Effector-phase IL-2 signals drive Th1 effector and memory responses dependently and independently of TCF-1

Abstract

Following viral infection, CD4⁺ T cell differentiaton is tightly regulated by cytokines and T cell receptor signals. While most activated CD4⁺ T cells express IL-2Rα after lymphocytic choriomeningtis virus infection, by day 3 post-infection only half of activated T cells maintain expression. IL-2Rα at this time point distinguishes precursors for terminally differentiated Th1 cells (IL-2Rα^hi) from precursors for Tfh cells and memory T cells (IL-2Rα^lo) and is linked to strong TCR signals. Here we test whether TCR-dependent IL-2 links the TCR to CD4⁺ T cell differentiation. We employ a cocktail of anti-IL-2 antibodies to neutralize IL-2 throughout the primary CD4⁺ T cell response to LCMV infection in mice or only after the establishment of lineage-committed effector cells (day 3 post-infection). We report that IL-2 signals drive the formation of Th1 precursor cells in the early stages of the immune response and sustain Th1 responses during its later stages (after day 3). Effector-stage IL-2 also shapes the composition and function of resulting CD4⁺ memory T cells. Although IL-2 has been shown previously to drive Th1 differentiation by reducing the activity of the transcriptional repressor TCF-1, we found that sustained IL-2 signals were still required to drive optimal Th1 differentiation even in the absence of TCF-1. Therefore, we concluded that IL-2 plays a central role throughout the effector phase in regulating the balance between Th1 and Tfh effector and memory cells via mechanisms that are both dependent and independent of its role in modulating TCF-1 activity.

Introduction

Upon exposure to pathogens, CD4⁺ T cells undergo a rapid program of differentiation into one of several T helper subsets. Cytokines control subset polarization and differentiation, although differences in TCR signals also play a role (1). T helper differentiation to acute viral infection, including lymphocytic choriomeningitis virus (LCMV)^2, is dominated by the T helper 1 (Th1) and T follicular helper (Tfh) subsets (2). Th1 cells, regulated by the transcription factor T-bet, mediate protection by secreting cytokines such as IFNγ and TNFα (3). Tfh cells, characterized by expression of the transcriptional repressors Bcl-6 and TCF-1, travel to germinal centers via expression of the homing surface receptor CXCR5 and help B cells mount antibody responses (4, 5). Following resolution of infection, most CD4⁺ effector T cells die. However, a proportion of effector T cells go on to form long-lived memory cells that are detectable throughout the life of a mouse (although with decline over time) and for many years in humans (2, 6–9). While some of these cells acquire the characteristics of central memory T cells that lack immediate effector qualities after reactivation, many CD4⁺ memory T cells retain commitment to either the Th1 or Tfh lineage at a phenotypic, functional, and epigenetic level and rapidly reacquire Th1 or Tfh effector functions following secondary activation (2, 4, 10–16).

While the role of cytokines in promoting Th1, Th2, and Th17 helper lineages is well-understood, it is now clear that cytokines regulate commitment to Th1 and Tfh lineages as well. Interleukin-2 (IL-2) plays a particularly central role in this process. IL-2 drives T cell clonal expansion as well as the development and function of regulatory T cells (Tregs) (17, 18). T cell-intrinsic IL-2 signaling, dependent on expression of the high-affinity IL-2 receptor (as marked by IL-2Rα (CD25) expression), prevents Tfh lineage commitment and instead drives commitment to the Th1 lineage. IL-2-driven T helper cell differentiation relies on the interplay between the transcriptional repressors TCF-1, Bcl-6 and Blimp-1. Under conditions of TCF-1 prevalence Blimp-1 is repressed, and Bcl-6 expression drives Tfh formation and function. Under conditions in which STAT5-mediated IL-2 signaling promotes the activity of the transcriptional repressor Blimp-1, TCF-1 and Bcl-6 are decreased, allowing for robust Th1 differentiation (12, 15, 19–23). IL-2 has also been implicated as an essential factor in the generation of CD4⁺ memory T cells, although in different model systems this role may be exerted in the early or late stages of the effector response (24).

We previously reported that most activated CD4⁺ T cells express IL-2Rα within 1–2 days after LCMV infection. However, by day 3 post-infection IL-2Rα expression became variable, with about half of activated T cells downregulating surface expression. Additionally, expression of IL-2Rα beginning at day 3 after LCMV infection distinguishes precursors for terminally differentiated Th1 cells (IL-2Rα^hi) from precursors for Tfh cells and memory T cells (IL-2Rα^lo) (22). Furthermore, TCR signal strength was linked to sustained IL-2Rα expression (22), suggesting the possibility that TCR-dependent IL-2 signals provide a link between the TCR and CD4⁺ T cell differentiation outcomes. In this study, we sought to address several unanswered questions suggested by these findings. First, what is the functional role of IL-2 in promoting distinct differentiation outcomes for IL-2Rα^hi and IL-2Rα^lo early effector T cells? Second, does IL-2 primarily exert its role during the priming and activation phase of the immune response (prior to day 3), when most activated CD4⁺ T cells undergo a period of IL-2Rα expression, or does it exert its role during the effector phase of the immune response (after day 3), when only some activated CD4⁺ T cells maintain IL-2Rα expression? Third, does IL-2 mediate its effect by regulating TCF-1 activity, or is it required to drive Th1 differentiation even after loss of TCF-1 expression?

To address these questions we employed an in vivo IL-2 neutralization method using a mix of two IL-2-specific antibodies to temporally block IL-2 signals to activated T cells throughout the primary response or only at the initiation of the effector phase (after day 3) (25). Using this approach, we found that effector phase IL-2 signals are required to sustain Th1 effector differentiation. While early blockade of IL-2 prevented the formation of IL-2Rα^hi Th1 precursor cells, IL-2 signals during the first three days of infection were insufficient to for a full Th1 response. Rather, IL-2 signals in the later stages of infection were required to sustain Th1 cell development and function. Additionally, only early blockade of IL-2 enhanced memory T cell numbers. However, sustained IL-2 signals throughout the effector response were required for the optimal formation of Th1-like memory T cell subsets. Lastly, while TCF-1 deficiency favored Th1 over Tfh differentiation as previously shown, sustained IL-2 signals were still required in that setting to drive optimal Th1 differentiation. Therefore, we concluded that IL-2 plays a central role throughout the effector phase of the immune response in regulating the balance between Th1 and Tfh effector and memory cells and that it accomplishes this task via mechanisms that are both dependent and independent of its role in regulating TCF-1 activity.

Material and Methods

Mice.

C57BL/6 mice (6 to 8 weeks old) were purchased from Jackson Laboratories. SMARTA TCR transgenic mice (on a Thy1.1 or CD45.1 background) were housed and maintained in our colony at the University of Utah (26). B6(Cg)-Tcf7^tm1Hhx/J mice, which express a EGFP fluorescent reporter under the control of the Tcf7 locus as well as LoxP sites flanking Exon 2, and B6.129-Gt(ROSA)26Sor^{tm1(cre/ERT2)Tyj}/J mice, which express a CRE-ERT2 transgene under the control of the endogenous ROSA26 promoter, were purchased from Jackson Laboratories and maintained in our mouse colony (27, 28). SMARTA TCR transgenic mice were bred to each of the other two strains to produce Tcf7 iKO SMARTA mice, or only to B6.129-Gt(ROSA)26Sor^{tm1(cre/ERT2)Tyj}/J mice to produce Tcf7 WT SMARTA mice. All mouse experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Utah.

Infections.

LCMV Armstrong 53b was propagated in baby hamster kidney (BHK) cells, titered using Vero cells and stored at −80 C in Dulbecco’s Modified Eagle Medium (DMEM) containing 10% fetal bovine serum (FBS) using established protocols (29). Mice were infected intraperitoneally with 2 × 10⁵ plaque-forming units (PFU) of LCMV Armstrong for harvest time points 8 and 42 days post-infection. For harvest at days 1–4 post-infection mice were infected with 2 × 10⁶ PFU LCMV Armstrong. Listeria monocytogenes expressing the LCMV Glycoprotein amino acids 61–80 (Lmgp61) was grown to log phase in brain heart infusion (BHI) media, and concentrations were measured at 600 nm optical density (O.D.) (16). Mice were injected intravenously (i.v.) with 2 × 10⁵ colony-forming units (CFU).

Cell isolations and flow cytometry.

Mouse spleens were collected into DMEM + 1% FBS media. Single-cell suspensions from spleens were prepared by mechanical disruption, lysis with ACK lysis buffer, and passaged through nylon filters into DMEM + 10% FBS media. Total splenocytes were calculated using a Countess 3 automated cell counter (Thermo Fisher) or hemocytometer. For detection of expression of cell surface molecules, 2 × 10⁶ splenocytes were placed in a 96 well plate, followed by addition of fluorescently-conjugated antibodies in antibody staining buffer (PBS + 1% FBS) for 30–45 minutes on ice. For detection of intracellular molecules, upon completion of surface staining cells were fixed and permeabilized using the Foxp3 Fixation/Permeabilization kit and accompanying protocol (Biolegend, San Diego, CA), followed by intracellular staining for 30–45 minutes on ice. For intracellular cytokine assays, splenocytes were restimulated for 4 hours with 1 μM of GP_61–80 peptide at 37°C in the presence of Brefeldin A (GolgiPlug 1μl/mL), fixed/permeabilized with buffers provided in a kit (Cytofix/Cytoperm, BD Biosciences) and stained with fluorescently labeled antibodies specific to the indicated cytokines. Flow cytometry was performed using a BD LSR Fortessa (BD Biosciences) and analyzed with FlowJo (Treestar, Mountain View, CA). The following antibody clones, purchased from Biolegend, BDBisociences or Thermofisher, were utilized in this study: anti-CD4 clone GK1.5, anti-CD45.1 clone A20, anti-CD45.2 clone 104, anti-CD44 clone IM7, anti-B220 clone RA3–6B2, anti-CD25 clone PC61.5, anti-Ly6C clone HK1.4, anti-CD62L clone MEL-14, anti-IL-2 clone JES6-SH4, anti-IFNγ clone XMG1.2, anti-TNFα clone MP6-XT22, anti-T-bet clone 4B10, anti-Granzyme B clone QA18A28, anti-Bcl-6 clone K112–91, anti-TCF-1 clones C63D9 and S33–966. Research reported in this publication was supported by the Flow Cytometry Core Facility at the University of Utah Health Sciences Center under the NCI award number 5P30CA042014–24.

Adoptive cell transfers.

CD4⁺ splenocytes from naïve donor SMARTA mice were injected i.v. into recipient mice. The percentage of SMARTA T cells was calculated by staining a small portion of cells in PBS + 1% FBS with anti-CD4 and anti-Va2 antibodies (BD Bioscience). SMARTA CD4⁺ T cells were administered via retro-orbital or tail vein injection into C57BL6 mice one day before LCMV infection. For analysis at day 8 and at memory time points we injected 1 × 10⁴ SMARTA cells, while for analysis at days 1–4 we injected 1 × 10⁵ SMARTA cells. To induce Cre translocation to the nucleus in Tcf7 iKO or Tcf7 WT SMARTA mice, tamoxifen emulsions in sunflower oil were administered via intraperitoneal (i.p.) injections at 2–3 mg/injection for four days prior to adoptive transfer.

Neutralizing anti-IL-2 Ab treatments.

S4B6–1 and JES6–1A12 (BioXCell, Lebanon, NH) anti-IL-2 antibody clones were injected i.p. into mice at indicated time points (0.25 mg/injection of each antibody). Each batch of antibody was verified in vivo by assessing decrease in regulatory T cells in uninfected mice and decrease in Th1 cells in LCMV-infected mice.

Statistical analyses.

Statisical analyses were performed using GraphPad Prism. Comparisons between two groups was done using a student’s t-test (unpaired; paired and unpaired for TCF-1 KO experiments), and comparisons between multiple groups was done using Dunnett’s multiple comparisons test One-way ANOVA.

Results

IL-2 drives Th1 formation throughout the effector CD4⁺ T cell response

While most antigen-specific CD4⁺ T cells express IL-2Rα (CD25) within the first 24–48 hours after lymphocytic choriomeningitis virus (LCMV) infection, ~50% of activated effector cells down-regulate IL-2Rα expression by day 3. Day 3 expression of IL-2Rα marks early effector cells fated for Th1 differentiation, while loss of IL-2Rα expression marks early effector cells fated for Tfh or memory T cell differentiation. Therefore, expression of the high-affinity IL-2 receptor, as determined by the surface expression of the alpha sub-unit (IL-2Rα), is a surrogate marker for viral-specific effector Th1 and Tfh fate decisions in T cells with high and low TCR signal strength, respectively (22).

We sought to clarify the role of IL-2 in specifying the fate of early effector T cells with differing patterns of IL-2R expression. To temporally block IL-2 we utilized a previously described approach in which a two-antibody cocktail neutralizes IL-2 in vivo (25). The first antibody clone (S4B6–1) binds to and stabilizes IL-2, preventing it from binding to the high-affinity IL-2R (IL-2Rαβγ) but enhancing its ability to signal through the intermediate affinity IL-2R (IL-2Rβγ). Conversely, the second clone (JES6–1A12) prevents binding to the intermediate affinity IL-2R but enhances IL-2-mediated signaling via the high-affinity IL-2R. When used together, binding to both the high and intermediate affinity IL-2R is prevented. Anti-IL-2 treatment beginning at the onset of infection with lymphocytic choriomeningitis virus (LCMV) resulted in a reduction of regulatory T cells and no change in the number of memory phenotype (CD44⁺IL-2Rβ⁺) CD8⁺ T cells in the spleen by day 3 post-infection (Supp. Fig. 1). To determine whether IL-2 blockade reduced Th1 polarization, as shown in previous studies (25, 30), we adoptively transferred 1 × 10⁴ TCR transgenic SMARTA (CD45.1⁺) T cells into recipient B6 mice (CD45.2⁺) followed one day later by LCMV infection. Neutralizing anti-IL-2 antibody administration was started one day prior to infection (Fig. 1A, “early anti-IL-2”). IL-2 neutralization did not affect the number of SMARTA cells in the spleen at the peak of the effector response (day 8) compared to the untreated group after infection, nor did treatment affect the kinetics of the SMARTA response throughout the effector phase (days 1–8) (Fig. 1B–C, Supp. Fig. 2). To identify Tfh effector cells at day 8 post-infection, we performed flow cytometric detection of the cell surface marker CXCR5 and intracellular flow cytometric detection for the transcription factors Bcl-6 and TCF-1, as previously (21, 23, 31, 32). To detect Th1 effector cells we stained for the cell surface marker Ly6C, the Th1 master transcription factor T-bet and the Th1-associated effector molecule Granzyme B (2, 4, 10, 11, 33). IL-2 neutralization impaired the emergence of Th1 cells expressing Ly6C, Granzyme B, and T-bet, leading to an increased proportion and number of Tfh cells expressing CXCR5 and Bcl-6 (Fig. 1D–J). These findings confirm previous studies demonstrating a T cell-intrinsic role for high-affinity IL-2R-mediated signaling in driving Th1 differentiation and provide additional evidence that combined antibody treatment effectively neutralizes in vivo IL-2 activity (12, 22, 25, 30).

Figure 1.

Both early and late IL-2 neutralization result in decreased Th1 and increased Tfh responses. 1 × 10⁴ SMARTA cells (CD45.1⁺) were injected i.v. into B6 mice (CD45.2⁺), followed by LCMV infection one day later. IL-2 neutralizing antibodies were administered beginning the day before infection (early anti-IL-2) or day 3 post-infection (late anti-IL-2). (A) Schematic showing a timeline of adoptive transfers, LCMV infection, treatments with anti-IL-2 antibodies, and time point analyses throughout the study. (B-C) Representative flow plots and graph show activated SMARTA T cell frequencies and numbers (CD4⁺CD44⁺CD45.1⁺) in untreated, early anti-IL-2 treated and late anti-IL-2 treated. (D) Representative flow plots indicate expression of CXCR5, Ly6C and Bcl-6 by SMARTA in the spleen at day 8 post-infection. (E-G) Graphs indicate the frequency and number of CXCR5⁺Ly6C⁻ and CXCR5⁺Bcl-6^hi GC Tfh SMARTA and the ratio of Ly6C⁺CXCR5⁻ to CXCR5⁺Ly6C⁻ Tfh SMARTA in the spleen at day 8 post-infection. (H) Representative flow plots indicate expression of CXCR5, Granzyme B and Tbet by SMARTA in the spleen at day 8 post-infection. (I-J) Graphs indicate the frequency and number of Tbet^hiGrazyme B^hi Th1 SMARTA and the ratio of Tbet^hiCXCR5⁻ Th1 to CXCR5⁺Tbet^lo Tfh SMARTA in the spleen at day 8 post-infection. Data displayed are n=5 per group and are reperesentative of three separate experiments. Error bars represent standard error of the mean (SEM). *p<0.05, ***p<0.001, ****p<0.0001

Because high-affinity IL-2R expression by activated T cells undergoes dynamic changes through the first three days after LCMV infection, we considered two possibilities: 1) IL-2 acts to specify Th1 lineage commitment in the early stages of the immune response; or 2) IL-2 sustains effector responses by committed Th1 cells, as indicated by high-affinity IL-2R expression at day 3. To determine the role of late IL-2 signals, we initiated IL-2 neutralization at day 3, then measured the day 8 effector response of SMARTA T cells responding to LCMV infection (Fig. 1A, “late anti-IL-2”). As with early anti-IL-2 treatment, late anti-IL-2 treatment did not impact overall SMARTA expansion (Fig. 1B–C, Supp. Fig. 2). However, late anti-IL-2 was associated with increased frequencies of CXCR5⁺Ly6C⁻ Tfh cells in the spleen (Fig. 1D–E), as well as a decrease in the ratio of the total number of Th1 (Ly6C⁺CXCR5⁻) compared to the total number of Tfh (CXCR5⁺Ly6C⁻) (Fig. 1F). While early anti-IL-2 also induced the formation of increased germinal center Tfh (GC Tfh: CXCR5⁺Bcl-6^hi), late anti-IL-2 significantly increased their numbers but not their frequency (Fig. 1G). Similarly, while early anti-IL-2 impaired the formation of Th1 effector cells expressing Granzyme B and Tbet, as determined by both frequency and number, late anti-IL-2 impaired their frequency but not their number (Fig. 1H–I). However, the relative ratio of Th1 (CXCR5⁻Tbet^hi) to Tfh (CXCR5⁺Tbet^lo) was significantly reduced for both treatment groups (Fig. 1J), indicating that IL-2 is required during the late stages of the effector phase (after day 3) to maintain the balance of Th1 to Tfh responses.

We further sought to assess the impact of IL-2 on CD4⁺ T cell function by measuring cytokine production after antigen stimulation. At day 8 post-infection, splenocytes were stimulated ex vivo with GP_61–80 peptide in the presence of Brefeldin A for 4–5 hours. Cells were then permeabilized and stained with antibodies specific for IFNγ, IL-2 and TNFα. Because previous studies have demonstrated that highly protective Th1 cells are polyfunctional, we focused on the frequency of SMARTA cells co-producing all three cytokines (3). We found that there were no significant differences in the frequency and number of triple-producing SMARTA T cells in the spleen at day 8 post-infection (Fig. 2A–B). Additionally, SMARTA T cells produced less IFNγ and TNFα on a per-cell basis after late anti-IL-2 treatment, providing evidence that late IL-2 drives both the emergence and function of Th1 responses (Fig. 2C–E).

Figure 2.

IL-2 blockade reduces cytokine production by Th1 cells. Day 8 SMARTA cells from the spleen were restimulated with GP_61–80 peptide for 4–5 hours in presence of Brefeldin A, followed by intracellular cytokine staining. (A) Representative flow plots show expression of IL-2, TNFα and IFNγ for each treatment group. (B-E) Graphs depict the frequency and number of triple producers (IFNγ⁺IL-2⁺TNFα⁺) as normalized to unstimulated controls and MFI of IFNγ, IL-2 and TNFα. Data displayed are n=5 per group and are reperesentative of three separate experiments. Error bars represent SEM. *p<0.05, ***p<0.001, ****p<0.0001

IL-2 signaling regulates formation of Th1 and Tfh precursor effector cells

To better understand the role of IL-2 in driving the formation of Th1 precursors during acute viral infection, we treated SMARTA T cell recipients with anti-IL-2 antibodies starting a day before LCMV infection and analyzed early T cell activation at day 3 post-infection. Neutralization of IL-2 did not significantly alter the number of SMARTA T cells in the spleen at day 3 (Fig. 3A). However, IL-2 blockade significantly reduced the early emergence of Th1 precursor cells, as measured by a reduced proportion and number of Tim-3⁺Ly6C⁺ SMARTA cells (Fig. 3B–C) (22, 34). These differences were also observed when we measured the early induction of transcription factors associated with Tfh and Th1 differentiation. Blockade of IL-2 induced a decrease in the ratio of the total number of Tbet^hiTCF-1^lo vs. TCF-1^hiTbet^lo SMARTA cells and an increase in the ratio of the total number of Bcl-6^hiTbet^lo vs. Tbet^hiBcl6^lo early effector SMARTA cells (Fig. 3D–F). We additionally assessed cytokine production (IFNγ, IL-2, TNFα) by early SMARTA effector cells. Following ex vivo peptide restimulation and intracellular cytokine staining, we did not observe significant increases in the frequency of SMARTA cells in the spleen producing any of the three cytokines singly or together. However, IL-2 production was significantly enhanced on a per-cell basis in anti-IL-2 treated groups, as shown by a significant increase in mean fluorescence intensity (MFI) (Fig. 3G–K). It is of interest to note that a previous study also demonstrated increased IL-2 production by precursor Tfh effector cells (35). While previous work has demonstrated a requirement for T cell-intrinsic IL-2 signaling via the high-affinity IL-2 receptor for the generation of Th1 at the peak of the effector response (12, 23), our data demonstrate that IL-2 is required for early Th1 fate commitment in the first 3 days of T cell activation in vivo.

Figure 3.

Early IL-2 drives the formation of day 3 Th1 precursors by day 3 of LCMV infection. 1 × 10⁵ SMARTA cells (CD45.1⁺) were injected i.v. into B6 mice (CD45.2⁺), followed by LCMV infection one day later. IL-2 neutralizing antibodies were administered beginning the day before infection. (A) Graph shows number of SMARTA cells in the spleen at day 3 post-infection. (B-C) Representative flow plots and graphs show the frequency and number of Th1 precursor cells (Ly6C⁺Tim3⁺) following anti-IL-2 treatment. (D) Representative flow plots indicate expression of Bcl-6, TCF-1 and Tbet by SMARTA in the spleen at day 3 post-infection. (E-F) Graphs display the ratio of Tbet^hiTCF-1^lo Th1 precursors to TCF-1^hiTbet^lo Tfh precursors (E) or Bcl-6^hiTbet^lo Tfh precursors to Tbet^hiBcl-6^lo Th1 precursors (F). (G) Representative flow plots depict cytokine production (IFNγ, IL-2, and TNFα) of day 3 effector SMARTA cells following ex vivo peptide restimulation as previously. (H-K) Graphs indicate the frequency and number of triple producing SMARTA (IFNγ⁺TNFα⁺IL-2⁺) and the MFI of cytokine staining for each. Data displayed are n=5 per group and are representative of three separate experiments. Error bars represent SEM. *p<0.05, ***p<0.001, ****p<0.0001

To better understand the impact of IL-2 on early activation of CD4⁺ T cells in vivo, we treated SMARTA recipients with anti-IL-2 and assessed their response on days 1, 2, and 3 post-infection. As we previously reported, most activated SMARTA cells expressed IL-2Rα by day 2 post-infection, while about half expressed IL-2Rα at day 3 (22). IL-2 neutralization minimally reduced IL-2Rα expression at day 2 but resulted in a substantial decrease in the proportion of IL-2Rα-expressing effector cells by day 3 (Fig. 4A). Similarly, both TCF-1 and Bcl-6 showed little change in expression at day 1 post-infection, as compared to host CD4⁺ T cells. Similarly to the expression of IL-2Rα, by day 2 post-infection, most activated SMARTA cells expressed TCF-1 and Bcl-6, with about half expressing each at day 3. In this case, however, IL-2 neutralization, while minimally impacting induction of TCF-1 and Bcl-6 expression at day 2, resulted in an increased proportion of TCF-1 and Bcl-6-expressing SMARTA cells at day 3 (Fig. 4A). Of note, TCF-1 and IL-2Rα were co-expressed (TCF-1^hiIL-2Rα^hi) at day 2 post-infection by most activated SMARTA cells, but by day 3, they were dis-coordinately expressed, with the majority of SMARTA cells segregating into IL-2Rα^hiTCF-1^lo and IL-2Rα^loTCF-1^hi subsets (Fig. 4B). Therefore, our earliest evidence of the emergence of Th1 and Tfh precursor cells is at day 3 after infection. Treatment with anti-IL-2 resulted in a decreased emergence of IL-2Rα^hiTCF-1^lo Th1 precursor cells (Fig. 4B–C) and significantly decreased expression of Tim-3, Granzyme B and T-bet (Supp. Fig. 3), indicating that IL-2 acts during the early phases of the immune response to regulate commitment to effector T cell lineages.

Figure 4.

IL-2 drives the emergence of Th1 precursors at day 3 post-infection. 1 × 10⁵ SMARTA cells (CD45.1⁺) were injected i.v. into B6 mice (CD45.2⁺), followed by LCMV infection one day later. IL-2 neutralizing antibodies were administered beginning the day before infection. (A) Histograms show expression of TCF-1, Bcl-6, and IL-2Rα by SMARTA T cells (black) and endogenous CD4⁺ T cells (grey) in the spleen in untreated and anti-IL-2 treated mice at days 1, 2 and 3 post-infection. (B) Representative flow plots depict expression of TCF-1 and IL-2Rα by SMARTA cells at days 1, 2 and 3 post-infection with or without IL-2 neutralization. (C) Graph depicts the ratio of IL-2Rα^hiTCF-1^lo Th1 precursors to TCF-1^hiIL-2Rα^lo Tfh precursors at days 2 and 3 post-infection. Data displayed are n=5 per group and are representative of three separate experiments. Error bars represent SEM. *p<0.05, ***p<0.001, ****p<0.0001

Early and late IL-2 impact memory T cell numbers and subset distribution

Because the IL-2Rα^lo effector cell subset at day 3 post-infection is the primary repository of memory precursor cells, we hypothesized that the increase in IL-2Rα^lo effector cells after IL-2 neutralization would lead to enhanced memory T cell formation (22). In support of this, we detected significantly higher numbers of SMARTA memory T cells (day 42 post-infection) in the spleen following early anti-IL-2, as compared to untreated controls. In contrast, late anti-IL-2 administration, which is not initiated until after the emergence of memory T cell precursors at day 3 post-infection, did not significantly impact memory T cell numbers (Fig. 5A). However, both early and late IL-2 neutralization impacted the distribution of memory T cell subsets. Anti-IL-2 neutralization resulted in a decreased ratio of Th1 like (Ly6C+CXCR5- or Tbet^hiCXCR5-) to Tfh-like (CXCR5+Ly6C- or CXCR5+Tbet^lo) memory T cell subsets regardless of whether IL-2 was neutralized throughout the effector response or beginning at day 3 post-infection (Fig. 5B–D).

Figure 5.

IL-2 shapes the number and function of CD4+ memory T cells. 1 × 10⁴ SMARTA cells (CD45.1⁺) were injected i.v. into B6 mice (CD45.2⁺), followed by LCMV infection one day later. Early and late IL-2 neutralization was performed as previously. (A) Graph depicts the number of SMARTA T cells in the spleen at day 42 post-infection. (B) Representative flow plots depict the expression of Ly6C, CXCR5 and Tbet by SMARTA memory cells (day 42 post-infection) in the spleen. (C-D) Graphs show ratios of Ly6C⁺CXCR5⁻ to CXCR5⁺Ly6C⁻ and Tbet^hiCXCR5⁻ to CXCR5⁺Tbet^lo SMARTA memory cells for each treatment group. (E) After the establishment of memory (day 42 post-infection) mice were rechalleneged with 2 × 10⁵ CFU of Lm-gp61. Graph shows number of secondary SMARTA effector cells in the spleen at day 7 after heterologous rechallenge. Data displayed are n=5 per group and are representative of three separate experiments. Error bars represent SEM. *p<0.05, ***p<0.001, ****p<0.0001

We additionally sought to determine whether neutralization of IL-2 during the primary T cell response impacted the recall ability of subsequent memory T cells, as activation of CD8⁺ T cells in the absence of IL-2 impairs their ability to form secondary T cell responses (36). To determine if CD4⁺ memory T cells are similarly impaired, we heterologously challenged LCMV memory mice with Listeria monocytogenes recombinantly expressing the LCMV GP_61–80 epitope (Lmgp-61), as previously (7, 16, 37). Because LCMV and Lm-gp61 share a single CD4⁺ T cell epitope, this approach allows for robust boosting of CD4⁺ memory T cells without rapid antigen clearance by CD8⁺ memory T cells. We measured the SMARTA recall response in the spleen at day 7 post-challenge. SMARTA memory T cells expanded similarly across all groups regardless of treatment (Fig. 5E). Overall, we concluded that early IL-2 neutralization enhanced the formation of memory precursor cells at the expense of the Th1 effector response, while both early and late IL-2 neutralization reduced the formation of Th1-like effector memory T cells. However, IL-2 neutralization did not significantly impact memory recall ability.

IL-2 is required for optimal Th1 precursor formation even in the absence of TCF-1

Other groups have previously shown that TCF-1 expression is opposed by IL-2 signaling (23, 38). We also demonstrated that IL-2 prevents the emergence of TCF-1-expressing T cells (Fig. 3). Because IL-2 neutralization boosted the emergence of TCF-1-expressing early effector cells as well as the formation of Tfh at the peak of the effector response, we hypothesized that IL-2 would mediate its effects on early fate commitment and effector T cell differentiation through regulation of TCF-1 expression. If this were the case, we would expect to see that T cell-intrinsic loss of TCF-1 compensates during IL-2 neutralization to enhance Th1 formation. Therefore, we crossed transgenic SMARTA (CD45.1⁺) mice to B6(Cg)-TCF7tm1Hhx/J mice (possessing LoxP sites flanking Exon 2 of Tcf7 and a LoxP-independent GFP reporter) and to ROSA26 Cre-ERT2 mice (tamoxifen-inducible) to generate Tcf7 inducible knockout (iKO) SMARTA mice (27, 28). Control SMARTA mice expressed the Tcf7 wildtype locus and the ROSA26 Cre-ERT2 allele (Tcf7 WT). We then treated Tcf7 iKO SMARTA mice or control Tcf7 WT SMARTA mice with tamoxifen intraperitoneally (i.p.) for four consecutive days to induce Cre-mediated excision of LoxP-flanked gene sequences. This approach resulted in the loss of TCF-1 expression in nearly all Tcf7 iKO SMARTA cells but did not impact TCF-1 levels in Tcf7 WT SMARTA cells (Supp. Fig. 4). We then adoptively co-transferred both Tcf7 iKO (CD45.1⁺) and Tcf7 WT (CD45.1⁺CD45.2⁺) SMARTA T cells into recipient B6 mice (CD45.2⁺) at a 1:1 ratio, followed by LCMV infection one day later. Some groups of mice also received IL-2-neutralizing antibodies prior to infection. At day 4 post-infection, Tcf7 iKO SMARTA cells in the spleen were reduced 2–3 fold in frequency and number as compared to their Tcf7 WT counterparts, and this reduction was not significantly impacted by IL-2 treatment (Fig. 6A–B). As previous labs have shown that TCF-1 is required for Tfh formation, we determined whether loss of TCF-1 would alter commitment to Th1 and Tfh lineages by day 4 effector cells. Loss of TCF-1 resulted in a significant increase in the ratio of Tim3⁺IL-2Rα WT to TCF-1 KO and Tbet^hiBcl-6^lo Th1 precursor cells, with an accompanying decrease in the frequency and number of Bcl-6^hiTbet^lo Tfh precursor cells (Fig. 6C–H). IL-2 neutralization partially reversed these effects in the absence of TCF-1, resulting in an increase in Tfh precursor cells and a decrease in Th1 precursor cells. However, IL-2 neutralization in the absence of TCF-1 still resulted in significantly more Th1 precursors than IL-2 neutralization in the WT setting, suggesting both TCF-1-dependent and -independent mechanisms of IL-2-driven Th1 commitment (Fig. 6C–H). Of note, baseline levels of Tim-3 were higher in Tcf7 iKO SMARTA cells (Fig. 6C). This may be due to direct repression of HAVCR2 (which encodes Tim-3) gene expression by TCF-1, as TCF-1 regulates Tim-3 expression and localizes to a regulatory region upstream of HAVCR2 (encoding Tim-3) (39). Overall, our findings suggest that TCF-1 plays a critical role in regulating the proportion of Th1 vs. Tfh precursors during the early part of the effector response, while IL-2 is required for the optimal emergence of Th1 precursor cells even in the absence of TCF-1.

Figure 6.

IL-2 is required for the emergence of Th1 precursor effector cells in both the presence and absence of TCF-1. Tcf7 iKO (CD45.1⁺) and Tcf7 WT (CD45.1⁺CD45.2⁺) SMARTA mice received four daily i.p. injections of tamoxifen, followed by transfer (1 × 10⁴ each) into B6 mice (CD45.2⁺) and infection with LCMV one day later. (A-B) Representative flow plots depict Tcf7 WT and Tcf7 iKO SMARTA T cells in the spleen at day 4 post-infection, gated on CD4⁺ T cells, whereas the graph indicates the ratio of WT to iKO SMARTA in anti-IL-2-treated and untreated conditions. (C) Representative flow plots, gated on SMARTA cells, depict expression of Tim-3 and IL-2Rα at day 4 post-infection. Graphs depict (D) WT to iKO ratio of the frequency of Tim-3⁺ SMARTA and (E-H) frequency of Tbet^hiBcl-6^lo and Bcl-6^hiTbet^lo in WT and iKO SMARTA cells in the spleen at day 4 post-infection. Data displayed are n=5 per group and are representative of two separate experiments. Error bars represent the SEM. *p<0.05, ***p<0.001, ****p<0.0001

Both early and late IL-2 repress TCF-1-independent Tfh differentiation

We next sought to determine the role of IL-2 in driving Th1 differentiation at the peak of the immune response (day 8) in Tcf7 iKO SMARTA. As above, we transferred Tcf7 iKO (CD45.1⁺) and Tcf7 WT (CD45.1⁺CD45.2⁺) SMARTA, both tamoxifen-treated, into B6 mice (CD45.2⁺), followed by LCMV infection one day later. Similar to prior studies, mice were administered the IL-2 neutralizing antibody cocktail prior to the start of infection (early) or beginning at day 3 (late). In the untreated group, we observed a 3-fold reduction in the frequency and number of Tcf7 iKO SMARTA T cells in the spleen at day 8 post-infection, as compared to Tcf7 WT SMARTA (Fig. 7A–B), similar to the early effector response at day 4 post-infection (Fig. 6A–B). IL-2 neutralization in both treatment groups resulted in an additional significant decrease in TCF-1 iKO SMARTA expansion, indicating a TCF-1-independent role for IL-2 in sustaining effector CD4⁺ T cell responses (Fig. 7A–B). Loss of TCF-1 almost entirely ablated the formation of CXCR5⁺Ly6C⁻ or CXCR5⁺Tbet^lo Tfh cells in the spleen, with little to no CXCR5 expression seen overall (Fig. 7C–I). These findings confirm previous work demonstrating a requirement for TCF-1 in Tfh differentiation and provide additional evidence that TCF-1 is effectively eliminated in Tcf7 iKO SMARTA cells (21, 23). Both early and late anti-IL-2 treatment reduced the proportion of CXCR5⁻Ly6C⁺ or Tbet^hiCXCR5⁻ Tcf7 iKO SMARTA Th1 cells while increasing the overall frequency of CXCR5⁺Ly6C⁻ or CXCR5⁺Tbet^lo Tfh cells (Fig. 7C–I). Therefore, we concluded that at least a portion of Tcf7 iKO SMARTA cells retained the ability to form Tfh but that this ability was effectively suppressed by IL-2 signals during both the early and late stages of the effector response. To assess which cells were rescued following anti-IL-2 treatment, we took advantage of the fact that the GFP reporter in Tcf7 iKO SMARTA remains intact even after excision of the LoxP-flanked region of the Tcf7 locus. Virtually all of the TCF-1-deficient Tfh cells that were rescued by anti-IL-2 treatment were GFP⁺ (Fig. 7F) while continuing to lack TCF-1 expression (Supp. Fig. 4). These findings indicate the persistence of a TCF-1-deficient lineage with active transcription at the Tcf7 locus that is rescued by both early and late anti-IL-2 independently of TCF-1. We conclude that IL-2 drives Th1 formation through mechanisms that are both dependent and independent of its ability to repress TCF-1 activity and that IL-2 mediates this function throughout the effector response.

Figure 7.

IL-2 neutralization cells partially restores the Tfh response while diminishing the Th1 response of Tcf7 iKO SMARTA. Tcf7 iKO (CD45.1⁺) and Tcf7 WT (CD45.1⁺CD45.2⁺). SMARTA mice received four daily i.p. injections of tamoxifen, followed by co-transfer into B6 mice (CD45.2⁺) and infection with LCMV as previously. (A-B) Representative flow plots depict Tcf7 WT and Tcf7 iKO SMARTA T cells in the spleen at day 8 post-infection, gated on CD4⁺ T cells, whereas the graph indicates the ratio of WT to iKO SMARTA in early anti-IL-2-treated, late anti-IL-2-treated and untreated conditions. (C-E) Representative flow plots, gated on SMARTA cells in the spleen, show expression of CXCR5 and Ly6C. Results are quantified in accompanying graphs. (F-H) Representative flow plots, gated on SMARTA cells in the spleen, show expression of CXCR5 and Tbet. GFP-expressing Tcf7 iKO SMARTA cells are displayed in green. Results are quantified in accompanying graphs. (I) Graphs depict WT to iKO ratio of Th1 (Tbet^hiCXCR5⁻) SMARTA cell numbers in the spleen at day 8 post-infection. Data displayed are n=5 per group and are representative of two separate experiments. Error bars represent SEM. *p<0.05, ***p<0.001, ****p<0.0001

Discussion

Our data demonstrate that IL-2 plays a crucial role in specifying the formation of Th1 precursor effector cells during the first few days after T cell activation and driving terminal Th1 differentiation throughout the later stages of the effector response. We previously reported that the TCR is a key determinant of effector versus memory T cell differentiation and that differential expression of the high-affinity IL-2R (as indicated by up-regulation of the IL-2Rα sub-unit) at day 3 post-infection distinguished memory precursor T cells from precursors for terminally differentiated Th1 cells. Specifically, activated CD4⁺ T cells that maintained IL-2Rα expression at day 3 experienced higher TCR signal strength and gave rise to terminally differentiated Th1 cells, whereas activated CD4⁺ T cells that down-regulated IL-2Rα expression by day 3 post-infection experienced comparatively lower TCR signal strength and gave rise to Tfh and memory cells (22, 40). TCRs that induced sustained IL-2Rα expression also induced increased NFAT activity (22, 41, 42). Another study also identified a key role for T cell-intrinsic signaling via the IL-2R in driving Th1 differentiation, although the role of TCR signal strength was not a focus (12). A role for strong TCR signals in favoring Th1 differentiation versus Th2 differentiation has previously been appreciated (40), and strong TCR signals induced by prolonged dwell time favor Th1 differentiation over Tfh during acute infection (33, 43). The unique contribution of the present study is that it links TCR signal strength to a specific mechanism driving Th1 commitment and polarization. Future studies will be required to assess the precise nature of the TCR signals that drive differential IL-2Rα expression, as well as biochemical measures such as TCR affinity, off-rate and bond lifetime.

It was previously reported that expression of IL-2 and IL-2Rα varies early after T cell activation, with IL-2-producing T cells often lacking IL-2Rα expression and giving rise to Tfh cells, and IL-2Rα-expressing T cells often lacking IL-2 expression and giving rise to Th1 cells (35). While we did not directly explore this finding in our study, we do note that after anti-IL-2 treatment early effector cells, that are enriched for Tfh precursors, produced IL-2 at a higher frequency and amount on a per cell basis. It seems likely, therefore, that IL-2 and IL-2Rα expression are dis-coordinately regulated in many T cells, despite being widely characterized as early activation genes dependent on NFAT activity (22, 41, 42). Additionally, expression of the IL2Rα and IL2 genes may be differentially impacted by TCR signals. Further study will be needed to understand how different TCR signal strength leads to differential expression of these genes, and how NFAT activity at these loci is regulated under settings of high and low TCR signal strength.

While IL-2 neutralization leads to a reduction of host Tregs, the number of SMARTA T cells induced by infection was not altered. Previous studies have shown that Tregs suppress the differentiation of Th1 cells in various settings while also promoting Tfh differentiation (44, 45). Here, we observed a reduction of Th1 and increase in Tfh differentiation even when Tregs were reduced, suggesting the role of IL-2 on effector differentiation is likely independent of it’s impact on Tregs in this setting. However, further study will be needed to assess role of IL-2-driven Treg function in regulating the balance of Th1 and Tfh responses.

An early role for IL-2 after T cell activation has previously been described in the generation of memory CD4⁺ T cells (24). In contrast, sustained IL-2 signals after day 4 post influenza infection have been shown to be required for the generation of memory T cells (25). For this reason, we considered the possibility of a role for both early and late IL-2 in our model system by using IL-2 blocking antibodies to determine the temporal role of IL-2 before and after the segregation of activated T cells based on IL-2Rα expression (day 3 post-infection). While sustained IL-2 signaling was required to drive Th1 polarization throughout the response, early and late IL-2 signals had different effects on memory T cell formation. Only early IL-2 blockade enhanced the number of memory CD4⁺ T cells, which correlated to an increased proportion of IL-2Rα^lo memory precursor T cells by day 3 post-infection (22). In contrast, both early and late IL-2 blockade impaired the formation of Th1-like memory T cells, highlighting the role of IL-2 in driving Th1 specification and polarization. Despite this, IL-2 blockade did not impair the CD4⁺ T cell recall response or the differentiation of secondary Th1 effector cells, as has been reported for CD8⁺ T cells (36). However, that study relied on genetic knockouts rather than antibody-mediated neutralization (which may not completely eliminate in vivo IL-2 signlaing). It has been previously shown that while low IL-2 signals may favor CD8+ memory T cell differentiation, high or persistent IL-2 signals drive terminal effector differentiation(46). Therefore, additional study will be needed to identify the precise role that IL-2 plays in this model system in shaping CD4⁺ memory T cell formation and function.

The transcription factor TCF-1 (Tcf7) is required for T cell development, Tfh differentiation and the development of CD4⁺ memory T cells in response to viral infection. Specifically, TCF-1 was found to generate and maintain Tfh cells while repressing the genes IL2ra and Prdm1 (encoding Blimp-1). Conversely, IL-2-induced Blimp-1 activation may repress the expression of TCF-1, favoring Th1 differentiation during viral infection (21, 23, 31, 38). Furthermore, TCF-1 deficiency compromised the generation of Tfh cells, and when the Th1 transcription factor Blimp-1 was deleted in TCF-1-deficient cells, Tfh generation was restored (21). Another study demonstrated that the overexpression of the Tfh transcription factor Bcl-6 could also restore Tfh differentiation in TCF-1-deficient cells (23).

We report that IL-2 is required for the optimal formation of Th1 cells even in the absence of TCF-1, indicating that the role of IL-2 is likely to involve additional mechanisms beyond repression of TCF-1 activity. Our study also found that limiting IL-2 signals in TCF-1-deficient cells restored some Tfh effector cells, though not completely. Almost all of the rescued Tfh cells expressed the GFP reporter that is knocked in to that locus, indicating that while Tcf7 expression was disrupted, transcription at this locus remained active. Therefore, we postulate that TCF-1-independent mechanisms of Tfh differentiation are likely to exist in settings where IL-2 is not abundant. One potential candidate for driving TCF-1-independent Tfh formation is Lef-1, which has been previously shown to play a vital role in the formation of central memory CD4⁺ T cells (15). Several mechanisms of chromatin modification have been ascribed to the activity of TCF-1 and Lef-1, including suppression of Blimp-1 through deacetylation of the Prdm1 locus (31). Further work will be needed to better understand how Lef-1 or other transcription factors may partner with or work independently of TCF-1 in suppressing Th1 differentiation and enabling the Tfh and memory T cell differentiation program.

Th1 cell differentiation was previously shown to be driven by enhanced expression of IL-2Rα, IL-2-dependent STAT5 activation and upregulation of Blimp-1 (12, 18, 20, 21, 47). In contrast, Tfh cell differentiation is driven by Bcl-6 and TCF-1 but is restricted by STAT5-mediated IL-2 signaling, IL-2Rα expression and Blimp-1 (5, 19). Blimp-1 strongly binds to TCF-1 in Th1 cells (23), and conversely, TCF-1 was also seen to bind to Blimp-1 in Tfh cells (21). In our study, we identified a mechanism by which TCR signal strength guides IL-2 signals in regulating the balance between Th1 and Tfh differentiation. Uncovering the specific TCR signals that favor either the IL-2- or TCF-1-dependent regulatory network is a critical goal for understanding how and why T cells activated in similar environments commit to distinct cellular fates and will have wide-ranging implications for interpreting and manipulating T cell activation and longevity in the setting of infection or vaccination, as well as immunotherapeutic settings for cancer.

Key Points.

• IL-2 sustains Th1 responses in the later stages of the anti-viral response.

• Differential exposure to IL-2 shapes the memory T cell compartment.

• IL-2 is required to sustain Th1 responses even in the absence of TCF-1.