The outcome of an immune response relies on the competitive capacities acquired through differentiation of CD4+ T cells into Th1 or Th2 effector cells. Because Stat4 and Stat6 proteins are implicated in the Th1 vs Th2 generation and maintenance, respectively, we compare in this study the kinetics of Stat4−/− and Stat6−/− CD4+ T cells during competitive bone marrow reconstitution and lymphopenia-driven proliferation. After bone marrow transplantation, both populations reconstitute the peripheral T cell pools equally well. After transfer into lymphopenic hosts, wild-type and Stat6−/− CD4+ T cells show a proliferation advantage, which is early associated with the expression of an active phospho-Stat4 and the down-regulation of Stat6. Despite these differences, Stat4- and Stat6-deficient T cells reach similar steady state numbers. However, when both Stat4−/− and Stat6−/− CD4+ T cells are coinjected into the same hosts, the Stat6−/− cells become dominant and out-compete Stat4−/− cells. These findings suggest that cell activation, through the Stat4 pathway and the down-regulation of Stat6, confers to pro-Th1 T cells a slight proliferation advantage that in a competitive situation has major late repercussions, because it modifies the final homeostatic equilibrium of the populations and favors the establishment of Th1 CD4+ T cell dominance.
The mechanisms involved in the differentiation of naive CD4+ T cells into Th1 and Th2 effector cells have been under scrutiny. Nevertheless, it is still not possible to establish a hierarchy of the factors that determine the final Th1 or Th2 pattern of an immune response. We know that the cellular environment of the initiating response influences the Th1/Th2 decision; an IL-12-rich environment induces the generation of Th1 cells (1, 2), whereas IL-4 promotes the generation of Th2 cells (3, 4). The type and concentration of Ag (5, 6) and/or the type of APCs involved in T cell priming, the genetic predisposition (7), the affinity and strength of the TCR signal conveyed (8, 9, 10), and/or other accessory signals (11, 12) also decide T cell fate (13). T cell differentiation may also be resolved by intrinsic properties of activated pro-Th1 and pro-Th2 T cells, such as their different survival abilities due to a modified susceptibility to Fas-mediated death signals (14, 15). Finally, the establishment of a Th1- or Th2-type immune response could also be predisposed by the frequency and/or rate of expansion (16, 17) of the differentiating T cell subsets.
Among the proteins concerned in CD4+ T cell differentiation, Stat4 has been implicated in the induction of Th1 T cells (3, 18, 19), whereas Stat6 has been shown to be involved in the generation of Th2 T cells (20, 21, 22). In the mouse, Stat4 and Stat6 proteins are activated upon IL-12 (23, 24) and IL-4 (25) cytokine simulation, respectively. Cytokine binding results in receptor dimerization and recruitment of Jak1/Jak3 for the IL-4R and Jak2/Tyk for the IL-12R, followed by phosphorylation of the receptor, which allows the binding of Stat proteins via their Src homology 2 domain. After tyrosine phosphorylation (26), the Stat proteins homodimerize and translocate to the nucleus, where they induce transcription of Th-specific proteins. Although Stat4 and Stat6 were originally considered the master regulators of Th1/Th2 induction, current data suggest that they play a role in maintaining Th cell phenotype and function. Stat6 increases the transcription of Gata3 (27), which remodels both IL-4 and IL-13 loci (28), rendering them accessible; Stat6 can then induce transcription of the IL-4 locus (29, 30, 31). Gata3 is found at basal levels before differentiation and can promote its own transcription, inducing Th2 differentiation in a Stat6-independent manner (32, 33). Despite this, Stat6 remains an important Th2 transcription factor also involved in lymphocyte proliferation and homeostasis (34, 35, 36, 37). The counterpart of Gata3 in Th1 cells is T-bet (38). T-bet is not dependent on Stat4 and can be induced by the homodimer of Stat1 after activation by IFN-γ. T-bet is responsible for the remodeling of the IFN-γ locus and the induction of the IL-12R β-chain (39, 40). Stat4 intervenes later by maintaining IL-12Rβ2 expression that mediates IL-12 induction of IFN-γ (36, 41). However, Stat4 independent pathways of IFN-γ production have been shown to exist (42, 43). Stat4’s central role in Th1 development has recently been reasserted by showing that Gata3 retroviral expression in developing Th1 cells disrupts their differentiation when T-bet is retrovirally coexpressed, but not when Stat4 is coexpressed (44).
We investigated in this study whether differences in the rate of expansion of the Th1 and Th2 precursor T cell subsets existed. Because Stat4 is involved in the induction and maintenance of the Th1 phenotype, whereas Stat6 is implicated in that of Th2, we used a lymphopenia-driven proliferation model consisting of the adoptivetransfer of T cells from wild-type (WT),4 Stat6-deficient (22) or Stat4-deficient (19) mice into RAG2-deficient hosts and studied the proliferation kinetics and competitive capacities of the various CD4+ T cells. Studying the CFSE dilution patterns, we found that Stat6−/− CD4+ T cells proliferate faster than Stat4−/− CD4+ T cells. When cotransferred into the same host, Stat6−/− CD4+ T cells are preferentially activated, become dominant, and out-compete Stat4−/− CD4+ T cells. Overall, our findings suggest that T cell activation using the Stat4 pathway, while down-regulating Stat6, confers on the cells a competitive advantage that might favor the final establishment of a Th1 response.
Materials and Methods
Stat6−/−Thy1.2 BALB/c (22) and Stat4−/−Thy1.2 BALB/c (19) micewere originally from The Jackson Laboratory, and RAG2−/−BALB/c mice were from the Centre de Développment des Techniques Avancées-Centre National de la Recherche Scientifique. Stat4−/−Thy1.2 mice were crossed with BALB/cThy1.1 to obtain Stat4−/−Thy1.1 BALB/c mice. Double-knockout mice were generated by crossing Stat4−/−Thy1.2 and Stat6−/−Thy1.2 mice. All mice where subsequently bred in our animal facilities. For each experiment, animals were matched for age and sex.
Bone marrow (BM) chimeras
Host 7- to 10-wk-old RAG2−/− mice, were lethally irradiated (850 rad) with a 137Ce source and received i.v. 2 × 106 T cell-depleted (<0.1%) BM cells from different donor mice, alone or mixed at different ratios. T cell depletion was performed using an AutoMACS (Miltenyi Biotec) after incubating BM cells with anti-CD3, followed by PE-MACSbeads. Eight weeks after transplant, thymus, spleen, and lymph node (LN) cell suspensions were prepared, and the number and phenotype of the cells from each donor population were evaluated.
LN cells from WT, Stat6−/−, Stat4−/−, and/or double-knockout donor mice differing in the Thy1 allotype were injected i.v., alone or mixed at different ratios, into RAG2−/− host mice. The host mice were killed at different time intervals after cell transfer. The number and phenotype of the cells found in the spleen and LN (inguinal, brachial, and mesenteric) were evaluated.
Labeling with CFSE and mathematical analysis
Donor LN cells were labeled with the cell division tracking die CFSE (45). Briefly, LN cell suspensions (107/ml) in PBS were incubated for 15 min at room temperature with CFSE (5 μM), washed in 5% FCS complete medium, and resuspended in RPMI 1640 medium for i.v. injection. To estimate cell division rates from the CFSE data, we analyzed the precursor cohort plots of the Stat4−/− and Stat6−/− cells. Precursor cohort plots compensate for the fact that cells that have divided are represented by twice as many cells in the next CFSE peak. The precursor population originating each CFSE peak was calculated by dividing the number of cells ni in each peak by 2i, where i represents the division number of the CFSE peak as originally proposed by Gett and Hodgkin (46). To compare the lag time of Stat4−/− and Stat6−/− cells to begin division, we calculated the percentage of cells from the precursor cohort that had made at least one division and plotted the percentage of dividing cells against time. The exponential regression curve fitted through each set of data gives an indication of the lag time elapsed before a significant number of cells achieved a first division. To obtain rough estimates of the subsequent division rates of proliferating Stat4−/− and Stat6−/− cells, the average division index of the precursor cohort at each time point (i.e., ∑(i × ni/2i)/∑(ni/2i)) was plotted, and the slope of the linear regression through these data was calculated. In the latter analysis, cells that had not made even one division were disregarded, i.e., i = 1–7.
Flow cytometric analysis
The following mAbs were used: anti-CD4 (L3T4/RM4-5), anti-CD3ε (145-2C11), anti-CD8α (53-6.7), anti-CD45RB (16A), anti-CD90.1 (Thy1.1, OX-7), anti-CD90.2 (Thy 1.2, 53-2.1), and anti-CD25 (784) from BD Pharmingen and anti-CD127 (A7R34) from eBioscience. Cell surface four-color staining was performed with the appropriate combinations of FITC-, PE-, PerCP-, and allophycocyanin-coupled Abs. Dead cells were excluded during analysis according to their light-scattering characteristics. All acquisitions and data analysis were performed with a FACSCalibur (BD Biosciences) interfaced to Macintosh CellQuest software (Apple Computer).
Purification of CD4+ cells
Host mice were killed at different time points after transfer, and their spleens were pooled. Cell suspensions were enriched for CD4+ cells by sorting in an AutoMACS (Miltenyi Biotec) with the help of labeled magnetic beads. Briefly, cells were first incubated with an allophycocyanin-labeled Ab mix (anti-B220 (RA3-6B2), anti-MacI (CD11b, M1/70), anti-CD11c (HL3), and anti-CD8α (53-6.7), all from BD Pharmingen), followed by incubation with anti-allophycocyanin-coated MACSbeads (Miltenyi Biotec). The negative fraction was additionally labeled with anti-CD3-PE, and anti-CD4-FITC, and the double-positive population was sorted to ≥98% purity on a FACStar Plus (BD Biosciences) for RT-PCR. Alternatively, the negative fraction was labeled with anti-CD4-coated MACSbeads (Miltenyi Biotec) and sorted for Western blotting analysis.
Intracellular staining for cytokine production
Approximately 5 × 106 spleen cells enriched for CD4+ cells (after negative selection) were incubated for 4 h at 37°C with brefeldin A (10 μg/ml) and with or without PMA (50 ng/ml) plus ionomycin (500 ng/ml). Harvested cells were washed and stained as described above. The cells were then fixed at room temperature with 1% paraformaldehyde. Thereafter cells were stained with allophycocyanin-coupled anti-IL-4 (11B11) and PE-coupled anti-IFN-γ (XMG1.2) in 0.1% saponin, 1% FCS, and PBS (all Abs were from BD Pharmingen).
Preparation of whole cell lysates and Western blots for Stat proteins
The purified CD4+ cells were washed once in PBS and resuspended in 30 μl (for 1.6 × 106 cells) of freshly prepared lysis buffer (1 mM MgCl2, 1 mM EGTA, 50 mM NaF, 1 mM Na3VO4, 10 mM Na4P2O710H2O, 150 mM NaCl, 1% (v/v) Nonidet P-40, 20 mM Tris (pH 7.5), 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 1 mM Pefabloc). The solution was kept on ice for 30 min, and the supernatant was cleared by centrifugation at 15,000 rpm. The protein solutions obtained were mixed with the loading buffer and boiled, and 40 μl was loaded into an 11% SDS-acrylamide gel (15 mm). The gel was run at 25 mA and blotted at 100 V for 90 min; blocking was performed with 5% BSA. The membrane was probed with anti-Stat4 (sc-486; Santa Cruz Biotechnology) or anti-Stat6 (sc-981; Santa Cruz Biotechnology) and anti-actin (sc-7210; Santa Cruz biotechnology) in 10 mM Tris-HCl, 150 mM NaCl, 0.05% Tween, and 1% BSA. Visualization was performed by a 1:1 mixture of Pierce’s Super Signal West Pico and exposure on a Kodak BioMax MR-1 film. The membranes were stripped with 0.1 M glycine/HCl, pH 2.3 (15 min, room temperature), followed by three washes with 1 M NaCl/PBS and four washes with 1% Tween/PBS. They were subsequently probed with anti-pStat6 (5A4; Cell Signaling Technology) or anti-pStat4 (ST4P; Zymed Laboratories) under the same conditions as before. The films were analyzed using a Kodak Image Station 440 interphased to a Kodak Digital Science 1D program. The Stat and p-Stat band intensities were normalized by actin, and the levels of protein present were quantified.
The purified CD4+ cells were washed twice in PBS before lysis, and RNA was purified using an RNeasy Mini Kit (Qiagen) according to the manufacturer’s indications. Thereafter, 1 μg of RNA was reverse transcribed into single-stranded cDNA with a SuperScript First-Strand Synthesis System (Invitrogen Life Technologies). The single-stranded cDNA (20 ng) was amplified with different TaqMan MGB probes from Applied Biosystems with a FAM reporter dye at the 5′: Stat-6 (Mm00447411-m1), IL-4 (Mm00445259-m1), IL-13 (Mm00434204-m1), T-box21 (Mm045096-m1), and hypoxanthine phosphoribosyltransferase (Mm00446968-m1). The samples were also amplified with designed probes for Gata3 and IFN-γ, which had a FAM reporter die at the 5′ end and a TAMRA quencher at the 3′ end. The sequences used were reported by Grogan et al. (17). All amplifications were performed under standard conditions as described by Applied Biosystems. Analysis was performed in an ABI PRISM 7000 Sequence Detection System (Applied Biosystems). Quantitation was performed by comparing the levels of expression of a given mRNA product to its levels found ex vivo (before injection) and normalizing to the levels of hypoxanthine phosphoribosyltransferase of the given sample:
In vitro Th differentiation
Sorted CD4+ LN cells from Stat6−/− or Stat4−/− donors were plated at 105 cells/well in complete medium (RPMI 1640, 10% FCS, 1/500 β-ME, 10 μl/ml HEPES, 1/100 penicillin/streptomycin, 10 U/ml IL-2, and 5 μg/ml anti-CD28). The wells were previously coated with 1 μg/ml anti-CD3 for 4 h at 37°C. To the wells containing Stat6−/− cells we added 20 μg/ml anti-IL-4, 10 ng/ml IL-12, and 10 ng/ml IFN-γ; to those containing Stat4−/− cells we added 10 μg/ml anti-IL-12, 10 μg/ml anti IFN-γ, and 20 ng/ml IL-4. The cells were incubated for 4 days at 37°C in 5% CO2. Thereafter, the cells were lysed for Western blotting analysis. Alternatively, they were washed and replated on uncoated wells with complete medium for another 3 days. They were then lysed for quantitative RT-PCR.
During lymphopenia-driven proliferation, Sta6−/− CD4+ T cells proliferate faster than Stat4−/− CD4+ T cells
The establishment of a Th1- or Th2-type immune response could be tailored by the rate of expansion of the different precursor T cell subsets. Because Stat4 is involved in the induction and maintenance of the Th1 phenotype, whereas Stat6 is implicated in that of Th2, we compared the proliferation rates of CD4+ T cells from Stat4−/− and Stat6−/− mice (19, 22). For this purpose we used a lymphopenia-driven proliferation model and followed the fate of CFSE-labeled LN T cells from WT, Stat4−/−, or Stat6−/− donors transferred into histocompatible RAG2-deficient hosts. According to the CD45RB expression, the ratio of effector/memory vs naive T cells was ∼1:3, and it was the same for WT, Stat6−/−, and Stat4−/− donor cells (data not shown). At various time points after transfer (days 1, 4, 5, 7, and 10), we analyzed in the host’s spleen the total number of CD4+ donor T cells (Fig. 1) and the number of cell divisions incurred (Fig. 2). On days 7 and 10 after transfer, we recovered more Stat6−/− than Stat4−/− CD4+CD3+ T cells in the host’s spleen (Fig. 1). On day 7, the majority of the Stat6−/− cells were CFSE-negative, whereas most Stat4−/− cells were still CFSE-positive (Fig. 2,A). We estimated the number of originally injected cells that gave rise to the cells present in each peak (see Materials and Methods). We found that by day 7, ∼9.5% of CD4+CD3+ Stat6−/− T cells had divided seven or more times (Fig. 2 B), whereas only 1.5% of Stat4−/− T cells had undergone the same number of cell divisions. By day 10, half the original injected Stat6−/− T cells had divided at least once, in contrast to only 30% of Stat4−/− cells. Interestingly, we found that T cells from WT BALB/c mice showed a pattern of cell division similar to that observed for Stat6−/− T cells (data not shown).
To determine whether the proliferative advantage of Stat6−/− was only due to a larger proportion of cells going into division or was also caused by a faster subsequent cell division rate, we plotted the percentage of cells from the precursor cohort that had made at least one division against time (Fig. 2,C). The percentage of cells that had divided was consistently larger for Stat6−/− cells than for Stat4−/− cells at any time after transfer. Subsequently, we analyzed the average division index of the cells that had made at least one division (Fig. 2,D). The number of divisions that proliferating Stat6−/− cells had made was consistently greater than that of cells of Stat4−/− origin. As a rough estimate of the division rate, we determined the slope of the regression line through the data (46). Fig. 2 D illustrates that the time interval between divisions was shorter for Stat6−/− (1.5 days) than for Stat4−/− (3.6 days) cells. As a reference, dividing T cells from BALB/c mice proliferated once every 1.8 days (data not shown). It is important to note that the above calculations provide only rough estimates of cellular division rates, because the average division index of proliferating cells is reduced by cells that start dividing relatively late, and because CFSE labeling does not allow any distinction between cells that have divided seven or more times. Despite these shortcomings, our findings suggest that Stat6−/− CD4+ T cells have a proliferation advantage over Stat4−/− CD4+ T cells during lymphopenia-driven proliferation; T cells of Stat6−/− origin show a shorter lag time to division and divide at a faster rate than cells of Stat4−/− origin.
During lymphopenia-driven proliferation of CD4+ T cells, Stat4 is phosphorylated before cell division, whereas Stat6 is down-regulated
We studied whether early activation during lymphopenia-driven proliferation could result in the differential regulation and phosphorylation of Stat4 and Stat6 proteins. CD4+ cells from WT, Stat6−/−, or Stat4−/− donor mice were injected into RAG2−/− mice. We analyzed the state of phosphorylation of the Stat proteins in donor-derived CD4+ T cells at 1, 3, and 4 days after transfer. After transfer, the expression of Stat4 by Stat6−/− and WT CD4+ T cells was first down-regulated (compared with the levels of expression before transfer) and increased thereafter (Fig. 3, A and C). Early after transfer, Stat4 is also transiently phosphorylated in both Stat6−/− and WT CD4+ T cells (Fig. 3, A and C). In contrast, the expression of Sta6 in both Stat4−/− and WT CD4+ T cells is abrogated early after transfer (Fig. 3, B and C). We should point out that P-Stat6 was detected in lysates of Stat4−/− CD4+ T cells activated in vitro under Th2 conditions (Fig. 3,B, activated lane). We confirmed the quantitative down-regulation of Stat6 expression at the mRNA level by real time RT-PCR, which showed that the Stat6 message only reappeared late after transfer (Fig. 3 D, right lane). In summary, early during lymphopenia-driven proliferation, WT CD4+ T cells express an active phospho-Stat4 that peaks on day 3 and is rapidly dephosphorylated thereafter; both events happen before the onset of expansion. Simultaneously, Stat6 is down-regulated.
Isolated Stat6−/− and Stat4−/− CD4+ T cells reach the same homeostatic plateau independently of their initial rate of proliferation
During lymphopenia-driven proliferation, donor T cells proliferate rapidly to reach equilibrium 3–4 wk after transfer. To investigate whether the initial faster rate of proliferation of Stat6−/− CD4+ cells would allow these cells to reach a higher homeostatic plateau, we injected different numbers (103, 104, and 105) of CD4+CD3+ T cells from Stat4−/−, Stat6−/−, or WT BALB/c donors into RAG2−/− mice. We found that 25 and 56 days after transfer (Fig. 4), when 104 cells or more were injected, all CD4+ populations from Stat4−/−, Stat6−/−, and BALB/c donors reached the same plateau at ∼2–3 million cells. It should be noted, however, that when the initial cell number injected was lower (103), the number of Stat6−/− T cells recovered was higher than the number of Stat4−/− and BALB/c T cells.
We next studied whether CD4+ T cell activation during lymphopenia-driven proliferation in the absence of intentional Ag stimulation would induce Th1 or Th2 cell differentiation. Early after transfer, a fraction of both Stat4−/− and Stat6−/− donor CD4+ T cells secreted IFN-γ after in vitro stimulation (Table I). IFN-γ-secretion correlated with cell division, because most of the IFN-γ-producing cells were among the CFSE-negative donor cells. Because Stat6−/− divided earlier and faster, the percentage of IFN-γ-secreting cells was higher for Stat6−/− than Stat4−/− CD4+ T cells (Table I). We were not able to detect IL-4 production by intracellular staining (not shown). Interestingly, soon after T cell transfer we detected a significant increase in the IFN-γ mRNA levels produced by nonlymphoid spleen cells of the RAG2−/− host (Table II), suggesting that T cell transfer induces activation and IFN-γ secretion by host cells. Once the donor T cells reached steady state, we quantified their levels of mRNA for different transcription factors and ILs involved in Th1 and Th2 differentiation. We found that although Stat4−/− cells in the spleen expressed increased mRNA levels for IFN-γ, T-bet, and IL-4, they maintained levels of Gata3 mRNA not significantly different from those found ex vivo (Table III). In contrast, Stat6−/− cells increased the levels of T-bet and IFN-γ and decreased IL-4 levels while maintaining Gata3 mRNA levels (Table III). BALB/c T cells increased mRNA levels for IFN-γ and T-bet. It should be noted, however, that these changes in mRNA levels were 5- to 10-fold lower than those obtained from fully polarized T cells stimulated in vitro under Th1 or Th2 conditions (Table III). The increased IFN-γ mRNA correlated with increased protein production, because ∼6–12% of Stat4−/− and Stat6−/− CD4+ spleen T cells contained IFN-γ in their cytoplasm after 4-h in vitro activation (not shown). We were unable to detect mRNAs for IL-13 in the ex vivo samples and in the expanded CD4+ T cells; likewise, we were unable to detect production of IL-4 or other cytokines by expanded CD4+ T cells whatever their phenotype, in contrast to what was observed at the mRNA level (Table III).
|.||% Producing IFN-γ .||.||.||.|
|.||Day 7 .||.||Day 10 .||.|
|.||CFSE− .||CFSE+ .||CFSE− .||CFSE+ .|
|.||% Producing IFN-γ .||.||.||.|
|.||Day 7 .||.||Day 10 .||.|
|.||CFSE− .||CFSE+ .||CFSE− .||CFSE+ .|
CFSE-labeled Stat6−/−, Stat4−/−, or BALB/c CD4+ cells were transferred into RAG2−/− hosts. The mice were killed at various time points, and the splenocytes were enriched for CD4+ cells. After 4-h in-vitro; activation, the cells were intracellularly marked for IFN-γ and analyzed by FACS.
|.||IFN-γ Production .||.||.|
|.||Day 1 .||Day 3 .||Day 4 .|
|.||IFN-γ Production .||.||.|
|.||Day 1 .||Day 3 .||Day 4 .|
Stat6−/−, Stat4−/−, or BALB/c CD4+ cells were transferred into RAG2−/− hosts. At different time points after transfer, the hosts were killed, and their nonlymphoyd splenocytes were purified. IFN-γ mRNA levels were determined.
|.||Stat6−/− .||Stat4−/− .||BALB/c .||In Vitro Th1 .||In Vitro Th2 .|
|.||Stat6−/− .||Stat4−/− .||BALB/c .||In Vitro Th1 .||In Vitro Th2 .|
Stat6−/−, Stat4−/−, and BALB/c CD4+ cells were transferred into RAG2−/− hosts. Five weeks after transfer, the mice were killed, and spleen CD4+CD3+ cells were purified. Their mRNA levels for T-bet, Gata3, IFN-γ, IL-4, and IL-13 were quantified by real-time RT-PCR. As a control, purified Stat6−/− or Stat4−/− CD4+CD3+ cells were incubated under Th2 or Th1 conditions, respectively, and similarly tested. Values > 1 mean that the mRNA expression found at the plateau was grater than in the ex vivo sample, values < 1 mean that the mRNA expression was lower. ND, not determined.
Thus, during lymphopenia-driven proliferation, both Stat4−/− and Stat6−/− CD4+ cells reach the same homeostatic plateau independently of their initially different rates of proliferation. Nevertheless, their pattern of Th-specific mRNA expression differs. Whereas Stat6−/− cells selectively increase their levels of T-bet and IFN-γ, Stat4−/− cells up-regulate T-bet, IFN-γ, and IL-4. The Gata3 levels were kept at ex vivo levels independently of the cell type studied.
During competitive expansion, Stat6−/− dominate Sta4−/− CD4+ T cells
We found that during lymphopenia-driven proliferation, Stat6−/− CD4+ T cells start dividing earlier and divide faster than Stat4−/− CD4+ T cells. Despite these differences, both cell populations reach identical homeostatic plateaus. We decided to investigate whether under competitive conditions the two cell types would behave differently. First, we used a competitive BM repopulation strategy, in which lethally irradiated RAG2−/− mice were reconstituted with T cell-depleted BM precursor cells from Stat6−/−Thy1.2 and Stat4−/−Thy1.1 donors mixed at different cell ratios (1:3, 1:1, and 3:1). The chimeras were killed 8 wk after reconstitution, and the numbers and phenotypes of the different donor-derived Thy1.1 and Thy1.2 T cell populations were assessed in thymus, spleen, and LN. We found that the original ratio of injected Stat6−/−Thy1.2 and Stat4−/−Thy1.1 donor cells was maintained among the thymus double-positive (not shown), thymus single-positive CD8+ (not shown), and thymus single-positive CD4+ cells (Fig. 5,A), as well as in the peripheral CD8+ (not shown) and CD4+ (Fig. 5 B) T cell pools. According to the level of CD45RB expression, the fraction of effector/memory T cells in the peripheral pool was ∼30% and was the same for Stat6−/− and Stat4−/− cells (data not shown). In these chimeras, we compared the competitive repopulation capacity of T cell populations during precursor differentiation to thymus maturation, export, and periphery colonization. Our findings indicate that during replenishment of peripheral pools by stable thymus production and export, Stat4−/− and Stat6−/− CD4+ T cells reach identical steady states and show similar competitive capacities.
Next, we compared the competitive capacities of Stat6−/−Thy1.2, Stat4−/−Thy1.1, and WT BALB/cThy1.1 or Thy1.2 CD4+ T cells during lymphopenia-driven expansion. RAG2−/− hosts received populations of LN CD4+ T cells, which were injected alone or coinjected two by two at different cell ratios (1:3, 1:1, and 3:1). Four weeks after transfer, the total number of CD4+ cells recovered was similar (∼1.5–3 × 106 cells) in all groups of host mice independently of the original ratio of injected cells. In the host mice receiving different mixtures of two donor cell types, we found that the CD4+ T cells from Stat6−/− donors became dominant and out-competed both WT and Stat4−/− T cells, whereas WT CD4+ T cells out-competed Stat4−/− T cells (Fig. 6,A). Thus, in mice injected with a mixture of LN cells containing 50% Stat6−/− and 50% Stat4−/− CD4+ T cells, the proportion of Stat6−/− CD4+ T cells reached 80% of the final number of cells recovered in the host (Fig. 6,A). The finding that WT cells were also out-competed by Stat6−/− T cells suggests that the presence of Stat6 may slow late T cell expansion. Therefore, we compared the competitive capacities of Stat6−/−Thy1.2, Stat4−/−Thy1.1, and WT CD4+ T cells with those of double-deficientStat6−/−Stat4−/− T cells (Fig. 6,B). In this experiments host mice were coinjected with 50:50 mixtures of these different cellular populations. We found that double-deficient Stat6−/−Stat4−/− T cells were out-competed by Stat6−/− T cells, but not by either Stat4−/− or WT T cells (Fig. 6 B). These findings suggest that the balance between Stat4 and Stat6 may be critical to the rates of CD4+ T cell proliferation and expansion.
To investigate whether Stat6−/− CD4+ T cells could have suppressed the early activation of the cotransferred Stat4−/− CD4+ T cells, we compared the early kinetics of proliferation of CFSE-labeled Stat4−/− CD4+ T cells transferred alone or with Stat6−/− CD4+ T cells. We found that the patterns of dilution of CFSE labeling were similar in both groups of mice and that on day 10 a significant fraction of the transferred cells were CFSE−, indicating that these cells underwent several rounds of division (not shown). These findings suggest that Stat4−/− CD4+ T cells are out-competed by failing to accumulate late after transfer, rather than during early proliferation. In the mice harboring the two T cell populations, we found that the levels of IFN-γ, T-bet, IL-4, and Gata3 mRNAs expressed by Stat4−/− and Sta6−/− CD4+ T cells at steady state (Table IV) were similar to those observed when these cells were injected alone (see Table III).
|.||Stat6−/− .||Stat4−/− .|
|.||Stat6−/− .||Stat4−/− .|
A 1:1 mixture of Stat6−/− Thy1.2 and Stat4−/− Thy1.1 CD4+ cells was cotransferred into RAG2−/− hosts. Five weeks after transfer, the mice were killed, and the spleen CD4+ CD3+ cells were purified. Their mRNA levels for T-bet, Gata3, IFN-γ, IL-4, and IL-13 were quantified by real-time RT-PCR. Values >1 mean that the mRNA expression found at the plateau was greater than that in the ex vivo sample; values <1 mean that the mRNA expression was lower.
Overall, these findings indicate that upon lymphopenia-driven expansion, in the absence of intentional Ag stimulation, Stat6−/− CD4+ T cells show a clear competitive advantage and by their presence regulate the accumulation of Stat4−/− CD4+ T cells. To determine whether the competitive advantage was exclusively due to the intrinsic properties of Stat6−/− T cells to expand or to their initial faster rate of proliferation and subsequent pre-emptive competition, we performed experiments using sequential cell transfers. Host mice first received a population of Stat4−/− CD4+ T cells and 9 days later were given a second population of Stat6−/− T cells, and vice versa. Host mice were killed 4 wk after transfer of the second population. We found that independently of the order the populations were injected, the size of the first population systematically exceeded that of the second population (Fig. 6 C). These findings suggest that the competitive advantage of Stat6−/− T cells is most likely due to their initial faster division rate that allows them to fill first the peripheral T cell compartments and then prevent the late accumulation of the more slowly dividing Stat4−/− T cells.
The establishment of a Th1- or Th2-type immune response relies on the cellular environment of the initiating response, but may also be influenced by the pace of differentiation and the rate of proliferation of the different CD4+ T cell subsets. Because Stat4 and Stat6 proteins have been implicated in the differentiation of Th1 and Th2 CD4+ T cells, respectively (3, 18, 19, 20, 21, 22), we studied whether the Stat4 and Stat6 pathways could play a role in the kinetics of CD4+ T cell proliferation and expansion. We used a model of T cell proliferation in the absence of intentional antigenic stimulation and followed the fate of Stat4- and Stat6-deficient T cells adoptively transferred into RAG2−/− immune-deficient hosts. Indeed, it has been shown that upon transfer into T cell-deficient hosts, mature T cells undergo a lymphopenia-driven proliferation and expansion to reach a stable homeostatic plateau 3–4 wk later (47). Different findings validate lymphopenia-driven proliferation as a model to study in vivo T cell activation and expansion. Lymphopenia-driven proliferation is TCR mediated (for review, see Ref.48): 1) T cell expansion is determined by the abundance of peptide/MHC epitopes and their interactions with the TCRs (49); 2) CD4+ T cells from normal donors fail to proliferate in H-2M-deficient hosts due to their inability to recognize the unique peptide presentation by the host’s MHC class II molecules (50, 51); 3) various populations of homogeneous TCR Tg T cells show different expansion capacities, a property related to the promiscuity or the strength of the TCR stimulation (52) (N. Legrand and A. A. Freitas, unpublished observations). Moreover, T cells undergoing lymphopenia-driven proliferation are capable of effector functions similar to those acquired upon Ag stimulation (48).
By comparing the fate of CFSE-labeled CD4+ T cells from WT, Stat4−/− (19), and Stat6−/− (22) BALB/c donor mice, we found that during early replenishment of the host’s peripheral T cell pool, Stat6−/− and WT CD4+ T cells have a shorter time lag to proliferation and divide faster than Stat4−/− CD4+ T cells. Because the ratio of naive to activated memory of the adoptive transferred cells was the same for these different populations, as judged by the level of expression of the surface marker CD45RB (not shown), the faster rate of proliferation of the population of Stat6−/− CD4+ T cells was not due to the presence of an enriched fraction of activated cells. Overall, our findings show that Stat6−/− cells show a proliferation advantage over Stat4−/− CD4+ T cells. This may not be an exclusive property of mature T cells, because it has been shown that Stat6-deficient bone marrow cells also show increased responses to growth factors compared with Stat4-deficient cells (53). The T cell proliferation of WT and Stat6−/− cells was preceded by an early transient phosphorylation of Stat4, as previously described (44, 54). The early activation of Stat4, lacking in Stat4−/− cells, may confer to WT and Stat6−/− CD4+ T cells a faster progression through the cell cycle. Stat4 can control cell proliferation by enhancing c-Myc transcription upon IL-2 stimulation of T cell expansion (55). Stat proteins also regulate p27kip1 expression (35, 37), a negative regulator of cytokine-stimulated T cell growth (43). Most interestingly, in WT and Stat4−/− CD4+ T cells, we observed for the first time a down-regulation of Stat6 protein and mRNA. The down-regulation of Stat6 associated with the simultaneous phosphorylation of Stat4 that occurs in Stat6-deficient and BALB/c cells, but is absent in Stat4-deficient T cells, may release CD4+ T cells from negative signaling, thus conferring an advantage to both Stat6−/− and WT CD4+ T cells.
The early phosphorylation of Stat4, accompanied by an early increase in the transcription of both T-bet and IFN-γ (not shown) indicates that lymphopenia-driven proliferation involves triggering of both TCR and IFN-γ receptors (56) and that the activated CD4+ T cells might be precommitted into a Th1 differentiation pathway. Interestingly, we found that early after T cell transfer, host nonlymphoid spleen cells increase the levels of expression of IFN-γ mRNA (Table II), showing that T cell transfer induces changes in the host environment. The increased production of IFN-γ by the host cells together with the probable existence in the host of basal levels of IL-12, but not IL-4, may contribute to the preferential phosphorylation of Stat4 in the transferred CD4+ T cells. Moreover, in the absence of specific exogenous signals, Stat6 is down-regulated, which probably reinforces T cell differentiation into the Th1 pathway. It is noticeable that in the presence of both IFN-γ receptor and TCR signaling, Stat4−/− T cells also increase the transcription of T-bet and the transcription and production of IFN-γ. However, in contrast to WT cells, Stat4−/− T cells also show a late increase in IL-4 transcription, which is not detected at the protein level. Overall, these findings suggest that Th1 precommitment may be the preferential pathway for WT CD4+ T cell differentiation during lymphopenia-driven proliferation. It would be of interest to generate BALB/c.RAG−/−.IL-12p40−/− mice to eliminate the major Stat4 ligands (57) and compare the patterns of lymphopenia-driven proliferation in these host mice. It should be noted, however, that the levels of mRNA expression for cytokines and transcription factors were 10-fold lower than those observed among fully polarized T cells stimulated in vitro in Th1 and Th2 conditions (Table III). This observation together with the facts that only ∼6–12% of the cells produce IFN-γ and secretion of IL-4 is not detected suggest that most CD4+ T cells did not complete effector function differentiation. At a population level, full commitment may require higher avidity TCR interactions and additional costimulatory signals (9), which are probably lacking in the absence of intentional antigenic stimulation. Only in conditions of optimal activation will T cells undergo the epigenetic changes necessary to complete polarization (39, 56).
It is well documented that during lymphopenia-driven proliferation donor T cells proliferate to eventually reach steady state equilibrium (47). Thus, we found that despite the initial faster rate of proliferation of Stat6−/− cells, CD4+ populations from WT, Stat4−/−, and Stat6−/− donors reach the same plateau, with similar CD4+ T cell numbers being recovered from the hosts 3–5 wk after transfer. At equilibrium, WT, Stat6−/−, and Stat4−/− CD4+ T cells expressed an activated phenotype, and the frequency of CD25+ cells was similar in both Stat6−/− and Stat4−/− donor cells (not shown). These findings imply that all these different CD4+ T cell types expand with similar efficiency, but with different kinetics. They also indicate that the final homeostatic control of T cell numbers is not determined by their rate of proliferation, but, rather, is dependent on the total space available for T cell expansion (58). However, when Stat4−/− and Stat6−/− CD4+ T cells were coinjected into the same hosts, the Stat6−/− CD4+ T cells became dominant and out-competed Stat4−/− CD4+ T cells from the peripheral T cells pools. Stat6−/− cells also out-competed WT CD4+ T cells, whereas WT cells out-competed Stat4−/− CD4+ T cells. The finding that Stat6−/− out-competed WT cells suggests that Stat6 may play a role in regulating T cell expansion. In WT T cells, Stat6 is down-regulated early after T cell transfer, which may facilitate early proliferation, and is later re-expressed, which may slow their rate of division compared with that of Stat6−/− cells. It should be noted that upon transfer of a limited number of cells (103), the number of Stat6−/− T cells recovered on day 25 is higher than that of WT cells (Fig. 4). Comparing the competitive capacities of Stat6−/−Thy1.2, Stat4−/−Thy1.1, and WT CD4+ T cells to that of double-deficient Stat6−/−Stat4−/− T cells, we found that double-deficient Stat6−/−Stat4−/− T cells were out-competed by Sta6−/− T cells, but not by either Stat4−/− or WT T cells (Fig. 6 B). On the whole, these findings suggest that the balance between the enhancing effects of Stat4 and the regulator effects of Stat6 may be critical to the rates of CD4+ T cell proliferation and expansion.
We have also used a competitive reconstitution strategy and compared the fates of Stat6−/− and Stat4−/− T cells developing simultaneously in BM chimeras. In these mice we found that in both thymus and peripheral pools, the relative representation of Stat6−/− and Stat4−/− CD4+ T cells recapitulates the ratio present in BM inoculum. Furthermore, in reconstituted BM chimeras, the representation of both Stat6−/− and Stat4−/− populations was the same among naive CD45RBhigh and CD45RBmed/low activated CD4+ T cells (not shown). Thus, during the replenishment of the periphery through stable thymus output, we did not observe a competitive advantage of Stat6−/− cells. We conclude that the competitive advantage of Stat6−/− CD4+ T cells can only be expressed upon activation and expansion of mature T cells in peripheral pools. These findings suggest that selection into activated T cell pools in steady state or during lymphopenia-driven proliferation may follow different rules. T cell transfer and activation in T cell-deficient mice may change the host environment and cytokine milieu, biasing the selection of cells using the Stat4 pathway. The present results raised the possibility that Stat6−/−CD4+ T cells might have suppressed the expansion of Stat4−/−CD4+ T cells. We have previously shown that CD25+CD4+ Treg cells can selectively inhibit the peripheral expansion of naive CD4+ T cells transferred into T cell-deficient hosts (59). CD25+CD4+ T cells suppressed the expansion of cotransferred naive CD4+ T cells, and total T cell recovery diminished according to the level of suppression; that is, overgrowth of coinjected CD25+CD4+ T cells did not compensate for the lack of expansion of naive T cells (59). However, in mice coinjected with CD4+ T cells from Stat6−/−CD4+ and Stat4−/− donors, the total T cell recovery was the same as that in mice injected with only one population. These observations demonstrate that the fate of Stat4−/− CD4+ T cells at the periphery is altered by cellular rivalry (58) and not by active suppressor mechanisms. We postulate that the faster rate of proliferation of Stat6−/− CD4+ T cells may allow them to occupy most of the peripheral T cell pool and thus, by pre-emptive competition, prevent further expansion of Stat4−/− CD4+ T cells (58). This hypothesis was confirmed by sequential transfer experiments, in which we showed that if transferred 9 days previously, Stat4−/− cells would dominate a second population of Stat6−/− cells. Thus, repopulation of the peripheral pool by lymphopenia-driven Stat-deficient T cells follows the rule “first come, first served.” These findings help to explain why once established, a Th1 or Th2 pattern of T cell response is so difficult to displace. They also demonstrate that minor initial differences in the rates of cell division may have later, large repercussions, e.g., in the establishment of clonal dominance during immune responses (60).
In conclusion, we show in this study that upon activation in the absence of intentional antigenic stimulation during lymphopenia-driven proliferation, Stat6-deficient CD4+ T cells show a competitive advantage that allows them to out-compete Stat4−/− CD4+ T cells. We speculate that in the host environment, T cell activation involving the Stat4 signaling pathway and the down-regulation of Stat6 confers upon the cells an expansion advantage that in a competitive situation has major late repercussions, because it modifies the final homeostatic equilibrium of the populations and favors the establishment of a Th1 response as a default pathway of T cell differentiation. We propose to investigate whether these same rules apply in the course of Ag stimulation by following the behaviors of Stat4- and Stat6-deficient TCR Tg CD4+ T cells with known antigenic specificities.
We thank Dr. Afonso Almeida for critical help, Dr. R. de Boer for helpful discussions, Dr. Caetano Reis e Sousa for anti-IL-12, Anne-Marie Balazuc for cell sorting (Plataforme de Cytométrie, Institut Pasteur), and Marie-Pierre Mailhe for technical help.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from Agence National de Recherche sur le SIDA, Association pour la Recherche sur le Cancer, Ligue contre le Cancer, Centre National de la Recherche Scientifique, the European Community, and the Institut Pasteur (PTR Pasteur/Necker 114). V.S.-G. is supported by the Danish Research Agency (Grant 642-02-0016) and benefits from half a study-loan from Consejo Nacional de Ciencia y Technología.
Abbreviations used in this paper: WT, wild type; BM, bone marrow; LN, lymph node.