be important in a T cell mediated colitis model (42) and chronic LCMV infection (20) through promoting CD4⁺ T survival. IL-6 deficient mice have not previously been found to have a defect in virus specific CD4⁺ T cell numbers during acute LCMV infection (17, 18), while in chronic LCMV infection we have previously shown that gp130 signaling drives survival of virus specific CD4⁺ T cells via IL-27, and not IL-6, signaling (18, 20). Taken together these data suggest that both intrinsic IL-27 and IL-6 likely play complementary roles in maintaining virus specific CD4⁺ T cells numbers at optimal levels in a variety of situations.

Elevated IL-21 production, predominantly by CD4⁺ T cells, has been associated with improved responses during chronic infections in both mice and humans (24–26, 43–45). Contrastingly IL-21 is thought to play a key role in inflammatory conditions such as diabetes, multiple sclerosis, Sjogren’s syndrome and brain injury after stroke (46–50). Previously we showed that IL-21 production in chronic infection is driven by gp130 signaling (20). Here we found that in the context of an acute stimulus, gp130 signaling on CD4⁺ T cells in part promotes IL-21 in vivo, although after restimulation ex vivo there was still significant IL-21 production. Nevertheless this indicates the potential of this pathway to modulate IL-21 production in a number of different contexts.

In particular IL-21 production by T_FH appeared to be compromised in the absence of gp130 signaling and no effect on Il21 expression was observed in non-T_FH cells. Fitting with this, while in T cell-gp130 deficient animals the generation of CXCR5⁺ Bcl6⁺ T_FH was not significantly reduced, the number of germinal center B cells and plasmablasts were dramatically reduced, potentially resulting from the reduced ICOS expression and IL-21 production by T_FH at day 12 p.i. Intriguingly in a competitive setting, gp130 deficiency adversely affect Bcl6 expression in T_FH cells indicating there may be a progressive requirement for gp130 signaling in optimal T_FH formation and function. Importantly, analysis of mice at 2 months p.i. indicated that the aforementioned defects are in great part carried on to the memory phase and that gp130 signaling is required for optimal lasting CD4 T cell immunity after acute viral infection.

Importantly in the absence of T cell-gp130 signaling, there was a significant shift in the type of antibodies produced, with increased IgG₁ and reduced IgG_2a/c and IgG_2b in Cd4-cre Il6st^fl/fl mice, and this defect also lasted up to 2 months post-infection. T cell specific gp130 signaling has been associated with an increase in Th2 cell formation, which might account for the skewed antibody response; here, however, we found that no apparent increase in Gata3 expression within the virus specific CD4⁺ T cell population, suggesting an alternative T cell specific role of gp130 in IgG₂ versus IgG₁ switch.

Gp130 signaling also regulated other T_FH specific features that could influence B cell responses including the expression of ICOS. Signaling through ICOS has been linked to T_FH migration and differentiation (51, 52). While we observed no difference in CXCR5 or Bcl6 expression, indicative that T_FH differentiation and migration were unaffected, in the non-chimeric setting, Bcl6 was down-regulated in total gp130 deficient T_FH cells when competing with WT cells in mixed chimeras. ICOS has also been implicated in c-Maf dependent transcription of IL-21 (53) and it is therefore possible that ICOS down-regulation contributes to the reduced expression of Maf and IL-21 observed in gp130 deficient CD4⁺ T cells. Recently it has also been shown that ICOS is required to act in a “feed-forward” loop that allows the T_FH to continuously provide B cell help once the germinal center has formed (54) and this could be related to the B cell defects observed in the T-cell-gp130 deficient mice after acute LCMV infection. Thus. Overall these results highlight the role of the gp130 pathway in modulating humoral immunity through functional changes in T_FH responses.

In addition to alteration in virus specific CD4⁺ T cell responses we also observed a preferential bias towards terminal effector KLRG1⁺CD127⁻ CD8⁺ T cells with reduced generation of virus specific KLRG1⁻CD127⁺ CD8⁺ T cells in the absence of gp130 signaling. Fitting with this there was reduced number of D^b NP_396–404 specific CD8⁺ T cells on re-challenge, despite similar numbers of total NP_396–404⁺ CD8⁺ T cells before secondary infection. In line with this, IL-21 has been linked to maturation of memory CD8⁺ T cells via an IL-10 and STAT3 enabled pathway, and therefore the reduction in memory CD8⁺ T cells could be a consequence of reduced IL-21 production by CD4⁺ T cells (55, 56). This hypothesis is supported by the reduced expression of Tcf7 observed in memory CD8⁺ T cells isolated from Cd4-cre Il6st^fl/fl mice. Alternatively IL-27 is known to directly promote production of IL-10 from CD8⁺ T cells and potentially drive memory precursor like effector cells (57, 58). In a mixed chimeric setting, however, no defects were observed in gp130 deficient memory CD8⁺ T cells early after acute LCMV infection, supporting a cell extrinsic gp130 signaling mechanism over an cell intrinsic one.

Overall, we have shown the capacity of gp130 signaling in T cells to orchestrate a variety of key T cell functions in vivo including CD8⁺ T cell memory formation and the functionality of T_FH in response to a prototypic acute viral infection.

Abbreviations used in this article

LCMV

lymphocytic choriomeningitis virus

ARM

Armstrong 53b

Cl13

Clone 13

T_FH

T follicular helper cells