Differential regulation of Th1 and Th2 functions of NKT cells by CD28 and CD40 costimulatory pathways

Valpha14 NKT cells produce large amounts of IFN-gamma and IL-4 upon recognition of their specific ligand alpha-galactosylceramide (alpha-GalCer) by their invariant TCR. We show here that NKT cells constitutively express CD28, and that blockade of CD28-CD80/CD86 interactions by anti-CD80 and anti-CD86 mAbs inhibits the alpha-GalCer-induced IFN-gamma and IL-4 production by splenic Valpha14 NKT cells. On the other, the blockade of CD40-CD154 interactions by anti-CD154 mAb inhibited alpha-GalCer-induced IFN-gamma production, but not IL-4 production. Consistent with these findings, CD28-deficient mice showed impaired IFN-gamma and IL-4 production in response to alpha-GalCer stimulation in vitro and in vivo, whereas production of IFN-gamma but not IL-4 was impaired in CD40-deficient mice. Moreover, alpha-GalCer-induced Th1-type responses, represented by enhanced cytotoxic activity of splenic or hepatic mononuclear cells and antimetastatic effect, were impaired in both CD28-deficient mice and CD40-deficient mice. In contrast, alpha-GalCer-induced Th2-type responses, represented by serum IgE and IgG1 elevation, were impaired in the absence of the CD28 costimulatory pathway but not in the absence of the CD40 costimulatory pathway. These results indicate that CD28-CD80/CD86 and CD40-CD154 costimulatory pathways differentially contribute to the regulation of Th1 and Th2 functions of Valpha14 NKT cells in vivo.     

Stimulation of the B cell receptor, CD86 (B7-2), and the beta 2-adrenergic receptor intrinsically modulates the level of IgG1 and IgE produced per B cell

Our findings using B cells from either wild-type, CD86-deficient, or beta 2-adrenergic receptor (beta2AR)-deficient mice suggest three mechanisms by which the level of IgG1 and IgE production can be increased on a per cell basis. Trinitrophenyl-specific B cells enriched from unimmunized mouse spleens were pre-exposed to Ag and/or the beta 2AR ligand terbutaline for 24 h before being activated by either a beta 2AR-negative Th2 cell clone or CD40 ligand/Sf9 cells and IL-4 in the presence or absence of an anti-CD86 Ab. Data suggest that the first mechanism involves a B cell receptor (BCR)-dependent up-regulation of CD86 expression that, when CD86 is stimulated, increases the amount of IgG1 and IgE produced in comparison to unstimulated cells. The second mechanism involves a BCR- and beta 2AR-dependent up-regulation of CD86 to a level higher than that induced by stimulation of either receptor alone that, when CD86 is stimulated, further increases the amount of IgG1 and IgE produced. The third mechanism is BCR-independent and involves a beta 2AR-dependent increase in the ability of a B cell to respond to IL-4. Flow cytometric and limiting dilution analyses suggest that the increase in IgG1 and IgE occurs independently from the isotype switching event. These findings suggest that the BCR, the beta 2AR, and CD86 are involved in regulating IL-4-dependent IgG1 and IgE production.     

Selective requirement for CD40-CD154 in drug-induced type 1 versus type 2 responses to trinitrophenyl-ovalbumin

CD154 is transiently expressed by activated T cells and interacts with CD40 on B cells, dendritic cells, macrophages, and monocytes. This costimulatory receptor-ligand couple seems decisive in Ag-driven immune responses but may be differentially involved in type 1 vs type 2 responses. We studied the importance of CD40-CD154 in both responses using the reporter Ag popliteal lymph node assay in which selectively acting drugs generate clearly polarized type 1 (streptozotocin) or type 2 (D-penicillamine, diphenylhydantoin) responses to a constant coinjected Ag in the same mouse strain. Treatment of mice with anti-CD154 reduced characteristic immunological parameters in type 2 responses (B and CD4(+) T cell proliferation, IgG1 and IgE Abs, and IL-4 secretion) and only slightly affected the type 1 response (small decrease in IFN-gamma production, influx of CD11c(+) and F4/80(+) cells, and prevention of architectural disruption of the lymph node, but no effect on IgG2a Ab and TNF-alpha secretion or B and CD4(+) T cell proliferation). The findings indicate that the CD40-CD154 costimulatory interaction is a prerequisite in drug-induced type 2 responses and is only marginally involved in type 1 responses. The observed expression patterns of CD80 and CD86 on different APC (B cells in type 2 and dendritic cells in type 1) may be responsible for this discrepancy.     

The CD28/CTLA-4-B7 signaling pathway is involved in both allergic sensitization and tolerance induction to orally administered peanut proteins

Dendritic cells are believed to play an essential role in regulating the balance between immunogenic and tolerogenic responses to mucosal Ags by controlling T cell differentiation and activation via costimulatory and coinhibitory signals. The CD28/CTLA-4-CD80/CD86 signaling pathway appears to be one of the most important regulators of T cell responses but its exact role in responses to orally administered proteins remains to be elucidated. In the present study, the involvement of the CD28/CTLA-4-CD80/CD86 costimulatory pathway in the induction of allergic sensitization and oral tolerance to peanut proteins was investigated. In both an established C3H/HeOuJ mouse model of peanut hypersensitivity and an oral tolerance model to peanut, CD28/CTLA-4-CD80/CD86 interactions were blocked using the fusion protein CTLA-4Ig. To examine the relative contribution of CD80- and CD86-mediated costimulation in these models, anti-CD80 and anti-CD86 blocking Abs were used. In the hypersensitivity model, CTLA-4Ig treatment prevented the development of peanut extract-induced cytokine responses, peanut extract-specific IgG1, IgG2a, and IgE production and peanut extract-induced challenge responses. Blocking of CD80 reduced, whereas anti-CD86 treatment completely inhibited, the induction of peanut extract-specific IgE. Normal tolerance induction to peanut extract was found following CTLA-4Ig, anti-CD86, or anti-CD80 plus anti-CD86 treatment, whereas blockade of CD80 impaired the induction of oral tolerance. We show that CD28/CTLA-4-CD80/CD86 signaling is essential for the development of allergic responses to peanut and that CD86 interaction is most important in inducing peanut extract-specific IgE responses. Additionally, our data suggest that CD80 but not CD86 interaction with CTLA-4 is crucial for the induction of low dose tolerance to peanut.     

Expression of CD80/86 on B cells is essential for autoreactive T cell activation and the development of arthritis

Depletion of B cells in rheumatoid arthritis is therapeutically efficacious. Yet, the mechanism by which B cells participate in the inflammatory process is unclear. We previously demonstrated that Ag-specific B cells have two important functions in the development of arthritis in a murine model of rheumatoid arthritis, proteoglycan (PG)-induced arthritis (PGIA). PG-specific B cells function as autoantibody-producing cells and as APCs that activate PG-specific T cells. Moreover, the costimulatory molecule CD86 is up-regulated on PG-specific B cells in response to stimulation with PG. To address the requirement for CD80/CD86 expression on B cells in the development of PGIA, we generated mixed bone marrow chimeras in which CD80/CD86 is specifically deleted on B cells and not on other APC populations. Chimeras with a specific deficiency in CD80/CD86 expression on B cells are resistant to the induction of PGIA. The concentration of PG-specific autoantibody is similar in mice sufficient or deficient for CD80/86-expressing B cells, which indicates that resistance to PGIA is not due to the suppression of PG-specific autoantibody production. CD80/86-deficient B cells failed to effectively activate PG-specific autoreactive T cells as indicated by the failure of T cells from PG-immunized CD80/86-deficient B cell chimeras to transfer arthritis into SCID mice. In vitro secondary recall responses to PG are also dependent on CD80/86-expressing B cells. These results demonstrate that a CD80/86:CD28 costimulatory interaction between B cells and T cells is required for autoreactive T cell activation and the induction of arthritis but not for B cell autoantibody production.     

Differential requirement for CD28/CTLA-4-CD80/CD86 interactions in drug-induced type 1 and type 2 immune responses to trinitrophenyl-ovalbumin

The use of mAbs to abrogate costimulatory interactions has attracted much attention with regard to prevention and modulation of adverse (auto)immune-like reactions. However, the role of costimulatory molecules and possible therapeutic use of Ab-treatment in drug-induced immunostimulation is poorly elucidated. In the present studies, we show that CD28/CTLA-4-CD80/CD86 costimulatory interactions differently regulate drug-induced type 1 and type 2 responses to an identical bystander Ag, TNP-OVA, in BALB/c mice using the reporter Ag popliteal lymph node assay. The antirheumatic drug D-Penicillamine, which may induce lupus-like side-effects, stimulated type 2 responses against TNP-OVA, characterized by the production of IL-4 and TNP-specific IgG1 and IgE. These responses were abrogated in CD80/CD86-deficient mice and in wild-type mice that were treated with anti-CD80 and anti-CD86, or CTLA-4-Ig. Anti-CTLA-4 intensively enhanced the D-Penicillamine-induced effects. In contrast, the type 1 response (IFN-gamma, TNF-alpha, IgG2a) to TNP-OVA induced by the diabetogen streptozotocin still developed in the absence of CD80/CD86 costimulatory signaling. In addition, it was demonstrated that coadministration of anti-CD80 and anti-CD86 mAbs slightly enhanced streptozotocin-induced type 1 responses, whereas the CTLA-4-Ig fusion protein completely abrogated this response. In conclusion, different drugs may stimulate distinct types of immune responses against an identical bystander Ag, which are completely dependent on (type 2) or independent of (type 1) the CD28/CTLA-4-CD80/CD86 pathway. Importantly, the effects of treatment with anti-CD80/CD86 mAbs and CTLA-4-Ig may be considerably different in responses induced by distinct drugs.     

Differential requirement for the CD40-CD154 costimulatory pathway during Th cell priming by CD8 alpha+ and CD8 alpha- murine dendritic cell subsets

Dendritic cells (DCs) regulate the development of distinct Th populations and thereby provoke appropriate immune responses to various kinds of Ags. In the present work, we investigated the role CD40-CD154 interactions play during the process of Th cell priming by CD8 alpha(+) and CD8 alpha(-) murine DC subsets, which have been reported to differently regulate the Th response. Adoptive transfer of Ag-pulsed CD8 alpha(+) DCs induced a Th1 response and the production of IgG2a Abs, whereas transfer of CD8 alpha(-) DCs induced Th2 cells and IgE Abs in vivo. Induction of distinct Th populations by each DC subset was also confirmed in vitro. Although interruption of CD80/CD86-CD28 interactions inhibited Th cell priming by both DC subsets, disruption of CD40-CD154 interactions only inhibited the induction of the Th1 response by CD8 alpha(+) DCs in vivo. CD40-CD154 interactions were not required for the proliferation of Ag-specific naive Th cells stimulated by either DC subset, but were indispensable in the production of IL-12 from CD8 alpha(+) DCs and their induction of Th1 cells in vitro. Taken together, in our immunization model of Ag-pulsed DC transfer, CD40-CD154 interactions play an important role in the development of CD8 alpha(+) DC-driven Th1 responses but not CD8 alpha(-) DC-driven Th2 responses to protein Ags.     

Multiple layers of CD80/86-dependent costimulatory activity regulate primary, memory, and secondary lymphocytic choriomeningitis virus-specific T cell immunity

The lymphocytic choriomeningitis virus (LCMV) system constitutes one of the most widely used models for the study of infectious disease and the regulation of virus-specific T cell immunity. However, with respect to the activity of costimulatory and associated regulatory pathways, LCMV-specific T cell responses have long been regarded as relatively independent and thus distinct from the regulation of T cell immunity directed against many other viral pathogens. Here, we have reevaluated the contribution of CD28-CD80/86 costimulation in the LCMV system by use of CD80/86-deficient mice, and our results demonstrate that a disruption of CD28-CD80/86 signaling compromises the magnitude, phenotype, and/or functionality of LCMV-specific CD8(+) and/or CD4(+) T cell populations in all stages of the T cell response. Notably, a profound inhibition of secondary T cell immunity in LCMV-immune CD80/86-deficient mice emerged as a composite of both defective memory T cell development and a specific requirement for CD80 but not CD86 in the recall response, while a related experimental scenario of CD28-dependent yet CD80/86-independent secondary CD8(+) T cell immunity suggests the existence of a CD28 ligand other than CD80/86. Furthermore, we provide evidence that regulatory T cells (T(REG)s), the homeostasis of which is altered in CD80/86(-/-) mice, contribute to restrained LCMV-specific CD8(+) T cell responses in the presence of CD80/86. Our observations can therefore provide a more coherent perspective on CD28-CD80/86 costimulation in antiviral T cell immunity that positions the LCMV system within a shared context of multiple defects that virus-specific T cells acquire in the absence of CD28-CD80/86 costimulation.     

CD28-independent costimulation of T cells by OX40 ligand and CD70 on activated B cells.

OX40 and its ligand (OX40L) have been implicated in T cell-dependent humoral immune responses. To further characterize the role of OX40/OX40L in T-B cell interaction, we newly generated an anti-mouse OX40L mAb (RM134L) that can inhibit the costimulatory activity of OX40L transfectants for anti-CD3-stimulated T cell proliferation. Flow cytometric analyses using RM134L and an anti-mouse OX40 mAb indicated that OX40 was inducible on splenic T cells by stimulation with immobilized anti-CD3 mAb in a CD28-independent manner, while OX40L was not expressed on resting or activated T cells. OX40L was inducible on splenic B cells by stimulation with anti-IgM Ab plus anti-CD40 mAb, but not by either alone. These activated B cells exhibited a potent costimulatory activity for anti-CD3-stimulated T cell proliferation and IL-2 production. Anti-CD80 and anti-CD86 mAbs partially inhibited the costimulatory activity, and further inhibition was obtained by their combination with RM134L and/or anti-CD70 mAb. We also found the anti-IgM Ab- plus anti-CD40 mAb-stimulated B cells exhibited a potent costimulatory activity for proliferation of and IL-2 production by anti-CD3-stimulated CD28- T cells from CD28-deficient mice, which was substantially inhibited by RM134L and/or anti-CD70 mAb. These results indicated that OX40L and CD70 expressed on surface Ig- and CD40-stimulated B cells can provide CD28-independent costimulatory signals to T cells.     

Requirement for donor and recipient CD40 expression in cardiac allograft rejection: induction of Th1 responses and influence of donor-derived dendritic cells

Costimulation through the CD40-CD40 ligand (CD40L) pathway is critical to allograft rejection, in that anti-CD40L mAb therapy prolongs allograft survival. However, the majority of studies exploring CD40-CD40L interactions have targeted CD40L. Less is known about the requirement for donor- and/or host-derived CD40 during rejection. This study assessed the relative contributions of donor and recipient CD40 expression to the rejection process. As the effectiveness of costimulatory blockade may be mouse strain dependent, this study explored the requirement for donor and recipient CD40 expression in BALB/c and C57BL/6 mice. Wild-type (WT) and CD40(-/-) BALB/c recipients readily rejected WT and CD40(-/-) C57BL/6 allografts, and rejection was associated with a prominent Th1 response. In contrast, CD40(-/-) C57BL/6 recipients failed to reject WT or CD40(-/-) BALB/c allografts and did not mount Th1 or Th2 responses. However, injection of donor CD40(-/-) dendritic cells induced both Th1 and Th2 responses and allograft rejection in CD40(-/-) C57BL/6 recipients. Finally, WT C57BL/6 mice rejected CD40(-/-) allografts, but this rejection response was associated with muted Th1 responses. These findings demonstrate that 1) CD40 expression by the recipient or the graft may impact on the immune response following transplantation; 2) the requirement for CD40 is influenced by the mouse strain; and 3) the requirement for CD40 in rejection may be bypassed by donor DC. Further, as CD40 is not required for rejection in BALB/c recipients, but anti-CD40L mAb prolongs graft survival in these mice, these results suggest that anti-CD40L therapy functions at a level beyond disruption of CD40-CD40L interactions.     

ICOS costimulates invariant NKT cell activation

It has been reported that costimulatory molecules, CD80/86-CD28 and CD154-CD40, critically contribute to activation of CD1d-restricted invariant NKT (iNKT) cells. Here we have demonstrated that ICOS, a new member of the CD28 family, plays a substantial role in iNKT cell activation. iNKT cells constitutively expressed ICOS as well as CD28 independently, and ICOS expression was further up-regulated 2-3 days after alpha-galactosylceramide (alpha-GalCer) treatment. Blockade of ICOS-mediated costimulation by administration of anti-ICOS ligand (B7RP-1) mAb or by ICOS gene knockout substantially inhibited alpha-GalCer-induced IFN-gamma and IL-4 production, cytotoxic activity, and anti-metastatic effect. Moreover, blockade of both B7RP-1-ICOS and CD80/86-CD28 interactions mostly abolished the alpha-GalCer-induced immune responses. These findings indicate that iNKT cell activation is regulated by CD28 and IOCS independently.     

Resting B lymphocytes as APC for naive T lymphocytes: dependence on CD40 ligand/CD40

Although resting B cells as APC are tolerogenic for naive T cells in vivo, we show here that they can provide all the costimulatory signals necessary for naive T cell proliferation in vivo and in vitro. In the absence of an activating signal through the B cell Ag receptor, T cell proliferation after Ag recognition on resting B cells depends on CD40 expression on the B cells, implying that naive T cells use the membrane-bound cytokine, CD40 ligand (CD154), to induce the costimulatory signals that they need. Induction of B7-1 (CD80) and increased or sustained expression of CD44H, ICAM-1 (CD54), and B7-2 (CD86) are dependent on the interaction of CD40 ligand with CD40. Transient expression (12 h) of B7-2 is T cell- and peptide Ag-dependent, but CD40-independent. Only sustained (>/=24 h) expression of B7-2 and perhaps increased expression of ICAM-1 could be shown to be functionally important in this system. T cells cultured with CD40-deficient B cells and peptide remain about as responsive as fresh naive cells upon secondary culture with whole splenic APC. Therefore, B cells, and perhaps other APC, may be tolerogenic not because they fail to provide sufficient costimulation for T cell proliferation, but because they are deficient in some later functions necessary for a productive T cell response.     

Murine B1 B cells require IL-5 for optimal T cell-dependent activation

T helper cell-driven activation of murine B cells has been shown to depend upon CD40-CD40 ligand (CD40L) interactions and a defined set of cytokines. These observations are primarily based on the use of conventional B cells obtained from the spleen. Therefore, it is presently unclear whether all mature B cell subsets found in the mouse have an equal dependence upon CD40-CD40L interactions and use the same T cell-derived cytokines. The present study tested the response of splenic follicular and marginal zone as well as peritoneal B2 and B1 B cells to Th cell stimulation. Splenic and peritoneal B cell subsets were sort purified based on CD23 expression, and cultured with rCD40L and cytokines or Th2 cells. The results demonstrate that follicular, marginal zone, and peritoneal B2 B cells require CD40-CD40L interactions and preferentially use IL-4 for optimal proliferation, differentiation, and isotype switching. In contrast, peritoneal B1 B cells use IL-5 in conjunction with CD40-CD40L interactions for maximal Th cell-dependent responses. Furthermore, B1 B cells are capable of proliferating, differentiating, and isotype switching in the absence of CD40-CD40L interactions. B1 B cells are able to respond to Th2 clones in the presence of anti-CD40L mAb as well as to Th2 clones derived from CD40L(-/-) mice. The CD40-CD40L-independent response of B1 B cells is attributable to the presence of both IL-4 and IL-5, and may explain the residual Ab response to T cell-dependent Ags in CD40L- or CD40-deficient mice, and in X-linked hyper-IgM (X-HIM) patients.     

The roles of CD28 and CD40 ligand in T cell activation and tolerance

Costimulation of T cell activation involves both the B7:CD28 as well as the CD40 ligand (CD40L):CD40 pathway. To determine the importance of these pathways to in vitro and in vivo T cell activation, a direct comparison was made of the responses of TCR transgenic T cells lacking either CD28 or CD40L. In vitro, CD28-/- T cells showed a greater reduction in proliferative responses to Ag than did CD40L-/- T cells. The absence of CD28 resulted in defective Th2 responses, whereas CD40L-/- T cells were defective in Th1 development. In vivo, CD28-/- T cells failed to expand upon immunization, whereas CD40L-/- T cells could not sustain a response. These results suggest that CD28 is critical for initiating T cell responses, whereas CD40L is required for sustained Th1 responses. The different functional roles of these costimulatory pathways may explain why blocking B7:CD28 and CD40L:CD40 interactions has an additive effect in inhibiting T cell responses.     

The role of CD40-CD154 interaction in antiviral T cell-independent IgG responses

Polyomavirus (PyV) infection elicits protective T cell-independent (TI) IgG responses in T cell-deficient mice. The question addressed in this report is whether CD40 signaling plays a role in this TI antiviral IgG response. Because CD40 ligand (CD40L) can be expressed on numerous cell types in addition to activated T cells, it is possible that cells other than T cells provide CD40L to signal through CD40 on B cells and hence positively influence the antiviral TI IgG responses. In this study we show, by blocking CD40-CD40L interactions in vivo with anti-CD40L Ab treatment in TCR betaxdelta-/- mice and by using SCID mice reconstituted with CD40-/- B cells, that the lack of CD40 signaling in B cells results in a 50% decrease in TI IgG secreted in response to PyV. SCID mice reconstituted with CD40L-/- B cells also responded to PyV infection with diminished IgG secretion compared with that of SCID mice reconstituted with wild-type B cells. This finding suggests that B cells may provide the CD40L for CD40 signaling in the absence of T cell help during acute virus infection. Our studies demonstrate that, although about half of the TI IgG responses to PyV are independent of CD40-CD40L interactions, these interactions occur in T cell-deficient mice and enhance antiviral TI Ab responses.     

The identification and characterization of a ligand for bovine CD5

CD5, a type I glycoprotein expressed by T cells and a subset of B cells, is thought to play a significant role in modulating Ag receptor signaling. Previously, our laboratory has shown that bovine B cells are induced to express this key regulatory molecule upon Ag receptor cross-linking. To date, a ligand has not been described for bovine CD5. Given the importance ligand binding presumably plays in the functioning of CD5 on this B cell subset and on T cells, we sought to characterize the ligand for this protein using a bovine CD5-human IgG1 (CD5Ig) fusion protein produced by both mammalian and yeast cells. As determined by CD5Ig binding, expression of this ligand is negative to low on freshly isolated lymphocytes, with low-density expression being limited to activated B cells. Activation with LPS, PMA, and calcium ionophore, or ligation of CD40 alone or in combination with anti-IgM, resulted in B cell-specific expression of this ligand. Interestingly, activation through B cell Ag receptor cross-linking alone, although able to induce CD5 expression, did not result in expression of CD5 ligand (CD5L). In addition, we demonstrate a functional role for CD5L as a costimulatory molecule that augments CD40L-stimulated B cell proliferation. Finally, immunoprecipitation with CD5Ig suggests that the ligand characterized in this study has a molecular mass of approximately 200 kDa. The data reported herein, as well as future studies aimed at further characterizing this newly identified bovine CD5L, will undoubtedly aid in understanding the role that the CD5-CD5L interaction plays in immune responses.     

Disruption of CD154:CD40 blocks generation of allograft immunity without affecting APC activation.

CD154 (CD40 ligand, gp39) interaction with its receptor CD40 has been shown to be critically important for the generation of cell-mediated as well as humoral immunity. It has been proposed that ligation of CD40 on APCs, presumably by activated Th cells, leads to increased APC function as defined by up-regulation of costimulatory molecules and enhancement of IL-12 production. In this report, we directly examined the contribution of the CD154:CD40 pathway in a murine model of allograft rejection. Generation of both the CTL and alloantibody responses following injection with allogeneic P815 tumor cells was severely compromised in CD154 knockout mice and wild-type C57BL/6 mice treated with the anti-CD154 mAb, MR1. Splenic production of IL-2, IFN-gamma, and TNF was significantly suppressed from CD154-deficient mice, indicating a lack of T cell priming. However, splenic cells from CD154 knockout mice induced comparable levels of CD86 expression and IL-12 production when compared with their wild-type littermates. The treatment of CD154-/- mice with the agonistic anti-CD40 mAb, FGK45, generated activated APCs yet failed to restore either the CTL or alloantibody responses to P815. Likewise, immunization with B7-transfected P815 tumor cells failed to generate expansion of the CTL effector population in CD154-/- mice. These results suggest that the generation of allograft immunity is dependent on the interaction of CD154 with CD40 but not primarily for the activation of APCs.     

4-1BB ligand, a member of the TNF family, is important for the generation of antiviral CD8 T cell responses.

4-1BB (CD137) is a costimulatory molecule expressed on activated T cells and interacts with 4-1BB ligand (4-1BBL) on APCs. To investigate the role of 4-1BB costimulation for the development of primary immune responses, 4-1BBL-deficient (4-1BBL-/-) mice were infected with lymphocytic choriomeningitis virus (LCMV). 4-1BBL-/- mice were able to generate CTL and eliminate acute LCMV infection with normal kinetics, but CD8 T cell expansion was 2- to 3-fold lower than in wild-type (+/+) mice. In the same mice, virus-specific CD4 Th and B cell responses were minimally affected, indicating that 4-1BB costimulation preferentially affects CD8 T cell responses. This result contrasts with our earlier work with CD40L-deficient (CD40L-/-) mice, in which the CD8 T cell response was unaffected and the CD4 T cell response was markedly impaired. When both 4-1BBL- and B7-dependent signals were absent, CD8 T cell expansion was further reduced, resulting in lower CTL activity and impairing their ability to clear LCMV. Altogether, these results indicate that T cells have distinct costimulatory requirements: optimal CD8 responses require 4-1BBL-dependent interactions, whereas CD4 responses are minimally affected by 4-1BB costimulation, but require CD40-CD40L and B7-dependent interactions.     

Mouse inducible costimulatory molecule (ICOS) expression is enhanced by CD28 costimulation and regulates differentiation of CD4+ T cells

The inducible costimulatory (ICOS) molecule is expressed by activated T cells and has homology to CD28 and CD152. ICOS binds B7h, a molecule expressed by APC with homology to CD80 and CD86. To investigate regulation of ICOS expression and its role in Th responses we developed anti-mouse ICOS mAbs and ICOS-Ig fusion protein. Little ICOS is expressed by freshly isolated mouse T cells, but ICOS is rapidly up-regulated on most CD4(+) and CD8(+) T cells following stimulation of the TCR. Strikingly, ICOS up-regulation is significantly reduced in the absence of CD80 and CD86 and can be restored by CD28 stimulation, suggesting that CD28-CD80/CD86 interactions may optimize ICOS expression. Interestingly, TCR-transgenic T cells differentiated into Th2 expressed significantly more ICOS than cells differentiated into Th1. We used two methods to investigate the role of ICOS in activation of CD4(+) T cells. First, CD4(+) cells were stimulated with beads coated with anti-CD3 and either B7h-Ig fusion protein or control Ig fusion protein. ICOS stimulation enhanced proliferation of CD4(+) cells and production of IFN-gamma, IL-4, and IL-10, but not IL-2. Second, TCR-transgenic CD4(+) T cells were stimulated with peptide and APC in the presence of ICOS-Ig or control Ig. When the ICOS:B7h interaction was blocked by ICOS-Ig, CD4(+) T cells produced more IFN-gamma and less IL-4 and IL-10 than CD4(+) cells differentiated with control Ig. These results demonstrate that ICOS stimulation is important in T cell activation and that ICOS may have a particularly important role in development of Th2 cells.