|Home | About | Journals | Submit | Contact Us | Français|
A single dose of CTLA4Ig, an inhibitor of CD28-mediated T cell costimulation, given 2 days after transplantation induces specific unresponsiveness to alloantigens in vivo. However, the mechanisms responsible are unknown. Using pigeon cytochrome c as a model Ag, we monitored the effect of CTLA4Ig on the fate of Ag-reactive T cells in normal mice and on pigeon cytochrome c-specific TCR transgenic cells adoptively transferred into congenic mice. CTLA4Ig significantly inhibits immunization with pigeon cytochrome c. In particular, ELISA and ELISPOT assays indicate an 80 to 90% reduction in Th1 (i.e., IL-2 and IFN-γ) cytokine production and in the numbers of cytokine-producing cells. Interestingly, despite this profound reduction in cytokine-producing cells, Ag-reactive T cells expand in CTLA4Ig-treated animals, although the degree of expansion is reduced by 50% compared with that in control Ig-treated animals. Thus, loss of Th1 cytokine production in CTLA4Ig-treated animals is not fully explained by the decreased expansion of Ag-specific T cells. These results suggest two mechanisms of action for CTLA4Ig in vivo: inhibition of expansion of Ag-reactive cells and induction of anergy in the residual population.
Naive T cell activation requires two signals (1). The first signal is transmitted following engagement of the TCR by Ag presented on a self-MHC molecule. The best characterized costimulatory signaling pathway is mediated by CD28, a surface molecule present on most human and all murine T cells (2). In vitro studies have demonstrated that blockade of T cell costimulation via the CD28 signaling pathway results in the development of Ag-specific T cell anergy (3–5). Consistent with this, preventing CD28 from binding to its ligands through the use of a soluble competitive inhibitor, CTLA4Ig, has been demonstrated to inhibit a variety of in vivo immune responses (6–9). However, the mechanism of immunosuppression induced by CTLA4Ig in vivo remains uncertain. Immunization with nominal Ag in the presence of CTLA4Ig resulted in inhibition of primary humoral responses to T cell-dependent Ags, but did not result in tolerance, as evidenced by a normal immune response upon rechallenge (6, 10). Thus, T cell-dependent B cell help was prevented, but the effects were not permanent. Similarly, administration of CTLA4Ig at the time of allogeneic transplantation, while prolonging allograft survival, ultimately failed to prevent rejection of the allogeneic tissues (8). These findings suggested that if anergy was induced by blockade of CD28 costimulation at the time of initial Ag exposure, it was transient, and affected T cells eventually recovered function. Delaying the administration of a single dose of CTLA4Ig until 2 days following transplantation, however, resulted in long-term survival of the allograft as well as tolerance toward subsequent challenge with allo-Ag in vivo (11, 12). Examination of functioning allografts removed from these tolerized animals demonstrated marked lymphocytic infiltrates that secreted IL-4 and IL-10 characteristic of a Th2 phenotype. Therefore, rather than T cell anergy, these findings suggested that a diversion of the immune response to allo-Ags toward a Th2 pathway might mediate the immune tolerance induced by delayed administration of CTLA4Ig. Yet another mechanism for CTLA4Ig has been suggested by recent studies showing that CD28 stimulation induces bcl-x, a homologue of bcl-2 that may be critical to T cell survival (13). This raises the possibility that CTLA4Ig administration could lead to deletion of Ag-specific T cells.
Distinguishing between these mechanisms of action requires the ability to identify and quantitate potentially Ag-reactive T cells through the course of an immune response. Specific populations of reactive T cells, identified by the restricted use of specific TCR-variable regions, have been shown to be stimulated and expanded in vivo following immunization with defined antigenic peptides (14–16). In one such model, immunization of B10.BR (H-2k) mice with a peptide consisting of residues 88 to 104 of pigeon cytochrome c (PCC)5 results in reproducible expansion of CD4+ T cells expressing TCRs that contain Vα11 and Vβ3 (14). Treatment of these animals with CTLA4Ig following immunization with PCC permits assessment of the in vivo effect of costimulatory blockade on a population of Ag-reactive T cells. Here, using the B10.BR model as well as adoptive transfer of T cells from TCR-transgenic animals to syngeneic hosts, we demonstrate that immunosuppression with CTLA4Ig has two effects: blunting the expansion of Ag-reactive T cells and induction of anergy in the residual population.
B10.BR mice (8–12 wk old) were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained in a pathogen-free facility. Mice were immunized with 100 µg of PCC fragment (amino acid residues 88–104) emulsified in CFA. In some experiments, mice also received an i.p. injection of staphylococcal enterotoxin A (SEA; 0.4 µg) at the time of immunization. Following immunization, mice were injected with either human CTLA4Ig (200 µg; generous gift from P. Linsley, Bristol-Myers-Squibb, Seattle, WA) or control human IgG (200 µg), given as a single i.p. dose 2 days after immunization.
B10.BR mice were immunized with 100 µg of keyhole limpet hemocyanin (KLH; Calbiochem, La Jolla, CA) and subsequently treated with CTLA4Ig or control human IgG as indicated above. The DTH response was assessed by the change in ear thickness on rechallenge with KLH 10 days postimmunization.
Serial dilutions of mononuclear cells from the draining lymph nodes were cultured with 4 × 105 mitomycin C-treated splenocytes in the presence or the absence of intact PCC (100 µg/ml) for 72 h. Each well was pulsed with 1 µCi of [3H]methyl-thymidine (ICN Biomedicals, Costa Mesa, CA) during the last 8 h of culture for determination of proliferation.
Purified lymph node cells were incubated with mAbs specific for Vα11 (FITC-conjugated; clone RR-18, PharMingen, San Diego, CA), Vβ3 (phycoerythrin-conjugated; clone KJ25, PharMingen), and CD4 (biotin-conjugated; clone L3T4, PharMingen). In experiments using transgenic T cells, FITC-conjugated anti-2B4 Ab (17) was substituted for Vα11. In experiments examining T cell activation, biotin-conjugated CD69 (clone H1.2F3, PharMingen) and phycoerythrin-conjugated CD4 (clone L3T4) were used. Cells were subsequently incubated with streptavidin-Red 670 (Life Technologies, Gaithersburg, MD). Three-color flow cytometry was performed, and 50,000 to 100,000 events were collected for each sample.
Lymph nodes and spleens were harvested from 2B4 transgenic mice (H-2k) whose T cells express a TCR specific for residues 88 to 104 of PCC (18). The mononuclear cell fraction was isolated, and an aliquot was analyzed by flow cytometry for the presence of transgenic TCR. The cells were resuspended in PBS at a concentration calculated to contain 20 × 106 transgenic T cells/ml. Naive nonirradiated B10.BR mice were injected i.v. with 250 µl of this suspension, and 24 h later were immunized with PCC as described above.
Isolated lymph node cells (2 × 106/ml) were cultured with mitomycin C-treated splenocytes (8 × 106/ml) in the presence or the absence of intact PCC (100 µg/ml). Culture supernatants were harvested after 24 or 48 h and analyzed by sandwich ELISA assay, as previously described (19). ELISPOT assays were performed using a gel substrate method, as previously described (20). For ELISPOT assays, serial dilutions of lymph node cells (8 × 104 to 2 × 106 cells/well) were stimulated with 100 µg/ml of intact PCC for 16 h before harvest. The Abs used for each cytokine were: IL-2, JES6–1A12; biotinylated IL-2, JES6–5H4; IFN-γ, R4–6A2; biotinylated IFN-γ, XMG1.2; IL-4, 11B11; biotinylated IL-4, BVD6–24G2; IL-10, JES5–2A5; and biotinylated IL-10, SXC-1. All Abs were purchased from PharMingen.
PCC peptide was purchased from Macromolecular Resources, Inc. (Fort Collins, CO). SEA was purchased from Toxin Technology (Sarasota, FL). CFA, mitomycin C, 5-bromo-4-chloro-3-indoly phosphate, 221 alkaline buffer solution, and intact PCC were purchased from Sigma Chemical Co. (St. Louis, MO).
In agreement with our previous results using allogeneic tissues, delayed administration of CTLA4Ig inhibited priming and in vivo responses following immunization with nominal Ags. As demonstrated in Figure 1, administration of a single dose of CTLA4Ig 2 days after immunization with PCC resulted in an 80% reduction in the proliferative response of draining lymph node cells to restimulation with intact PCC. Similarly, as shown in Figure 2, administration of CTLA4Ig 2 days following immunization with KLH significantly inhibited the development of DTH to rechallenge with this Ag (change in ear thickness at 48 h, 9.7 mm in CTLA4Ig-treated animals vs 18.5 mm in control Ig-treated animals; percent inhibition, 47.5%; n = 3). Thus, a single dose of CTLA4Ig given 48 h after immunization is sufficient to inhibit both in vitro and in vivo responses to Ag rechallenge. Further, as DTH responses are largely mediated by Th1 cells (21), inhibition of this response by delayed CTLA4Ig administration is consistent with the diminished immune reactivity to solid organ allografts previously seen (12).
Diminished DTH responses could result from a failure of stimulated T cells to secrete IL-2 and IFN-γ, thus limiting proliferation and subsequent differentiation to a Th1 phenotype. Immunization using CFA as an adjuvant preferentially results in the secretion of Th1 cytokines (22, 23). Naive B10.BR animals were immunized with PCC in CFA, and mononuclear cells of the draining lymph nodes from CTLA4Ig-treated and control animals were isolated after 10 days and restimulated in vitro to determine the cytokine profiles of the reactive cells (Fig. 3). The levels of IL-2 and IFN-γ in culture supernatants from CTLA4Ig-treated animals were reduced by at least 66% and 80%, respectively, compared with those in controls. Consistent with other reports, neither IL-4 nor IL-10 could be detected in primary restimulation cultures (24–26).
To determine whether the diminished levels of Th1 cytokines reflected a diminished number of cytokine-producing T cells or simply diminished secretion of these cytokines by reactive T cells in the culture, ELISPOT assays were performed. As shown in Figure 4, treatment with CTLA4Ig resulted in a reduction in the numbers of T cells secreting IL-2 (>75%) and IFN-γ (>88%). The reduction in cytokine levels detected in culture supernatants, therefore, reflects a loss in the number of Ag-specific T cells capable of secreting Th1 cytokines. Moreover, as the level of reduction in cytokine secretion detected by ELISA is similar to the decline in the numbers of cytokine-secreting cells detected by ELISPOT, the results suggest that the functional capacity of any individual cell that escapes immunosuppression by CTLA4Ig is unimpaired.
One potential mechanism for the observed reduction in cytokine-secreting cells would be deletion of Ag-reactive T cells following immunization as a result of CTLA4Ig treatment. Previous studies have demonstrated that PCC-responsive T cells in B10.BR mice are restricted primarily to CD4+ cells expressing TCRs containing Vα11 and Vβ3 regions (14). While immunization with PCC results in expansion of the fraction of CD4+Vβ3+ T cells that also express the Vα11 region from a mean of 8.3 to 15.8%, immunization with a different peptide, residues 34 to 45 of hen egg lysozyme, has no effect on the percentage of these cells (data not shown). Mice treated with 200 µg of CTLA4Ig 2 days after immunization with PCC also expanded this population (to 12.5%), although the increase in CD4+Vβ3+Vα11+ cells was only 43% compared with that in the controls.
As not all CD4+Vβ3+Vα11+ T cells in B10.BR mice are responsive to PCC, a background of non-PCC-reactive cells could obscure important quantitative changes in PCC-reactive cells. To improve the detection of potential differences between untreated animals and animals treated with CTLA4Ig, we used superantigen-mediated deletion to selectively eliminate non-PCC-specific CD4+Vβ3+Vα11+ T cells. Previous work has demonstrated that treatment of B10.BR mice with SEA results in deletion of CD4+Vβ3+ T cells (27). Further, coadministration of PCC to these animals at the time of SEA injection blocks deletion of PCC-reactive Vβ3+ T cells (14). Therefore, this model permits a selective enhancement of PCC-specific T cells within the draining lymph nodes of the immunized animals. As shown in Table I, the proportion of CD4+Vβ3+Vα11+ T cells in the draining lymph nodes was increased to 37.5% by concomitant administration of SEA to PCC-immunized animals. In animals also treated with CTLA4Ig on day 2, the proportion of CD4+Vβ3+Vα11+ T cells was 25.3%, an expansion of only 53% compared with that in control IgG-treated animals (p < 0.05), confirming the results seen in the absence of SEA. Administration of CTLA4Ig on day 0 yielded similar results. CTLA4Ig did not have any significant effect on the overall size or the cellularity of draining lymph nodes (19.1 ± 8.1 × 106 cells in control IgG-treated animals (n = 23) vs 21.0 ± 12.1 × 106 cells in CTLA4Ig-treated animals (n = 18); p > 0.05). Furthermore, analysis of animals killed on day 28 showed equivalent levels of CD4+Vβ3+Vα11+ T cells in control Ig- and CTLA4Ig-treated mice (Table I). Thus, we found no evidence for late changes, either expansion or exaggerated loss of Ag-reactive T cells, in CTLA4Ig-treated mice compared with controls.
As an additional test of the effect of CTLA4Ig administration on T cell expansion following immunization with nominal Ag, T cells from 2B4 Tg were adoptively transferred into naive B10.BR mice that were subsequently immunized with PCC. These T cells, which bear a TCR specific for PCC in the context of H-2k, were identified by a specific Ab to the transgenic α-chain. Draining lymph nodes of mice adoptively transferred with 2B4 Tg cells contained 0.08 ± 0.02% transgenic T cells 7 days after control immunization with PBS. When mice were immunized with PCC, 2B4+ T cells comprised 1.98 ± 0.02% of cells from draining lymph nodes, indicating a 25-fold expansion in vivo, consistent with previously reported data using adoptive transfer of OVA-specific T cells (28). Treatment with CTLA4Ig resulted in the expansion of 2B4 cells to 0.80 ± 0.06% of the mononuclear cell population, an expansion of 10-fold, which is only 41% of that seen in the PCC-immunized animals treated with control IgG (p < 0.01). A representative experiment is depicted in Figure 5. As the total numbers of mononuclear cells in the draining lymph nodes of control and CTLA4Ig-treated animals were not significantly different (control, 27.5 ± 5.2 × 106 cells; CTLA4Ig, 22.5 ± 10.8 × 106 cells), these results again demonstrate that treatment with CTLA4Ig significantly inhibited the expansion of Ag-reactive 2B4+ T cells without evidence for deletion below base line levels. In addition, the failure of 2B4 cells to expand to control levels in animals treated with CTLA4Ig cannot be attributed solely to a lack of T cell activation, as CD69 was expressed on over 75% of Ag-specific T cells in CTLA4Ig-treated animals (data not shown).
Functional assays revealed that the inhibition of IL-2 and IFN-γ production, as assessed by ELISPOT following in vitro restimulation with PCC, was greater (75% and 95%, respectively) than the diminution of 2B4+ T cells. As with the experiments in normal B10.BR mice (Fig. 3 and Fig. 4), neither IL-4 nor IL-10 was detected. These results are similar to those found using the non-Tg system (Fig. 4). As the adoptively transferred transgenic T cells are the overwhelmingly dominant Ag-reactive population, these results provide further evidence for functional inhibition of Ag-reactive cells following CTLA4Ig administration.
T cell costimulation via the CD28 signaling pathway plays a critical role in the establishment of resistance to anergy induction (3). In addition, stimulation of the CD28 pathway is important for the secretion of cytokines necessary for T cell proliferation and differentiation of naive T cells into cytotoxic T cells (29, 30). Initially described as a system involving two components, B7 and CD28, successive studies have demonstrated a second receptor highly homologous to CD28, called CTLA-4, and multiple ligands for both CD28 and CTLA-4 (2). At present, two ligands for CD28/CTLA-4, B7–1 (CD80) and B7–2 (CD86), have been identified and cloned, and a third ligand, B7–3/BB-1, has been postulated on the basis of Ab binding studies (31, 32). Differences in temporal expression of B7–1 and B7–2 as well as differential up-regulation of CD28 and CTLA-4 following T cell activation suggest that a considerable degree of complexity exists within this signaling pathway. This complexity is further highlighted by recent studies indicating that CTLA-4 transduces a negative signal to the T cell (33). The ability of CTLA4Ig to competitively inhibit binding of both B7 molecules to either CD28 or CTLA-4, therefore, will interrupt multiple potential interactions that occur over the first 48 h following T cell stimulation. The net effect of CTLA4Ig and its mechanism of action will be a function of the roles of the CD28/CTLA-4 and B7 pathways in an immune response, the relative ability of CTLA4Ig to block each pathway, and the timing of CTLA4Ig administration.
The present study demonstrates that the effect of CTLA4Ig administration in vivo, using a strategy that induces transplantation tolerance, results in at least two alterations in normal T cell priming. Following immunization, PCC-reactive T cells from animals treated with a single dose of CTLA4Ig expand to 50% of the level achieved in control animals. This finding is consistent with initial results reported by Vella et al. using PCC in otherwise unmanipulated animals (34). However, the concomitant use of SEA and the adoptive transfer of transgenic T cells allowed us to increase the sensitivity of this assay, with a detectable in vivo expansion of PCC-reactive cells up to 25-fold in response to immunization. Comparison of in vivo expansion of Ag-reactive T cells from animals treated with a single dose of CTLA4Ig at the time of immunization with those treated 2 days after immunization revealed no difference in the absolute number of PCC-reactive cells within the draining lymph node or the kinetics of expansion. This suggests that the expansion observed in the day 2 treated group is not the result of unimpeded costimulation through the CD28 pathway during the first 48 h of Ag encounter. The restricted expansion of PCC-specific T cells in these animals is correlated with diminished in vitro and in vivo responses on rechallenge with Ag and is probably a major mediator of CTLA4Ig-induced tolerance.
Nonetheless, blockade of expansion by only 50% is unlikely by itself to be responsible for tolerance. This strongly suggests that other mechanisms must be active as well. Indeed, our studies show that despite expansion and activation of Ag-reactive cells in CTLA4Ig-treated animals, Th1 cytokine production by these cells is almost completely suppressed. Interestingly, only a 67% blockade in Ag-reactive T cell expansion was observed recently in animals treated daily with high dose cyclosporine following immunization with OVA, while complete blockade was seen with daily CTLAIg (35). Yet, this regimen of cyclosporine treatment induces transplantation tolerance, while daily CTLA4Ig does not (8, 11). Perhaps, the expansion of residual regulatory populations of T cells in CTLA4Ig-treated animals may be important in maintaining tolerance. Consistent with this, induction of transplantation tolerance by day 2 of treatment with CTLA4Ig is associated with an increase in Th2-secreting cells within the renal allografts in these animals (12). Moreover, despite repetitive doses of exogenous IL-2 at the time of CTLA4Ig administration, 50% of animals did not reject their grafts, and those that did exhibited a markedly delayed tempo of rejection. This would suggest that the functional alterations induced by CTLA4Ig are not solely due to a failure of adequate IL-2 secretion during the initial expansion phase of the T cell response. However, a preferential loss of Th1 cell function as a result of either specific T cell anergy or immune deviation to a Th2 profile (which may require repetitive Ag stimulation to detect in these models) would be consistent with the effects of CTLA4Ig on allograft survival (12) and the data presented in this manuscript.
The failure of CTLA4Ig to block expansion completely may be due to the potent effects of small amounts of residual IL-2 or to other cytokines that are comitogenic but are unable to prevent anergy induction, as recently demonstrated in vitro (36, 37). Our studies also suggest that T cell proliferation and avoidance of anergy may be two distinct biologic phenomena in vivo, and that Ag-specific immunosuppression following the administration of CTLA4Ig probably reflects an alteration in T cell effector function, perhaps as a result of interference with B7–1- or CTLA-4-mediated signaling, as both of these molecules are maximally up-regulated 24 to 48 h after activation of APCs or T cells, respectively. The enhanced inhibition observed in animals treated with delayed CTLA4Ig may also indicate that unimpaired early signaling via B7–2:CD28, perhaps promoting a Th2 phenotype, is important for this result.
We thank Guoli Zheng for technical assistance, and Jonni Moore and Steven Eck for helpful discussion.
1This work was supported by an AGA Industry Scholar Award (to T.A.J.) and National Institutes of Health Grant AI37691 (to L.A.T.).
5Abbreviations used in this paper: PCC, pigeon cytochrome c; SEA, staphylococcal enterotoxin A; DTH, delayed-type hypersensitivity; KLH, keyhole limpet hemocyanin; 2B4 Tg, T cells expressing 2B4 transgenic TCR; ELISPOT, enzyme-linked immunospot assay.