IL-2 Modulates the TCR Signaling Threshold for CD8 but Not CD4 T Cell Proliferation on a Single-Cell Level

This article is featured in In This Issue, p.2213

Adaptive immune responses rely upon robust clonal expansion of rare Ag-specific lymphocyte populations. CD8 T cells can divide up to 15 times in vivo, as often as every 4–8 h, to give rise to more than 10⁴ progeny within 7 d (1). Lymphocytes integrate information about the concentration, affinity, and avidity of Ags, in addition to the presence of cytokines such as IL-2, to inform the decision to divide and how many times (2, 3). Tight regulation of this process is critical for both host defense and immune tolerance. T cell proliferation has been amenable to study in vitro through the use of TCR transgenic models and polyclonal stimulation with TCR cross-linking Abs. Yet even uniformly stimulated populations of genetically identical lymphocytes exhibit enormous heterogeneity in proliferative responses, making it challenging to dissect how input signals are interpreted on a single-cell level.

Nur77 is an orphan nuclear hormone receptor encoded by the primary response gene Nr4a1, and is robustly and rapidly induced by Ag receptor signaling (4–7). We have recently characterized a novel Nur77-eGFP BAC transgenic mouse line that recapitulates endogenous Nr4a1 expression (6). Due to the long half-life of eGFP, reporter expression in T cells serves as a faithful and specific marker of intensity and duration of TCR signaling. Nevertheless, we observed that reporter T cells that have divided in response to TCR stimulation express a high and invariant amount of GFP regardless of Ag dose and modulation of TCR signal strength (7). These observations suggested that individual T cells must reach a minimal TCR signaling threshold for cell division. However, it is not known whether extensive titration of Ag affinity alters the minimal TCR signaling threshold for proliferation. T cell proliferation is also modulated by cytokines such as IL-2 (2, 3), yet it is not clear how such environmental cues about the inflammatory milieu influence the TCR signaling threshold for T cell proliferation on a single-cell level.

The Nur77-eGFP reporter facilitates dissection of the contribution of TCR signals to proliferation in different contexts, as it is sensitive to TCR but not cytokine-dependent JAK/STAT signals (4, 6, 7). Prior work has shown that Nur77 expression can be induced not only by TCR together with CD28 stimulation, but also by either ionomycin or PMA, and is both calcineurin- and protein kinase C–dependent (6, 8). However, Nur77 reporter expression could not be induced by a constitutively active STAT5 construct, consistent with insensitivity of the reporter to IL-2 (4, 7).

In this study we show that expression of Nur77-eGFP reporter sensitively shows relative differences in peptide affinity and dose in CD8 and CD4 cells. However, proliferating Ag-specific CD8 and CD4 T cells exhibit high and invariant GFP expression in response to broad titration of peptide potency and concentration, identifying a minimal threshold of cumulative TCR signaling required for proliferation. Though IL-2 is a known T cell growth factor, in this context it is surprising that provision of exogenous IL-2 markedly reduces the minimal TCR signaling threshold required by CD8, but not CD4, T cells to proliferate, particularly under conditions of suboptimal TCR stimulation. We show that CD4 T cells take more time to become responsive to low doses of IL-2 upon activation with Ag than CD8 T cells, and propose that this temporal delay accounts for the differential effects of IL-2 upon this TCR threshold. Finally, we show that expression of the primary response gene c-Myc is coordinately regulated downstream of both antigenic and cytokine stimuli, and defines a minimal threshold for mitogenic stimulation that is invariant among individual proliferating lymphocytes. We propose that graded population proliferative responses are achieved in part by altering the fraction of responding cells that have received sufficient integrated mitogenic signaling to cross this threshold.

Nur77-eGFP mice [both low- and high-expressing BAC transgenic (Tg) lines] have been previously described (6). All T cell assays were performed using the low-GFP line except that shown in Fig. 8A, which used the high-GFP line. BoyJ and C3H/HeJ mice were obtained from the Jackson Laboratory. OTI, and OTII mice were previously described (9, 10). AND Tg mice on the C57BL/6 genetic background were generously provided by P. Allen (Washington University), and have been previously described (11, 12). All strains were fully backcrossed to C57BL/6 genetic background. Mice were used at 5–9 wk of age for all functional and biochemical experiments. All mice were housed in a specific pathogen-free facility at University of California, San Francisco (UCSF) according to the University Animal Care Committee and National Institutes of Health (NIH) guidelines.

Abs for FACS: CD4, CD8, TCR vα2, CD25, CD69, CD122, CD127, CD132, lambda1, CD45.1, CD45.2 conjugated to FITC, biotin, PE, PerCP-Cy5.5, PE-Cy7, Pacific Blue, APC or Alexa 647 (eBiosciences or BD Biosciences); Abs for intracellular staining: c-MYC Ab (clone D84C12; Cell Signaling), STAT5 pY694 (BD, clone 47; BD Biosciences), AKT pS473 (193H12 clone; Cell Signaling); cytokines and other stimuli: anti-CD3ε (clone 145–2C11; Harlan), anti-CD28 (clone 37.5; UCSF flow core; dose 2 μg/ml), anti–IL-2 (clone JES6-5H4; UCSF flow core; dose 10 μg/ml), recombinant human IL-2 (rhIL-2) was from the NIH AIDS Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, NIH (M. Gately; Hoffmann-La Roche), recombinant IL-7 (variable doses for signaling), and IL-15 (100 ng/ml) (R&D Systems). Peptides: OTII-specific OVA (OVA_323–339) peptide; OTI-specific OVA (OVA_257−264) peptide (SIINFEKL/N4), as well as altered peptide ligands Q4R7, and G4 have been previously described (13), as have moth cytochrome c peptide (MCC) and altered peptide ligands K99E and T102L (14) (Genscript). Inhibitor: Jak3i (15).

Lymphocytes were loaded with CellTrace Violet (Invitrogen) per the manufacturer’s instructions except with 1 ml/5 × 10⁶ cells rather than 1 ml/1 × 10⁶ cells.

CD8 or CD4 T cells were purified from single-cell suspensions of spleen and lymph nodes from male and female mice aged 6–12 wk by negative selection with biotinylated Abs to CD4 (GK1.5) or CD8 (53-6.7) respectively, CD19 (6D5), B220 (RA3-6B2), CD11b (M1/70), CD49b (DX5), Ly-76 (Ter119), CD24 (M1/69) (all BioLegend), and CD11c (N418; Tonbo) and magnetic anti-biotin beads (MACSi Beads; Miltenyi Biotec) as previously described (15). Regulatory T cells (Tregs) were additionally depleted from CD4 T cells by adding 1:100 anti CD25 (PC61.5).

In vitro cultured cells were stained to detect dead cells using fixable live/dead stain (Near IR, Life Technologies) per manufacturer’s instructions, compatible with both CellTrace Violet and MeOH permeabilization as needed.

Cells were stained with indicated Abs and analyzed on a Fortessa (Becton Dickson) as previously described (16). Data analysis was performed using FlowJo (v9.7.6) software (Treestar, Ashland, OR). Division index and % divided parameters were calculated from CellTrace Violet-loaded lymphocytes using FlowJo (17). Statistical analysis and graphs were generated using Prism v6 (GraphPad Software). Figures were prepared using Illustrator CS6 v16.0.0.

Staining for intracellular c-MYC expression was performed as follows: after stimulation under various conditions as described below (with or without CellTrace Violet loading), 10⁶ cells per well were resuspended in 96-well plates, stained with fixable live/dead dye as above, fixed in 2% PFA, washed in FACS buffer, and permeabilized with ice-cold 90% methanol at −20°C overnight, or for at least 30 min on ice. Cells were then washed in FACS buffer, stained with anti-MYC Ab (1:200) × 45 min, washed in FACS buffer, and then stained with Alexa 647-conjugated secondary goat anti-rabbit Ab (Jackson ImmunoResearch) along with directly conjugated Abs to detect surface markers. Samples were washed and then refixed with 2% PFA prior to acquisition on BD Fortessa. Staining for intracellular pSTAT5 was performed as previously described (15), with the addition of pAKT staining (S473 193H12 Cell Signaling, diluted 1:50, and detected with anti-rabbit-PE 1:200; Jackson ImmunoResearch).

CD8 or CD4 T cells were purified from either OTI-Nur77-eGFP or AND-Nur77-eGFP mice as described above and loaded with CellTrace Violet. OTI cells were mixed in a 5:1 ratio with peptide-loaded splenocytes from Zap70-deficient mice to serve as APCs, plated at a concentration of 10⁵ cells/200 μl in round-bottom 96-well plates in complete media (RPMI 1640). AND cells were plated at a 1:1 ratio with peptide-loaded splenocytes from C3H/HeJ mice. Cells were stained and analyzed as described above after 1 or 3 d of culture. For all intracellular c-Myc staining assays, lymphocytes from OTI-Nur77-eGFP mice were loaded with CellTrace Violet, and directly incubated with peptides at varying concentrations. Cells were plated at a concentration of 10⁶ cells/200 μl in round-bottom 96-well plates for incubations ≤48 h, and 10⁵ cells/200 μl for 72 h assays. For plate-bound T cell stimulation, flat-bottom plates were coated in 50 μl PBS with varied doses of anti-CD3ε ± anti-CD28 overnight at 4°C, and washed in PBS prior to plating cells at a concentration of 2 × 10⁵ cells/200 μl.

CD8 T cells were purified from CD45.2 OTI-Nur77-eGFP reporter mice, and loaded with CellTrace Violet as described above. Cells (5 × 10⁵) were transferred into CD45.1 BoyJ hosts via tail vein injection on d-1. APCs (total splenocytes) from Zap-70^−/− mice were loaded with N4, Q4R7, G4, or no peptide (10⁻⁸ M) via 2 h incubation at 37°C. On d 0, 10⁶ loaded APCs were transferred into recipients via tail vein injection. Three days later splenocytes from recipients were surface stained to detect vα2, CD45.1, CD45.2, and CD8 expression, and analyzed by FACS to detect GFP expression, CellTrace Violet dilution, and surface marker expression.

RNA was isolated from 1 to 2 × 10⁶ CD4⁺ T cell blasts per condition using an RNeasy kit (Qiagen), and cDNA was synthesized using qScript (Quanta Biosciences). mRNA was detected by Primetime (IDT) or Taqman (Life Technologies) predesigned quantitative PCR assays. Assay ID: Taqman: il2ra (Mm01340213_m1); IDT: il2rg (Mm.PT.58.43694030), il2rb (Mm.PT.58.15857480), socs1 (Mm.PT.58.11527306.g), socs2 (Mm.PT.58.5195465), socs3 (Mm.PT.58.7804681), and cish (Mm.PT.58.11557699). Data are from three replicates collected on a QuantStudio 12k system (Life Technologies), plotted with 95% confidence intervals as calculated by QuantStudio (Life Technologies).

We previously reported that individual T cells integrate TCR signaling over time and require an invariant minimal amount and duration of signaling to divide, as detected by Nur77-eGFP reporter expression (7). Moreover, this threshold is not altered by stimulus dose or modulation of Zap70 kinase activity. In the current study, we determined whether peptide affinity, which can vary broadly, influences this threshold in T cells. We introduced the Nur77-eGFP reporter onto the OVA-specific OTI TCR Tg model, and took advantage of a well-characterized set of OVA-derived altered peptide ligands with varying affinity for this TCR (10, 13). Purified OTI-Nur77-eGFP mature CD8 T cells were incubated with APCs loaded with varying concentrations of high-, medium-, and low-affinity peptides (OVA, or N4/SIINFEKL, Q4R7, and G4, respectively), and Nur77-eGFP induction was assessed after 24 h culture. Low concentrations of N4 peptide (10⁻¹² M) were sufficient to induce GFP expression, whereas 10-fold and 1000-fold higher concentrations of Q4R7 and G4 peptides, respectively, were required to do so (Fig. 1A, 1B), consistent with previous reports (13, 18). Although all OT1 T cells upregulate GFP at high doses of G4 peptide (10⁻⁶ M), GFP fluorescence intensity among responding cells remains low due to lower peptide affinity (Fig. 1A, 1C).

We next sought to determine whether a minimal TCR signaling threshold in vivo was influenced by peptide affinity. To do so, we adoptively transferred OTI-Nur77-eGFP CD8 T cells that had been loaded with the dilutional dye CellTrace Violet into CD45.1⁺ congenic hosts. We subsequently adoptively transferred APCs loaded with a fixed concentration of N4, Q4R7, or G4 peptides, selected based on in vitro titration, and recipient spleens were analyzed 3 d later. Low-affinity G4 peptide did not induce either GFP upregulation or proliferation at the concentration used, whereas N4 robustly drove all transferred OTI cells to divide (Fig. 1D). Intermediate affinity peptide Q4R7 induced GFP upregulation even among undivided OTI cells, and drove a small fraction of those cells to undergo one to three divisions. Importantly, despite the vast difference in potency between N4 and Q4R7, both peptides drive high GFP expression in proliferating OTI cells (Fig. 1E, 1F). This is consistent with a high minimal threshold of accumulated TCR signaling that is required for T cell proliferation and is independent of peptide affinity. Interestingly, GFP mean fluorescence intensity (MFI) decreased with sequential divisions, but not by 50% as would be predicted by simple 2-fold dilution as seen with CellTrace Violet (Fig. 1F). This result suggests that cells may receive ongoing TCR signals during early cell divisions in vivo.

We showed previously that the threshold for CD4 T cell proliferation in vitro is not altered by neutralization or addition of exogenous IL-2 (7). CD8 T cells, in contrast to CD4 T cells, produce lower amounts of autocrine IL-2 and are exquisitely sensitive to IL-2 for proliferative responses (2, 19, 20). We therefore wanted to probe the effect of IL-2 on the minimal TCR signaling threshold for CD8 T cell proliferation. Purified OTI-Nur77-eGFP CD8 T cells loaded with CellTrace Violet were incubated in vitro with APCs loaded with varying concentrations of N4, Q4R7, or G4 peptides. Cells were cultured in the presence or absence of 50 U/ml IL-2, a concentration that maximally amplifies proliferative responses of OTI cell cultures (2). After 24 h, we observed peptide affinity- and dose-dependent eGFP induction that was unaffected by exogenous IL-2 (Supplemental Fig. 1A). This is consistent with previous work suggesting that Nur77-eGFP reporter expression is insensitive to IL-2–dependent JAK-STAT signaling (4, 7). Upregulation of the high-affinity IL-2Rα-chain (CD25) was sensitive to both peptide concentration and, somewhat, to IL-2 treatment as expected (Supplemental Fig. 1B) (21).

We assessed CellTrace Violet dilution after 72 h and quantified proliferation by the division index (average number of cell divisions undergone by the entire population) and % divided (fraction of initial population estimated to have divided at least once) to capture information about both the responder population and the amplitude of the response under these conditions (Fig. 2A, Supplemental Fig. 1C) (17). Addition of IL-2 to cultures reduced the peptide EC₅₀ for these indices, and this effect was most evident at low peptide doses and with low-affinity peptide (Fig. 2A, Supplemental Fig. 1C).

Although GFP expression in OT1 T cells varied broadly across several orders of magnitude in proportion to stimulus intensity after 24 h of stimulation before any cell divisions have occurred (Fig. 1A, 1B, Supplemental Fig. 1A), in the absence of exogenous IL-2, GFP expression among divided cells was consistently high regardless of peptide dose and affinity (Fig. 2B, 2C, Supplemental Fig. 1D, 1E). Under no stimulatory conditions, in the absence of IL-2, did we detect any T cell that had undergone one cell division with GFP fluorescence substantially lower than 10,000 (arbitrary fluorescence units), suggesting that a minimal threshold of integrated TCR signaling essential for proliferation corresponded to this value (Fig. 2B, 2C, Supplemental Fig. 1D, 1E). Importantly, high doses of N4 or Q4R7 drove GFP expression above the perceived threshold among divided cells, suggesting that this minimal signaling threshold could be superseded, and might in turn account for subtle differences in GFP observed in vivo (Figs. 1E, 2B, 2C, Supplemental Fig. 1D, 1E). By contrast, CD25 expression in divided cells correlated with peptide dose and affinity, but not with proliferation per second, suggesting that CD25 expression did not impose a minimal boundary condition (Supplemental Fig. 1D).

We next wanted to explore the impact of exogenous IL-2 on the minimal TCR signaling threshold required for CD8 T cell proliferation. Low-dose IL-2 drives minimal proliferation in the absence of peptide stimulation (<1% dividing cells), but synergistically enhances the proliferation of T cells stimulated with suboptimal peptide dose and/or affinity (Fig. 2A, 2B, Supplemental Fig. 1C, 1D). Importantly, in IL-2 treated cultures with low-affinity and/or low doses of peptide, we observed division of T cells with markedly lower GFP expression relative to matched cultures without IL-2 supplementation (closer to 2000 MFI) (Fig. 2B, 2C, Supplemental Fig. 1D, 1E). As before, T cells from cultures treated with high doses and/or high affinity peptide expressed high GFP well above this minimal threshold (Fig. 2B, 2C, Supplemental Fig. 1D, 1E). Because IL-2 signaling does not directly modulate GFP expression (Supplemental Fig. 1A), we reasoned that this was due to a reduced minimum of integrated TCR signaling required to drive cell division in the presence of IL-2.

CD8 T cells can produce IL-2 in response to TCR stimulation sufficient to drive quorum-sensing behavior (22). To isolate the contribution of autocrine IL-2 production on the minimal TCR signaling threshold required for CD8 T cell proliferation, we took advantage of a novel and specific JAK3 inhibitor (15). We incubated CellTrace Violet-loaded Nur77-eGFP OTI T cells together with APCs pulsed with varying concentrations of N4 and treated the cultures with either inhibitor or DMSO. Indeed, inhibitor-treated samples exhibit reduced CD25 expression consistent with inhibition of IL-2–dependent signaling, and concomitant increase in GFP expression among divided lymphocytes (Fig. 3A, 3B). This shows that even autocrine production of IL-2 by CD8 T cells is sufficient to reduce the minimal TCR signaling threshold required for proliferation.

We next wanted to determine whether another critical common γ-chain cytokine, IL-15, could also alter the minimal TCR signaling threshold for CD8 T cell proliferation. Like IL-2, IL-15 also signals through the common γ-chain, JAK1/3, and STAT5 to drive T cell proliferation (3). Importantly, the IL-15Rα-chain is expressed on naive T cells along with the common γ-chain (3). We incubated CellTrace Violet–loaded Nur77-eGFP OTI T cells together with APCs pulsed with varying concentrations of N4 in the presence or absence of low doses of IL-2 or IL-15. As expected, IL-2 and IL-15 could enhance T cell proliferation at suboptimal doses of N4 (Fig. 3C). Under these conditions, we observed reduced GFP expression among dividing T cells, suggesting that IL-15 and IL-2 could both reduce the minimal amount of integrated TCR signaling required for proliferation.

In light of observations with the OTI model system (Fig. 2), we hypothesized that analogous modulation of the minimal TCR signaling threshold for proliferation by IL-2 might be revealed in an MHC class II–restricted system using low-affinity peptides. To test this hypothesis, we took advantage of the AND TCR Tg specific for MCC and a series of altered peptide ligands with varying potency (11, 12, 14). AND Tg CD4 T cells harboring the Nur77-eGFP reporter were incubated with APCs that had been loaded with varying concentrations of high-, medium-, or low-potency peptides (MCC, K99E, or T102L respectively). After 24 h of culture with the addition of either exogenous IL-2 or anti-IL-2 neutralizing Abs (to block the autocrine effects of endogenously produced IL-2), we observed a dose response of GFP induction that corresponded to peptide potency, such that MCC was ∼10-fold more potent than K99E, and 1000-fold more potent than T102L (Fig. 4A). Consistent with prior observations, IL-2 induced CD25 expression (most evident with weaker potency peptides), but did not potentiate GFP (Fig. 4A, 4B).

After 72 h of culture, we observed clear titration of proliferative responses across dose and affinity of peptide (Fig. 4C). Moreover, GFP expression in undivided T cells varied with dose and peptide affinity as seen after 24 h of culture. However, among divided cells, GFP expression was relatively uniform and high irrespective of peptide affinity or concentration, confirming in an independent model system the existence of a high and invariant minimal TCR signaling threshold for proliferation (Fig. 4C–E). Most strikingly, addition of IL-2 to these cultures markedly upregulated CD25 as expected, but did not alter the high GFP threshold for proliferation, even with low peptide concentration or weak peptide affinity (Fig. 4C–F).

We next wanted to determine whether insensitivity of the TCR signaling threshold to IL-2 in CD4 T cells could be observed in an independent system. We therefore took advantage of OTII OVA-specific TCR Tg mice harboring the Nur77-eGFP reporter. We incubated CellTrace Violet-loaded lymphocytes from OTII-Nur77-eGFP mice with varying concentrations of OVA peptide in the presence of either 50 U/ml exogenous IL-2 or anti-IL-2 neutralizing Abs. We found that anti-IL-2 treatment profoundly reduced surface expression of CD25 and inhibited proliferation on a population level, particularly at near-threshold doses of OVA peptide (Supplemental Fig. 2A–C). However, among proliferating OTII T cells, GFP expression was high and invariant on a single-cell level for any given cell division, regardless of IL-2 signaling (Supplemental Fig. 2A–C), consistent with observations in the AND Tg model system (Fig. 4C–E). Taken together, these data suggest that IL-2 treatment modulates the TCR signaling threshold for proliferation of CD8 (OTI) but not CD4 (AND, OTII) T cells.

We have shown that the minimal TCR signaling threshold required for T cell proliferation is independent of Ag dose and affinity. To dissect the role of CD28 costimulation in regulating this threshold, we studied polyclonal Nur77-eGFP reporter T cells stimulated with plate-bound anti-CD3ε with or without concomitant anti-CD28 stimulation. We further treated cultures with either 50 U/ml IL-2 or anti-IL-2 neutralizing Abs to isolate the contribution of CD28 signaling independent of autocrine IL-2 production. We again observed that in the presence of IL-2, the TCR signaling threshold for proliferation was reduced in CD8, but not CD4 T cells, suggesting that this difference is generalizable across a range of TCR specificities (Supplemental Fig. 2D–F). However, we found that CD28 costimulation, although enhancing proliferation, had no impact on this threshold independent of IL-2 (Supplemental Fig. 2D). Furthermore, IL-2 can modulate this threshold in the presence or absence of CD28 costimulation. Finally, we confirmed the biological activity of reagents used to modulate IL-2 by assessing CD25 expression in cultured T cells (Supplemental Fig. 2G, 2H).

To understand why proliferation of CD8 and CD4 T cells were differentially responsive to IL-2, we first assessed IL-2 signaling competence in naive OTI and OTII T cells. Consistent with the lack of high affinity IL-2 receptor complex expression (comprised of IL-2Rα, β, and γ) (Fig. 5A) (21, 23), neither naive OTI nor naive OTII T cells phosphorylated STAT5 in response to the low dose of IL-2 used in our proliferation assays (Fig. 5B, 5C). In contrast to IL-2 stimulation, low doses of the common γ-chain cytokine IL-7 drove STAT5 phosphorylation in both naive CD4 and CD8 T cells at a comparable EC₅₀, albeit to different maximal levels (Fig. 5D, 5E). This argues that the common γ-chain is expressed and not limiting for signaling, consistent with prior reports (23). Higher doses of IL-2 (>500 U/ml) stimulate the intermediate affinity IL-2 receptor complex of IL-2Rβ (CD122) and IL-2Rγ (CD132) (21), which we observed when stimulating naive CD8 OTI T cells with IL-2 doses over 300 U/ml. However, naive CD4 OTII T cells did not respond even to extremely high doses of IL-2, which may reflect minimal expression of CD122, in contrast to naive OTI cells (Fig. 5A), as previously reported (23). Taken together, these data suggest that neither naive CD4s nor naive CD8s can respond to the dose of IL-2 used in our proliferation assay, and presumably require upregulation of CD25 (as well as CD122 in CD4 T cells) expression to do so.

We next asked whether induction of CD25 or signaling through the high affinity IL-2 receptor complex differed in activated CD4 and CD8 T cells. We stimulated polyclonal CD4 and CD8 T cells with plate-bound anti-CD3ε for 24 h, and assessed IL-2 receptor chain expression. CD25 and CD132 were robustly induced in both cell types at the transcript and protein levels (Fig. 6A, Supplemental Fig. 3A–C). By contrast, although Il2rb transcript was highly induced by TCR stimulation in both cell types, surface expression of CD122 (IL-2Rβ) was not upregulated in CD8 T cells. Although CD122 surface expression was slightly increased by TCR ligation in CD4 T cells, it remained reproducibly lower than in CD8 T cells (Fig. 6A, Supplemental Fig. 3A–C). Although it is possible that CD122 expression remains limiting on activated CD4 T cells, both activated CD4 and CD8 T cells express the components of the high affinity IL-2 receptor.

We restimulated activated CD4 and CD8 T cells with a broad range of IL-2 doses, and observed high basal and robust inducible STAT5 phosphorylation among CD25^hi CD8 T cells even at low IL-2 doses. CD25^lo CD8 T cells also responded, but at doses of 200 U/ml or higher (Fig. 6B–D). By contrast, IL-2 treatment induced only weak STAT5 phosphorylation in a subset of CD25^hiCD4 T cells. We assessed phosphorylation of Akt at Ser473 in response to titration of IL-2, and observed a very similar pattern of response (Supplemental Fig. 3D–F). We propose that the distinct IL-2 response in CD4 and CD8 T cells at early time points (≤24 h of culture) may account for the differential sensitivity of the TCR signaling threshold to cytokine supplementation in these cell types.

Because activated CD4 and CD8 T cells both express the high affinity IL-2 receptor chains, we next explored whether differential expression of a negative regulator of JAK-STAT signaling could account for distinct sensitivity to IL-2. We assessed expression of the SOCS transcripts, Socs1–3 and Cish in naive and TCR-stimulated CD4 and CD8 T cells (Supplemental Fig. 3G, 3H). Socs1, 2, and 3 mRNA levels are downregulated by TCR stimulation, whereas Cish mRNA is robustly induced. However, we found no marked differences in transcript expression in activated CD4 and CD8 T cells. It remains possible that CD4 and CD8 T cells exhibit posttranscriptional differences in abundance or activity of these gene products.

Because Nur77-eGFP reporter expression is sensitive to TCR, but not JAK-STAT signaling, we can visualize Ag-dependent signals in the context of various cytokine milieus on a single-cell level. This approach has revealed that costimulatory signals through IL-2 in CD8 T cells but not CD4 T cells markedly reduce the minimal threshold of Ag receptor signaling required for cell division on a single-cell level. We next wanted to identify a regulator of cell division that was cooperatively controlled by multiple mitogenic inputs and might mark a truly invariant mitogenic signaling threshold for cell division. Nr4a genes are among a small set of rapidly induced primary response genes (PRGs) that are regulated by multiple mitogenic inputs (24). Among these PRGs are the well-studied cell cycle regulators c-Myc, c-Fos, and JunB. Cantrell and colleagues (25) have recently shown that c-Myc is induced in a digital manner by TCR signals in OTI T cells via transcriptional regulation, but argue that IL-2 markedly increases c-Myc protein expression in a graded manner by enhancing protein translation. We focused attention on c-Myc as a likely integrator of TCR and IL-2-dependent signaling because it drives metabolic reprogramming and cell cycle progression that is essential for rapid clonal expansion (26–28).

Intracellular staining enables simultaneous single-cell analysis of c-Myc protein expression and Nur77-eGFP expression in a heterogeneous population of proliferating lymphocytes stimulated with plate-bound anti-CD3ε. We observed initial c-Myc induction in a small subset of CD4 and CD8 T cells as early as 2 h after stimulation, with the proportion of c-Myc⁺ cells increasing over the next several hours (Fig. 7A, 7B). The distributions of GFP and c-Myc expression in stimulated T cells were each bimodal and tightly correlated (Fig. 7A, 7B).

We next sought to determine whether regulation of c-Myc expression reflected integration of both TCR and IL-2 signaling, and under what conditions this could be observed. Induction of c-Myc in a subpopulation of OTI T cells was detectable after 4 h of stimulation, and the MFI varied correspondingly with peptide dose and affinity, (Fig. 7C–E). As expected, low-dose IL-2 treatment had no effect on c-Myc induction after 4 h of stimulation as CD25 upregulation is essential and occurs significantly later (Fig. 7E). At later time points (24 and 48 h), c-Myc protein expression was not sustained in cultures treated with peptide alone, except with high doses of high affinity of N4 peptide (Fig. 7E). Upon addition of IL-2 to these cultures, high c-Myc protein expression was maintained at later time points (Fig. 7E). Of note, IL-2 induced CD25 upregulation at 24 and 48 h as expected (Fig. 7E), and c-Myc protein expression strongly correlated with CD25 expression (Supplemental Fig. 4A). This pattern bears a striking resemblance to stimulatory conditions under which IL-2 enhances proliferative responses. These observations suggest that c-Myc protein expression in CD8 T cells is sustained by IL-2 and accounts for the reduced TCR signaling threshold for proliferation (Fig. 2B, 2C).

We hypothesized that c-Myc was representative of a gene expression program that integrates Ag and costimulatory inputs to impose a minimal mitogenic signaling threshold for proliferation on a single-cell level. We sought to determine whether c-Myc protein expression marked an invariant minimal mitogenic signaling threshold among proliferating OTI T cells cultured with a broad range of peptide dose and affinity, in the presence or absence of IL-2. As we reported earlier, we again observed markedly enhanced proliferation in the presence of IL-2 at low-dose and low-affinity peptide conditions, with a concurrently reduced expression of GFP in dividing cells (Fig. 8A, 8B). Like GFP, c-Myc expression could be induced to supra-threshold levels by high-affinity, high-dose peptide. Most importantly, in G4-treated samples with weak Ag stimulation, c-Myc expression did not drop below an invariant minimal threshold among dividing CD8 T cells irrespective of IL-2 supplementation (Fig. 8A, 8C).

To confirm that our observations were not specific to the OT1 TCR transgenic model system, we assessed Nur77-eGFP and c-Myc expression in proliferating polyclonal CD8 T cells cultured in the presence or absence of exogenous low dose IL-2. IL-2 supplementation allowed T cells with low integrated TCR signaling as read out by GFP expression to enter the first division, but all cells in the first division expressed high and invariant levels of c-Myc protein irrespective of IL-2 treatment (Supplemental Fig. 4B–E).

These data suggest that expression of c-Myc, and perhaps coordinately regulated induction of other primary response genes downstream of mitogenic inputs, might function to enforce a minimal threshold of integrated mitogenic signaling and reduce the minimal requirement for Ag receptor signaling in CD8 T cells on a single-cell level in the presence of costimulation (see model, Fig. 9).

It is not clear how early biochemical events triggered in lymphocytes by Ag and other stimuli are translated into complex cellular responses such as cell division that require hours or days, rather than minutes. We explore how lymphocyte proliferation is regulated in response to cumulative TCR signaling integrated over prolonged periods of time. We show in this study, and in prior work, that Nur77-eGFP reporter expression—by virtue of its rapid induction and long half-life—serves as a specific and faithful read-out of TCR signaling intensity summed over time in primary lymphocytes (6, 7). This reflects a broad gene expression program that is sensitive to dose and affinity of peptide–MHC stimulation on a single-cell level (29). Yet, even uniformly stimulated, genetically identical lymphocytes with a single TCR express a broad distribution of Nur77-eGFP, consistent with stochastic cell-to-cell variation in peptide–MHC contact and signal transduction. Despite this, we have observed a consistent sharp lower boundary or minimal threshold of cumulative TCR signaling for proliferation, independent of Ag abundance, affinity, or avidity. Although TCR signal strength-dependent thresholds for proliferation (as well as effector functions) have long been postulated (30), the Nur77-eGFP reporter uniquely allows such thresholds to be visualized. We propose that graded population responses to titration of antigenic stimuli are achieved in part by altering the fraction of responding cells that have received sufficient TCR signaling to cross this minimal threshold (see model, Fig. 9). Only the cells in a population that have accumulated sufficient TCR signaling to cross this putative threshold are licensed to divide; conversely, lymphocytes that accumulate subthreshold amounts of TCR signaling are entirely excluded from participating in proliferative responses.

Using the Nur77-eGFP reporter, we further demonstrate that the single-cell TCR signaling threshold for proliferation is markedly reduced by mitogenic cytokines and signals in CD8 T cells, but not CD4 T cells. We propose that cytokine input disproportionately enhances lymphocyte proliferation in the context of suboptimal Ag abundance and/or potency by permitting individual cells with low accumulated Ag receptor signaling to cross a reduced minimal signaling threshold (see model Fig. 9). Mitogenic cytokines may license low-affinity CD8 T cells to mount productive immune responses even in response to low amounts of cumulative AgR signaling that would fall below threshold in isolation. Conversely, by sequestering IL-2, Tregs may do the opposite, thereby restricting primary CTL responses by low-affinity clones or in response to low abundance Ags (31). Indeed, the Rudensky laboratory has recently shown that capture of IL-2 by mature Tregs is dispensable for the control of CD4 T cells, but is vital for limiting the activation of CD8 T cells (32). Our observations may help explain why CD8 but not CD4 T cells begin to mount responses to presumably low-affinity self-antigens in this context. A model for quorum sensing by T cells that is mediated by secreted IL-2 and modulated by IL-2 sequestration has been proposed to operate at precisely such suboptimal stimulation conditions with low-affinity or low-abundance Ag (33). Recent work by Altan-Bonnet and colleagues (19, 22) similarly demonstrates in vitro bystander activation (or co-optation) of low-affinity CD8 T cell clones mediated by IL-2 secretion from strongly activated high affinity CD8 T cells. Supply of costimulatory signals by either CD4 or CD8 T cells could thus alter not only the magnitude but also the quality of resulting immune responses by recruiting low-affinity clones to participate (30, 34). By contrast, we find that CD4 T cells maintain a high TCR signaling threshold for Ag-dependent proliferation that is not reduced by cytokine input. In this manner, low-affinity, self-antigen–responsive CD4 T cells are restrained from clonal expansion via bystander elaboration of cytokines in the context of a physiologic immune response. This could provide a critical added layer of protection against autoimmunity.

Hodgkin and colleagues (2) have proposed a calculus of T cell responses in which average number of cell divisions in a population before quiescence or death (mean division destiny) sum linearly when TCR and costimulatory signals are combined. However, this model of linear summation focuses on analysis of strong TCR stimuli (N4 and Q4), which only differ in potency by a factor of 40 (13), and consequently it does not capture an important feature of costimulatory signals: synergistic effects with weak stimuli. We show that a wider range of peptide affinity and dose unmasks non-linear synergy between IL-2 and suboptimal TCR stimulation. We propose that a vital function of IL-2 is to transform suboptimal stimuli into productive immune responses. We further argue that this effect permits weakly stimulated lymphocytes to participate in an immune response. In a physiologic context with T cell clones of varying affinity, IL-2 should permit low-affinity clones to participate significantly in an immune response from which they would otherwise be excluded.

Mitogenic stimuli in lymphocytes trigger expression of an overlapping set of immediate-early or PRGs, such as c-myc, c-fos, and Jun-B, which are essential to direct cell-cycle progression (24). Not surprisingly, protein expression driven by the primary response genes is under exquisite posttranscriptional and posttranslational regulation. Recent work has identified a robust layer of posttranscriptional regulation of c-Myc by cytokine signaling (25). We show that expression of c-Myc protein integrates AgR and cytokine inputs in T cells. We suggest that c-Myc protein accumulation may serve as an integrator of mitogenic input, and in turn marks a seemingly invariant minimal threshold essential for initiation of cell division. The Hodgkin group has recently proposed that c-Myc expression declines over time in proliferating lymphocyte populations, and when it drops below a minimal threshold, cell division ceases (35). Similarly, the PRG c-fos has been proposed to serve as a counter of intermittent TCR signaling events in vivo, and phospho-c-Jun is sensitive to strength and duration of TCR signaling, although it is not known whether their expression marks a similar threshold for proliferation (36, 37). c-Myc and related PRGs, such as c-Fos and Jun, may collectively enforce such a threshold by directing a broad program of metabolic reprogramming and proliferation (28). Indeed, recent work shows that the amount of c-Myc protein expressed in a CD8 T cell population correlates with proliferative potential (26, 27). It has been proposed that the PI3K pathway links mitogenic inputs with c-Myc and PRG expression, but this remains an open area of investigation (19, 38).

How is a high and invariant minimal threshold of TCR signaling maintained by CD4 but not CD8 T cells in the face of costimulatory IL-2? We show that neither CD4 nor CD8 naive T cells phosphorylate STAT5 in response to low doses of IL-2. However, despite comparable strong CD25 upregulation on activated CD4 and CD8 T cells after 24 h of stimulation, CD8 but not CD4 T cells robustly phosphorylate STAT5 at this time point. Further, this difference is not attributable to differences in CD132 expression at this time point. Because CD4 T cells can respond sensitively to IL-2 by 72 h as evidenced by increased expression of CD25 at that time point, we propose that CD4 T cells acquire IL-2 responsiveness more slowly than CD8 T cells, perhaps because of limiting expression of CD122, or differential expression of a regulator of JAK-dependent signaling. We do not identify differential expression of Socs gene transcripts in CD4 as compared with CD8 T cells, but it remains possible that posttranscriptional differences in abundance or activity of these gene products may contribute to our observations. Such a wiring scheme would prolong the duration of TCR-dependent regulation of CD4 T cell proliferation, and thereby impose a high TCR signaling threshold on CD4 T cells, which is not modulated by mitogenic cytokine supplementation (Fig. 9).

Taken together, our model provides a conceptual framework for understanding synergistic regulation of clonal expansion by subthreshold Ag receptor signaling in the context of mitogenic costimulatory signals, thereby rendering CD8 T cells exquisitely dependent upon environmental cues such as IL-2. By contrast, we show that CD4 T cells maintain an invariant minimal TCR signaling threshold for proliferation that is independent of Ag affinity, abundance, and IL-2 supply, thereby restricting responses to low-affinity self-antigens irrespective of IL-2 supply.

We thank Al Roque for assistance with animal husbandry.

The authors have no financial conflicts of interest.