Reprints
PubMed
CAVEAT LECTOR
Cees van Kooten and Jacques Banchereau
Dept of Nephrology, Leiden University Hospital, Leiden, the Netherlands and Schering-Plough, Laboratory for Immunological Research, Dardilly, France
Received 11/14/96; Accepted 12/02/96; On-line 01/01/97
3. IMMUNE REGULATION BY CD40-CD40-L INTERACTIONS
3.1. Structure and expression of CD40 and its ligand
The CD40 antigen is a 45-50 kDa glycoprotein of 277 AA, which belongs to the Tumor Necrosis Factor Receptor family (7). The 193 AA extracellular domain is composed of four imperfect repeats of ~40 residues, anchored by a superimposable pattern of six cysteines. This organization is found in the other members of the family including: the p75 low-affinity Nerve Growth Factor Receptor; the p55/CD120a and p75/CD120b receptors for Tumor Necrosis Factor; the receptor (TNFR rp) for the Iymphotoxin alpha / Iymphotoxin ß membrane complex; CD27; CD30; OX 40; 4-1-BB; FAS/CD95; two viral homologs of the TNF receptors. The X-ray crystal structure of the complex formed by soluble binding domains of p55 TNFR and a Iymphotoxin alpha trimer suggests that CD40 looks like a slightly bent rod. Three CD40 subunits are likely to engage the CD40-L trimer and scrupulously avoid contact between each other. The mouse CD40 gene, composed of nine exons that span a 16.3 kb of genomic DNA, is located on the distal region of chromosome 2 which is syntenic to human chromosome 20q11-q13 where the human CD40 gene is located.
CD40 is expressed by multiple cell types. In the hematopoietic system, it is expressed on CD34+ hematopoietic progenitors, B cell progenitors, mature B Iymphocytes, plasma cells, monocytes, dendritic cells, eosinophils, basophils and on some T Iymphocytes. CD40 is also expressed on non-hematopoietic cells such as endothelial cells, fibroblasts and epithelial cells (5)(figure 1).
Figure 1: Functional consequences of CD40 triggering on CD40 expressing cell types. Activation includes increase in cell size and alteration of phenotype.
In 1992, expression cloning using a CD40-Fc fusion protein allowed the isolation of a CD40-ligand (CD40-L) from activated T cells (8). The human CD40-L is a polypeptide of 261 AA including a 215 AA extracellular domain with five cysteins. CD40-L is a member of the Tumor Necrosis Factor family that includes TNFalpha, LTalpha, LTß, CD27-L/CD70, CD30-L, 4-1BB-L, OX40-L, FAS-L and TRAIL/APO2 (7, 9, 10). The gene for CD40-L is located on the X-chromosome at position q26.3-q23.1. It spans over 12-13 kb and consists of five exons. CD40-L is expressed on CD4+ and CD8+ T cells readily after activation through pathways that are inhibited by Cyclosporin A (11). Basophils, eosinophils and activated B Iymphocytes have also been reported as expressing CD40-L. In patients with SLE, CD40-L has been found hyperexpressed on both T and B lymphocytes (12, 13).
Like all other members of the TNF-R family, CD40 has no kinase domain and no known consensus sequence for binding to kinases. Yet, CD40 ligation activates protein-tyrosine kinases, including Lyn and Syk and induces the tyrosine phosphorylation of multiple substrates including phosphatidylinositol 3-kinase and phospholipase Cgamma2. The phosphorylation of the latter is consistent with the anti-CD40 induced IP3 production. The inhibitory effects of protein-tyrosine kinase inhibitors and of CD45 crosslinking demonstrate the functional relevance of the CD40-induced tyrosine phosphorylation (14). CD40 ligation also appears to activate the serine-threonine protein kinases and most particularly the stress-activated protein kinases (SAPK, also known as JNK for c-jun NH2-terminal kinase)(15, 16). Finally, these different activation pathways result in the activation of various transcription factors including NF-kappaB, NF-kappaB-like molecules (such as p50, p65 (rel A), cRel), c-jun and NF-AT.
In recent years it has been established that members of the TNF- family associate intracellularly with different families of signalling molecules, including the 'death domain' family and the TRAF family (17). With the two-hybrid system, such protein-protein interactions have been demonstrated for TNF-R1, TNF-R2, Fas/CD95, CD30 and CD40. Interestingly, although there is no cross reactivity between the extracellular ligands of the TNF-R family (with the exception of TNF and LT), the intracellular ligands seem to be much more promiscuous and form a complex network of homo- and hetero-dimers (figure 2).
Figure 2: Schematic representation of molecules involved in CD40 signal transduction. X,Y and Z represent as yet unidentified molecules associated with the signalling complexes of TNF-R members. TRAF2 and TRAF3, which are indicated by hatched figures, associate with CD40, but are also shared by other members of TNF-R family (ie CD30 and TNF-R2). For more details, readers are referred to the text.
CD40 interacts with TRAF3 (TNF-R Associated Factor-3), also identified under the name CRAF1, CD40bp, LAP1 and CAP1 (18-21). TRAF3 is a 62 kD intracellular protein which is expressed in almost all cell types. The protein contains several functional domains which might be involved in signal transduction (RING finger domain, Isoleucine zipper, Zinc finger) or which play a role in protein-protein interactions either to associate with CD40 or to form homo- / heterodimers. TRAF3 appears to be essential for CD40 signalling inasmuch as a mutant CD40 with a point mutation at position 234 Thr > Ala, which is functionally inactive, does not bind TRAF3 (19). TRAF3 has been demonstrated to associate with CD30 (22). Interestingly, TRAF3 also binds to the EBV (Epstein Barr Virus) transforming gene product LMP1 (latent infection membrane-protein-1), which indicates that EBV utilizes the CD40 signalling pathway to activate and immortalize B lymphocytes (20).
In addition, CD40 has been demonstrated to associate with TRAF2, a molecule which also associates with TNF-R2. The induction of NF-kappaB activation via CD40 crosslinking (and also via TNF-R2) could be attributed to TRAF2 signalling (23), because a truncated TRAF2 protein, lacking an amino-terminal RING finger domain behaves as a dominant negative inhibitor. Finally, a novel 23 kD cell surface protein has been demonstrated to associate with CD40, but the functional relevance of this protein awaits more experimental data (24).
3.3. Functions of CD40 in vitro.
CD40 ligation activates resting B cells as shown by increase in size and expression of new surface molecules involved in homotypic and heterotypic aggregation (CD23, VLA-4), as well as T cell costimulation (CD80/CD86). Furthermore, CD40-activated B cells secrete a panel of cytokines which may act as autocrine and paracrine growth and differentiation factors (25, 26).
CD40-activated B cells enter into proliferation which is further stimulated by addition of cytokines such as IL4, IL13 or IL10 and their combination (1). Cytokines can also induce CD40-activated B cells to secrete immunoglobulins, with IL10 inducing the secretion of large quantities of Igs as a consequence of induced plasma cell differentiation. In addition to being a differentiation factor, IL10 acts as a factor inducing switch towards IgG1 and IgG3, as demonstrated at the molecular level by the appearance of 'switch circles (27). Combination of IL10 and TGFb induce IgA production, and IL4/lL13 induce CD40-activated B cells to switch towards IgG4 and IgE.
Interestingly, CD40 activation of B cells also results in the induction of Fas expression, and renders cells susceptible to Fas-induced apoptosis (28-30). In fact, together with BCR crosslinking, these three receptors generate a complex network of positive and negative signals, where the response of the B cell, activation or death, is determined by its differentiation stage (31-33). Importantly, dual triggering of resting B Iymphocytes through their CD40 and antigen receptor induces a phenotype characteristic of cells from germinal centers (34, 35), the anatomical site where B cells undergo somatic mutation, selection, isotype switching and become either plasma or memory cells. Prolonged triggering of CD40 skews the maturation of B cells into memory cells while interruption of CD40 signalling allows plasma cell differentiation (36, 37).
Although most normal and malignant B cells proliferate in response to CD40 engagement, plasma cells appear unresponsive (38). Furthermore, CD40 ligation appears to inhibit the proliferation of diffuse B cell Iymphomas, thus illustrating a possible negative role of CD40 on cell proliferation (39). A similar observation of negative signalling has been made in several transformed cell lines of both mesenchymal and epithelial origin (40).
3.3.2. Monocytes and dendritic cells
The expression of CD40 on professional antigen presenting cells like monocytes and dendritic cells is well established now. In addition, the functional role of CD40 on these cell types is becoming well documented both by in vitro and in vivo experiments (41, 42). Low spontaneous expression of CD40 is detected on freshly isolated monocytes, which can be upregulated by cytokines such as GM-CSF, IL3 and IFNgamma (43). In contrast, CD40 is expressed at high levels on dendritic cells isolated from different tissues or generated in vitro by culturing hematopoietic progenitors (44-49). Interestingly, CD40 expression has also been detected on monocytes infiltrating brain lesions of patients with Multiple Sclerosis (50).
CD40 ligation of monocytes and dendritic cells results in the secretion of multiple proteins including cytokines such as IL1, IL6, IL8, IL10, IL12, TNFalpha, MiP1alpha (43, 45, 51, 52), as well as enzymes such as matrix metalloproteinase (MMP)(53). Importantly, the secretion of IL12 allows a skewing of T cell maturation towards the Th1 pathway (52, 54-57). It has been suggested that CD40 specifically induces the expression of p40 (54). In addition, CD40 ligation considerably alters the phenotype of these APCs by upregulating the expression of costimulatory molecules such as CD54 (ICAM-1), CD58/LFA-3, CD80/B7-1, CD86/B7-2 (45, 46, 49, 57, 58). Interrupting CD40/CD40-L interactions during T cell/dendritic cell cocultures results in reduced T cell proliferation (47, 59), possibly as a consequence of both altered CD40 signalling to the APCs (reduced expression of costimulatory membrane molecules and cytokines) and altered CD40-L signalling to the T cells (figure 3).
Figure 3: Functional consequences of CD40-CD40-L interactions between activated T cells and Antigen Presenting Cells (monocytes and dendritic cells). The functional consequences of CD40 crosslinking on B cells are not included in this figure. Although most pathways are operational at the same time, for reasons of clarity, processes are separated in: A) activation of the APC; B) direct T cell activation via CD40-L; C) T cell costimulation via costimulatory molecules; D) Th1 skewing via cytokine production.
Ligation of CD40 enhances the survival of dendritic cells and monocytes (45, 49, 60, 61). Of note, CD40 ligation of CD34+ hematopoietic progenitors induces their proliferation and differentiation into cells with prominent dendritic cell attributes (62).
The importance of this receptor-ligand pair for the cellular immune response, has been demonstrated by the diminished immunity against several pathogens in CD40 and CD40-L knockout mice (see section 3.4). In keeping with this, CD40 ligation turns on monocyte tumoricidal activity as well as NO synthesis (63, 64).
3.3.3. Endothelial cells, epithelial cells and fibroblasts.
Immunohistology performed on various tissue sections shows that anti-CD40 antibodies stain vascular endothelium, epidermal basal membrane, scattered fibroblasts, thymic epithelium and follicular dendritic cells. Recently, CD40 expression was demonstrated in lesions of Kaposi's Sarcoma, as well as on vascular endothelium in areas adjacent to the tumors (65). Consistently, primary lines of endothelial cells, thymic and kidney tubular epithelial cells, keratinocytes, skin and synovial fibroblasts as well as follicular dendritic cells express CD40 at a relatively low density (5, 66). Yet, these adherent cells express a functional CD40 whose ligation induces phenotypic alterations, cytokine secretion as well as stimulation or inhibition of proliferation. In particular, endothelial cells display marked upregulation of CD54/lCAM-1, CD106/VCAM-1 and CD62E/E-Selectin, thus, resulting in increased ability to bind leukocytes (67-69). On the other hand, endothelial cells are also capable to enhance the expression of CD40-L on activated T cells, probably by a CD2/LFA3 mediated pathway (70).
3.4. In vivo functions of CD40/CD40-L interactions.
The first demonstration of the critical role of CD40/CD40-L interactions in vivo came from the discovery that the hyper IgM syndrome, an X-linked immunodeficiency, is due to a genetic alteration of the CD40-L (71). This disease is characterized by a severe impairment of T cell dependent antibody responses with no B cell memory and no circulating IgG, IgA and IgE. Initial experiments performed with antagonists to CD40/CD40-L interactions such as antibodies to CD40-L or CD40-Fc fusion proteins demonstrated impairment of B cell memory (72). The generation of CD40 and CD40-L knockout mice confirmed and further extended these observations and revealed a phenotype comparable to that of the patients suffering from the hyper IgM syndrome. In particular, these mice display decreased IgM responses to thymus dependent antigens, no antigen specific IgG1, IgA and IgE but normal T and B Iymphocyte numbers. As expected, these mice respond normally to thymus independent antigens with increased IgM and IgG3 levels.
Patients with hyper-IgM syndrome have an enhanced susceptibility to opportunistic infections, such as Pneumocystis carinii pneumonia and Cryptosporidium diarrhea. This indicates a role for CD40-CD40-L interactions in cell-mediated immune responses. Indeed, CD40-L knockout mice display a considerable impairment of antigen specific T cell priming (73) and appear particularly susceptible to Leishmania infection (74-76). This most likely results from a defective Th1 response which is related to an impaired production of IL12 by antigen presenting cells.
3.4.2. Formation of germinal centers
Germinal centers are the anatomical sites in lymphoid organs where isotype switching and somatic mutations are initiated (77). In a recent study, it was demonstrated that patients with the hyper-IgM syndrome lack germinal centers in their lymphoid organs (78). In accordance, somatic mutations within VDJ transcripts could not be detected in the circulating B cells from 5/6 patients (79, 80). Interestingly, the patient who had almost normal levels of somatic mutations, displayed a mutation within the transmembrane region of CD40-L, which results in a very transient CD40-L expression on activated T cells (80). These data suggest that a minimal CD40-L expression is sufficient to allow the activation of events leading to the introduction of somatic mutations in vivo. However, CD40 activation of B cells per se, is not sufficient to induce somatic mutations in vitro (81), while activated T cells can do so (82)(Razajanaona et al. and Denepoux et al., unpublished data).
Studies with a CD40-Fc fusion protein have highlighted the important role of T cell signalling through CD40 ligand in the development of helper function. In particular, administration of soluble CD40 in vivo to CD40 knockout mice initiates germinal center formation and T cells primed in the absence of CD40 are unable to help normal B cells to class switch and to form germinal centers (83). In this context, ligating CD40 ligand of human activated T cells considerably enhances their production of IL4 (84, 85).
3.4.3. CD40/CD40-L: treatment of autoimmunity
Administration of antibodies to CD40-L has been shown to prevent the establishment of autoimmune symptoms in various murine models including: 1) collagen type II-induced arthritis; a model for human rheumatoid arthritis; 2) lupus nephritis in lupus prone mice that represent models for the systemic lupus erythematosus; 3) proteolipoprotein induced experimental encephalomyelitis; a model of human multiple sclerosis. Importantly, in this latter case, the antibody could induce an important reduction of the disease even when administered after onset. Consistently, activated helper T cells expressing CD40-L surface protein are detected in multiple sclerosis patient brain sections where CD40 bearing antigen presenting cells can be found.
3.4.4. CD40/CD40-L: induction of transplantation tolerance
Administration of anti-CD40-L antibodies prevents the development of graft versus host disease (GVHD) that occurs as a major complication of allogeneic bone marrow transplantation (86). This treatment affects both the acute GVHD, essentially mediated by cytotoxic T cells, and the chronic GVHD, mainly due to the polyclonal B cell activation and the production of self reactive antibodies. The addition of anti-CD40-L antibody appears to considerably enhance the tolerogenic effect of B cells (87). Furthermore, a combination of allogeneic B cells and anti-CD40-L antibody considerably decreases host reactivity of both CD4+ and CD8+ T Iymphocytes thereby allowing efficient transplantation of allogeneic pancreatic ß islet cells (88). Anti-CD40-L antibodies markedly extend the survival of cardiac allografts in both naive and sensitized hosts when administered at the time of transplantation (89). More importantly, even long term acceptance of skin and cardiac allografts can be obtained by a simultaneous blocking of the CD40 and CD28 pathways (90).