Repeated Antigen Exposure Is Necessary for the Differentiation, But Not the Initial Proliferation, of Naive CD4⁺ T Cells

Mice

The 3A9 mice are transgenic for a TCR specific for the hen egg lyzosyme (HEL)³ peptide 46–61 presented by I-A^k (19). The 3A9 mice were maintained on a CBA/J background or, for adoptive transfer, crossed with B10.BR Ly5.1⁺ congenic mice.

Adoptive transfer

(CBA/J × B10.BR)F₁ recipient mice were injected i.v. with 2.5 × 10⁶ naive or in vitro-activated CD4⁺ T cells from the 3A9-transgenic mice. Twenty-four hours before adoptive transfer, the mice were immunized by s.c. injection, in the backs of recipient mice, of 200 μg of native HEL protein (Sigma-Aldrich, St. Louis, MO) emulsified in CFA (Sigma-Aldrich) or CFA alone in a 200-μl volume, as previously described (20).

Abs and FACS staining

The Abs anti-CD4 (RM4-5), anti-IFN-γ (XMG1.2), anti-IL-2 (JES6-5H4), anti-IL-4 (11B11), and anti-Ly5.1 (A20) were purchased from BD PharMingen (San Jose, CA). These Abs were directly coupled to allophycocyanin, PE , FITC , or biotin, in which case staining was revealed using streptavidin-allophycocyanin (BD PharMingen). TO-PRO-3 (Molecular Probes, Eugene, OR) was used at a final concentration of 0.5 nM. For cytokine staining, cells were stimulated for 4 h with 50 ng/ml PMA and 500 ng/ml ionomycin (Sigma-Aldrich), and 10 μg/ml brefeldin A (Sigma-Aldrich) was added for the last 2 h or for the entire 4-h stimulation period when ex vivo T cells were analyzed. Cells were harvested, stained for appropriate surface markers, fixed with 2% paraformaldehyde, and incubated in permeabilization buffer containing the anti-cytokine Ab as previously described (1).

Cell preparation and in vitro activation

CD4⁺ T cells were purified from the lymph nodes of 3A9-transgenic mice by negative selection using a mixture of Ab composed of an anti-I-E^k (M5/114.15.2), an anti-B220 (RA36B2), an anti-CD8 (H-59-101-2), an anti-FcγRII/III (2.4.G2), and an anti-CD11b (M1/70.15.11.5HL) followed by an incubation with sheep anti-rat IgG magnetic beads (Dynal Biotech, Oslo, Norway). On average, the population recovered was composed of >90% CD4⁺ T cells, of which <3% were CD44⁺. The cells were labeled with CFSE (Molecular Probes) as previously described (21). T cell-depleted splenic APCs were prepared by Ab-mediated complement lysis using an anti-Thy1.2 Ab (J1J) and were gamma irradiated at 24 Gy. Naive T cells (5 × 10⁵) were stimulated in 24-well plates with APCs (1 × 10⁶) and 0.6 μg/ml HEL 46–61 peptide. Forty-eight hours later, dividing CFSE⁺CD4⁺ T cells were FACS sorted by excluding cells that were in the most brightest peak (0 division). FACS-sorted populations were >98% pure. Sorted CD4⁺ T cells (7.5 × 10⁴) were seeded either in the upper or lower compartment of a 12-well Transwell plate with a 0.4-μm membrane pore size (Costar, Cambridge, MA) or in a standard 24-well-plate. When indicated, 5 × 10⁵ APCs were added along with 0.6 μg/ml HEL 46–61 peptide. For anti-CD3 stimulation, plates were coated overnight with 10 μg/ml anti-CD3 (145.2C11) and 1 μg/ml soluble anti-CD28 Ab (37.51) was added during the culture period. When indicated, neutralizing anti-IL-2 (2 μg/ml, JES6-1A12; BD PharMingen), anti-IL-12 (5 μg/ml,C17-6, a generous gift from G. Trinchieri, Schering-Plough Research Institute, Dardilly, France), anti-IL-4 (5 μg/ml, 11B11; BD PharMingen), or anti-IFN-γ (5 μg/ml, XMG1.2; BD PharMingen) Ab were added. Recombinant murine cytokines were IL-2 (10 or 50 U/ml; PeproTech, Princeton, NJ), IL-12 (3.5 ng/ml; PeproTech), or IFN-γ (50U/ml; Genzyme, Cambridge, MA).

RNA preparation and RT-PCR

Total RNA was extracted using the High Pure RNA Isolation kit (Roche Diagnostic, Mannheim, Germany) according to the manufacturer’s instructions, treated with DNase I (Roche Diagnostic) and reverse transcribed using random primers and Superscript II RT (Life Technologies, Grand Island, NY). Real-time PCR was performed on cDNA samples using the Taqman Sybr Green system (PE Biosystems, Warrington, U.K.). Primers used were hypoxanthine phosphoribosyltransferase (HPRT) sense 5′-AGCCCTCTGTGTGCTAAGG-3′, HPRT antisense 5′-CTGATAAAATCTACAGTCATAGGAATGGA-3′; T-bet sense 5′-CAACAACCCCTTTGCCAAAG-3′, and T-bet antisense 5′-TCCCCCAAGCAGTTGACAGT-3′. Cycling conditions were 1 cycle at 50°C for 2 min, 1 cycle at 95°C for 10 min, and 40 cycles each corresponding to 15 s at 95°C and 1 min at 60°C. Analysis used the sequence detection software supplied with the instrument. The relative quantitation value is expressed as 2^−ΔCT and percentage of maximal 2^−ΔCT, where ΔC_T is the difference between the mean C_T value of triplicates of the sample and of the endogenous HPRT control.

Role of TCR re-engagement in sustaining T cell division

We analyzed the response of CD4⁺ T cells isolated from the 3A9 TCR-transgenic mice that are specific for the HEL peptide presented by the class II MHC molecule I-A^k (referred to hereafter as 3A9 CD4⁺ T cells (19)). To follow T cell division, we used the vital dye CFSE that is equally partitioned to daughter cells at each division (21). To analyze the role of TCR re-engagement or cytokine signaling in further driving the division and differentiation of cycling CD4⁺ T cells, we stimulated the 3A9 CD4⁺ T cells for 48 h and FACS sorted the 3A9 CD4⁺ T cells that had accomplished at least one division. At this time point, on average, 90% of the CD4⁺ T cells had achieved one to three divisions (data not shown) and did not produce significant levels of cytokines such as IL-2, IL-4, or IFN-γ without restimulation (Fig. 1⇓). Most of these activated 3A9 CD4⁺ T cells are undifferentiated since they produce IL-2 but little IFN-γ or IL-4 following restimulation with PMA and ionomycin or upon TCR re-engagement (Fig. 1⇓, b and d). We first determined whether TCR re-engagement and cytokine signaling were necessary to sustain cycling of activated T cells. To mimic as closely as possible the Ag and cytokine environment that may occur during an in vivo response, we used a Transwell system with a membrane pore size of 0.4 μm that allows the passage of soluble factors such as cytokines but not cells. Sorted CD4⁺ T cells were seeded in the upper and lower chambers of the Transwell while APCs and Ag were only added to the upper chamber. The upper well thus contained CD4⁺ T cells that had re-engaged their TCR and, as a consequence of this TCR stimulation, may produce cytokines available for themselves and for the CD4⁺ T cells present in the lower well that did not re-engage their TCR. As shown in Fig. 2⇓a, CD4⁺ T cells present in the upper and lower chambers of the Transwell divided equally within the 48 h of reculture. The continuation of cell division in the lower chamber did not result from carryover of Ag-loaded APC since the CD4⁺ T cells were FACS sorted to >98% purity. Furthermore, similar results were obtained when the upper and lower wells contained CD4⁺ T cells recognizing distinct MHC-peptide complexes (data not shown). Finally, naive 3A9 CD4⁺ T cells when added to the lower chamber were not activated nor did they divide (data not shown). These results indicate that although entry into the cell cycle requires TCR engagement, further progression through cell division, during the 48 h of analysis, does not require TCR re-engagement.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Cytokine production by 48-h activated 3A9 CD4⁺ T cells. CFSE-labeled 3A9 CD4⁺ T cells were stimulated with APCs plus HEL peptide for 48 h, and dividing CD4⁺ T cells were FACS sorted based on their CFSE profile. The ability of these activated cells to produce IL-2, IL-4, and IFN-γ either spontaneously (a), after 4-h stimulation with PMA/ionomycin (b), after 8-h incubation with IL-2, IL-12, and IFN-γ (c), or after 8-h stimulation with APC and HEL peptide (d) was determined. Similar results were obtained in three independent experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Neither TCR re-engagement nor cytokines are necessary to sustain the proliferation of activated CD4⁺ T cells. CFSE-labeled 3A9 CD4⁺ T cells were stimulated with APCs plus HEL peptide for 48 h, and dividing CD4⁺ T cells, having accomplished one to three divisions, were FACS sorted (sorted cells). a, Dividing cells were reincubated in the two compartments of a Transwell plate containing APCs plus HEL peptide in the upper well only. When indicated, an anti-IL-2-blocking Ab was added to the culture medium. b, Dividing CD4⁺ T cells were reincubated in medium alone or with APCs either alone or with HEL peptide or with 10 U/ml IL-2 as indicated. The CFSE profile of CD4⁺ T cells was analyzed at 24 h (open histograms) or 48 h (filled histograms) of secondary culture. The interval between the two dotted lines corresponds to one to three subsequent divisions in addition to the three divisions accomplished by the 48-h activated T cells. Viability was determined by TO-PRO-3 staining and the percentage of viable cells corresponding to TO-PRO-3-negative CD4⁺ T cells at 48 h after reincubation is indicated. Similar results were obtained in three independent experiments.

We further determined whether subsequent cycling of activated T cells was dependent on the production, by the restimulated T cells, of cytokine such as IL-2 that is produced following TCR re-engagement (Fig. 1⇑d). Anti-IL-2 blocking Ab was therefore added to the Transwell system. Cycling of CD4⁺ T cells in the upper chamber of the Transwell was similarly sustained whether IL-2 was present or not during the reincubation period, indicating that cycling of cells induced by TCR re-engagement may be IL-2 independent (Fig. 2⇑a). Although continuous IL-2 signaling was not necessary to sustain the proliferation of activated 3A9 CD4⁺ T cells, IL-2 greatly contributed to the survival of these activated T cells in vitro (Fig. 2⇑a). Indeed, 73% of the cycling CD4⁺ T cells died when restimulated in the absence of IL-2 as evidenced by TO-PRO-3 uptake, a dye that specifically stains dead cells. 3A9 CD4⁺ T cells present in the lower well also progressed through further division whether IL-2 was present or not (Fig. 2⇑a). Indeed, in the absence of IL-2, the 48-h activated CD4⁺ T cells made an additional three divisions within 24 h (Fig. 2⇑a). Past this 24-h time period, when deprived of IL-2, the CD4⁺ T cells present in the lower well did not further divide and most of them died (Fig. 2⇑a). We further tested that no other cytokines, such as IL-4, produced by activated T cells or IL-7 or -15 produced by the APC present in the upper well, may contribute to cycling of the CD4⁺T cells present in the lower chamber. We reincubated sorted cycling CD4⁺ T cells in medium alone, a condition where none of these cytokines are produced or, as control, in combination with syngeneic APCs with or without IL-2 or HEL peptide. Activated CD4⁺ T cells further divided in all four culture conditions (Fig. 2⇑b). As observed in the Transwell system, activated 3A9 CD4⁺ T cells made three to four additional divisions within 24 h regardless whether cytokines or Ag were added or not during the restimulation period. Past this 24-h period, 3A9 CD4⁺ T cells maintained in medium alone stopped cycling and most of them died, in contrast to cells reincubated with IL-2 or Ag-loaded APCs that continued to divide (Fig. 2⇑b).

Collectively, these results indicate that once into cycle, CD4⁺ T cells proceed through several rounds of divisions in the absence of subsequent TCR and cytokine signaling. We further determined whether in vivo cycling of activated CD4⁺ T cells also proceeded in an Ag-independent fashion. For these experiments, CFSE-labeled 3A9 CD4⁺ T cells expressing the Ly5.1 allotype were activated for 48 h in vitro, purified to remove any Ag-loaded APCs, and then transferred into syngenic mice expressing the Ly5.2 allele. Twenty-four hours before adoptive transfer, the recipient mice were injected s.c. with CFA with or without HEL, or left unimmunized. Activated T cells further divided when transferred into unimmunized animals (Fig. 3⇓a). Under those conditions, they performed four additional divisions within 48 h and some of them further divided. As observed in vitro, most activated 3A9 CD4⁺ T cells stopped dividing after 48-h posttransfer into unimmunized hosts, although under those in vivo conditions the CD4⁺ T cells did not die since the number of 3A9 CD4⁺ T cells recovered at 48 and 72 h were comparable (Fig. 3⇓b). Cell cycle arrest in this case is therefore not resulting from starvation or death of the CD4⁺ T cells but instead reflects the extent of the programming ofproliferation induced by initial Ag encounter. To ensure that the proliferation of activated 3A9 CD4⁺ T cells when transferred into unimmunized hosts occurred in the absence of Ag, we cotransferred naive 3A9 CD4⁺ T cells along with purified activated 3A9 CD4⁺T cells. In this case, the 3A9 CD4⁺ T cells expressing the host Ly5.2 molecule were stimulated in vitro and cotransferred with CFSE-labeled naive 3A9 CD4⁺ T cells expressing the Ly5.1 allele. The proliferation of naive 3A9 CD4⁺ T cells that served as sensors for the presence of residual Ag was then evaluated. As shown in Fig. 3⇓c, no proliferation of the sensor cells was detected over the 3-day period of analysis, indicating that indeed, proliferation of the 3A9 CD4⁺ T cells transferred into unimmunized hosts proceeded in the absence of Ag.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Repeated exposure to Ag is not necessary to sustain the division of activated CD4⁺ T cells but is required for the differentiation of effector Th1 cells. CFSE-labeled 3A9 Ly5.1⁺CD4⁺ T cells were activated for 48 h as described in Fig. 1⇑ legend. Activated CD4⁺ T cells were purified and adoptively transferred into syngenic Ly5.2 recipient mice that were left unimmunized or injected 24 h before transfer with CFA alone or CFA/HEL as indicated. Two and 3 days later, the draining lymph nodes were removed and the CFSE profile (a) and the absolute number (b) or IFN-γ production at 85 h after a 4-h stimulation with PMA and ionomycin (d) of the recovered CD4⁺Ly5.1⁺ T cells were analyzed. d, IFN-γ production by host CD4⁺ T cells was also analyzed (endogenous). c, Naive CFSE-labeled 3A9 Ly5.1⁺CD4⁺ T cells (sensors) were transferred into syngenic Ly5.2⁺ recipient mice either alone or along with dividing 3A9 Ly5.2⁺CD4⁺ T cells. The CFSE profile of the Ly5.1⁺CD4⁺ T sensor cells was analyzed 72 h after transfer. One experiment of two performed with two mice per group is presented.

Transient exposure to Ag therefore induces a program of cell division that further develops in an Ag- and cytokine-independent fashion. This programmed proliferation is however limited, lasting for only 24–48 h and allowing only three to four additional divisions. Division can however be reinitiated by restimulation with Ag-loaded APC.

Repeated Ag exposure is necessary for the differentiation of Th1 cells

Having shown that TCR re-engagement was not required to sustain the initial phase of proliferation of activated T cells, we wished to determine whether the same applied for the differentiation of activated T cells into IFN-γ-producing Th1 cells. We adoptively transferred 48-h activated 3A9 CD4⁺ T cells into mice injected 24 h previously with CFA with or without HEL, as described above. The ability of the 3A9 CD4⁺ T cells to produce IFN-γ following transfer was determined by intracellular FACS staining upon 4-h in vitro stimulation with PMA and ionomycin. Importantly, the 48-h activated 3A9 CD4⁺ T cells used for the transfer are mainly undifferentiated, most of them producing no IFN-γ (Fig. 1⇑). The 3A9⁺CD4⁺ T cells that were transferred into mice previously immunized with HEL further differentiated into IFN-γ-producing Th1 cells (Fig. 3⇑d). In contrast, 3A9 CD4⁺ T cells transferred in the absence of Ag did not differentiate into IFN-γ-producing cells (Fig. 3⇑d).

Altogether, these results indicate that, in contrast to the early burst of proliferation, the full differentiation of effector Th1 cells requires repeated contacts with Ag-loaded APCs.

Sustained TCR and cytokine signaling are necessary for optimal differentiation of effector Th1 cells

The above experiment indicated that the differentiation of CD4⁺ T cells into Th1 effector cells required repeated contact with Ag-loaded APCs. Since both TCR engagement and IL-12 signaling may result from this interaction, it was essential to determine which of those signals was critical for the further differentiation of activated CD4⁺ T cells. To address this question, we used the Transwell system described above. Forty-eight-hour activated CD4⁺ T cells were sorted and reincubated in the upper and lower chambers of the Transwell system. Forty-eight hours later, the fraction of IFN-γ- and IL-2-producing cells was determined after a 4-h stimulation with PMA and ionomycin. Since the transcription of the IL-2 locus is progressively lost as CD4⁺ T cells differentiate into IFN-γ-producing Th1 cells (22), we could then more precisely evaluate the differentiation of the distinct CD4⁺ T cells populations. Importantly, cell division was identical under all culture conditions (Fig. 2⇑a and data not shown). As observed in vivo, efficient Th1 differentiation was only observed for cells that re-engaged their TCR. Indeed, 47 ± 3.5% (n = 3) of the 3A9 CD4⁺ T cells in the upper well further differentiated into effector Th1 cells producing IFN-γ but not IL-2 (Fig. 4⇓Ab). In contrast, only 16 ± 4.7% (n = 3) of the 3A9 CD4⁺ T cells in the lower well were fully mature Th1 cells producing IFN-γ but not IL-2 and 14 ± 4.2% (n = 3) remained undifferentiated producing IL-2 and IFN-γ (Fig. 4⇓Ac). Similar results were obtained with higher doses of HEL peptide, indicating that the poor differentiation of the 3A9 CD4⁺ T cells in the lower well did not result from suboptimal stimulation of the naive CD4⁺ T cells (data not shown). In addition, the limited differentiation of 3A9 CD4⁺ T cells present in the lower well into mature Th1 cells was not due to low levels of IL-12 or IFN-γ, as similar profiles where obtained when cycling CD4⁺ T cells were reincubated with large amounts of these cytokines (Fig. 4⇓Cb). IL-12 and IFN-γ were nonetheless essential for the limited maturation of these cells since their differentiation was almost completely abolished when anti-IL-12 and anti-IFN-γ Ab were added during the 48-h reincubation period (Fig. 4⇓A, compare a, c, and e). The differentiation of the 3A9 CD4⁺ T cells in the upper compartment was also affected by blocking IL-12 and IFN-γ (Fig. 4⇓Ad). Indeed, under those conditions, 25 ± 9.1% (n = 3) of the 3A9 CD4⁺ T cells differentiated into effector Th1 cells as compared with 47 ± 3.5% when IL-12 and IFN-γ were present (Fig. 4⇓A, compare b and d). Importantly, however, under those conditions, very few cells had an immature IL-2⁺IFN-γ⁺ phenotype. To ensure that under those conditions IL-12 signaling did not occur, we restimulated the 48-h activated 3A9 CD4⁺ T cells with plate-bound anti-CD3, in the absence of APCs that may produce IL-12 and further added anti-IL-12 and anti-IFN-γ blocking Ab. Under those conditions, differentiation of the 3A9 CD4⁺ T cells into mature Th1 cells also occurred, although less efficiently than when cytokines were also present (Fig. 4⇓C, c and d). As expected, in the total absence of IL-12 and IFN-γ, that is when anti-IL12 and anti-IFN-γ blocking Ab were added in both the primary and secondary stimulation periods, no IFN-γ-producing cells developed (Fig. 4⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Optimal differentiation of naive T cells into effector Th1 cells requires TCR re-engagement and continuous exposure to cytokines. CFSE-labeled 3A9 CD4⁺ T cells were stimulated as described in Fig. 1⇑ legend. Forty-eight hours later, dividing CD4⁺ T cells having accomplished at least one division were FACS sorted. A and B, Sorted cells were reincubated for 48 h in the two compartments of a Transwell plate containing APCs and HEL peptide in the upper well only. Cultures were performed in the presence of anti-IL-4 Ab alone or along with anti-IL-12 and anti-IFN-γ Ab as indicated. C, Sorted cells were reincubated in the presence or absence of plate-bound anti-CD3 Ab as indicated. The medium was supplemented with a combination of IL-2 (50 U/ml), IL-12 (3.5 ng/ml), IFN-γ (50 U/ml), anti-IL-4, anti-IL-12, and anti-IFN-γ Ab as indicated. The production of IL-2 and IFN-γ was revealed by intracellular FACS staining after a 4-h stimulation with PMA and ionomycin. The numbers in each quadrant correspond to the percentage of cells producing IL-2 and IFN-γ or both, expressed as the mean percentage of three independent experiments.

These results confirm our in vivo study showing that differentiation of effector Th1 cells is a two-step process requiring multiple contact with Ag-loaded APCs. They further suggest that IL-12/IFN-γ and TCR signaling have complementary and synergistic roles in driving the full differentiation of effector Th1 cells. Signaling through the TCR seems however more efficient in driving full differentiation of effector Th1 cells that produce IFN-γ but not IL-2. Interestingly, it was recently shown that the transcription factor T-bet was essential for Th1 cell differentiation and could induce IFN-γ gene expression and repress IL-2 gene expression (14). We therefore analyzed T-bet expression in the different populations showing distinct cytokine profile.

Sustained high level of T-bet gene expression correlates with Th1 differentiation

T-bet gene expression was measured by quantitative RT-PCR in the different culture conditions described above. The level of expression of T-bet mRNA was very low in naive 3A9 CD4⁺ T cells and was up-regulated by a 48-h stimulation in the presence but not in the absence of IL-12 and IFN-γ, suggesting that induction of T-bet gene expression in naive CD4⁺ T cells requires both TCR and cytokine signaling (Fig. 5⇓). Sustained high levels of T-bet mRNA expression correlated with full differentiation of Th1 effector cells (Fig. 5⇓). Indeed, cycling CD4⁺ T cells restimulated with anti-CD3 in the presence of IL-12 and IFN-γ that did efficiently differentiate into effector Th1 cells express a high level of T-bet mRNA. Likewise, cells restimulated with anti-CD3 in the absence of cytokines, of which 33% produce IFN-γ also express high levels of T-bet mRNA. In contrast, 3A9⁺ T cells that remained undifferentiated, i.e., when maintained for an additional 48 h in medium containing or not polarizing cytokines, had reduced levels of T-bet mRNA expression. In addition, this study shows that both TCR and cytokine signaling are necessary for the induction of T-bet expression in naive T cells. In committed CD4⁺ T cells however, expression of T-bet can be maintained by TCR signaling only.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

The efficiency of Th1 differentiation correlates with the level of T-bet mRNA expression. CFSE-labeled 3A9 CD4⁺ T cells were stimulated as described in Fig. 1⇑ legend in the presence (48 h: TCR) or absence (48 h: TCR + cytokines) of anti-IL-12 and anti-IFN-γ Ab. Forty-eight hours later, dividing CD4⁺ T cells were FACS sorted based on their CFSE profile. Sorted cells were reincubated for 48 h with IL-2 only (96 h), with IL-2/IL-12/IFN-γ (96 h: cytokines), with anti-CD3, anti-IL-12, and anti-IFN-γ Ab (96 h: TCR) or anti-CD3 + IL-12 + IFN-γ (96 h: TCR + cytokines). RNA was extracted from the different populations and the level of T-bet mRNA expression was determined by quantitative RT-PCR. The relative mean value (left axis) of three independent measurements was calculated as indicated in Materials and Methods. In addition, the relative value expressed as percentage of the maximal value (right axis) is also presented. The percentage of IFN-γ-producing cells in the different populations of cells analyzed at 96 h is indicated. One experiment of two with similar results is presented.