NF-κB transcription factors play an important role in the activation of the IL-2 gene in response to TCR ligation. The release of NF-κB factors to the nucleus requires phosphorylation and degradation of the inhibitory κ-B proteins (IκBs). IκBα and IκBβ phosphorylation is dependent on dual signaling by the TCR and the CD28 accessory receptor. This pathway involves a multisubunit IκB kinase (IKK) complex, which includes the IKKα (IKK-1) and IKKβ (IKK-2) kinases. We demonstrate that stimulation of primary human CD4+ T cells by CD3/CD28 activates two distinct endogenous IKK complexes, a heterodimeric IKKα/β and a homodimeric IKKβ complex. IKKβ overexpression in a Jurkat cell line resulted in the formation of a constitutively active IKK complex, which was CD3/CD28 inducible. In contrast, ectopic expression of IKKα assembled into a complex with negligible IκB kinase activity. Moreover, IKKβ, but not IKKα, overexpression enhanced transcriptional activation of the CD28 response element in the IL-2 promoter. Conversely, only kinase-inactive IKKβ interfered in the activation of the IL-2 promoter. Sodium salicylate, an inhibitor of IKKβ, but not IKKα, activity, inhibited IL-2 promoter activation as well as IL-2 secretion and interfered in activation of both the heterodimeric as well as the homodimeric IKK complexes in primary CD4+ T cells. Taken together, these data demonstrate the presence of an IKKβ-mediated signaling pathway that is activated by TCR and CD28 coligation and regulates IL-2 promoter activity.
The outcome of TCR stimulation is regulated by the simultaneous engagement of costimulatory molecules, including the CD28 receptor (1, 2). Engagement of CD28 by its ligands, CD80 or CD86, which are expressed on APCs, promotes T cell proliferation and enhances the production of various cytokines and chemokines (3, 4, 5). Stimulation of naive T cells in the absence of CD28 coligation may lead to a state of nonresponsiveness, known as anergy (6). One of the prominent features of anergic T cells is their inability to produce IL-2; CD28 plays an important role in preventing anergy induction through transcriptional activation of the IL-2 gene (6). Mutational analysis of the minimal IL-2 promoter has revealed the presence of a CD28 response element (CD28RE)3 located at −160 to −150 relative to the start site (7, 8). CD28 cooperates with TCR/CD3 in the activation of AP-1 and NF-κB transcription factors, which have cognate binding sites in the CD28RE domain (9, 10, 11).
The NF-κB pathway, in addition to its role in TCR activation, also plays a role in T cell responsiveness to TNF-α, IL-1, and phorbol esters (12, 13). In resting cells, NF-κB transcription factors are sequestered in the cytoplasm by a group of inhibitory proteins known as the IκBs (12, 13, 14). Upon stimulation, IκBα and/or IκBβ are phosphorylated on specific serine residues and are subsequently degraded through the 26S proteosome (12, 13, 14, 15, 16). This allows the release and nuclear localization of NF-κB transcription factors. The phosphorylation of IκBs is mediated by a recently identified multicomponent signalsome, known as the IκB kinase (IKK) complex (16, 17, 18, 19, 20, 21, 22). This complex contains two serine kinase subunits, IKKα (IKK-1) and IKKβ (IKK-2), with relative molecular masses of 85 and 87 kDa, respectively (17, 18, 19, 20, 21, 22). IKKα and IKKβ genes have been cloned, and their expressed proteins have been purified (23, 24, 25). IKKα and IKKβ, which are nearly 50% homologous, are able to homo- and heterodimerize in vitro and in vivo through their C-terminal leucine zipper motifs (23, 24, 25). Although the active kinase complex purified from TNF-α- and IL-1-stimulated cells contains IKKα and IKKβ heterodimers, homodimerized versions of each IKK are capable of phosphorylating IκB proteins in vitro (23, 24, 25). Interestingly, purified IKKβ has substantially higher basal IκBα kinase activity than IKKα (23, 25). Furthermore, in transient transfection studies, expression of a kinase-inactive IKKα mutant has minimal effects on the kinase activity compared with the potent inhibitory effect of kinase-inactive IKKβ (19). These findings raise the possibility that, in addition to IKK heterodimers, homodimers may contribute to NF-κB signaling in vivo. Mercurio et al. have recently demonstrated the existence of heterodimeric IKKα/IKKβ as well as homodimeric IKKβ complexes that can be activated by TNF-α stimulation in HeLa cells (23). In contrast, SLB cells contained IKKα/IKKβ heterodimers but no homodimers (23). This suggests the existence of heterogeneous IKK complexes that may play differential roles in NF-κB activation in different cell types. The possible existence of homo- or heterodimeric IKK complexes in T cells has not been addressed.
Activation of the IKK complex is dependent on the action of upstream kinases (9, 18, 19, 20, 21, 22, 25, 26, 27, 28). NF-κB-inducing kinase (NIK) and MEKK1, members of the mitogen-activating protein kinase kinase kinase (MAP3K) family, have been shown to activate the IKK complex (9, 20, 21, 25, 26, 27, 28, 29). NIK copurifies with the IKK complex and has been shown to be involved in IKK activation by IL-1 and TNF-α (20, 21, 26, 27). Although overexpression of NIK leads to stimulation of IKKα and IKKβ activities in vivo, NIK preferentially phosphorylates IKKα and has little activity on IKKβ (26). At least one study has suggested that NIK is involved in CD28 costimulation (27). While MEKK1 also copurifies with IKK complexes (18), its overexpression leads to differential activation of IKKβ in Jurkat and other cell types (9, 29). HTLV-I tax protein also activates the NF-κB complex by interacting with MEKK1 and preferentially activating IKKβ kinase activity (30). However, in some cell types, MEKK1 has been linked to both IKKα and IKKβ activation (25, 31). Taken together, it is possible that several MAP3Ks may be involved in the activation of IKK complexes in T cells.
We were interested in determining whether signaling by TCR/CD3 and CD28 activates different IKK complexes in primary human T lymphocytes. In addition, we were interested in whether there is a functional relationship between IKKs and in the activation of the CD28RE in the IL-2 promoter. Here we describe that CD3 and CD28 costimulation results in the activation of IKKα/IKKβ heterodimeric and IKKβ homodimeric complexes in primary human CD4+ T cells. We also provide evidence that IKKβ, rather than IKKα, is critical for activation of the CD28RE in the IL-2 promoter.
Materials and Methods
OKT3 (anti-CD3) was obtained from Ortho Pharmaceuticals (Raritan, NJ), and the 9.3 mAb (anti-CD28) were provided by Bristol-Meyer Squibb (Princeton, NJ). The primary stimulating Abs were cross-linked with mAb 187.1 (Bristol-Meyer Squibb). For Western blotting and immunoprecipitation experiments, polyclonal anti-IKKα, anti-IKKβ, anti-IκBα, and anti-IκBβ Abs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Flag (M2) Abs were obtained from Sigma (St. Louis, MO). Anti-IKKγ (IKAP1) Abs and GST-IκBα1–54 have been described by us (23). HRP conjugated to protein A was obtained from Amersham (Arlington Heights, IL). Tosyl-activated magnetic beads and M-450 anti-CD4 beads were purchased from Dynal (Great Neck, NY). PMA and ionomycin were purchased from Sigma.
Jurkat T cells (clones BMS2, Jurkat-hIL-2-Luc) were grown in RPMI medium, supplemented with 10 mM HEPES (pH 7.4), 10% FCS, 2 mM glutamine, 100 U of penicillin, and 100 μg of streptomycin/ml. The Jurkat-hIL-2-luciferase cell line was provided by Dr. A. Weiss (Howard Hughes Medical Institute, University of California, San Francisco, CA), while Jurkat BMS2 was a gift from Dr. B. Mittler (Bristol Myers Squibb) (32).
Preparation of primary CD4+ T cells
Mononuclear cells were isolated from human peripheral blood by density centrifugation and depletion of adherent cells on plastic culture dishes. CD4+ T cells were positively selected with anti-CD4 Dynabeads according to the manufacturer’s instructions (Dynal). This yielded a T cell subset that was >98% positive for the CD4 marker as determined by dual-color CD4/CD8 flow cytometry. Isolated CD4+ T cells were stimulated with anti-CD3 and anti-CD28 mAb coupled to tosyl-activated magnetic beads as previously described (21). Cells were replenished with 40 U/ml rIL-2 (Chiron, Emeryville, CA) on day 3 and allowed to complete their growth cycle over the course of 12–13 days. At this point, IL-2 and beads were removed, and cells were allowed to return to their resting state over a 48-h period. Resting vs activated state was assessed by cell size analysis in a Coulter Counter (Hialeah, FL). Cells were subjected to flow cytometry for a second time to confirm that >95% of the cells remained CD4+.
Gene constructs and cellular transfection
IκB kinase constructs were previously described and characterized (21). Flag-tagged wild-type IKKα and IKKβ and kinase-inactive mutants (K44A) cloned in pRK5 vector were provided by Dr. M. Rothe (Tularik, San Francisco, CA) (21). Full-length kinase-inactive MEKK1 (K1253 to M), a gift from Dr. M. Karin (University of California, San Diego, CA) (33), was cloned in pcDNA3 (Invitrogen, Carlsbad, CA). The JNK-interacting protein 1 (JIP-1), cloned in CMV5 vector (34), was a gift from Dr. R. Davis (Howard Hughes Medical Institute, Worcester, MA) (33). The CD28RE/AP-1 luciferase reporter and related mutants, provided by Dr. A. Weiss (Howard Hughes Medical Institute) were cloned in pAEODLO vectors (8). CD28RE/AP-1 Luc contains four tandem copies of the sequence 5′-tttaaagaaattccaaagagtcatca-3′, which is situated 150–160 bp upstream from the start site of the IL-2 gene (8). For cellular transfection, we used the stated amounts of an individual or combination of plasmids for electroporation (240 V, 950 μF) into 1 × 107 Jurkat cells as previously described (9, 35).
Purification of heterogeneous IKK complexes by sequential immunoprecipitation and Western blot analysis
CD4+ T cells (2 × 107) were lysed in kinase lysis buffer and cleared by centrifugation as previously described (23). For immunoprecipitation, 200 μg of precleared lysate was treated with 2 μg of anti-IKKα Abs, bound to protein A-Sepharose, and rocked for 2 h at 4°C (9, 17, 23). To completely remove all IKKα, this step was repeated once. The remaining supernatants were subjected to further immunoprecipitation using 2 μg of the antisera to either IKKβ or IKKγ. Immunoprecipitated complexes were washed extensively and subjected to immunoblotting as previously described (9, 23). IKKα, IKKβ, and IKKγ Abs were used at a 1/1000 dilution for primary staining. Protein A-conjugated HRP was used at a 1/2000 dilution, and blots were developed by enhanced chemiluminescence. Western blotting of cellular lysates was performed as described previously (9).
Immune complex kinase assays
Jurkat cells or CD4+ resting T cells (1 × 107) in 1 ml of RPMI were left unstimulated or were stimulated with 2 μg/ml anti-CD3 or a combination of 2 μg/ml anti-CD3 and 2 μg/ml anti-CD28 mAb, secondarily cross-linked with 10 μg/ml mAb 187.1. In another group of experiments, primary human CD4+ T cells were treated with 20 mM sodium salicylate (36) for 2 h before stimulation as described above. All kinase assays were performed as described previously (23). Briefly, after cell lysates were precleared with protein A-Sepharose beads, 200 μg of lysate was treated with 2 μg of the specific anti-IKK protein A-Sepharose for 2 h. In a third variation, the immunoprecipitation was performed sequentially, first by adding anti-IKKα in two rounds, and then adding either anti-IKKβ or anti-IKKγ to the remaining supernatants (23). Immune complexes were washed and equilibrated in kinase buffer as previously described (23). Kinase reactions were initiated by the addition of 10 μCi of [γ-32P]ATP and 3 μg GST-IκB1–54 substrate. The reaction was conducted for 30 min at 30°C. Products were analyzed on SDS-PAGE and detected by autoradiography.
Twelve micrograms of the indicated reporter gene constructs were transiently transfected into 107 Jurkat cells (9). The cells were rested for 24 h and then stimulated for 6 h with 2 μg/ml anti-CD3, a combination of 2 μg/ml anti-CD3 and 2 μg/ml anti-CD28 mAb, or a combination of 100 nM PMA and 1 μg/ml ionomycin (PMA+I). The cells were washed and lysed in luciferase buffer (Analytical Luminescence, Ann Arbor, MI), and luciferase activity was measured in 50 μg of lysate in a Monolight 2010 luminometer (Analytical Luminescence). Transfection efficiency was monitored by cotransfection of a β-galactosidase plasmid (CMV-β-gal); β-galactosidase activity was used for adjusting luciferase values among cell populations transfected with different vector combinations (9).
Jurkat T cells (1 × 106) in 2 ml of RPMI were treated with either anti-CD3 or anti-CD3 plus anti-CD28 mAb in the presence or the absence of 20 mM sodium salicylate for 24 h. Cells were removed by centrifugation, and the supernatants were collected. Triplicate aliquots were analyzed by ELISA for the presence of IL-2 (UMAB Cytokine Core Laboratory, Baltimore, MD).
Electrophoretic mobility shift assays
Nuclear protein was extracted from 107 Jurkat T cells, harvested at the indicated time to evaluate DNA binding of NF-κB as previously described (37). Briefly, after washing, the cell pellet was suspended in 1 ml of buffer A (10 mM HEPES (pH 7.9), 1.5 mM MgCl2, 10 mM KCl, and 1 mM DTT) containing 0.1% Triton X-100. After incubating for 10 min on ice, the lysates were centrifuged, and the nuclei were resuspended in 20–40 μl of buffer C (20 mM HEPES (pH 7.9), 25% glycerol, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, and 1 mM DTT). This suspension was incubated for 30 min on ice followed by centrifugation at 10,000 × g for 20 min. Double-stranded oligonucleotides containing a consensus NF-κB sequence (Promega, Madison, WI) were end labeled with T4 polynucleotide kinase in the presence of [γ-32P]dATP (Amersham). The DNA binding reaction was performed at room temperature for 30 min in a final volume of 15 μl, which contained 3–5 μg of nuclear extract, oligonucleotide probe (40 fmol), and binding buffer containing 100 μg/ml poly(dI-dC) as nonspecific competitor. Reactions were subjected to electrophoresis on nondenaturing 5% polyacrylamide gels in 0.5× TBE at 125 mA for 4 h at 4°C.
CD28 coligation leads to CD3-induced activation of IKK complexes in primary and Jurkat T cells
CD28 is an important costimulatory receptor that regulates cellular activation by the TCR/CD3 complex (1, 2, 3, 4, 5). While CD28 can independently signal T cells (38), this receptor also participates in signaling events that require TCR/CD3 coligation. This has been best demonstrated in the activation of the N-terminal c-Jun kinase (JNK) cascade, which regulates transcriptional activation of AP-1 response elements in the IL-2 promoter (9, 35, 39). More recently, CD28 has been shown to contribute to the activation of a composite CD28RE in the IL-2 promoter (8). Since NF-κB signaling is mediated through a multicomponent IκB kinase (IKK) complex, we wanted to determine whether immunoprecipitation of IKKα or IKKβ shows dual receptor requirement for the activation of IKK complexes. In vitro kinase assays using GST-IκBα as a substrate showed CD3- and CD28-inducible kinase activity in IKKα and IKKβ immunoprecipitates obtained from Jurkat T cells (Fig. 1,A). In contrast, ligation of CD3 alone had no effect on IKK activation, indicating that CD28 cooperates with CD3 in IKK activation. P+I treatment had the same effect as dual receptor ligation (Fig. 1 A). Similar cooperation between PMA and anti-CD28 mAb in IKK activation in Jurkat T cells has been demonstrated by Harhaj and Sun (11).
IKK activation results in phosphorylation of IκBα and IκBβ (12, 13, 14, 15, 16), which, in turn, leads to ubiquitination and proteolytic degradation of these proteins and release of NF-κB transcription factors into the nucleus (12, 13, 14, 15, 16). While recombinant IκBα as well as IκBβ were substrates for CD3- and CD28-inducible IKK complexes, the former substrate was more robustly phosphorylated than the latter in vitro kinase assays (Fig. 1,B, top two panels). This finding agrees with previous studies that showed that IκBα is a better substrate for the IKKs than IκBβ (17). The kinetics of IκBα and IκBβ phosphorylation were coincident with the degradation of endogenous IκBα as well as the appearance of DNA-binding NF-κB complexes in the nucleus at approximately 15–30 min (Fig. 1,B, bottom three panels). Although de novo IκBα expression was detected at 60 min (due to the transcriptional activation of its gene by NF-κB), the intranuclear translocation of NF-κB factors still continues at this time point (Fig. 1,B, bottom panel). Since IκBβ degradation was still seen at 60 min, IκBβ probably plays a role in prolonged NF-κB activation in T cells. Taken together with the findings in Fig. 1 A, these data show cooperation between CD3 and CD28 in activating IKK kinases and NF-κB transcriptional activity in T cells.
To investigate IKK activation in primary human T lymphocytes, we generated CD4+ T cell blasts, which were rested for 48 h before restimulation. This was accomplished by using magnetic bead separation of a CD4+ subset from human peripheral blood, followed by expansion in tissue culture with anti-CD3- and anti-CD28 (9.3 mAb)-conjugated beads (32). After allowing these cells to return to the resting state, CD3 and CD28 coligation or P+I stimulation induced IKK activity (Fig. 2), which could be precipitated by either anti-IKKα or -IKKβ Abs. These data confirm that endogenous IKK complexes in primary T cells are activated by dual CD3 and CD28 coligation. While IKK activation was detected with both anti-IKKα and anti-IKKβ immune complexes, it was not clear from these assays whether CD3/CD28 coligation activated heterodimeric or homodimeric IKK complexes.
Evidence that different IKK complexes, which differ in their IKKα and IKKβ composition, are induced by CD3 and CD28 coligation in primary T cells
Recombinant IKKα or IKKβ form homo- and heterodimeric kinase complexes in vitro, suggesting that heterogeneous IKK com- plexes may exist in vivo (23, 24). In this regard, Mercurio et al. have recently shown that HeLa cells contain two distinct IKK complexes, one consisting of IKKα/IKKβ heterodimers, and the other containing IKKβ homodimers (23). Moreover, both complex types could be activated by TNF-α, with heterodimeric complexes showing more abundant IKK activity than homodimeric complexes (23). However, not all cell types contain heterogeneous complexes (23). We were interested in determining the composition of IKK complexes in primary human T cells. First, we examined primary human CD4+ as well as Jurkat T cell lysates for expression of different IKK subunits. Western blot data indicated that IKKβ and IKKα were abundantly expressed in Jurkat and primary T cells (Fig. 3,A, lanes 1–4). The relative amounts of IKKα or IKKβ messages were not affected by cellular stimulation (data not shown). Secondly, we used a sequential immunoprecipitation approach, previously used in HeLa cells (23), to examine primary T cells for assembly of IKKα/IKKβ heterodimers and homodimers (23). This required preclearing of IKKα-containing complexes from CD4+ cell lysates with excess anti-IKKα Abs before precipitating any possible remaining IKKβ from the supernatant with IKKβ-specific antiserum (23). These immunoprecipitates were transferred to immunoblotting membranes, which were sequentially overlaid with IKKα and IKKβ antisera. As demonstrated in Fig. 3 B (lanes 1 and 4), anti-IKKα Abs precipitated both IKKα and IKKβ from the cell lysates, indicating the presence of a heterodimeric complex. The remaining supernatant was devoid of IKKα (lane 2), but still contained IKKβ protein, which could be precipitated by the anti-IKKβ antiserum (lane 6). These complexes were devoid of IKKα (lane 3). The residual amount of IKKβ protein after IKKα clearance (lane 6) was more abundant than the amount of IKKβ that associates with IKKα (lane 4). These results demonstrate the presence of a homodimeric IKKβ complex in primary human CD4+ T cells. No evidence was obtained, however, for the existence of IKKα homodimers in experiments in which we attempted to immune precipitate IKKα from lysates that were precleared with anti-IKKβ Abs (data not shown).
Nonlymphoid cells have recently been shown to include a third noncatalytic IKK subunit, which has been designated IKKγ or IKK-associated protein (IKKAP1) (23, 40). Using an antiserum to IKKγ, we could demonstrate that Jurkat as well as primary human CD4+ T cells, in addition to IKKα and IKKβ, express IKKγ (Fig. 3,A, lanes 5 and 6). Since IKKγ has selective affinity for IKKβ (40), we asked whether IKKγ is present in the homodimeric IKKβ complexes shown in Fig. 3,B. In a further sequential immunoprecipitation experiment, anti-IKKγ Abs were used to immunoprecipitate this subunit from an IKKα-depleted CD4+ cell supernatant. Subsequent immunoblotting showed that IKKγ interacts with IKKβ complex in vivo in the absence of IKKα (Fig. 3,C, lanes 4 and 8). However, heterodimeric complexes also contained IKKγ (40) (not shown). Therefore, IKKγ is a common component of heterodimeric and homodimeric IKK complexes in T cells. Also notice that immunoprecipitation of IKKβ after IKKα depletion (Fig. 3,C, lanes 3 and 7), yielded positive blotting for IKKβ alone. This confirms the existence of IKKβ homodimers, as noted in Fig. 3 B.
Since the data in Fig. 3, B and C, indicate the existence of heterogeneous IKK complexes in primary human CD4+ T cells, we asked whether one or both complexes could be activated by CD3 and CD28 or P+I stimulation. This experiment was conducted in exactly the same way as shown in Fig. 3,C, except that the washed immune complexes were incubated with GST-IκBα and [γ-32P]ATP. Our data show that heterodimeric complexes, precipitated with anti-IKKα (Fig. 3,D, lanes 1–4), and homodimeric complexes, precipitated with either anti-IKKβ (lanes 5–8) or anti-IKKγ (lanes 9–12) after IKKα/IKKβ heterodimer depletion, harbored CD3- and CD28-inducible as well as PMA- and ionomycin-inducible IKK activities. Neither type of complex could be activated by anti-CD3 alone (Fig. 3 D, lanes 2, 6, and 10). These results demonstrate that two distinct IKK complexes are activated upon CD3/CD28 and P+I costimulation in primary human CD4+ T cells.
IKKβ activation is required for transcriptional activation of the CD28RE in the IL-2 promoter
To study the biological significance of IKK activation by TCR/CD28, we concentrated on the IL-2 gene, since its promoter contains a composite c-Rel/AP-1 element, previously characterized as the CD28RE (7, 8, 9). CD28RE plays a critical role in CD3- and CD28-induced activation of the IL-2 gene (7, 8, 9). Moreover, CD28RE is a key integrator of TCR/CD3 and CD28 signaling pathways (7, 8, 9). In a Jurkat cell line transfected with a CD28RE/AP-1 luciferase gene construct, IKKα overexpression had no independent stimulatory effect (Fig. 4,A). In contrast, IKKβ overexpression dramatically enhanced CD3- and CD28-stimulated reporter activity. Similar trends were seen in PMA- and ionomycin-activated cells (Fig. 4,A). It is interesting to note that IKKα and IKKβ overexpression in Jurkat cells generate anti-Flag-precipitable complexes with different levels of IKK activity (Fig. 4,B). While immunoprecipitation of exogenously expressed IKKβ yielded constitutively active complexes, immunoprecipitation of Flag-tagged IKKα yielded a complex with negligible levels of kinase activity (Fig. 4,B). These differences were not due to different amounts of these kinases being expressed, since anti-Flag immunoblotting showed roughly equal amounts of the kinase proteins in the cellular lysate (Fig. 4,C). Since overexpressed IKKβ does not coprecipitate significant amounts of endogenous IKKα (not shown), the data in Fig. 4,B probably reflect the formation of active IKKβ homodimers. The activity of these complexes could be further increased by CD3 and CD28 or P+I stimulation (Fig. 4 B, lanes 5 and 6). While overexpressed IKKα was not biochemically active in Jurkat cells, we have previously shown that overexpression of Flag-tagged IKKα in a macrophage cell line can result in the formation of an LPS-inducible IKK complex (41).
To determine whether overexpressed IKKα and IKKβ also yield differential effects on the full-length IL-2 promoter, we compared the effects of wild-type and kinase-inactive IKKs on a cotransfected IL-2 Luc promoter (Fig. 5). While kinase-inactive IKKα had a minor effect, kinase-inactive IKKβ was a potent (76% decrease) inhibitor of IL-2 Luc activity (Fig. 5). Wild-type IKKα and IKKβ had minimal effects (Fig. 5). The effect of kinase-inactive IKKβ was comparable to the effect of kinase-inactive MEKK1, which acts as a MAP3K in the JNK cascade (Fig. 5). We have recently shown that the JNK cascade plays a critical role in regulating the CD28RE/AP-1 element in the IL-2 (9). The role of JNK in this activation of this promoter is also demonstrated by the inhibitory effect of the JIP-1 (33), which sequesters endogenous JNK in the cell (Fig. 5). Taken together, these results indicate that MEKK1 and IKKβ play critical roles in the activation of the IL-2 promoter.
To determine whether endogenous, rather than exogenous, IKKβ exerts similar effects on the IL-2 promoter, we made use of a recent demonstration that sodium salicylate inhibits IKKβ, but not IKKα, activity in vivo and in vitro (36). At a dose (20 mM) that was previously shown to inhibit IKKβ in vivo (36), sodium salicylate completely inhibited the activation of a full-length IL-2 promoter luciferase construct stably transfected into a Jurkat cell line (Fig. 6,A). Furthermore, sodium salicylate suppressed induction of IL-2 secretion in Jurkat cells stimulated by CD3 and CD28 or P+I (Table I). This drug was also effective in interfering in CD28RE/AP-1 activation during CD3 and CD28 or P+I stimulation, even in the presence of overexpressed IKKβ (Fig. 4 A).
|Stimulations .||IL-2 (pg/ml) .||.|
|.||Untreated .||Sodium salicylate (20 mM) .|
|Control||23.4 ± 2.8||17.4 ± 2.3|
|αCD3||40.4 ± 2.9||1.5 ± 0.3|
|αCD3+ αCD28||1346 ± 24.2||50.5 ± 3.1|
|P+ I||1710 ± 1.0||32.2 ± 3.3|
|Stimulations .||IL-2 (pg/ml) .||.|
|.||Untreated .||Sodium salicylate (20 mM) .|
|Control||23.4 ± 2.8||17.4 ± 2.3|
|αCD3||40.4 ± 2.9||1.5 ± 0.3|
|αCD3+ αCD28||1346 ± 24.2||50.5 ± 3.1|
|P+ I||1710 ± 1.0||32.2 ± 3.3|
Jurkat T cells were left untreated or preincubated with 20 mM sodium salicylate for 2 h. Stimulations were carried out for 24 h as described. Supernatant from each sample was used to test for IL-2 levels by ELISA performed by University of Maryland at Baltimore Cytokine Core Laboratories (Baltimore, MD). Results represent the average of triplicate measurements.
In primary human CD4+ lymphocytes, we asked whether sodium salicylate had an effect on the activation of the homo- or heterodimeric IKK complexes demonstrated in Fig. 3,D. Prior salicylate treatment of intact lymphocytes almost completely inhibited activation of the homodimeric complex (Fig. 6,B, lanes 9–12), but also had a significant inhibitory effect on IKKα/β heterodimers (Fig. 6 B, lanes 1–4), although not to the same extent as the homodimeric complex. Similar to Jurkat cells, sodium salicylate interfered in the induction of IL-2 secretion during CD4+ cell stimulation by CD3 and CD28 (not shown). These data confirm that IKKβ plays a critical role in regulating NF-κB transcription factors that leads to activation of the IL-2 promoter.
In a previous study, we demonstrated that the c-Jun kinase and the NF-κB pathways synergize in the activation of the IL-2 gene in Jurkat T cells (9). In the present study we extend those findings by showing that IKK complexes play a critical role in primary T cells stimulated via the CD3 and CD28 receptors (Figs. 1 and 2). Moreover, biochemical analysis revealed the presence of at least two types of IKK complexes in CD4+ T lymphocytes, i.e., a heterodimeric complex containing equivalent amounts of IKKα and IKKβ, and a homodimeric complex containing IKKβ alone (Fig. 3). Both complexes included IKKγ, a recently discovered non kinase component of the IKK signalsome (Fig. 3,A) (23, 40). These findings agree with previous demonstration of heterogeneous IKK complexes in HeLa cells (23) and suggest that different IKK complexes may contribute to differential NF-κB signaling in human primary CD4+ T cells. Although the roles of different IKK complexes still need to be clarified, it is interesting that kinase-inactive IKKβ exerts a stronger effect on the IL-2 promoter than IKKα (Fig. 5 A). While there may be a number of explanations for this finding, our results suggest one of two possibilities. The first is that IKKβ homodimers rather than IKKα/β heterodimers play the dominant role in regulating the IL-2 promoter in response to CD3/CD28 costimulation. The second possibility is that homo- and heterodimers both contribute to the activation of the IL-2 promoter, but the regulation of these complexes involves a CD3/CD28-inducible component that primarily signals IKKβ. IKKγ, which binds IKKβ selectively (40), may play a role in this process.
Several lines of evidence support the idea that IKKβ is the predominant kinase regulated by CD3/CD28 coligation in primary human CD4+ T cells. First, IKKα depleted lysates from CD4+ cells contained an active kinase complex that could be precipitated by anti-IKKβ alone (Fig. 3). This finding is consistent with observations in other cell types that ectopically expressed, kinase-inactive IKKα minimally interfered in IKK activation, while overexpressed, kinase-inactive IKKβ was a potent inhibitor of IKK activation (19, 42). The contribution of IKKβ to CD3/CD28 costimulation was further demonstrated by pharmacological interference in IKK activation by a specific IKKβ inhibitor, sodium salicylate (36). Not only did this drug inhibit IKK activation, but it also abrogated transcriptional activation of the CD28RE as well as the full-length IL-2 promoter (Fig. 6). Although heterodimeric IKK complexes were also activated by CD3/CD28 costimulation, it is possible that initially this event is triggered by IKKβ. This will explain why sodium salicylate also interfered in the activation of the heterodimeric complex (Fig. 6,B). What role, if any, IKKα plays in the phosphorylation of the IκB proteins remains to be determined. In fact, IKKα has been shown to phosphorylate IκBα at carboxyl-terminal residues that are distinct from the serines (S32/S36) that play a role in proteolytic degradation (24, 25); the significance of that phosphorylation event still needs to be determined. Recently, several additional reports have appeared that indicate that IKKβ has a dominant effect in regulating IκB kinase activity (43, 44, 45, 46). Mutational alteration of the regulatory and catalytic domains of IKKα and IKKβ showed that IKKβ is the main target for TNF-α and IL-1 stimulation (43). Moreover, embryonic fibroblasts obtained from IKKα-deficient mice exhibit reasonably good IKK activation, while embryonic cells from IKKβ-deficient mice do not express significant IKK activity (44, 45, 46). These findings together with our observations that kinase-inactive IKKβ, but not kinase-inactive IKKα, interfered in the activation of the IL-2 promoter (Fig. 5) support our conclusions that IKKβ is the principal kinase involved in CD3/CD28-mediated NF-κB activation.
New information on the regulation of IKK complexes is emerging with the discovery of new IKK subunits. One example is IKKγ (IKKAP-1), a nonkinase subunit that was recently identified as an essential component of the heterodimeric IKK complex (23, 40). We show that IKKγ is also a component of the homodimeric IKKβ complexes in T cells (Fig. 3). This is probably due to the specific binding affinity of IKKγ for IKKβ (23, 40). Moreover, an anti-IKKγ antiserum immunoprecipitated active IKK complexes from IKKα-depleted lysates in primary human T cells (Fig. 3). While the role of IKKγ still needs to be clarified, one suggestion has been that it relays signals to IKKβ from afferent components in the IKK complex (40). A second novel protein that influences IKKβ activity is IKAP (note that IKAP is distinct from IKKAP-1/IKKγ), which appears to function as a scaffold protein regulating the assembly of IKK complexes (42). While in unstimulated cells most endogenous IKKβ appears to be IKAP associated (42), IL-1 stimulation results in the dissociation of IKKβ from IKAP (42). Similar to the role of JIP-1 in the JNK cascade, overexpression of IKAP abrogates IKK activity (42). While the role of IKAP in T cells still needs to be clarified, one possibility is that IKAP may play a role in the assembly of specific IKK complexes.
The apparent selectivity of IKKβ in the activation of the IL-2 promoter is compatible with our recent finding that a second, but interconnected, signaling cascade is activated in a synergistic fashion by CD3/CD28 coligation. Activation of the Jun kinase cascade is medicated by a MAP3K, MEKK1, that additionally activates IKKβ (9). This selectivity for IKKβ, but not IKKα, has also been demonstrated by other investigators (29, 30). The inhibitory effect of dominant-negative MEKK1 on the IL-2 promoter (Fig. 5) may therefore reflect a role of MEKK1 in both the JNK and NF-κB signaling pathways. That idea is compatible with the stimulatory effects of dominant-active MEKK1 on both the AP-1 and c-Rel binding sites in the CD28RE/AP-1 (9). While we do not know at this stage whether MEKK1 selectively activates homo- or heterodimeric IKK complexes, further studies to address that question are ongoing. The idea that heterogeneous IKK complexes may be selectively activated by different upstream kinases is further strengthened by the recent demonstration that the proto-oncogene, Cot/Tp1, a MAP3K-related serine-threonine kinase, also participates in IKK activation by CD3/CD28 (27). Interestingly, Cot interacts directly with IKKα and NIK, and it has been proposed that Cot acts upstream of NIK (27). Although ectopic expression of a kinase-inactive version of NIK inhibits IKKα and IKKβ activities, NIK only phosphorylates IKKα in vitro and in vivo (26). Since this implies that Cot may selectively signal IKKα, it is possible that this MAP3K may be regulating the activities of those complexes that contain IKKα, while MEKK1 may regulate the homodimeric complex. We are in the process of testing that hypothesis.
Finally, the data presented in this paper indicate that the composition of IKK signalsome may not be rigid, and its assembly is subject to regulation by the afferent signal as well as the cell type. Such regulation may be important for selective gene expression by the NF-κB pathway. We are currently pursuing the role of heterogeneous IKK complexes in selective gene activation during CD3/CD28 costimulation of primary human CD4+ T cells.
We thank Ivonne Castaneda and Peggy Shih for secretarial assistance and Martin Kaszubowski for technical support.
This work was supported by grants from the U.S. Public Health Service AG14992 and the Southern California Chapter of the Arthritis Foundation (to A.E.N.) and a National Institute of Allergy and Infectious Diseases Immunology training grant.
Abbreviations used in this paper: CD28RE, CD28 response element; IKK, IκB kinase; MAP3K, mitogen-activated protein kinase kinase kinase; MEKK1, mitogen-activated protein kinase kinase kinase 1; JNK, N-terminal Jun kinase; NIK, NF-κB- inducing kinase; IKAP, IKK complex-associated protein; IKKAP-1, IKK-associated protein 1; JIP-1, JNK-interacting protein; DA, dominant active; DN, dominant negative; P+I, PMA plus ionomycin; Tpl, tumor progression locus; Cot, cancer Osaka thyroid; Luc, luciferase.