The development of effector and memory CD4 cell populations depends upon both T cell receptor (TCR) engagement of peptide/major histocompatibility complex (MHC) class II complexes and ligation of costimulatory molecules with counter receptors on antigen-presenting cells (APCs). We showed previously that sustained interactions with APCs could be crucial for optimal expansion of CD4 cells and for development of effectors that secrete cytokines associated with Th2 cells. Using an adoptive transfer model with TCR transgenic CD4 cells, we now show that responses of CD4 cells primed in B cell–deficient mice become aborted, but are fully restored upon the transfer of activated B cells. Although B cells have the capacity to secrete multiple cytokines that could affect CD4 priming, including IL-4, we were unable to distinguish a role for cytokines that are secreted by B cells. However, B cell costimulation via the OX40L/OX40 pathway that has been implicated in CD4 cell expansion, survival, and Th2 development was required. Th2 but not Th1 responses were impaired in OX40L-deficient recipients and normal responses were restored with OX40L sufficient B cells. The results suggest that without engagement of OX40L on B cells, CD4 cell responses to many protein Ag would be dominated by Th1 cytokines. These data have important implications for strategies to achieve optimal priming of CD4 subsets.
The development of immunity depends upon the generation of an expanded population of CD4 effectors which mediate inflammatory processes that lead to clearance of foreign Ag and the persistence of a cohort of primed cells with the capacity to mount a rapid recall response. Initial activation of naive CD4 cells depends upon prolonged signaling through TCR engagement of peptide/MHC complexes as well as ligation of costimulatory molecules on APC. Productive interactions can be achieved by sustained as well as serial encounters of T cells with APC (1). The duration of stimulation can determine the commitment of CD4 cells to division which typically correlates with the development of effector functions (2). Once division commences, effectors can expand and differentiate independently of Ag (3, 4). Cytokine production and optimal induction of costimulation are two major mechanisms that contribute to the growth, differentiation, and survival of CD4 cells. Through these mechanisms, APCs are integral not only for initiating a CD4 response but also for determining magnitude of expansion, and development of effector populations that secrete cytokines associated with Th1 or Th2 phenotypes (e.g., TNF-α and IFN-γ, or IL-4, IL-5, and IL-13, respectively). In this context, APCs can be viewed as orchestrating the outcome of a primary immune response which determines the frequency of persisting memory cells and their cytokine commitment.
Naive CD4 cells initially interact with peptide Ag presented by dendritic cells (DCs)* which express MHC class II and B7 family molecules that provide the key costimulation for the induction of the T cell response via CD28 (5). As a consequence, additional costimulatory receptors are up-regulated on CD4 cells, including CD40L (CD154), OX40 (CD134), and ICOS (6, 7), and become engaged with counter-receptors on DC to promote optimal T cell responses. Once activated, CD4 cells also form conjugates with B cells (8) which become competent to act as APC and provide costimulation after activation by uptake of Ag, ligation of costimulatory/adhesion molecules, or interaction with cytokines. Our previous studies and others have shown that B cells can be necessary in vivo to support T cell expansion (9) and in this capacity can play a crucial role in the development of effector responses and memory (4, 10). We also found that sustained CD4 responses are necessary for the generation of high frequencies of effectors that secrete the Th2 cytokines, IL-4, IL-5, and IL-13, and that B cells can function to regulate the processes of differentiation and expansion in vivo (10). Although subsets of DCs can also promote development of type I or type II CD4 cells (11), recruitment of DCs diminishes after the initial phase of a response (1), and DCs retain the capacity to present Ag and produce cytokines that affect T cell differentiation for only a limited duration before further maturing and exiting lymphoid sites (12). Thus, a switch in APC usage from DCs to B cells can be favored as a response proceeds, and in many instances may be necessary to sustain progression to Th2 cytokines.
In this study, we investigated mechanisms that underlie the capacity of B cells to support CD4 effector development in vivo. As our previous work pointed to Ag-independent processes, we focused on the involvement of cytokines and costimulation in an adoptive transfer model where naive TCR transgenic CD4 cells are dependent upon provision of B cells as APCs for optimal expansion and differentiation to a Th2 response. Although several studies suggest that activated B cells produce cytokines that can regulate DC function as well as T cell survival and effector development (13, 14), we were unable to distinguish roles for key cytokines that could participate in this response, i.e., IL-4, IL-6, and IL-10. However, costimulation via OX40/OX40L, a TNF/TNFR-family pathway which regulates CD4 cells (6, 7) was an essential contribution of B cells.
Materials And Methods
Normal and splenectomized C57BL/6, C57BL/6 Igh-6 μMT (immunoglobulin μ chain knock out mice); and B6.PL Thy1a/Cy, IL-4, -6, and -10 deficient mice were from The Jackson Laboratory. OX40L-deficient mice were provided by Dr. B. Paigen (The Jackson Laboratory, Bar Harbor, ME) with kind permission from Dr. K. Sugamura (Tohoku University, Sendai, Japan; reference 15) and bred in our vivarium. OT-II TCR (vα2, vβ5) CD4 transgenic mice that respond to peptide 323–339 of OVA were obtained from W. Heath (Walter and Eliza Hall, Melbourne, Australia; reference 16) and bred in our colony to B6.PL-Thy1a/Cy mice to facilitate tracking by Thy 1.1 in C57BL/6 and μMT recipients that express Thy 1.2.
T and B Cell Isolation and Adoptive Transfer.
CD4 cells were isolated from the spleens and pooled lymph nodes of OT-II Thy 1.1 mice by negative selection using magnetic sorting with an enrichment cocktail from Stem Cell Technologies Inc. according to the manufacturer's instructions. Vα2, Vβ5+ cells were quantitated by flow cytometry and 5 × 106 transgenic CD4 cells were injected intravenously into C57BL/6, μMT, or OX40L−/− mice. For some experiments, CD4 cells were labeled before injection with the fluorescent dye, CFSE (5-and-6)-carboxyfluorescein diacetate succinimidyl ester; Molecular Probes). Mice were immunized on the day of cell transfer by intraperitoneal injection of 100 μg alum precipitated OVA (Sigma-Aldrich) with B. pertussis vaccine organisms (4). In some experiments, mice were also given a subcutaneous injection of the same dose of Ag at the base of the tail. B cells were isolated from the spleens of C57BL/6 mice or various knockout mice by magnetic separation with an enrichment cocktail (Stem Cell Technologies Inc.) and incubated overnight in RPMI 1640 media (Irvine Scientific) containing 7% FCS (Irvine Scientific), 200 μg/ml penicillin, 200 U/ml streptomycin, 4 mM l-glutamine, 10 mM Hepes, and 5 × 10−5 M 2-ME. The cultures were also supplemented with 10 ng/ml rIL-4, 10 μg/ml dextran sulfate (Sigma-Aldrich), 5 μg/ml LPS (Difco Laboratories), and 100 μg/ml OVA. 5 × 106 B cells were injected intravenously into T cell recipient mice 24 h after immunization with OVA.
The following reagents from BD Biosciences were used for staining of T and B cells: FITC-conjugated rat IgG2a (R35–95), -anti-MHC class II (AF6–120.1), -CD40 (3/23), and -B7–2 (GL1); PE-conjugated anti-Vα2 (B20.1); and biotinyl-anti-Vβ5 TCR (MR9–4). Peridiunin chlorophyll protein (PerCP)-conjugated streptavidin was from Becton Dickinson. APC-conjugated-anti-Thy1.1 (HIS51) and FITC-conjugated-anti-rat IgG were from eBioscience. FITC-labeled rat anti-OX40L (RM134L) was a gift from Dr. Michael Croft (La Jolla Institute of Allergy and Immunology, La Jolla, CA). Cell division of transgenic cells was analyzed by CFSE staining after gating on Vα2, Vβ5, and Thy1.1+ cells. Samples were analyzed with a FACSCalibur™ flow cytometer using CELLQuest software (Becton Dickinson). 10,000 gated events were analyzed for surface marker expression and a minimum of 20,000 gated events was collected for CFSE analysis.
In Vitro Responses of CD4 Cells.
Donor CD4 cells were enriched from spleens of recipient mice at various times after cell transfer as indicated in the text for individual experiments. Anti-Thy 1.2 labeled magnetic beads (Miltenyi Biotech) were added to CD4 enrichment mAb from Stem Cell Technologies before magnetic sorting. Resulting populations of Thy 1.1+ cells were exclusively transgenic as indicated by vα2 and vβ5 expression. For in vitro restimulation, varying numbers of Thy 1.1+ cells from individual mice were cultured in triplicate in 250 μl in 96-well flat-bottomed plates (Costar) with 8 μg/ml OVA peptide, and with 2 × 105 APCs that were C57BL/6 spleen cells activated by overnight culture with dextran sulfate and LPS, and treated with mitomycin C (4). For limiting dilution analysis (LDA) of IL-2 and IL-4 secreting cells, serial dilutions of CD4 cells were plated in 36-well replicates in combination with APC and peptide. For ELISPOT analysis to enumerate IFN-γ–producing effectors, varying numbers of cells were cultured in duplicate with APCs and peptide for 24 h in Immunospot M200 plates (Cellular Technology Limited) that were coated with 4 μg/ml anti-IFN-γ capture antibody (R46A2 from BD Biosciences).
IL-2 and IL-4 secretion were assessed by bioassay (4). The sensitivity of detection is 1 pg/ml for IL-2 and 0.5 pg/ml for IL-4. For LDA, wells with cpm values greater than three standard deviations above the mean of values from 36 wells containing APCs were considered positive. Frequencies were calculated using maximum likelihood analysis. IL-1, -5, -6, -10, -12, -13, and IFN-γ were measured by ELISA as described (10) using capture and detecting mAb pairs from BD Biosciences. The sensitivity of detection for these cytokines is 10–30 pg/ml. Recombinant cytokines from R&D Systems were used for quantitation. IFN-γ ELISPOTs were developed by sequential incubation of plates with biotin-anti-IFN-γ (XMG1.2; BD Biosciences), strepavidin-alkaline phosphatase, and nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate substrate (Kirkegaard & Perry Laboratories) as described (17). ELISPOTs were counted with an ImmunoSpot Analyzer (Cellular Technology Limited).
Deficient CD4 Cell Expansion in B Cell–deficient Mice.
Our previous studies showed that CD4 cell responses are suboptimal in B cell–deficient mice as measured by a low frequency of Ag-specific cells (4) and limited development of effectors that secrete Th2 cytokines (10). To evaluate mechanisms that might affect CD4 priming at the level of Ag-specific CD4 cells, we established an adoptive transfer model using OT-II vα2, vβ5 TCR transgenic CD4 cells that are specific for OVA. The response of naive transgenic CD4 cells was followed in C57BL/6 or B cell deficient (μMT) recipients by flow cytometry after immunization with OVA protein. An increase in the number of transgene+ CD4 cells in the spleen was observed by d4 after immunization (Fig. 1 A). While similar kinetics were observed with both groups of recipients, a much lower level of donor cell accumulation occurred in the absence of B cells. The data suggest that the response of CD4 cells was not fully supported in μMT mice and was aborted. To determine if the reduced recovery of OT-II cells reflected more limited expansion, cell division of transgenic donor cells in recipients was analyzed after CFSE-labeling. OT-II cells did not begin to proliferate until d3 after immunization (Fig. 1 B). Substantially fewer cells divided in μMT mice, but division was at the same rate seen in control animals. By d7, the majority of cells in both groups of recipients had undergone several divisions, although a larger proportion of the population remains undivided in μMT mice (14%) than in C57BL/6 mice (3%). Together the results suggest that fewer cells become initially committed to division in the absence of B cells, and that their expansion is not sustained.
Defective Development of Th2 Cytokine-secreting Effectors in the Absence of B Cells.
Although the capacity of CD4 cells to produce effector cytokines does not depend upon division, several studies indicate DNA synthesis and entry into cell cycle does typically correlate with effector development, particularly for secretion of Th2 cytokines (2, 18). Although our previous study showed limited development of Th2 effectors in the absence of B cells, we could not directly compare Ag-specific populations from the intact animals (10). With the adoptive transfer model we could test whether cytokine production by Ag-specific cells was altered as a consequence of priming in μMT mice. Donor CD4 cells isolated from C57BL/6 and μMT mice on the basis of Thy1.1 expression on d4 after immunization of recipients were compared for their capacity to produce IL-2 and several effector cytokines after restimulation with peptide and APC. As seen in Fig. 2 A, while OT-II cells produced similar quantities of IL-2, IL-10, and IFN-γ irrespective of whether they were primed in normal or B cell–deficient mice, those from μMT mice produced markedly lower levels of IL-4 and IL-13. IL-5 was not detectable in this model.
To determine if limited production of Th2 cytokines reflected differences in the levels of secretion by individual cells and/or the frequencies of cytokine producers, we used LDA to evaluate IL-2 and IL-4 production by OT-II cells isolated from recipient mice. As shown in Fig. 2 B, at the single cell level, OT-II cells produced similar levels of IL-2 and IL-4/cell whether primed in normal or μMT mice. The frequencies of IL-2–producing cells were similar in both groups of recipients, but fivefold fewer IL-4 producers were detected among OT-II cells from μMT mice. These data establish that cells with the capacity to produce normal levels of IL-4 do develop in the absence of B cells, but that fewer cells are present in the effector populations from μMT mice.
Activated B Cells Restore CD4 Cell Responses in μMT Mice.
Our studies showed that splenic DC from μMT mice are not defective in their ability to initiate or support CD4 cell responses but are likely to become limiting (4, 10). To confirm that B cells can be responsible for deficient priming of CD4 effectors, C57BL/6 and μMT mice were reconstituted with activated, Ag-pulsed B cells. Provision of B cells fully restored expansion of OT-II cells in μMT mice (Fig. 3 A) as well as the capacity for Th2 cytokine secretion (Fig. 3 B). We showed previously that transferred B cells persisted throughout CD4 effector development in μMT mice (10) and here we find that the optimal time period for reconstitution of the response was during the initial 48 h (Fig. 3 C). While the numbers of IL-4–producing cells were augmented by the presence of B cells in μMT mice, frequencies of IFN-γ–producing effectors were not significantly altered (Fig. 3 D), as we showed previously for innate CD4 cells from μMT mice (10). The results suggest that the IFN-γ response which can develop before that of IL-4 (2, 19) can be optimally primed without a contribution from B cells.
Limited CD4 Responses in μMT Mice Are Not Restricted to the Spleen.
A previous study showed that B cell–deficient mice have reduced numbers of T cells and DCs in their spleens as well as defective splenic T zone architecture due to the absence of B cells during development (20). Although we find that OT-II cells localize normally in the spleens of μMT mice (unpublished data), their limited accumulation/expansion after priming might reflect not only limited availability of APC, but also defects in the microenvironment that include reduced numbers of stromal cells and expression of the lymphoid tissue chemokine, CCL21. However, T zone development as well as chemokine expression in LN is B cell independent and μMT mice have normal T cell and DC numbers in these tissues. Thus, if B cells directly rather than indirectly contribute to CD4 effector responses, their requirement should also be evident in the LN.
To test this hypothesis, C57BL/6 or μMT recipients were injected with CFSE-labeled OT-II cells and immunized with OVA and adjuvant (subcutaneously and intraperitoneally). An additional group of μMT recipients also received activated B cells. Division of the transgenic donor cells was assessed in LN at 2 1/2 d after priming. Fewer transgene+ cells were dividing at this time in μMT mice than in C57BL/6 mice, and the presence of B cells improved both expansion (Fig. 4 A) and recovery (Fig. 4 B). The data suggest that, as previously observed in studies mice rendered B cell deficient by antibody treatment (21), B cells are necessary for CD4 cell responses in LN as well as spleen. To determine if migrating cells from the spleen might contribute to the defective response in LN, we also evaluated cytokine secretion of OT-II cells primed in splenectomized recipients. As shown in Fig. 4 C, a more limited Th2 cytokine response was seen in μMT mice unless reconstituted with B cells, and there were no differences relative to the controls.
B Cell–derived Cytokines Do Not Have an Apparent Role in Sustaining CD4 Responses.
One mechanism by which B cells could contribute to Th2 differentiation is through cytokine production. We found that B cells that were activated in culture produced readily detectable levels of several cytokines that included IL-1, -2, -4, -6, -10, and -13, as well as IFN-γ (Fig. 5 A). In view of a major role for IL-4 in CD4 cell differentiation, and a potential in vivo role for B cell-derived IL-4 in Th2 development (13, 22), we isolated B cells from IL-4 knockout mice and assessed their capacity to restore responses of μMT mice. We also examined B cells from IL-6–deficient mice, since this cytokine can support differentiation and survival of Th2 cells (23) and IL-10–deficient mice, as IL-10–mediated down-regulation of IL-12 synthesis by DCs can facilitate Th2 responses (24). However, B cells from the cytokine-deficient animals were fully capable of restoring development of IL-13–producing effectors in μMT mice (Fig. 5 B). The levels of IL-4 secretion were also normal when μMT mice were reconstituted by B cells from the cytokine deficient animals (unpublished data). The data suggest either that redundancy of cytokine functions can compensate for a single deficiency, or that B cell production of IL-4, -6, or -10 is not obligatory.
A Role for Costimulation via B Cell Expression of OX40L in Sustaining CD4 Responses.
Like DC, activated B cells express multiple costimulatory molecules that are involved in the development of CD4 responses as revealed by the dramatic consequences of gene knockout studies (5, 7, 15, 25). However, activated T cells are less dependent upon costimulation. It is unclear whether continued availability is necessary once a response is underway when cytokines that can drive expansion and survival become produced in the local milieu. The OX40L/OX40 pathway that has been implicated in both T cell priming and T–B interactions (26–28) appears to also have the potential to contribute to effector responses in vivo (15, 29). We therefore determined if this pathway might play a role in B cell regulation of CD4 effector development in our model. We used B cells from the OX40L gene knockout animals (15) to reconstitute OT-II responses in μMT mice. These B cells expressed normal levels of class II MHC, as well as B7–2, and CD40 (Fig. 6 A). However, activated B cells that lacked OX40L were incapable of restoring the IL-4 and IL-13 responses whereas IFN-γ responses were normal in the absence of B cells and not further augmented by their addition (Fig. 6 B).
It is noteworthy that CD4 cell proliferation is defective in mice that lack OX40 (30), suggesting that a major contribution of costimulation through this pathway is to promote T cell expansion. OX40 deficient CD4 cells have been shown to become initially activated to secrete IL-2 and divide normally, but they are unable to sustain proliferation (29). As OX40L deficient mice have normal numbers of T and B cells, we asked whether OX40L signals from B cells affect the development of Th2 responses. OT-II cells were transferred into C57BL/6 or OX40L−/− mice and primed with OVA protein/adjuvant. An additional group of OX40L−/− recipients were then given C57BL/6 B cells as in the experiments with μMT mice. On d4 after immunization, the cytokine responses of reisolated transgenic donor cells were measured. As shown in Fig. 7, OT-II cells primed in OX40L−/− mice were defective in their production of IL-4 and IL-13 but not of IFN-γ or IL-2 (unpublished data). Importantly, B cells from normal mice fully restored defective Th2 priming in OX40L−/− recipients. Together with the experiments showing that OX40L−/− B cells do not restore the responses of μMT mice, the data reveal that OX40L expression on B cells can be crucial for the expansion/survival of Th2 cells.
Many studies indicate that B cell and T cell responses are coordinated to orchestrate cross regulation of expansion, survival, and differentiation via cooperative interactions. Although the essential role of CD4 cell–derived cytokines in B cell differentiation and responses has been widely studied, mechanisms that underlie the regulation of CD4 cell responses by B cells have not yet been fully explored. In this study, we determined that B cells can directly affect CD4 effector responses through costimulation by OX40L. We show that in the absence of B cells that express OX40L, Ag-induced expansion of CD4 cells and development of effectors that secrete Th2 cytokines are impaired.
Several studies have identified a link between B cell Ag-presentation and the induction of IL-4 secretion by CD4 cells (31, 32) but did not reveal whether additional signaling pathways were required. Although a recent report suggests that B cells can provide a source of IL-4 for differentiation of Th2 cells (13), this does not appear to be the sole mechanism to support Th2 responses since our results show that IL-4–deficient B cells were fully capable of eliciting Th2 cytokines. However, it is striking that several costimulation molecules have profound effects on the development of CD4 cells. It has been proposed that three major costimulatory pathways, CD28/B7, CD40L/CD40, and OX40/OX40L, act sequentially and synergistically to regulate CD4 cell responses from initial activation to optimal expansion, survival, and migration into B cell follicles (6). Although CD4 cell signaling through CD28, CD40L, or OX40 are associated with IL-4 induction (6), in vivo studies suggest that IL-4 production and responses that require IL-4 may be particularly dependent upon the OX40/OX40L pathway (27, 33, 34). Our data reveal that B cells can be an essential population to provide signals via OX40L to CD4 cells.
We envision that following CD4 cell activation by TCR-interactions with MHC/peptide complexes and CD28 engagement of B7 molecules, receptors such as CD40L and OX40 are induced and serve to amplify TCR signaling, as well as enable responding T cells to interact with B cells which constitutively express CD40 and up-regulate OX40L in response to activation. Temporally, OX40 expression on T cells coincides with induction of OX40L on B cells (27). Although interactions with DCs can initially support the response, they become limiting as T cells proliferate, and without further recruitment of DC, usage of B cells can become necessary to sustain the response. Our data suggest that CD4 cells can be maintained in this physiologic context by OX40 interactions with OX40L on B cells. Although this pathway can also contribute to the early activation phase of CD4 responses via DCs (15), our study of the knockout animals as recipients of Ag-specific CD4 cells shows that although priming occurs, Th2 responses are limited. Importantly, normal B cells are sufficient to restore Th2 effector development.
Our previous studies suggest that other APC can fulfill the role of B cells in CD4 cell expansion and Th2 development if made available in sufficient numbers during effector priming (4, 10). In addition, provision of costimulation to T cells via signaling with an agonist antibody to OX40 (29) is associated with increased frequencies of Ag-specific CD4 cells. Recent studies also implicate OX40 in survival of activated CD4 cells (35), suggesting that expansion together with increased lifespan may allow effectors to participate in a response for an extended period. Our results where a high number of Ag-specific naive CD4 cells are present at the time of immunization show that the capacity to secrete several effector cytokines, with the notable exception of Th2 cytokines, can develop independently of expansion that is sustained via OX40/OX40L interactions.
In view of our finding that IL-4 producing effectors, though reduced in number appear to be primed normally, and previous reports that OX40/OX40L interactions maintain later rather than initial CD4 cell expansion, we envision an indirect role of costimulation in the development of Th2 responses through effects on survival (35) and division (29). The data support the concept that sustained responses are necessary for optimal Th2 development (19). As the OX40 pathway has been associated with CD4 cell accumulation in germinal centers (36), it is attractive to speculate that B cell regulation of CD4 cell survival and response by OX40 in turn facilitates a sustained B cell response (37). Although cytokines produced by APC including IL-13, IL-1, and IL-6 can support Th2 differentiation, IL-4 is the key mediator to function in this capacity (38). The lag in optimal priming for IL-4 synthesis in our model may in part be due to dependence of responding CD4 cells upon autocrine secretion and usage of IL-4 to direct Th2 development (39) through induction of transcription factors that include Stat6, Gata3, and Gfl-1 (38, 40). Conditions of strong TCR stimulation and costimulation may be necessary to achieve sufficient endogenous IL-4 levels for commitment to a Th2 phenotype and achieve stable secretion after extended chromatin remodeling that accompanies cell division (41).
Our study shows that in vivo, OX40L expressed on B cells can play a crucial role in determining the outcome of a response in terms of the frequency and cytokine commitment of effectors and, from our previous work and that of others, in turn on persisting memory. In the absence of B cell involvement, CD4 cell responses, at least to many protein Ag would most typically be biased toward Th1 cytokines, as is often observed in responses where Ag is limiting or of low affinity. The data indicate that modulation of costimulation through OX40 can provide a key checkpoint in the regulation of CD4 cell responses.
This work was supported by National Institutes of Health grants AI32978 and AI46530 to L.M. Bradley and AG01743 and AG19249 to P.-J. Linton.
Abbreviations used in this paper: CFSE, 5-(and-6)-carboxyfluorescein diacetate succinimidyl ester; DC, dendritic cell; LDA, limiting dilution analysis.