Cutting Edge: Nuclear Factor of Activated T Cells and AP-1 Are Insufficient for IL-2 Promoter Activation: Requirement for CD28 Up-Regulation of RE/AP¹

The activation of T cells is critical for the generation of an immune response. Minimally, two signals are required to convert resting T cells into effector cells: one signal through the TCR in response to appropriate Ag bound to a MHC molecule, and a second “costimulatory” signal (1). TCR activation alone may result in a state of unresponsiveness or anergy (2). CD28, a 44-kDa glycoprotein expressed as a homodimer on most T cells, can mediate the costimulatory signal, and its stimulation can prevent anergy (3). In vivo, CD28 binds to one of its ligands, B7-1 and B7-2, on the surface of APC (4). A CD28-related molecule, CTLA-4, can also bind to B7-1 and B7-2, but with a higher affinity than CD28 (4). Recombinant soluble CTLA-4 expressed as an Ig fusion protein, CTLA4Ig, can effectively block CD28-mediated signaling by preventing association of CD28 with its ligands (3).

One important consequence of T cell stimulation is the production of cytokines, including IL-2, due to both transcriptional up-regulation and stabilization of the mRNA (5). T cell activation induces the association of NF of activated T cells (NFAT),³ AP-1, and NF-κB family members to sites within the IL-2 promoter (reviewed in Refs. 6 and 7). Mutagenesis studies from this laboratory have previously revealed that a CD28 response element (CD28RE) located within the IL-2 promoter is necessary for CD28-mediated transcriptional activation (8). However, we and others recently showed that CD28 responsiveness within the IL-2 promoter is actually conferred by a composite element (“RE/AP”) containing both the CD28RE and adjacent NF-IL-2B AP-1 site (9, 10, 11). The CD28RE is a poor consensus NF-κB binding site, which has been shown to associate with NF-κB family members, primarily c-Rel (11, 12, 13). However, it has also been reported that NFAT and not NF-κB family members binds to the CD28RE (14). If NFAT and AP-1 mediate transcriptional activation of the RE/AP site within the IL-2 promoter, then it would be expected that RE/AP should be activated under conditions where both NFAT and AP-1 are up-regulated.

It has been previously shown that activation of the IL-2 promoter in Jurkat T cells in response to superantigen requires both TCR and CD28 signals (15). Here, we show that NFAT and AP-1 activation occurs in Jurkat T cells in response to superantigen independent of CD28 signaling. In contrast, the activation of RE/AP, like the intact IL-2 promoter, is abrogated when CD28 engagement is blocked. Therefore, transcriptional regulation of RE/AP is not mediated by NFAT/AP-1 family members, because RE/AP activation does not occur when both NFAT and AP-1 are up-regulated. In addition, our findings suggest that activation of both NFAT and AP-1 is insufficient to induce IL-2 promoter transcription.

The NFAT, AP-1, and RE/AP luciferase reporters used have been previously described (11, 16). The IL-2 promoter luciferase construct was provided by Gerry Crabtree (Stanford University, Palo Alto, CA). CTLA4Ig was provided by David Daikh and David Wofsy (University of California, San Francisco, CA). Staphylococcal enterotoxin E (SEE) and D (SED) were purchased from Toxin Technology (Sarasota, FL). PMA and ionomycin were purchased from Calbiochem (La Jolla, CA). C305 is an anti-clonotypic Ab recognizing the TCR of Jurkat T cells (17). 9.3 is a mAb specific for human CD28 and was provided by Bristol-Myers-Squibb Pharmaceutical Research Division (Seattle, WA).

Jurkat transfections and all luciferase assays were performed as previously described (11). Transfections into human PBLs were performed as described (18), except that 25 μg of either RE/AP or NFAT luciferase reporter were used. Briefly, PBLs were isolated from whole blood by Ficoll gradient (Sigma, St. Louis, MO) and washed several times. PBLs were seeded at 5 × 10⁶ cells/ml and stimulated with 1 μg/ml PHA (Abbott Laboratories, Chicago, IL) before transfection. Cells were washed and resuspended at 2 × 10⁷ cells/ml. A total of 0.25 ml cells and 25 μg DNA were electroporated in a 0.4-cm gap cuvette at 250 V, 960 μF, using a Bio-Rad gene pulser (Bio-Rad, Richmond, CA).

Jurkat cells were stimulated 20–28 h after transfection with Raji/SED as previously described (15). FK506 was used at a final concentration of 100 ng/ml and was added to the Jurkat cells just before the addition of Raji/SED. CTLA4Ig or control Ab MOPC195 was used at a final concentration of 10 μg/ml and added to Raji cells 30 min before the addition of Jurkat cells and SED. Then, 16 h after stimulation by Raji/SED, the cells were harvested for luciferase assays. Stimulations of Jurkat cells by C305, PMA, and 9.3 were performed as previously described (11).

PBLs were stimulated 2 h after transfections, using 2 × 10⁵ live PBL/sample as determined by trypan blue exclusion and a superantigen mixture containing 100 ng/ml of SEA, SEB, and SEC3 as previously described (15). Raji cells, preincubation with CTLA4Ig and timing of stimulation were performed as described above for Jurkat cells. The data described are the result of three independent transfections. All p values were determined by student’s t test.

Jurkat cells were stimulated as above. After stimulation, the cultures were stained using anti-human CD69-FITC and anti-human CD5-phycoerythrin (Becton Dickinson, San Jose, CA). Jurkat, but not Raji, cells expressed CD5. Cells were gated for expression of CD5 and examined for expression of CD69 using a FACScan (Becton Dickinson). Cell staining was performed as recommended by the manufacturer.

IL-2 promoter activation in Jurkat T cells in response to superantigen requires both stimulation through the TCR as well as CD28 costimulation (15). Jurkat T cells can be activated to secrete IL-2 by the superantigen SED presented by the B cell line Raji (8, 19). We wanted to determine the requirements for IL-2 promoter transcription in response to superantigen by examining the activation of various reporter constructs containing transcriptional elements involved in IL-2 promoter activation, NFAT, AP-1, and RE/AP, and their responses in the presence or absence of CD28 costimulation. Jurkat cells were transiently transfected with luciferase reporters containing either the IL-2 promoter or multimerized copies of RE/AP, NFAT, or AP-1, and were stimulated with Raji cells, SED, or a combination of both. Incubation with either Raji or SED alone was insufficient for activation of any of the reporter constructs (Fig. 1,A). However, activation of all four reporters was induced by the combination of this APC and SED. Additionally, the combination of Raji and SED induced expression of the cell surface activation Ag CD69 on Jurkat cells (Fig. 1 B). Use of a different superantigen, SEE, produced similar effects (data not shown).

To determine the contribution of CD28 to the up-regulation of each of these reporters during this T cell-APC interaction, Raji cells were preincubated with CTLA4Ig before the addition of SED and Jurkat cells to the cultures. CTLA4Ig recognizes the same ligands as CD28, B7-1 and B7-2, but with higher affinity, thereby blocking their association with and subsequent activation of CD28 (3). As shown in Fig. 2,A, CTLA4Ig inhibited the activation of the IL-2 promoter in Jurkat cells stimulated with Raji/SED by >80%. Similarly, CTLA4Ig also inhibited activation of the RE/AP composite element by >90%. No significant effect of CTLA4Ig was observed on either the NFAT or AP-1 reporters (Fig. 2,A) or on the induction of the activation Ag CD69 (Fig. 2,B). Thus, CTLA4Ig did not globally disrupt activation of Jurkat T cells by Raji/SED. Identical amounts of an irrelevant mAb had no effect on the induction of any reporter or CD69 (Fig. 2, A and B). Therefore, CD28 signaling is critical for the activation of the IL-2 promoter in response to superantigen and APC.

Because activation of NFAT and AP-1 was unaffected by the blockade of CD28, it does not appear that these elements are responsible for the defect in IL-2 promoter transcription produced by the block of the CD28-B7 interaction by CTLA4Ig. Moreover, these data show that activation of NFAT and AP-1 is insufficient to activate the IL-2 promoter. However, activation of the RE/AP composite element is dependent on CD28 signaling. Together with our previous studies in which the CD28RE was mutated (8), we conclude that RE/AP is the critical element within the IL-2 promoter that is activated by CD28 and is required for activation of the IL-2 gene by superantigen. In addition, because RE/AP activation was blocked by CTLA4Ig even though NFAT and AP-1 reporters were unaffected, it does not appear that activation of RE/AP is mediated by NFAT/AP-1 as had been previously proposed (14). Rather, these data are consistent with the findings that RE/AP activation is mediated by c-Rel and AP-1 (10, 11). Consistent with a major role for RE/AP activation in IL-2 promoter regulation, the c-Rel knockout mouse is defective in IL-2 production in response to CD3/CD28 stimulation (20).

The findings that RE/AP and NFAT differ in their requirement for CD28 activation were confirmed by transfecting normal human PBLs. As previously reported (15), the transcriptional response of PBLs stimulated by a pool of three staphylococcal enterotoxins presented by Raji cells can be monitored with reporter constructs. Stimulation with enterotoxins presented by Raji B cells activated RE/AP and NFAT reporter constructs by 22- and 23-fold, respectively. RE/AP activation in PBLs stimulated by enterotoxins presented by Raji cells was inhibited by almost two-thirds by the addition of CTLA4Ig to the culture (62.0 ± 1.3%). Significantly less inhibition of the NFAT reporter response was observed when CTLA4Ig was added (30.5 ± 3.3%; p < 0.01). The addition of a control Ab did not affect either RE/AP or NFAT activation (data not shown). Therefore, these findings with normal PBLs support our data obtained with Jurkat cells; RE/AP is not a NFAT site, and it is the major site of CD28 activation in the IL-2 promoter. We should note that the inhibition of RE/AP by CTLA4Ig was not as complete in normal PBLs as was observed in Jurkat cells. This may relate to the prestimulation that must be used to render the PBLs transfectable with the DNA reporter constructs. They must be prestimulated for 20 h with mitogenic concentrations (1 μg/ml) of PHA before transfection. This may alter the requirements for costimulation after transfection. Consistent with this, the background luciferase activity of RE/AP in PBLs is considerably higher than in Jurkat cells (∼6400 in PBLs compared with ∼400 in Jurkat cells). Also, there may be additional costimulatory molecules present and activated on PBLs as compared with Jurkat cells. However, the data obtained with PBLs is consistent with the data in Jurkat cells that RE/AP is different from NFAT and is the major site of CD28 costimulation in the IL-2 promoter.

Activation of RE/AP in response to superantigen required CD28 signaling (Fig. 2,A). Our previous data suggested that RE/AP was a site of signal integration in T cells; any pairwise combination of either TCR, CD28, or PMA can be used to activate RE/AP-dependent transcription (11). TCR-mediated activation of RE/AP is sensitive to inhibition by the calcineurin inhibitor FK506, while CD28-mediated activation of RE/AP is resistence to FK506 (11). To determine whether RE/AP also integrates signals through the TCR in a physiological response to superantigen, FK506 was used in the Raji/SED system. As shown in Fig. 3,A, FK506 blocked activation of RE/AP, NFAT, and the IL-2 promoter in response to Raji/SED almost completely (>95% inhibition). However, activation of AP-1 and the expression of CD69 remained intact (Fig. 3, A and B), indicating that although activation of the IL-2 promoter has been inhibited, Jurkat cell activation in response to superantigen has not been globally interrupted. Therefore, in a response to a physiological stimulus, SED superantigen presented by Raji B cells, the activation of RE/AP minimally requires two signals: one through the TCR as demonstrated by inhibition by FK506 and the second through CD28 as demonstrated by inhibition by CTLA4Ig. It is interesting that CTLA4Ig specifically interrupted only CD28 signals, presumably leaving other adhesive and costimulatory interactions between the Jurkat T and Raji B lymphoblastoid cells during Ag presentation intact, including LFA-1/ICAM, LFA-2/LFA-3, etc. However, it does not appear that these molecules on the surface of Jurkat T cells can compensate for the loss of CD28 signaling. These findings are consistent with studies that demonstrated a central role for CD28 stimulation in IL-2 production (21, 22). However, these data do not preclude the notion that other costimulatory pathways (not present or activated between Raji B cells and Jurkat T cells) exist that can substitute for CD28 engagement for RE/AP and IL-2 promoter activation. Alternate mechanisms for activation of RE/AP are currently under investigation.

We thank Dr. Keith Yamamoto and members of the Weiss laboratory for thoughtful discussion of the data. We also thank Dr. Mike Shapiro for statisical analyses.