CD40/CD40 Homodimers Are Required for CD40-induced Phosphatidylinositol 3-Kinase-dependent Expression of B7.2 by Human B Lymphocytes*

Preformed CD40/CD40 homodimers were initially observed on human Burkitt lymphoma cell lines, normal B cells, and transitional bladder carcinoma cell lines. However, the nature and the biological relevance of these homodimers have not yet been investigated. In the present study, we demonstrated that Epstein-Barr virus-transformed B cells and CD40-transfected HEK 293 cells constitutively expressed disulfide-linked CD40/CD40 homodimers at low levels. Oligomerization of CD40 leads to a rapid and significant increase in the disulfide-linked CD40/CD40 homodimer formation, a response that could be prevented using a thiol-alkylating agent. Formation of CD40/CD40 homodimers was found to be absolutely required for CD40-mediated activation of phosphatidylinositol 3-kinase, which, in turn regulated B7.2 expression. In contrast, CD40 monomers provided the minimal signal emerging from CD40, activating p38 MAP kinase and inducing homotypic B cell adhesion. CD40/CD40 homodimer formation was totally independent of TRAF1/2/3/5 associations with the threonine at position 254 in the cytoplasmic tail of the CD40 molecules. This finding may be vital to better understanding the molecular mechanisms that govern cell signaling triggered by CD40/CD154 interactions.

The CD40 molecule is a 45-50-kDa type I phosphorylated glycoprotein that belongs to the tumor necrosis factor receptor superfamily and was initially considered to be a growth factor receptor on B cells (1). The interaction of CD40 with its natural ligand, CD154, is a crucial step in T cell-dependent B cell activation and differentiation as demonstrated in patients suffering from the X-linked hyper-IgM syndrome (2) and then confirmed in CD154 (3) and CD40 knock-out mice (4). Although CD40 has no kinase domain, its ligation induces the activation of protein tyrosine kinases (including lyn, fin, and syk) (5), the adaptor protein jak3 (6), the phosphatidylinositol-3 kinase (PI 3-kinase) 1 (7), the phospholipase C␥2 (7), and the mitogen-activated protein (MAP) kinases p38, c-Jun N-terminal kinase/ stress-activated protein kinase (JNK/SAPK), and extracellular signal-regulated kinase 1 or 2 (ERK1/2) (8 -10). Most of these early events are mediated by the association of CD40 via its cytoplasmic domain with several members of the tumor necrosis factor receptor-associated factor (TRAF) family (11)(12)(13)(14)(15)(16)(17).
The cellular events triggered in CD40-positive cells following CD40/CD154 interactions are not restricted to B cells only but can also be triggered in dendritic cells, monocytes, epithelial cells, endothelial cells, and fibroblasts (1). Thus, CD40 ligation mediates a broad variety of immune and inflammatory responses (1,18). Abnormal CD40 signaling seems to be directly associated with the pathogenesis of chronic inflammatory diseases such as autoimmune diseases (19), neurodegenerative disorders (20), graft-versus-host diseases (21), cardiovascular diseases (22), and cancer (23). Consequently, elucidation of the molecular mechanisms that govern CD40-mediated signal transduction is critical not only in explaining how CD40 exerts its multifunctional outcome but also in developing new therapeutic agents and strategies leading to blockage of specific CD40-induced responses.
Although it is well established that fully active CD154 is mainly present in a trimeric form at the cell surface and/or following release from the cell membrane (1,18), little research has been conducted on the possible existence of CD40/CD40 homodimers and their role in immune responses. It was originally reported in 1989 that bladder carcinoma cell lines and human B cell lines constitutively express low levels of CD40/ CD40 homodimers (24). Since this discovery, there has only been one report demonstrating that CD40/CD40 homodimers are formed as a consequence of B cell activation (25). However, a functional link between these homodimers and the downstream signaling events affecting key B cell functions has not yet been elucidated.
Here, we report that, although low levels of disulfide-linked CD40/CD40 homodimers were present on EBV-positive and CD40-transfected cells, they rapidly increased after CD40 oligomerization. Whereas engagement of CD40 monomers generated the minimal signal required to activate the MAP kinase p38 and induce homotypic B cell adhesion, the formation of CD40/CD40 homodimers was required for CD40-mediated activation of PI 3-kinase and the subsequent up-regulation of B7.2 expression. CD40/CD40 homodimer formation did not depend on TRAF1/2/3/5 association with the threonine at position 254 of the CD40 cytoplasmic tail. These findings suggest that disulfide-linked CD40/CD40 homodimers play a pivotal role in CD40-induced responses MATERIALS AND METHODS Antibodies, Chemicals, and Reagents-The hybridomas producing the mouse mAbs directed against human CD40 (G28-5; IgG1) and human CD154 (5C8; IgG2a) were obtained from American Type Culture Collection (Manassas, VA). The monoclonal anti-B7.2 (Bu86; IgG1) Ab was kindly provided by Dr. Michel Tremblay (Infectiology Unit Research, Laval University, Québec, Canada). The irrelevant IgG1 and IgG2a isotype controls were produced in our laboratory. Polyclonal rabbit anti-CD40, goat anti-mouse IgG, anti-p38, anti-AKT, and their anti-phosphorylated counterparts as well as the horseradish peroxidase-conjugated anti-mouse and anti-rabbit IgG were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Trimeric CD154 was a generous gift from Immunex Corp, Seattle, WA (26). Cell culture agents and electrophoresis grade chemicals were purchased from Sigma. The MAP kinase p38 inhibitor SB203580 and the PI 3-kinase inhibitor LY294002 were from Calbiochem.
Plasmid Constructs and Site-directed Mutagenesis-The CD40-T254A mutant was generated by site-directed mutagenesis using the QuikChange® XL kit according to the manufacturer's instructions (Stratagene). Wild-type CD40 and mutant CD40-T254A constructs were subcloned into the pCEP4 expression vector (Clontech). The mutation T254A was confirmed by the entire sequencing of the inserts.
Cell Culture and Stable Transfection-All B cell lines, mouse LTK Ϫ fibroblast cells, and HEK 293 cells were obtained from the American Type Culture Collection. Human tonsillar B cells were isolated as described previously (27). The mouse fibroblast LTK Ϫ cells were stably transfected with a gene encoding for the wild-type form of the human CD154 protein, as described previously (28). HEK 293 cells were cultured at 75% confluency in 6-well plates and transiently transfected with the expression vectors pCEP4 (mock-transfected cells), pCEP4 wild-type CD40, and pCEP4 CD40-T254A using the Clontech calcium phosphate precipitation kit. Fluorescence-activated cell sorter analysis using the anti-CD40 mAb G28-5 was performed 24 h post-transfection (29).
Induction of CD40/CD40 Homodimers-B cells and HEK 293 transfectants were diluted to 10 7 cells/ml and 1.5 ϫ 10 6 cells/ml, respectively. G28-5 was added in the culture media to a final concentration of 5 g/ml, and the cells were incubated on ice for 15 min. Washed cells were then resuspended with 5 g/ml anti-mouse IgG and incubated for 10 min at 37°C. In some experiments, B cells were stimulated with trimeric, soluble CD154 (5 g/ml) for 30 min at 37°C. For co-incubation with mouse fibroblasts, 7.5 ϫ 10 5 LTK Ϫ -untransfected (m-CD154Ϫ) and LTK Ϫ -CD154-transfected cells (m-CD154ϩ) were left to adhere for at least 24 h before adding 3 ϫ 10 6 serum-deprived (3 h) B cells. B cells were gently collected using ice-cold Hanks' buffer after a 30-min coincubation period at 37°C.
Isolation of mRNA and Northern Blot Analysis-After stimulation, B cells were homogenized in Trizol reagent (Invitrogen). The homogenates were extracted with chloroform, and total RNA was precipitated with 80% isopropanol. Twenty micrograms of total RNA was fractionated in 1.2% agarose gels and transferred to nylon membranes (Amersham Biosciences). Blots were probed with 32 P-labeled human B7.2 followed by glyceraldehyde-3-phosphate dehydrogenase (GAPDH).
Cell-Cell Adhesion-CD40-induced homotypic adhesion of human B cells (BJAB) was monitored in co-culture and anti-CD40 mAb systems. In the co-culture system, 4 ϫ 10 5 B cells were co-incubated with 1 ϫ 10 5 LTK Ϫ -untransfected (m-CD154Ϫ) or LTK Ϫ CD154-transfected (m-CD154ϩ) cells seeded previously in 24-well plates. In the mAb system, 2 ϫ 10 5 B cells were seeded in 96-well plates in the presence or absence of 1 g/ml G28-5. Cell-cell adhesion was assessed by phase-contrast microscopy at the times indicated in the Fig. 5 legend.

RESULTS
CD40 Engagement Rapidly Induces CD40/CD40 Homodimer Formation on Human B Cells-Studies on CD40/CD154 interactions demonstrated that various degrees of CD40 cross-linking result in different cellular responses (30), an outcome that could be attributed to the ability of CD154 to trigger the formation of CD40 trimers and/or oligomers of different extents. We first investigated the possible existence of CD40 trimers or oligomers in different human B cell lines and in resting B cells. For this purpose, cell lysates were analyzed by Western blot under non-reducing conditions. As shown in Fig. 1, all inactivated B cell lines expressed the typical 43-and 48-kDa CD40 protein profiles, whereas EBV-positive LG2 B cells also expressed low levels of an 85-90 kDa protein band, a molecular mass range that corresponds to the expected molecular mass range of CD40/CD40 homodimers (24). Interestingly, a significant increase in CD40/CD40 homodimers on LG2 B cells was induced after 30 min of stimulation with m-CD154 (in the co-culture system), and soluble CD154 (s-CD154). The formation of CD40 dimers was slightly enhanced when B cells were stimulated for 15 min with a bivalent anti-CD40 mAb alone, whereas CD40 dimerization was induced at similar levels to that of CD154 when anti-CD40 mAbs were cross-linked by secondary Abs (Fig. 1, A and B). CD40/CD40 homodimer formation was not restricted to EBV-positive B cells alone, because CD40 dimerization was also rapidly induced in EBV-negative BJAB, Ramos, and human tonsillar B cells (Fig. 1B). In all cases, the formation of CD40/CD40 homodimers was mirrored by a decrease in the pool of CD40 monomers. The absence of CD40 complexes with molecular masses equal to or greater than 120 -150 kDa, even after a long exposure of the autoradiogram (data not shown), indicated that oligomerization of CD40 induces the formation of only stable CD40/CD40 homodimers.

CD40/CD40 Homodimer Formation Is Mediated by a Disulfide Bond and Completely Prevented by Alkylation of Free Sulf-
hydryl Groups-Receptor homodimerization is not unique for CD40, as it has been observed in the case of type II tumor necrosis factor receptors (31). In this case, the homodimer is mediated by disulfide bond formation and pretreatment with N-ethylmaleimide, which alkylates free sulfhydryl groups, inhibits dimer formation of these receptors on live human erythroleukemia K562 cells (31). Based on these observations and those reported in Fig. 1 showing that CD40/CD40 homodimers are very detergent-(2-3% SDS) and heat-stable (100°C for 5 min), we hypothesize that homodimer formation is mediated by disulfide bonds in living B cells. To confirm our hypothesis, two complementary approaches were used. In the first approach, B cells were pretreated with 1 mM IA or 1 mM DTT, a well known disulfide-reducing agent, for 5 min at 37°C before stimulation. Cells were washed extensively and stimulated with membranebound CD154. Cell lysates were analyzed under non-reducing conditions by SDS-PAGE and Western blot analysis using rabbit anti-CD40 polyclonal antibody. Fig. 2A shows a representative result for all B cells tested. The formation of CD40/CD40 homodimers was prevented when the cells were pretreated with the irreversible thiol-alkylating agent IA but not when they were pretreated with DTT. In the second approach, once CD40/CD40 homodimerization had been induced, cell lysates were heat-treated in the presence of 1 mM DTT or 1 mM IA and analyzed by SDS-PAGE and Western blot as described above. Fig. 2B shows that treatment of cell lysates with DTT completely disrupted constitutive as well as induced CD40/CD40 homodimerization. In contrast, CD40/CD40 homodimers were detected at approximately the same levels regardless of whether or not the thiol-alkylating agent IA was added to the samples. Similar results were obtained with cells (BJAB, Ramos, and LG2 B) stimulated with cross-linked anti-CD40 mAbs (Fig. 2C). These results indicated that CD40/CD40 homodimerization was mediated by disulfide bond formation and that alkylation of free sulfhydryl groups completely inhibited homodimer formation.
CD40/CD40 Homodimer Formation Is Required for CD40induced Phosphatidylinositol-3 Kinase Activation in Human BJAB B Cells-Based on the above results and recent observations that different degrees of CD40 cross-linking result in different cellular responses (30,32), we explored the possibility that the engagement of CD40/CD40 homodimers could be involved in the activation of at least one classic CD40-induced signaling pathway. BJAB B cells were left untreated or pretreated with 1 mM IA and then activated in the co-culture system. Protein samples from control and IA-pretreated cell lysates were Western blotted and probed with polyclonal Abs directed against the activated, phosphorylated form of AKT, a specific PI 3-kinase substrate, and p38. Our results showed that, in control BJAB B cells, CD40/CD40 homodimer formation ( Fig. 2) was correlated with PI 3-kinase and p38 activation (Fig. 3A). Interestingly, in conditions in which IA pretreatment inhibited CD40/CD40 homodimer formation, only the PI 3-kinase pathway was completely abolished, as shown by the almost total lack of phosphorylation of AKT and the slightly increased phosphorylation of p38 in IA-pretreated B cells (Fig.  3A). The specificity of IA treatment with regard to CD40/CD40induced responses was determined by analyzing the effect of IA on oxidative stress stimulation by hydrogen peroxide (H 2 O 2 ), an agent that is known to induce a similar response in BJAB B cells (33). Cells were pretreated with IA, washed, and stimulated with 1 M H 2 O 2 for 10 min at 37°C. Our results (Fig. 3B) demonstrated that hydrogen peroxide-induced activation of PI 3-kinase and p38 was unaffected by IA pretreatment, confirming the specificity of the IA treatment with regard to CD40/ CD40-induced responses.
CD40/CD40 Homodimer Formation Is an Absolute Requirement for B7.2 Expression-To further explore the biological outcomes of CD40/CD40 homodimer engagement, we analyzed the transcription of B7.2, a strong T-cell co-stimulatory molecule regulated following CD40 ligation on B cells (34,35). Our results indicate that CD40 ligation by m-CD154 enhanced B7.2 transcription on BJAB B cells after a 2-h co-incubation period (Fig. 4A) and B7.2 protein expression at the cell surface after 24 h of co-incubation (Fig. 4B). Pretreatment of BJAB B cells with IA completely inhibited CD40-mediated up-regulation of B7.2 expression at both the mRNA and protein levels, strongly suggesting that the CD40/CD40 homodimer formation is an absolute requirement for B7.2 expression. tion of B7.2 mRNA (Fig. 5A). The activation of PI 3-kinase and the expression of B7.2 were unaffected following m-CD154 stimulation of BJAB B cells incubated in the presence of 10 M SB203580, a specific inhibitor of p38 phosphorylation (data not shown). To further support the role of CD40/CD40 homodimers in CD40-induced, PI 3-kinase-dependant expression of B7.2, BJAB B cells were stimulated with G28-5 alone or G28-5 followed by a secondary cross-linker Ab. As shown in Fig. 5B, stimulation with G28-5 slightly increased the activation of PI 3-kinase after a 15-min period of stimulation. In the same experiment, BJAB B cells were stimulated with G28-5 for 5 min, washed, and incubated in the presence of secondary Abs for 10 min. Under these conditions, PI 3-kinase activation was enhanced at level similar to that with m-CD154 (Fig 5, A versus  B) and, as shown in Fig. 1B, the levels of CD40 dimers were significantly increased in BJAB B cells (data not shown). On the other hand, the expression of B7.2 mRNA was induced when G28-5 was cross-linked, but not in the presence of G28-5 alone (Fig. 5B). On the basis of these results, we concluded that CD40-mediated activation of PI 3-kinase and the subsequent up-regulation of B7.2 mRNA expression are critically dependent on CD40/CD40 dimer formation.
CD40-induced Homotypic B cell Adhesion Was Totally Independent of CD40/CD40 Homodimer Formation-Engagement

FIG. 2. CD40/CD40 homodimer formation is prevented on B cells pretreated with thiol-alkylating agents and mediated by a disulfide bond. A,
LG2 B cells were pretreated with media alone or media containing 1 mM IA or 1 mM DTT for 5 min at 37°C, washed, cocultured with mouse untransfected (m-CD154Ϫ) or CD154-transfected fibroblasts (m-CD154ϩ) for 30 min. LG2 cells were gently harvested and then lysed in a Tris buffer containing 2% SDS and 10% glycerol. B, untreated LG2 B cells were co-cultured for 30 min with mouse untransfected (m-CD154Ϫ) or CD154-transfected fibroblasts (m-CD154ϩ), harvested, and diluted in a Tris buffer containing 2% SDS and 10% glycerol in the absence or presence of 5 mM DTT or 5 mM IA. C, CD40/CD40 homodimerization on BJAB, Ramos, and LG2 B cells induced by cross-linked anti-CD40 antibodies (G28-5 mAb plus goat anti-mouse IgG (␣IgG)) was also disrupted by DTT but not IA. Monomeric (M) and dimeric (D) forms of CD40 molecules were revealed using polyclonal rabbit anti-CD40 antibody. Similar results were obtained in three other experiments.

FIG. 3. CD40/CD40 homodimers formation is required for CD40-induced phosphatidylinositol-3 kinase activation in human BJAB B cells.
A, m-CD154-mediated PI 3-kinase but not p38 activation was inhibited on IA-pretreated BJAB B cells. B, IA treatment did not affect PI 3-kinase or p38 activation on BJAB B cells induced within 10 min of stimulation with 1 M hydrogen peroxide. Similar results were obtained in three other experiments. AKT-P, phosphorylated AKT; p38-P, phosphorylated p38. of CD40 molecules induces B cells to proliferate, differentiate, and form homotypic cellular aggregates. Cell-cell adhesion events play critical roles in the sequential migrations and multiple specific cell-cell interactions that B cells undergo during normal development and function. A recent study demonstrated that the interference of CD40 signaling in spontaneous or B cell antigen receptor (BCR)-mediated apoptosis of human B cells requires minimal cross-linking, whereas promoting homotypic cell-cell adhesion and proliferation requires intermediate and high levels of CD40 oligomerization, respectively (32). Because strong B cell adhesion can be generated using bivalent anti-CD40 mAbs, such as G28-5, and given that this mAb does not induce high levels of CD40/CD40 homodimerization ( Figs. 1 and 2), we hypothesized that activation via noncovalent ligation of CD40 monomers would provide the minimal signal to induce homotypic cell-cell adhesion. Strong cell-cell adhesion of BJAB B cells was induced after a 6-h incubation period in the co-culture system (Fig. 6A), a cell response that was inhibited in the presence of anti-CD154 5C8 Abs (data not shown). The formation of these cellular aggregates was unaffected by IA-pretreatment (1 mM) of BJAB B cells (Fig. 6A). These results were confirmed using IA-pretreated BJAB B cells stimulated with anti-CD40 mAb G28-5 (1 g/ml). Strong cellular aggregates at levels similar to untreated cells were observed after a 3-h stimulation period (Fig. 6B). Taking together, these results indicate that B cell homotypic adhesion in response to CD40 ligation is totally independent of disulfide-linked CD40/ CD40 homodimer formation.
CD40/CD40 Homodimer Formation Was Totally Independent of TRAF1/2/3/5 Association with Cytoplasmic Threonine 254 of CD40 Molecules-CD40 multimerization during cell signaling seems to be a critical event in TRAF recruitment (36,37). Indeed, the threonine 254 residue mainly mediated CD40-induced signaling and CD40-association with different TRAF members, in particular TRAF1/2/3 and 5 (11). To determine the involvement of this residue in CD40/CD40 dimer formation, HEK 293 cells expressing CD40 wild-type or CD40 mutated at residue T254 (CD40-T254A) were used. Cells were stimulated with G28-5 alone or G28-5 cross-linked with a secondary Ab and then analyzed for CD40/CD40 homodimer formation as above. Our results (Fig. 7) clearly demonstrate that CD40/ CD40 homodimer formation was induced at similar levels on cells expressing the CD40 wild-type and the T254A mutant. This indicated that CD40/CD40 homodimer formation did not depend on a CD40-TRAF association, at least for TRAF1/2/3 and 5, which interact with the cytoplasmic threonine at position 254 in the CD40 molecule. DISCUSSION Similarly to other ligand/receptor interactions, the binding of various members of the tumor necrosis factor family to their receptors triggers dimerization or even oligomerization of the relevant receptors (37,38). It is well established that such an event is commonly required for maximal signal transduction (39). However, little is known about the stability and nature of these complexes. Data obtained in the course of the present investigation show that low levels of preformed and stable CD40/CD40 homodimers are consistently present on some human B cell lines and CD40-transfected cells. CD40 ligation with trimeric CD154 (soluble and membrane-bound) or with cross-linked anti-CD40 mAbs leads to a rapid and significant increase in the levels of the stable CD40/CD40 homodimers in different cell types, including normal tonsillar B cells. The CD40/CD40 homodimer was mediated by disulfide bond formation, and alkylation of free sulfhydryl groups completely inhibited homodimer formation. The failure to detect CD40 complexes with molecular masses Ն120 -150 kDa, even after a long exposure of the autoradiogram, indicated that ligation of CD40 leads to the formation of only stable CD40/CD40 homodimers. Together, these results strongly suggest that free sulfhydryl groups are already present in the extracellular domain of CD40 monomers and that specific residues are involved in the formation of disulfide bounds between two CD40 monomers. The ligand-induced formation of a disulfide-linked dimer is not unique for CD40 molecules, as it has also been observed in the case of tumor necrosis factor receptor II in the human erythroleukemia K562 cells following stimulation with the tumor necrosis factor (31), suggesting that the rapid formation of disulfide-linked homodimers is a feature that is also shared by other members of tumor necrosis factor receptor superfamily.
Consequently, we hypothesized that ligation of disulfidelinked CD40 dimers could induce or overlap the biological responses obtained after ligation of CD40 monomers in a noncovalent way. Our results demonstrated that activation of the MAP kinase p38 and induction of homotypic cell adhesion were triggered by m-CD154 or G28-5 mAb alone under conditions in which sulfhydryl groups were blocked leading to the inhibition of CD40 dimerization. These findings suggest that such biological functions do not require high levels of CD40 oligomerization to be triggered. On the other hand, our results clearly demonstrated that CD40-mediated activation of PI 3-kinase, which regulated B7.2 expression, required the formation and engagement of CD40/CD40 homodimers. This is supported by the observation that significant levels of CD40 dimer formation, PI 3-kinase activation, and B7.2 gene expression were induced only when anti-CD40 Abs were cross-linked with a secondary Ab. In contrast, ligation of two CD40 monomers with G28-5 mAb alone induced little CD40 dimerization and PI 3-kinase activation and, subsequently, did not increase B7.2 gene expression.
The different responses obtained with a bivalent ligand (anti-CD40 mAb) compared with those obtained with a trimeric CD154 suggest that increased CD40 dimer formation and PI 3-kinase activation are dependent on the extent of receptor oligomerization. In this context, our results favor a model in which CD40 dimers could provide the structural requirement for the rapid extension of receptor oligomerization and maximal signal transduction. This idea is compatible with initial reports (reviewed in Ref. 30) demonstrating that proliferation of resting B cells is significantly enhanced by trimeric CD154 in a membrane-bound or a soluble form and, to a lesser extent by dimeric CD154 or immobilized anti-CD40 mAbs. To mimic the effects of trimeric CD154 on B cell proliferation, soluble anti- However, because of their inability to cross-link CD40 molecules, monomeric CD154 and F(ab) fragments of CD40 mAbs are unable to trigger B cell proliferation, but they do maintain the ability to inhibit spontaneous or B cell antigen receptorinduced apoptosis of B cells (30). More recently, Pound et al. (32) reported that CD40-dependent rescue of human B cells from apoptosis requires minimal cross-linking, whereas the promotion of homotypic adhesion and cell proliferation requires a more stringent level of receptor cross-linking. These authors also demonstrated that anti-CD40 mAbs alone did not effect proliferation but that the cross-linking of anti-CD40 mAbs by a secondary Ab strongly enhances proliferation of peripheral blood B cells (32). As a general conclusion, it is inferred that formation of high ordered CD40 oligomers on the B cell surface is a prerequisite for entry into the cell cycle following CD40 ligation. In an effort to test this hypothesis, a trimeric and a dodecameric form of CD154 were engineered by Haswell et al. (35), and the function of each was compared with its ability to induce B cell proliferation and the expression of the co-stimulatory molecules ICAM-1, B7.2, and MHC class II. Although both soluble CD154 forms bond CD40 with similar affinity and kinetic parameters, the dodecameric form (potentially able to bind 12 CD40 molecules) was more potent than the trimeric form in inducing B cell proliferation and ICAM-1 and B7.2 expressions but had similar activity in the induction of MHC class II expression. These results suggest that the signaling pathways governing the up-regulation of MHC class II are distinct from those required for triggering B cell proliferation or from the expression of ICAM-1 and B7.2. Based on our results, it is tempting to speculate that the activity of high ordered CD154 oligomers is directly linked with their ability to induce CD40 dimer formation and, subsequently, the activation of PI 3-kinase, which, in turn could regulate B cell proliferation (40, 41) and the expression of T cell co-stimulatory molecules such as B7.2. Although B7.2 is known to be regulated by CD40 ligation on B cells (34), our results demonstrated for the first time that CD40-mediated B7.2 expression depends on PI 3-kinase activation. Thus, CD40 dimer formation by CD154 in vivo could provide an additional mechanism for regulating CD28/B7.2 interactions that are essential for initiating antigen-specific T cell responses, up-regulating cytokine expression, and promoting T cell expansion and differentiation (42)(43)(44)(45).
In an attempt to determine whether TRAF1/2/3 and 5-induced signal is involved in the dimer formation, HEK 293 cells transfected with CD40 wild-type or CD40 mutant (Thr-254 was substituted by Ala) were stimulated with the anti-CD40 mAb followed by a second Ab and then analyzed for the formation of CD40/CD40 dimers. Our results clearly demonstrated that CD40 dimer formation was induced to similar levels on both CD40 wild-type and the mutant T254A transfectants. In contrast with our conclusions, Baker et al. (25) has shown that soluble, trimeric CD154 was unable to induce CD40 homodimer formation in a mutant CD40-T254A Rat-1 fibroblast. The discrepancy between these two studies could be explained by the different expression level of CD40 at the cell surface as well as by the duration of the stimuli. In our studies, the HEK 293transfected cells expressed similar levels of CD40 wild-type and CD40 mutant, and the cells were stimulated for 10 -30 min. In the studies by Baker et al. (25), however, cells transfected with the CD40 wild-type expressed significantly high levels of CD40 as compared with CD40 mutant, and dimer formation was analyzed after 72 h of stimulation. Taken together, it is highly likely that the levels of expression of CD40 at the cell surface plays an important role in the dimer formation and that CD40/CD40 dimer formation is independent of CD40-TRAF association, at least for TRAF1/2/3 and 5 that interacted with the cytoplasmic threonine at position 254 in CD40 molecules.
CD40 clustering during cell signaling seems to be a critical event for the juxtaposition of the cytoplasmic domains of the receptors and, subsequently, in the recruitment of adaptor signaling molecules such TRAFs and Jak3 (6,11,37). Important B cell effector functions, such as expression of B7 molecules and secretion of IL-6 and immunoglobulins, are mainly mediated by the association of CD40 with TRAF-6 (46). Interestingly, the association of CD40 with TRAF-6 requires high levels of CD40 oligomerization as provided by membranebound CD154 stimulation but not by anti-CD40 mAbs (46,47). In the same way, Pullen et al. (37) has recently demonstrated that stimulation with trimeric CD154 was sufficient for optimal NF-B and p38 activation through wild-type CD40 transiently transfected in HEK 293 cells. In contrast, a higher degree of CD40 multimerization, as obtained by ligation with hexameric CD154, was necessary for maximal signaling in HEK 293 cells expressing a mutated CD40 (T254A) that signaled only through TRAF-6 (36,37). From these observations and our results, we suggest that CD40 homodimer formation is an upstream event that could be a prerequisite to recruit and trigger cell signaling mediated through TRAF-6 binding. This idea is supported by a recent study demonstrating that hexameric CD154 induces a physical interaction of CD40 molecules with PI 3-kinase and the adaptor molecule Cbl-b on human B cells (48). TRANCE and CD154-mediated PI 3-kinase-dependent Akt activation on dendritic cells and B cells required the association of CD40 with Cbl-b and, probably, the related protein c-Cbl with TRAF-6 and the Src kinase family to form an efficient signaling complex (48). The possible involvement of Src kinase family, TRAF-6, and other mechanisms (domains and/or specific residues) that lead to CD40/CD40 homodimers formation are actually tested in our laboratory.
Based on our results, we propose a mechanism to explain the multifunctional properties of CD40/CD154 interactions. First, ligation of CD40 monomers by trimeric CD154 is involved in early signaling that is independent of CD40/CD40 homodimer formation and can lead to increased adhesion, favoring interactions of CD40-positive APCs with CD154-expressing T cells and the p38 MAP kinase pathway that may be involved, for example, in early rescue signals in B cells. Second, recruitment of CD40 monomers favors the formation of CD40 complexes, which leads to the formation of CD40/CD40 homodimers (which induce TRAF recruitment), the activation of signaling pathways such as PI 3-kinase, and the translocation/activation of a subset of nuclear transcription factors (such as nuclear factor B (NF-B), cAMP-responsive element (CRE), nuclear factor of activated T cells (NF-AT), and AP-1) that, in turn, govern the expression of some genes (B7.2, for example). In this context, the type of TRAF associated with CD40, the extent of TRAF recruitment, and the magnitude of the biological responses would be highly dependent on the formation of CD40/CD40 homodimers. Blockade of CD40 dimerization could be an attractive target of inhibition of specific biological outcomes following CD40 activation, known to be implicated in several human diseases, such as rheumatoid arthritis and atherosclerosis.