Glucocorticoid-Induced TNF Receptor, a Costimulatory Receptor on Naive and Activated T Cells, Uses TNF Receptor-Associated Factor 2 in a Nov
http://www.100md.com
免疫学杂志 2005年第12期
Abstract
Glucocorticoid-induced TNFR (GITR) has been implicated as an essential regulator of immune responses to self tissues and pathogens. We have recently shown that GITR-induced cellular events promote survival of naive T cells, but are insufficient to protect against activation-induced cell death. However, the molecular mechanisms of GITR-induced signal transduction that influence physiologic and pathologic immune responses are not well understood. TNFR-associated factors (TRAFs) are pivotal adapter proteins involved in signal transduction pathways of TNFR-related proteins. Yeast two-hybrid assays and studies in HEK293 cells and primary lymphocytes indicated interactions between TRAF2 and GITR mediated by acidic residues in the cytoplasmic domain of the receptor. GITR-induced activation of NF-B is blocked by A20, an NF-B-inducible protein that interacts with TRAFs and functions in a negative feedback mechanism downstream of other TNFRs. Interestingly, in contrast with its effects on signaling triggered by other TNFRs, our functional studies revealed that TRAF2 plays a novel inhibitory role in GITR-triggered NF-B activation.
Introduction
Signaling pathways triggered by TNFR family members globally influence immune responses (1). Glucocorticoid-induced TNFR (GITR)3 has recently been demonstrated to augment activation, proliferation, and cytokine production of T cells (2, 3, 4). Furthermore, we have previously shown that survival during the early phases of T cell activation is promoted by GITR-induced signaling events (5). Signaling through GITR abrogates the suppressive function of naturally occurring regulatory T cells (Treg cells) and renders naive effector T cells refractory to Treg cell-mediated suppression (6, 7, 8). Although the impact of GITR on T cell function is being elucidated, little is known about the molecular mechanisms of signal transduction induced by GITR.
TNFR-related proteins use members of a cytoplasmic adapter protein family, termed TNFR-associated factors (TRAFs), to transmit extracellular signals (9, 10, 11, 12). TRAF2 is the prototypical member of the TRAF family, which is defined by a conserved TRAF domain that mediates self-association and interactions with receptors and other cytoplasmic proteins (11). At its N terminus, TRAF2 contains a RING finger and several zinc finger motifs (13, 14). The RING finger is necessary and sufficient for TRAF2-mediated signaling, because its deletion results in a dominant-negative molecule and enforced oligomerization of the motif triggers downstream signaling events (14, 15, 16). TRAF2 initiates the formation of multiprotein complexes and controls activation of kinases, such as the IB kinases and the JNKs (15, 17). These kinases trigger the activation of transcription factors of the NF-B and AP-1 families that modulate expression of pro- and antiapoptotic proteins (18, 19). Further, TRAF2 regulates the production of reactive oxygen species, which can sustain cell survival by promoting NF-B activation but can also trigger apoptosis by damaging lipids and proteins (20, 21). Highlighting the importance of TRAF2 in the immune system, targeted gene disruption of the adapter protein causes postnatal lethality, profound disruption of immune homeostasis and development, and defects in immune responses to the TNFR-related protein CD40 (22, 23). Furthermore, TRAF2 deficiency shifts the balance between cell survival and apoptotic pathways triggered by TNFR1 resulting in hypersensitivity to TNF--induced cell death.
We and others have observed that GITR triggers NF-B activation (5, 24, 25). Studies of the human GITR homologue, termed activation-induced TNFR-related receptor (AITR), suggest that TRAFs can play a role downstream of GITR (26, 27). However, little is known about molecular signaling mechanisms triggered by GITR. Our findings demonstrate that TRAFs can interact directly with GITR and reveal a novel inhibitory function of TRAF2 in GITR-induced NF-B activation.
Materials and Methods
Saccharomyces cerevisiae PJ69-4A was cotransformed with pAS1[GITRcp] and pACT expressing full-length TRAF2 fused to the AD of GAL4 or the AD only using lithium acetate (34). Specific interactions were determined by yeast growth on histidine-deficient plates supplemented with the competitive HIS3 inhibitor 3-aminotriazole (3-AT) to lower background growth due to the leakiness of the HIS3 promoter. Filter lift assays were performed to monitor for -galactosidase activity.
Results
NF-B activation is triggered by GITR
The transcription factor NF-B plays a vital role in coordinating the expression of genes that are central to inflammatory responses and is activated by members of the TNFR family (35, 36). NF-B activation can be measured by detecting protein levels of IB, which binds NF-B subunits in the cytoplasm and is targeted for proteosomal degradation in response to agonistic stimuli before resynthesis as an NF-B-dependent gene product (37). To determine whether GITR induces NF-B activation, IB levels were analyzed from lysates of activated T cells treated with agonistic GITR-specific Ab. Compared with control treatments, GITR cross-linking resulted in time-dependent degradation and subsequent resynthesis of IB, indicating NF-B activation (Fig. 1A). To determine whether engagement of GITR by its ligand can trigger NF-B activation, GITR and GITR ligand (GITR-L) were coexpressed in HEK293 cells and NF-B activation was determined by a luciferase reporter assay. While expression of either the receptor or its ligand alone was insufficient to activate NF-B, coexpression of GITR and GITR-L caused a significant induction in NF-B activity (Fig. 1B). Consistent with previously published studies by Yu et al., our findings indicate that GITR triggers signaling pathways that lead to NF-B activation (38).
GITR interacts with distinct members of the TRAF family
TRAFs have been implicated in NF-B activation triggered by TNFR family members (9, 10, 12). The cytoplasmic domain of GITR contains acidic residues that are conserved between human and mouse and may serve as TRAF binding sites (Fig. 2A) (27, 39, 40). To determine whether TRAFs interact with GITR, the cytoplasmic domain of GITR was fused to the GAL4 DNA-binding domain (DBD). Interactions between GITR-DBD and distinct TRAFs fused to the GAL4 activation domain (AD) were tested in a yeast two-hybrid system. Equivalent transformation efficiencies were confirmed by monitoring growth of yeast on plates lacking leucine and tryptophane (data not shown). To increase the stringency of the assay, 3-AT, a competitive inhibitor of HIS3, was added to media deficient in leucine, tryptophane, and histidine (LTH–). Coexpressing TRAF1-, TRAF2-, or TRAF3-AD with GITR-DBD allowed growth of yeast on LTH– plates containing at least 3 mM 3-AT, whereas the AD alone or TRAF4-AD did not (data not shown). -gal activity in yeast coexpressing GITR-DBD with TRAF1-, TRAF2-, or TRAF3-AD confirmed that yeast growth on LTH– plates was due to GITR-TRAF interactions (Fig. 2B). These results indicate that distinct members of the TRAF family interact directly with the cytoplasmic domain of GITR and suggest that TRAFs serve as signaling intermediates downstream of this receptor.
To determine whether acidic residues, which are conserved in human and mouse, are required for interactions between TRAFs and GITR, deletion and alanine substitution mutants of the cytoplasmic domain of GITR were tested as baits in yeast two-hybrid assays. Yeast coexpressing TRAF2-AD and a C-terminal deletion mutant of GITR (211STOP) did not grow on LTH– plates and were negative for -gal activity, suggesting that residues in the C terminus of GITR play a role in TRAF2 recruitment (Fig. 2C and data not shown). Further studies with alanine substitution mutants replacing acidic residues at either 202/203 (202ED>2A) or 211–213 (211EEE>3A) resulted in disruption of TRAF2 interaction with the receptor, whereas mutation of the glutamate residues 219/220 (219EE>2A) had no effect, indicating that amino acids 202/203 (ED) and 211–213 (EEE) are crucial for TRAF interactions.
Conserved acidic residues in the cytoplasmic domain of GITR are critical for TRAF interaction and NF-B activation
To confirm the yeast two-hybrid analysis in mammalian cells, coimmunoprecipitation and functional studies were performed after transfecting HEK293 cells with expression constructs of TRAF2, GITR-L and full-length GITR with either a wild-type cytoplasmic domain or the mutations described above (Fig. 2, A and C). Flow cytometric analysis confirmed that all GITR mutants were expressed at similar levels on the cell surface and did not affect the expression levels of GITR-L (Fig. 3A). Similar to the results obtained in the yeast two-hybrid studies, wild-type GITR and mutant 219EE>AA interact with TRAF2 in mammalian cells (Fig. 3B). Interactions between GITR and TRAF2 were significantly reduced when acidic residues at positions 202/203 and 211–213 were replaced by alanine residues. Western blot analysis revealed equal expression levels and similar amounts of immunoprecipitated receptor. The distinct migration patterns of the GITR mutants were most likely due to the replacement of charged amino acids with alanine residues. Taken together, these data indicate that the cytoplasmic domain of GITR contains a single TRAF binding site where acidic residues 202/203 and 211–213 are critical for this interaction.
To investigate whether TRAF recruitment is involved in GITR-triggered NF-B activation, HEK293 cells were transfected with expression constructs of GITR-L and full-length GITR or receptor mutants used to define the TRAF binding site. Like the intact receptor, the GITR mutant (219EE>2A) capable of interacting with TRAF caused NF-B activation in HEK293 cells (Fig. 3C). In contrast, expression of GITR mutants that were impaired in TRAF binding (202ED>2A and 211EEE>3A) resulted in significantly decreased NF-B activation in response to ligand engagement by the receptor. This observation provides specific indications that NF-B activation is being directly triggered by GITR and argues against reverse signaling occurring through GITR-L under these conditions. Furthermore, these studies revealed the relevance of the TRAF binding site for GITR-induced NF-B activation, suggesting that recruitment of TRAFs to GITR is critical for signaling events downstream of the receptor.
TRAF2 recruitment to GITR is ligand-dependent
To establish that GITR and TRAF2 interact in T cells expressing physiologic levels of the proteins, co-IP analysis was performed using Ag-stimulated DO11.10 splenocytes. FACS analysis indicated that activated cells were >90% CD4+ T cells and expressed high levels of GITR (data not shown). After treatment of primary T cells with an agonistic GITR Ab for 1 or 15 min to cross-link the receptor before cell lysis, TRAF2 was coimmunoprecipitated with GITR (Fig. 4A). Only a small fraction of the total cellular TRAF2 was detected after immunoprecipitation of GITR, suggesting either a low affinity or transient interaction of the adapter protein with the receptor. To gain further insights into the kinetics of TRAF2 recruitment to GITR after ligand engagement, primary T cells were incubated with GITR-L, which caused GITR to interact with TRAF2 rapidly and indicated that receptor interaction with the adapter protein is ligand dependent (Fig. 4B). Comparing the cellular pool of TRAF2 with the amount in GITR-containing protein complexes is consistent with a transient or weak GITR-TRAF2 interaction and argues that TRAF2 molecules not engaged in GITR signaling can interact with other adapter proteins or TNFR-related proteins. These results confirm that TRAF2 interacts with GITR under physiologic conditions and may regulate early proximal stages of receptor-mediated signaling pathways.
The cytoplasmic domain of GITR causes TRAF2 relocalization
Previously, we and others have shown that interactions of TRAFs and TNFR-related molecules cause TRAF relocalization to a detergent-insoluble cellular compartment (41, 42, 43). Further, we have shown that dimerization of the cytoplasmic domains of TNFR-related proteins results in TRAF intracellular relocalization and constitutive NF-B activation (20, 31). To examine whether GITR also regulates the intracellular localization of TRAF2, TRAF2 levels in the detergent-soluble and -insoluble fractions of cell lysates were determined by Western blot analysis after cotransfection of HEK293 cells with TRAF2 and CD28-GITR, a chimeric receptor consisting of the extracellular and transmembrane regions of the homodimer CD28 fused to the cytoplasmic domain of GITR. Similar to CD28-CD30, which has been previously shown to translocate TRAF2 (41), CD28-GITR caused TRAF2 redistribution from the detergent-soluble to the -insoluble cellular fraction, suggesting that GITR controls the intracellular trafficking of TRAF2 similar to other TNFR-related proteins (Fig. 5). CD28Tail, a receptor lacking a cytoplasmic domain, did not cause TRAF2 relocalization and served as a negative control. Consistent with a weak or transient interaction between TRAF2 and GITR, expression of CD28-GITR caused less efficient relocalization of TRAF2 than CD28-CD30, which resulted in complete depletion of TRAF2 from the detergent-soluble compartment. These data argue that the functional relationship between GITR and TRAF2 is distinct from other TNFRs, such as CD30.
The zinc finger protein A20 blocks NF-B activation triggered by GITR
Multimerization of TNFR-related proteins results in the assembly of TRAF-containing signaling complexes at the cytoplasmic domains of these receptors as a proximal step in signal transduction (9). Regulation of the composition as well as disassembly of TRAF-containing protein complexes is not well understood. A20 is an anti-apoptotic adapter protein that interacts with TRAFs and inhibits NF-B activation in a negative feedback loop (44, 45, 46). When A20 was coexpressed with GITR and GITR-L in HEK293 cells, receptor-induced NF-B activation was significantly impaired (Fig. 6). This observation points to A20 as an inhibitory signaling intermediate downstream of GITR and further implies that TRAF-containing protein complexes are involved in GITR-induced signaling events.
TRAF2 inhibits NF-B activation triggered by GITR
Genetic analyses argue that TRAF2 promotes signaling events triggered by TNFR-related proteins, such as NF-B activation (22, 23, 47). To test whether TRAF2 plays a similar role downstream of GITR, HEK293 cells were cotransfected with GITR, GITR-L, and either full-length TRAF2 or TRAF2DN, a dominant-negative mutant of TRAF2 lacking the N-terminal RING finger (15, 31). TRAF2DN interacts with receptors containing TRAF-binding domains and is thought to interfere with the recruitment of other TRAF molecules to these sites (15). As expected, coexpression of TRAF2DN abolished GITR-induced NF-B activation measured by a luciferase reporter assay, consistent with the proposed role of TRAFs downstream of GITR (Fig. 7A). However, rather than amplifying GITR-induced NF-B activation, TRAF2 significantly reduced NF-B activation triggered by GITR, similar to the N-terminal deletion mutant of TRAF2. Expression of TRAF2 or its N-terminal deletion mutant alone did not affect NF-B activity indicating the specificity of the finding. As an additional specificity control, coexpression of TRAF2 with GITR-L did not alter NF-B activation (data not shown). We and others have previously shown that the ratios of receptor molecules to adapter components are pivotal to downstream signaling events triggered by TNFR-related proteins (29, 48, 49, 50). HEK293 cells did not express amounts of endogenous TRAF2 that were detectable by Western blot (Fig. 5). To rule out that the inhibitory effect of TRAF2 was caused by suboptimal molecular ratio of GITR and the adapter protein, the amount of plasmid encoding TRAF2 used to transfect HEK293 cells was titrated. Consistent with the previous result, amplification of GITR-induced NF-B activation was not observed; rather, the inhibition of GITR-induced NF-B activation varied directly with TRAF2 expression levels (Fig. 7B). These results suggest a novel inhibitory function of TRAF2 in signaling pathways triggered by GITR.
Discussion
Adapter proteins, such as TRAFs, are critical signaling intermediates downstream of TNFR-related molecules and serve as both convergent and divergent platforms triggering activation of kinases, such as IB kinases that lead to activation of transcription factors, such as NF-B. Therefore, we hypothesized that TRAFs function as crucial signaling intermediates downstream of GITR. Previously, we and others have used yeast two-hybrid systems to identify proteins interacting with the cytoplasmic domains of TNFR-related molecules and to map amino acids required for these interactions (13, 14, 31, 33). Although TRAF binding sequences from different receptors diverge, acidic residues of two weak consensus motifs–(P/S/T/A)x(Q/E)E and PxQxxD–seem critical for their interactions (39, 40). In agreement with these motifs, residues 200 to 203 (SAED) and 210 to 213 (PEEE) in the cytoplasmic domain of GITR contain a serine or proline at position P0 and acidic residues at positions P2 and P3 (Fig. 2A). In contrast, the two glutamic acid residues at positions 219 and 220 of GITR do not fit either of the TRAF-binding motifs. Consistently, residues 202/203 and 211–213 of GITR are critical for TRAF2 binding, but residues 219/220 are not (Figs. 2 and 3). Distinct TRAFs can compete for receptor-binding motifs, which agrees with the observation that TRAFs other than TRAF2 interact with the cytoplasmic domain of GITR (Fig. 2B and Refs. 39 and 40).
Co-IP analyses verified the interactions between GITR and TRAF2 in primary lymphocytes expressing endogenous levels of the proteins (Fig. 4). These studies indicate that GITR rapidly recruits TRAF2 in a ligand-dependent manner and suggest that the adapter protein can regulate early stages of GITR-induced signal transduction under physiologic conditions. Similar to the engagement of TNFR-related proteins with their ligands, chimeric receptors containing the extracellular and transmembrane domains of CD28 and the cytoplasmic domains of either CD30, Ox40, or 4-1BB trigger signaling events including intracellular TRAF relocalization, which are dependent on interactions between the TNFRs and TRAFs (31, 41). When HEK293 cells were cotransfected with CD28-GITR and TRAF2, relocalization of TRAF2 from a detergent-soluble to -insoluble fraction of lysates was observed, arguing further for TRAF2 interaction with the cytoplasmic domain of GITR (Fig. 5). Consistent with our co-IP analyses that suggested a weak or transient interaction between TRAF2 and GITR, translocation of TRAF2 induced by GITR does not significantly reduce the TRAF2 level in the detergent-soluble compartment, suggesting that the adapter protein is still available to elaborate signaling events triggered by other TNFR-related proteins expressed in the cell.
Similar to previously published reports, GITR cross-linking with its ligand or an agonistic Ab activated NF-B (Fig. 1, and Refs. 2 , 5 , and 24). To test whether TRAF interactions are critical for NF-B activation, HEK293 cells were transfected with GITR mutants sufficient or defective in TRAF2 interaction (Fig. 3C). Mutation of residues in GITR needed for TRAF interaction correlated with defects in receptor-mediated NF-B activation. Because TRAFs bind overlapping sites in TNFR-related proteins, this result argues that TRAFs in addition to TRAF2 may elaborate GITR-induced signaling pathways, which is currently being examined. Consistent with this hypothesis, A20, an adapter protein that interacts with and regulates TRAF function, inhibited GITR-induced NF-B activation (Fig. 6 and Refs. 41 , 45 , and 51). Moreover, we recently showed that TRAF4 enhances NF-B activation induced by GITR (29). The observations that TRAF4 did not interact with GITR in yeast, yet required the TRAF interaction site to augment NF-B activation further implies that other TRAFs or adapter proteins initiate GITR-induced signaling events (Fig. 2B and Ref. 29).
Importantly, functional studies revealed intriguing differences between TRAF2-mediated signaling triggered by GITR and other members of the TNFR superfamily including AITR (26, 27). In contrast to previously published findings that suggest a role of TRAF2 as a positive regulator of NF-B activation triggered by a variety of TNFR-related molecules, TRAF2 significantly inhibited GITR-mediated activation of NF-B similar to the expression of a dominant-negative mutant of TRAF2 (Fig. 7 and Refs. 22 , 23 , and 47). TRAF2 is known to dampen specific signaling pathways triggered by TNFR1 (23). Furthermore, TRAF2 has been implicated as an inhibitor of Th2 responses through its interaction with NFAT-interacting protein 45 and as a negative regulator of the noncanonical pathway of NF-B activation in mature B cells (52, 53). Whether TRAF2 negatively regulates signaling beyond competing with other signaling proteins for GITR binding is unclear. Intriguingly, TRAF2 has been shown recently to function as an E3 ubiquitin ligase and to undergo ubiquitination itself (42, 54, 55). Future studies will test whether ubiquitination of TRAF2 or other proteins plays a role in the novel regulatory role of TRAF2 in GITR-induced NF-B activation.
Taken together, our findings indicate that TRAF2 is a proximal signaling intermediate downstream of GITR and suggest a novel regulatory role for TRAF2 in NF-B activation triggered by GITR. As GITR is a critical regulator of the interface between Treg cells and immune effector cells, gaining further insight into the mechanisms of signal transduction pathways triggered by GITR will allow the development of therapeutic approaches targeting inflammatory immune responses and autoimmunity.
Footnotes
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.
1 This research was supported in part by National Institutes of Health Grants 5 RO1 HL67312-02 and T32 HL007317 and an Investigator Award of the Cancer Research Institute (to R.H.A.).
2 Address correspondence and reprint requests to Dr. Robert H. Arch, School of Medicine, Washington University, Campus Box 8052, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address: arch{at}wustl.edu
3 Abbreviations used in this paper: GITR, glucocorticoid-induced TNFR; Treg, regulatory T; TRAF, TNFR-associated factor; AITR, activation-induced TNFR-related receptor; IP, immunoprecipitation; GITR-L, GITR-ligand; HEK, human embryonic kidney; DBD, DNA-binding domain; AD, activation domain; 3-AT, 3-aminotriazole; -gal, -galactosidase.
Received for publication February 11, 2005. Accepted for publication April 18, 2005.
References
Croft, M.. 2003. Co-stimulatory members of the TNFR family: keys to effective T-cell immunity?. Nat. Rev. Immunol. 3: 609-620.
Tone, M., Y. Tone, E. Adams, S. F. Yates, M. R. Frewin, S. P. Cobbold, H. Waldmann. 2003. Mouse glucocorticoid-induced tumor necrosis factor receptor ligand is costimulatory for T cells. Proc. Natl. Acad. Sci. USA 100: 15059-15064.
Ronchetti, S., O. Zollo, S. Bruscoli, M. Agostini, R. Bianchini, G. Nocentini, E. Ayroldi, C. Riccardi. 2004. Frontline: GITR, a member of the TNF receptor superfamily, is costimulatory to mouse T lymphocyte subpopulations. Eur. J. Immunol. 34: 613-622.
Kanamaru, F., P. Youngnak, M. Hashiguchi, T. Nishioka, T. Takahashi, S. Sakaguchi, I. Ishikawa, M. Azuma. 2004. Costimulation via glucocorticoid-induced TNF receptor in both conventional and CD25+ regulatory CD4+ T cells. J. Immunol. 172: 7306-7314.
Esparza, E. M., R. H. Arch. 2005. Glucocorticoid-induced TNF receptor functions as a costimulatory receptor that promotes survival in early phases of T cell activation. J. Immunol. 174: 7869-7874.
Shimizu, J., S. Yamazaki, T. Takahashi, Y. Ishida, S. Sakaguchi. 2002. Stimulation of CD25+CD4+ regulatory T cells through GITR breaks immunologic self tolerance. Nat. Immunol. 3: 135-142.
McHugh, R. S., M. J. Whitters, C. A. Piccirillo, D. A. Young, E. M. Shevach, M. Collins, M. C. Byrne. 2002. CD4+CD25+ immunoregulatory T cells: gene expression analysis reveals a functional role for the glucocorticoid-induced TNF receptor. Immunity 16: 311-323.(Edward M. Esparza and Rob)
Glucocorticoid-induced TNFR (GITR) has been implicated as an essential regulator of immune responses to self tissues and pathogens. We have recently shown that GITR-induced cellular events promote survival of naive T cells, but are insufficient to protect against activation-induced cell death. However, the molecular mechanisms of GITR-induced signal transduction that influence physiologic and pathologic immune responses are not well understood. TNFR-associated factors (TRAFs) are pivotal adapter proteins involved in signal transduction pathways of TNFR-related proteins. Yeast two-hybrid assays and studies in HEK293 cells and primary lymphocytes indicated interactions between TRAF2 and GITR mediated by acidic residues in the cytoplasmic domain of the receptor. GITR-induced activation of NF-B is blocked by A20, an NF-B-inducible protein that interacts with TRAFs and functions in a negative feedback mechanism downstream of other TNFRs. Interestingly, in contrast with its effects on signaling triggered by other TNFRs, our functional studies revealed that TRAF2 plays a novel inhibitory role in GITR-triggered NF-B activation.
Introduction
Signaling pathways triggered by TNFR family members globally influence immune responses (1). Glucocorticoid-induced TNFR (GITR)3 has recently been demonstrated to augment activation, proliferation, and cytokine production of T cells (2, 3, 4). Furthermore, we have previously shown that survival during the early phases of T cell activation is promoted by GITR-induced signaling events (5). Signaling through GITR abrogates the suppressive function of naturally occurring regulatory T cells (Treg cells) and renders naive effector T cells refractory to Treg cell-mediated suppression (6, 7, 8). Although the impact of GITR on T cell function is being elucidated, little is known about the molecular mechanisms of signal transduction induced by GITR.
TNFR-related proteins use members of a cytoplasmic adapter protein family, termed TNFR-associated factors (TRAFs), to transmit extracellular signals (9, 10, 11, 12). TRAF2 is the prototypical member of the TRAF family, which is defined by a conserved TRAF domain that mediates self-association and interactions with receptors and other cytoplasmic proteins (11). At its N terminus, TRAF2 contains a RING finger and several zinc finger motifs (13, 14). The RING finger is necessary and sufficient for TRAF2-mediated signaling, because its deletion results in a dominant-negative molecule and enforced oligomerization of the motif triggers downstream signaling events (14, 15, 16). TRAF2 initiates the formation of multiprotein complexes and controls activation of kinases, such as the IB kinases and the JNKs (15, 17). These kinases trigger the activation of transcription factors of the NF-B and AP-1 families that modulate expression of pro- and antiapoptotic proteins (18, 19). Further, TRAF2 regulates the production of reactive oxygen species, which can sustain cell survival by promoting NF-B activation but can also trigger apoptosis by damaging lipids and proteins (20, 21). Highlighting the importance of TRAF2 in the immune system, targeted gene disruption of the adapter protein causes postnatal lethality, profound disruption of immune homeostasis and development, and defects in immune responses to the TNFR-related protein CD40 (22, 23). Furthermore, TRAF2 deficiency shifts the balance between cell survival and apoptotic pathways triggered by TNFR1 resulting in hypersensitivity to TNF--induced cell death.
We and others have observed that GITR triggers NF-B activation (5, 24, 25). Studies of the human GITR homologue, termed activation-induced TNFR-related receptor (AITR), suggest that TRAFs can play a role downstream of GITR (26, 27). However, little is known about molecular signaling mechanisms triggered by GITR. Our findings demonstrate that TRAFs can interact directly with GITR and reveal a novel inhibitory function of TRAF2 in GITR-induced NF-B activation.
Materials and Methods
Saccharomyces cerevisiae PJ69-4A was cotransformed with pAS1[GITRcp] and pACT expressing full-length TRAF2 fused to the AD of GAL4 or the AD only using lithium acetate (34). Specific interactions were determined by yeast growth on histidine-deficient plates supplemented with the competitive HIS3 inhibitor 3-aminotriazole (3-AT) to lower background growth due to the leakiness of the HIS3 promoter. Filter lift assays were performed to monitor for -galactosidase activity.
Results
NF-B activation is triggered by GITR
The transcription factor NF-B plays a vital role in coordinating the expression of genes that are central to inflammatory responses and is activated by members of the TNFR family (35, 36). NF-B activation can be measured by detecting protein levels of IB, which binds NF-B subunits in the cytoplasm and is targeted for proteosomal degradation in response to agonistic stimuli before resynthesis as an NF-B-dependent gene product (37). To determine whether GITR induces NF-B activation, IB levels were analyzed from lysates of activated T cells treated with agonistic GITR-specific Ab. Compared with control treatments, GITR cross-linking resulted in time-dependent degradation and subsequent resynthesis of IB, indicating NF-B activation (Fig. 1A). To determine whether engagement of GITR by its ligand can trigger NF-B activation, GITR and GITR ligand (GITR-L) were coexpressed in HEK293 cells and NF-B activation was determined by a luciferase reporter assay. While expression of either the receptor or its ligand alone was insufficient to activate NF-B, coexpression of GITR and GITR-L caused a significant induction in NF-B activity (Fig. 1B). Consistent with previously published studies by Yu et al., our findings indicate that GITR triggers signaling pathways that lead to NF-B activation (38).
GITR interacts with distinct members of the TRAF family
TRAFs have been implicated in NF-B activation triggered by TNFR family members (9, 10, 12). The cytoplasmic domain of GITR contains acidic residues that are conserved between human and mouse and may serve as TRAF binding sites (Fig. 2A) (27, 39, 40). To determine whether TRAFs interact with GITR, the cytoplasmic domain of GITR was fused to the GAL4 DNA-binding domain (DBD). Interactions between GITR-DBD and distinct TRAFs fused to the GAL4 activation domain (AD) were tested in a yeast two-hybrid system. Equivalent transformation efficiencies were confirmed by monitoring growth of yeast on plates lacking leucine and tryptophane (data not shown). To increase the stringency of the assay, 3-AT, a competitive inhibitor of HIS3, was added to media deficient in leucine, tryptophane, and histidine (LTH–). Coexpressing TRAF1-, TRAF2-, or TRAF3-AD with GITR-DBD allowed growth of yeast on LTH– plates containing at least 3 mM 3-AT, whereas the AD alone or TRAF4-AD did not (data not shown). -gal activity in yeast coexpressing GITR-DBD with TRAF1-, TRAF2-, or TRAF3-AD confirmed that yeast growth on LTH– plates was due to GITR-TRAF interactions (Fig. 2B). These results indicate that distinct members of the TRAF family interact directly with the cytoplasmic domain of GITR and suggest that TRAFs serve as signaling intermediates downstream of this receptor.
To determine whether acidic residues, which are conserved in human and mouse, are required for interactions between TRAFs and GITR, deletion and alanine substitution mutants of the cytoplasmic domain of GITR were tested as baits in yeast two-hybrid assays. Yeast coexpressing TRAF2-AD and a C-terminal deletion mutant of GITR (211STOP) did not grow on LTH– plates and were negative for -gal activity, suggesting that residues in the C terminus of GITR play a role in TRAF2 recruitment (Fig. 2C and data not shown). Further studies with alanine substitution mutants replacing acidic residues at either 202/203 (202ED>2A) or 211–213 (211EEE>3A) resulted in disruption of TRAF2 interaction with the receptor, whereas mutation of the glutamate residues 219/220 (219EE>2A) had no effect, indicating that amino acids 202/203 (ED) and 211–213 (EEE) are crucial for TRAF interactions.
Conserved acidic residues in the cytoplasmic domain of GITR are critical for TRAF interaction and NF-B activation
To confirm the yeast two-hybrid analysis in mammalian cells, coimmunoprecipitation and functional studies were performed after transfecting HEK293 cells with expression constructs of TRAF2, GITR-L and full-length GITR with either a wild-type cytoplasmic domain or the mutations described above (Fig. 2, A and C). Flow cytometric analysis confirmed that all GITR mutants were expressed at similar levels on the cell surface and did not affect the expression levels of GITR-L (Fig. 3A). Similar to the results obtained in the yeast two-hybrid studies, wild-type GITR and mutant 219EE>AA interact with TRAF2 in mammalian cells (Fig. 3B). Interactions between GITR and TRAF2 were significantly reduced when acidic residues at positions 202/203 and 211–213 were replaced by alanine residues. Western blot analysis revealed equal expression levels and similar amounts of immunoprecipitated receptor. The distinct migration patterns of the GITR mutants were most likely due to the replacement of charged amino acids with alanine residues. Taken together, these data indicate that the cytoplasmic domain of GITR contains a single TRAF binding site where acidic residues 202/203 and 211–213 are critical for this interaction.
To investigate whether TRAF recruitment is involved in GITR-triggered NF-B activation, HEK293 cells were transfected with expression constructs of GITR-L and full-length GITR or receptor mutants used to define the TRAF binding site. Like the intact receptor, the GITR mutant (219EE>2A) capable of interacting with TRAF caused NF-B activation in HEK293 cells (Fig. 3C). In contrast, expression of GITR mutants that were impaired in TRAF binding (202ED>2A and 211EEE>3A) resulted in significantly decreased NF-B activation in response to ligand engagement by the receptor. This observation provides specific indications that NF-B activation is being directly triggered by GITR and argues against reverse signaling occurring through GITR-L under these conditions. Furthermore, these studies revealed the relevance of the TRAF binding site for GITR-induced NF-B activation, suggesting that recruitment of TRAFs to GITR is critical for signaling events downstream of the receptor.
TRAF2 recruitment to GITR is ligand-dependent
To establish that GITR and TRAF2 interact in T cells expressing physiologic levels of the proteins, co-IP analysis was performed using Ag-stimulated DO11.10 splenocytes. FACS analysis indicated that activated cells were >90% CD4+ T cells and expressed high levels of GITR (data not shown). After treatment of primary T cells with an agonistic GITR Ab for 1 or 15 min to cross-link the receptor before cell lysis, TRAF2 was coimmunoprecipitated with GITR (Fig. 4A). Only a small fraction of the total cellular TRAF2 was detected after immunoprecipitation of GITR, suggesting either a low affinity or transient interaction of the adapter protein with the receptor. To gain further insights into the kinetics of TRAF2 recruitment to GITR after ligand engagement, primary T cells were incubated with GITR-L, which caused GITR to interact with TRAF2 rapidly and indicated that receptor interaction with the adapter protein is ligand dependent (Fig. 4B). Comparing the cellular pool of TRAF2 with the amount in GITR-containing protein complexes is consistent with a transient or weak GITR-TRAF2 interaction and argues that TRAF2 molecules not engaged in GITR signaling can interact with other adapter proteins or TNFR-related proteins. These results confirm that TRAF2 interacts with GITR under physiologic conditions and may regulate early proximal stages of receptor-mediated signaling pathways.
The cytoplasmic domain of GITR causes TRAF2 relocalization
Previously, we and others have shown that interactions of TRAFs and TNFR-related molecules cause TRAF relocalization to a detergent-insoluble cellular compartment (41, 42, 43). Further, we have shown that dimerization of the cytoplasmic domains of TNFR-related proteins results in TRAF intracellular relocalization and constitutive NF-B activation (20, 31). To examine whether GITR also regulates the intracellular localization of TRAF2, TRAF2 levels in the detergent-soluble and -insoluble fractions of cell lysates were determined by Western blot analysis after cotransfection of HEK293 cells with TRAF2 and CD28-GITR, a chimeric receptor consisting of the extracellular and transmembrane regions of the homodimer CD28 fused to the cytoplasmic domain of GITR. Similar to CD28-CD30, which has been previously shown to translocate TRAF2 (41), CD28-GITR caused TRAF2 redistribution from the detergent-soluble to the -insoluble cellular fraction, suggesting that GITR controls the intracellular trafficking of TRAF2 similar to other TNFR-related proteins (Fig. 5). CD28Tail, a receptor lacking a cytoplasmic domain, did not cause TRAF2 relocalization and served as a negative control. Consistent with a weak or transient interaction between TRAF2 and GITR, expression of CD28-GITR caused less efficient relocalization of TRAF2 than CD28-CD30, which resulted in complete depletion of TRAF2 from the detergent-soluble compartment. These data argue that the functional relationship between GITR and TRAF2 is distinct from other TNFRs, such as CD30.
The zinc finger protein A20 blocks NF-B activation triggered by GITR
Multimerization of TNFR-related proteins results in the assembly of TRAF-containing signaling complexes at the cytoplasmic domains of these receptors as a proximal step in signal transduction (9). Regulation of the composition as well as disassembly of TRAF-containing protein complexes is not well understood. A20 is an anti-apoptotic adapter protein that interacts with TRAFs and inhibits NF-B activation in a negative feedback loop (44, 45, 46). When A20 was coexpressed with GITR and GITR-L in HEK293 cells, receptor-induced NF-B activation was significantly impaired (Fig. 6). This observation points to A20 as an inhibitory signaling intermediate downstream of GITR and further implies that TRAF-containing protein complexes are involved in GITR-induced signaling events.
TRAF2 inhibits NF-B activation triggered by GITR
Genetic analyses argue that TRAF2 promotes signaling events triggered by TNFR-related proteins, such as NF-B activation (22, 23, 47). To test whether TRAF2 plays a similar role downstream of GITR, HEK293 cells were cotransfected with GITR, GITR-L, and either full-length TRAF2 or TRAF2DN, a dominant-negative mutant of TRAF2 lacking the N-terminal RING finger (15, 31). TRAF2DN interacts with receptors containing TRAF-binding domains and is thought to interfere with the recruitment of other TRAF molecules to these sites (15). As expected, coexpression of TRAF2DN abolished GITR-induced NF-B activation measured by a luciferase reporter assay, consistent with the proposed role of TRAFs downstream of GITR (Fig. 7A). However, rather than amplifying GITR-induced NF-B activation, TRAF2 significantly reduced NF-B activation triggered by GITR, similar to the N-terminal deletion mutant of TRAF2. Expression of TRAF2 or its N-terminal deletion mutant alone did not affect NF-B activity indicating the specificity of the finding. As an additional specificity control, coexpression of TRAF2 with GITR-L did not alter NF-B activation (data not shown). We and others have previously shown that the ratios of receptor molecules to adapter components are pivotal to downstream signaling events triggered by TNFR-related proteins (29, 48, 49, 50). HEK293 cells did not express amounts of endogenous TRAF2 that were detectable by Western blot (Fig. 5). To rule out that the inhibitory effect of TRAF2 was caused by suboptimal molecular ratio of GITR and the adapter protein, the amount of plasmid encoding TRAF2 used to transfect HEK293 cells was titrated. Consistent with the previous result, amplification of GITR-induced NF-B activation was not observed; rather, the inhibition of GITR-induced NF-B activation varied directly with TRAF2 expression levels (Fig. 7B). These results suggest a novel inhibitory function of TRAF2 in signaling pathways triggered by GITR.
Discussion
Adapter proteins, such as TRAFs, are critical signaling intermediates downstream of TNFR-related molecules and serve as both convergent and divergent platforms triggering activation of kinases, such as IB kinases that lead to activation of transcription factors, such as NF-B. Therefore, we hypothesized that TRAFs function as crucial signaling intermediates downstream of GITR. Previously, we and others have used yeast two-hybrid systems to identify proteins interacting with the cytoplasmic domains of TNFR-related molecules and to map amino acids required for these interactions (13, 14, 31, 33). Although TRAF binding sequences from different receptors diverge, acidic residues of two weak consensus motifs–(P/S/T/A)x(Q/E)E and PxQxxD–seem critical for their interactions (39, 40). In agreement with these motifs, residues 200 to 203 (SAED) and 210 to 213 (PEEE) in the cytoplasmic domain of GITR contain a serine or proline at position P0 and acidic residues at positions P2 and P3 (Fig. 2A). In contrast, the two glutamic acid residues at positions 219 and 220 of GITR do not fit either of the TRAF-binding motifs. Consistently, residues 202/203 and 211–213 of GITR are critical for TRAF2 binding, but residues 219/220 are not (Figs. 2 and 3). Distinct TRAFs can compete for receptor-binding motifs, which agrees with the observation that TRAFs other than TRAF2 interact with the cytoplasmic domain of GITR (Fig. 2B and Refs. 39 and 40).
Co-IP analyses verified the interactions between GITR and TRAF2 in primary lymphocytes expressing endogenous levels of the proteins (Fig. 4). These studies indicate that GITR rapidly recruits TRAF2 in a ligand-dependent manner and suggest that the adapter protein can regulate early stages of GITR-induced signal transduction under physiologic conditions. Similar to the engagement of TNFR-related proteins with their ligands, chimeric receptors containing the extracellular and transmembrane domains of CD28 and the cytoplasmic domains of either CD30, Ox40, or 4-1BB trigger signaling events including intracellular TRAF relocalization, which are dependent on interactions between the TNFRs and TRAFs (31, 41). When HEK293 cells were cotransfected with CD28-GITR and TRAF2, relocalization of TRAF2 from a detergent-soluble to -insoluble fraction of lysates was observed, arguing further for TRAF2 interaction with the cytoplasmic domain of GITR (Fig. 5). Consistent with our co-IP analyses that suggested a weak or transient interaction between TRAF2 and GITR, translocation of TRAF2 induced by GITR does not significantly reduce the TRAF2 level in the detergent-soluble compartment, suggesting that the adapter protein is still available to elaborate signaling events triggered by other TNFR-related proteins expressed in the cell.
Similar to previously published reports, GITR cross-linking with its ligand or an agonistic Ab activated NF-B (Fig. 1, and Refs. 2 , 5 , and 24). To test whether TRAF interactions are critical for NF-B activation, HEK293 cells were transfected with GITR mutants sufficient or defective in TRAF2 interaction (Fig. 3C). Mutation of residues in GITR needed for TRAF interaction correlated with defects in receptor-mediated NF-B activation. Because TRAFs bind overlapping sites in TNFR-related proteins, this result argues that TRAFs in addition to TRAF2 may elaborate GITR-induced signaling pathways, which is currently being examined. Consistent with this hypothesis, A20, an adapter protein that interacts with and regulates TRAF function, inhibited GITR-induced NF-B activation (Fig. 6 and Refs. 41 , 45 , and 51). Moreover, we recently showed that TRAF4 enhances NF-B activation induced by GITR (29). The observations that TRAF4 did not interact with GITR in yeast, yet required the TRAF interaction site to augment NF-B activation further implies that other TRAFs or adapter proteins initiate GITR-induced signaling events (Fig. 2B and Ref. 29).
Importantly, functional studies revealed intriguing differences between TRAF2-mediated signaling triggered by GITR and other members of the TNFR superfamily including AITR (26, 27). In contrast to previously published findings that suggest a role of TRAF2 as a positive regulator of NF-B activation triggered by a variety of TNFR-related molecules, TRAF2 significantly inhibited GITR-mediated activation of NF-B similar to the expression of a dominant-negative mutant of TRAF2 (Fig. 7 and Refs. 22 , 23 , and 47). TRAF2 is known to dampen specific signaling pathways triggered by TNFR1 (23). Furthermore, TRAF2 has been implicated as an inhibitor of Th2 responses through its interaction with NFAT-interacting protein 45 and as a negative regulator of the noncanonical pathway of NF-B activation in mature B cells (52, 53). Whether TRAF2 negatively regulates signaling beyond competing with other signaling proteins for GITR binding is unclear. Intriguingly, TRAF2 has been shown recently to function as an E3 ubiquitin ligase and to undergo ubiquitination itself (42, 54, 55). Future studies will test whether ubiquitination of TRAF2 or other proteins plays a role in the novel regulatory role of TRAF2 in GITR-induced NF-B activation.
Taken together, our findings indicate that TRAF2 is a proximal signaling intermediate downstream of GITR and suggest a novel regulatory role for TRAF2 in NF-B activation triggered by GITR. As GITR is a critical regulator of the interface between Treg cells and immune effector cells, gaining further insight into the mechanisms of signal transduction pathways triggered by GITR will allow the development of therapeutic approaches targeting inflammatory immune responses and autoimmunity.
Footnotes
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.
1 This research was supported in part by National Institutes of Health Grants 5 RO1 HL67312-02 and T32 HL007317 and an Investigator Award of the Cancer Research Institute (to R.H.A.).
2 Address correspondence and reprint requests to Dr. Robert H. Arch, School of Medicine, Washington University, Campus Box 8052, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address: arch{at}wustl.edu
3 Abbreviations used in this paper: GITR, glucocorticoid-induced TNFR; Treg, regulatory T; TRAF, TNFR-associated factor; AITR, activation-induced TNFR-related receptor; IP, immunoprecipitation; GITR-L, GITR-ligand; HEK, human embryonic kidney; DBD, DNA-binding domain; AD, activation domain; 3-AT, 3-aminotriazole; -gal, -galactosidase.
Received for publication February 11, 2005. Accepted for publication April 18, 2005.
References
Croft, M.. 2003. Co-stimulatory members of the TNFR family: keys to effective T-cell immunity?. Nat. Rev. Immunol. 3: 609-620.
Tone, M., Y. Tone, E. Adams, S. F. Yates, M. R. Frewin, S. P. Cobbold, H. Waldmann. 2003. Mouse glucocorticoid-induced tumor necrosis factor receptor ligand is costimulatory for T cells. Proc. Natl. Acad. Sci. USA 100: 15059-15064.
Ronchetti, S., O. Zollo, S. Bruscoli, M. Agostini, R. Bianchini, G. Nocentini, E. Ayroldi, C. Riccardi. 2004. Frontline: GITR, a member of the TNF receptor superfamily, is costimulatory to mouse T lymphocyte subpopulations. Eur. J. Immunol. 34: 613-622.
Kanamaru, F., P. Youngnak, M. Hashiguchi, T. Nishioka, T. Takahashi, S. Sakaguchi, I. Ishikawa, M. Azuma. 2004. Costimulation via glucocorticoid-induced TNF receptor in both conventional and CD25+ regulatory CD4+ T cells. J. Immunol. 172: 7306-7314.
Esparza, E. M., R. H. Arch. 2005. Glucocorticoid-induced TNF receptor functions as a costimulatory receptor that promotes survival in early phases of T cell activation. J. Immunol. 174: 7869-7874.
Shimizu, J., S. Yamazaki, T. Takahashi, Y. Ishida, S. Sakaguchi. 2002. Stimulation of CD25+CD4+ regulatory T cells through GITR breaks immunologic self tolerance. Nat. Immunol. 3: 135-142.
McHugh, R. S., M. J. Whitters, C. A. Piccirillo, D. A. Young, E. M. Shevach, M. Collins, M. C. Byrne. 2002. CD4+CD25+ immunoregulatory T cells: gene expression analysis reveals a functional role for the glucocorticoid-induced TNF receptor. Immunity 16: 311-323.(Edward M. Esparza and Rob)