In recent years, reagents have been developed that specifically target signals critical for effective T cell activation and function. Manipulation of the CD28/CD80/86 and CD40/CD154 pathways has exhibited extraordinary efficacy, particularly when the pathways are blocked simultaneously. Despite the reported efficacy of anti-CD154 in rodents and higher models, its future clinical use is uncertain due to reported thromboembolic events in clinical trials. To circumvent this potential complication, we developed and evaluated a chimeric Ab targeting CD40 (Chi220, BMS-224819) as an alternative to CD154. Although Chi220 blocks CD154 binding, it also possesses partial agonist properties and weak stimulatory potential. The anti-CD40 was tested alone and in combination with a rationally designed, high affinity variant of CTLA4-Ig, LEA29Y (belatacept), in a nonhuman primate model of islet transplantation. Although either agent alone only modestly prolonged islet survival (Chi220 alone: 14, 16, and 84 days; LEA29Y alone: 58 and 60 days), their combination (LEA29Y and Chi220) dramatically facilitated long term survival (237, 237, 220, >185, and 172 days). We found that the effects of Chi220 treatment were not mediated solely through deletion of CD20-bearing cells and that the combined therapy did not significantly impair established antiviral immunity.
The discovery of the multistep process required for effective T cell activation has elucidated potential targets for immunosuppression that are highly T cell specific (1, 2). In particular the need for additional costimulatory signals after interaction of the TCR with its cognate Ag has generated considerable excitement. Two of the best-known and -studied costimulatory pathways are the interactions between the Ig superfamily member CD28 and its ligands CD80 and CD86 and the TNF/TNF receptor family pairing CD40/CD154. Although manipulation of either of these pathways alone has produced encouraging effects in a myriad of models, it appears that even more potent effects can be achieved when these pathways are blocked simultaneously (3, 4, 5).
One of the most widely studied reagents used to block CD28/CD80/86 interactions is the Ig fusion protein construct CTLA4-Ig (abatacept). CTLA4 (CD152) is expressed on T cells and also binds to CD80 and CD86, but with higher affinity. In an effort to increase the avidity of CTLA4-Ig for CD80 and CD86, a novel mutant form, LEA29Y, was constructed. LEA29Y (belatacept) differs by two amino acid residues within the regions that bind to CD80 and CD86. These amino acid substitutions confer increased binding affinity to CD86 (slower rate of dissociation) without appreciably altering the interaction with CD80. The resultant improvement in binding affinity led to potent immunosuppressive properties in vitro and in vivo (6).
Perhaps the most effective single therapy tested to prevent transplant rejection in animal models has been Ab blockade of the CD40/CD154 pathway (7, 8, 9, 10). Questions remain, however, about the exact mechanism of action. Recent work has suggested that anti-CD154 therapy might not be simply acting to block costimulatory signals, but, rather, the binding of the Ab to CD154 on activated T cells may actually induce a signaling event that could promote cell death or, alternatively, it could be marking the subset of activated T cells expressing CD154 for complement-dependent or cell-mediated lysis (11, 12). Additional work is needed to fully elucidate the relevance of these findings.
Although treatment with anti-CD154 has proven extremely efficacious in rodent and nonhuman primate models of transplantation, there are significant barriers to its clinical development. In particular, early clinical trials in both transplantation and autoimmunity have observed an increased incidence of thromboembolic complications. Recent work has suggested that this could be due to the expression of CD154 by platelets and the need for soluble CD154 to stabilize thrombi; however the exact mechanism is not fully understood (13). An alternative to CD154 therapy would be targeting of the receptor, CD40. Although Abs to CD40 traditionally have been used to augment immune responses, recent work in rodent systems has shown that agonistic anti-CD40 Ab treatment may in some cases attenuate immune responses (14, 15, 16). In this report a chimeric mouse anti-human CD40 mAb (Chi220) was developed, which both blocks CD154 binding and has partial agonist properties, potentially circumventing the thromboembolic complications associated with the inadvertent targeting of platelets during anti-CD154 therapy.
In this study we describe the development and characterization of Chi220, including its efficacy, in models of humoral and cellular immunity. We also test the effects of LEA29Y and Chi220 therapy alone and in combination on islet allograft survival in nonhuman primates. We found that although the individual monotherapies moderately prolonged allograft survival, simultaneous blockade using the combination of LEA29Y and Chi220 dramatically extended islet allograft survival. In addition, we report that the effects of anti-CD40 treatment were not mediated by deletion of CD20-bearing cells and that this therapy did not significantly impair established antiviral immunity.
Materials and Methods
BALB/c mice (The Jackson Laboratory) were immunized with a human CD40-murine IgG2b chimeric protein. The initial injection was given in CFA, followed by subsequent booster injections in IFA or PBS. Spleen and lymph nodes were harvested and fused with X63-Ag8-653 mouse myeloma cells using standard methods (17). Cell suspensions were seeded into 96-well culture plates at a plating density of ∼170,000 total cells (preinfusion)/well. Cell culture supernatants were screened as described in the text, and ∼200 master wells were identified. Based on in vitro inhibition of B cell proliferation, the most strongly inhibiting master wells were selected for cloning. Cloning of the appropriate Ab-secreting cells was accomplished in a two-step process. First, cells from each master well were minicloned at a seeding density of 10 cells/well, after which the highest titrated CD40-specific miniclone well was formally cloned by a limiting dilution method.
Cells from the human B cell line Raji (American Type Culture Collection) were incubated with 2 or 20 μg/ml of the various anti-CD40 mAbs, followed by a second incubation in undiluted COS cell supernatant containing mCD8-CD154 fusion protein (sCD154). Bound sCD154 was detected by additional incubation of the cells with an FITC-labeled anti-CD8 mAb (BD Pharmingen) and analysis of the samples on a FACScan flow cytometer (BD Biosciences). The percent inhibition of binding was calculated by dividing the mean fluorescence of samples incubated with the Ab by the mean fluorescence of samples without Ab in the first incubation.
The VL and VH regions of the anti-CD40 mAb 2.220 were obtained by PCR cDNA, which was generated from RNA isolated from the hybridoma expressing the 2.220 mAb using an IgG1-specific or a Ck-specific antisense primer to obtain the VH or VL region, respectively. A poly-G tail was added to these single-stranded cDNAs. The variable regions were then amplified by PCR using as a sense primer an oligonucleotide containing a poly-C sequence, complimentary to the poly-G tail, and a nested set of antisense primers. The PCR product obtained was then inserted into a bacterial vector using restriction sites included in the primers. To generate the chimeric form of the Ab, the variable regions of the Ab were amplified by PCR using primers with a sequence encoding the signal sequence of the human Ab found to most closely match the 2.220 sequence. These PCR products were inserted into vector containing sequences encoding the constant regions of human κ or human γ1 to generate a complete L or H chain, respectively. These vectors contained the appropriate drug resistance genes for the generation and amplification of stable lines expressing the protein. Protein was produced by transient expression from COS cells, followed by protein A purification.
Primate immunization studies
Captive bred adolescent cynomolgus monkeys (Macaca fascicularis) were used in accordance with institutional guidelines. On study day 1, animals were immunized with SRBCs (1.7 ml/kg of a 10% solution i.v.) immediately before receiving a single i.v. dose of Chi220 at 10, 40, or 100 mg/kg or sterile PBS as a control (n = 4/group). On study day 149, animals were reimmunized with SRBCs and, in addition, received keyhole limpet hemocyanin (KLH;4 10 mg i.m.). Serum chemistry and hematologic parameters as well as PBL populations were monitored at selected time points. To assess efficacy, specific IgM and IgG Ab formation against the SRBCs and KLH immunogens was assessed from serum samples obtained just before immunization and weekly thereafter. Geometric mean Ab end-point titers were used when comparing Ab responses between groups, where the end-point titer is equivalent to the reciprocal of the greatest dilution of serum with an absorbance more than two times the mean plate background.
Captive bred adolescent rhesus monkeys (Macaca mulatta) were used as recipients and donors. The absence of preformed, donor-specific Abs in the recipient was confirmed before transplant. All potential donors and recipients were tested for anti-CMV Abs, and seronegative recipients were not paired with seropositive donors. Donor-recipient pairings were defined based on molecular typing using a panel of previously defined MHC alleles (18, 19, 20). Pairings maximized disparity at both class I and class II loci. Rejection was defined as two consecutive fasting blood glucose measurements >130 mg/dl on subsequent days.
Donor pancreatectomy and islet isolation
Donor pancreatectomy was performed under general anesthesia 1 day before transplantation. The splenorenal and splenocolic ligaments were divided so as to mobilize the spleen and tail of the pancreas. The head of the pancreas and second portion of the duodenum were mobilized using the Kocher maneuver. After administration of heparin (200 U/kg), the aorta was cannulated just above its bifurcation, and the animal was exsanguinated. Cold slush was immediately placed in the lesser sac and behind the body of the pancreas. The body and neck of the pancreas were carefully excised, taking care not to violate the pancreatic capsule.
Rhesus monkey islet isolation was completed via minor modifications of the automated method for human islet isolation (21, 22) using Liberase (Roche) at a concentration of 0.47–0.71 mg/ml. A three-layer, discontinuous Euroficoll gradient (densities 1.108, 1.097, and 1.037; Mediatech) and a Cobe 2991 blood cell processor (Gambro) were used for purification of islets from the pancreatic digest. Samples of the final islet preparation were stained with dithizone (Sigma-Aldrich), and the preparation was assessed by counting the number of islets in each of the following size ranges: 50–100, 100–150, 150–200, 200–250, 250–300, 300–350, and 350–400 μm. The data were mathematically converted to determine the number of islets with an average diameter of 150 μm and were expressed as islet equivalents (IEQ) (23).
Recipient pancreatectomy and intrahepatic islet cell transplantation
Total pancreatectomy, without duodenectomy or splenectomy, was performed at least 1 wk before transplant. The splenic, inferior, and superior mesenteric, middle colic, and portal veins were identified and preserved during dissection of the body of the pancreas. The duodenum was mobilized, and branches of the pancreaticoduodenal vessels that entered the pancreas were ligated and divided, leaving the duodenal branches intact. The common bile duct was identified and preserved. The main and accessory pancreatic ducts were ligated and divided, and the pancreas was removed from the abdominal cavity. The procedure was well tolerated, and two forms of pancreatic enzyme replacement were administered postoperatively, Pancrease, pancrelipase enteric coated microspheres (Ortho-McNeil Pharma), and Viokase V (Fort Dodge Animal Health), pancreatic enzyme supplement powder.
Overnight cultured islets were washed in medium and counted to determine the number of IEQ. Islets were then resuspended in 20 ml of medium supplemented with 200 U of heparin. Intrahepatic islet transplantation was performed via gravity drainage of islets into a sigmoid or branch of the left colic vein through a 22-gauge i.v. catheter.
Blood glucose monitoring and insulin administration
Fasting and postprandial blood glucose levels were monitored (Glucometer Elite; Bayer) twice daily (prebreakfast and postlunch) via ear-stick. Insulin (NPH, Ultralente; Eli Lilly) was administered three times daily in an attempt to maintain fasting blood glucoses <300 mg/dl in pretransplant pancreatectomized animals or those that had rejected their allografts.
Experimental groups and immunosuppressive protocols
Four protocols were tested: 1) LEA29Y alone, 2) Chi220 (anti-CD40), 3) LEA29Y combined with Chi220, and 4) LEA29Y combined with anti-CD20. LEA29Y was administered i.v. intraoperatively (20 mg/kg); on postoperative days 4, 7, and 14; then every 2 wk until day 100. Additional doses (20 mg/kg) were administered monthly through 6 mo. The chimeric anti-human CD40 mAb (20 mg/kg i.v.) was administered intraoperatively and on postoperative days 2, 4, 7, 9, and 11. LEA29Y and Chi220 are proprietary reagents and were provided by Bristol-Myers Squibb. The anti-human CD20 (rituximab) was purchased from the Emory University Hospital pharmacy and administered in an identical fashion to the anti-CD40 mAb. Survival of the islet grafts among experimental groups was compared using Mann-Whitney U test.
Detection of anti-donor Abs
The presence of detectable donor-specific alloantibody was determined using flow cytometry. PBLs served as the target cells for the pretransplant analysis. Leukocytes isolated from mesenteric lymph nodes obtained at the time of transplant were the target cells for the post-transplant assays.
Development of Chi220, a chimeric anti-human CD40 mAb
A murine mAb directed against human CD40 was generated using standard hybridoma selection and purification techniques (17, 24). A potentially important characteristic of any therapeutic Ab for transplantation would be to not only bind to human CD40, but also to interfere with ligand (CD154) binding. Consequently, Abs from all potential clones were assessed for their ability to inhibit the binding of a soluble recombinant murine CD8-human CD154 fusion protein, sCD154, to plate immobilized human CD40-Ig. This process led to the selection of four Abs (1.66-IgG2b, 2.36-IgG2a, 2.174-IgG1, and 2.220-IgG2a) capable of completely inhibiting the binding of sCD154 to CD40 bearing Raji cells (Fig. 1 a). After additional testing, including assessment of cross-reactivity with nonhuman primate cells, clone 2.220 was chosen for further development and evaluation.
Although mouse mAbs exhibit exquisite specificity, they also, by nature, possess immunogenic properties when used in distinct species. Humanized Abs are less likely to be immunogenic, but mutations introduced during the humanization process can affect Ag binding. Indeed, no humanized derivatives of 2.220 were identified as candidates for further development. In contrast, chimeric Abs, composed of unaltered murine variable regions and human constant regions, minimize immunogenicity while retaining the Ag-binding properties of the parent Ab. Using standard molecular techniques, a chimeric version (Chi220) of the murine 2.220 clone was generated by grafting the binding region of the murine Ab to a human IgG1 tail. To assess the activity of the chimeric Ab, resting human tonsillar B cells were incubated with sCD154, rabbit anti-human IgM-coated immunobeads, and the indicated concentration of anti-CD40 mAb or medium control. Chi220 effectively blocked sCD154-mediated B cell proliferation in a similar fashion to the parent murine Ab (Fig. 1 b).
Attempts at therapeutic manipulation of CD40 have generally focused on generating Abs with agonistic properties to augment immune responses (e.g., to enhance vaccines or boost antitumor activity) and identifying antagonists of CD40 signaling to inhibit immune responses. There are, however, recent reports of immune attenuation after treatment with an agonistic anti-CD40 mAb (14, 15, 16). Although we have shown that Chi220 can effectively block CD154-CD40 interactions, it was not known whether the binding of Chi220 elicited any signaling activity through CD40 or if it acted as a complete antagonist. To test this, freshly isolated human B cells were cultured in the presence of anti-IgM beads in combination with control or experimental Abs, including Chi220. Medium alone, control Ab (3C6), or an Ab specific for CD154 (h106) did little to stimulate B cell proliferation (Fig. 1 c). Interestingly both the murine and chimeric versions of clone 2.220 significantly facilitated proliferation of the stimulated B cells. The level of proliferation, however, was ∼10-fold lower than that induced by an anti-human CD40 Ab frequently used to activate CD40-bearing cells (clone G28.5), suggesting that clone 2.220 is a weak agonist. Although the effects of attenuated signaling through CD40 at the time of Ag exposure are unknown in vivo, it may predispose these cells to death, as has been shown in other systems.
Anti-CD40 therapy inhibits both primary and secondary humoral responses
Because CD40-CD154 interactions have been shown to be critical for the development of a productive humoral immune response, we next tested the ability of Chi220 to suppress primary and secondary T cell-dependent Ab responses in vivo. Groups of cynomolgus monkeys were immunized with SRBCs and received a single i.v. dose of either Chi220 at 10, 40, or 100 mg/kg or sterile PBS as a control (n = 4/group). Substantial suppression of the primary humoral immune response against SRBCs was observed at all three dose levels. On the peak day of the control primary anti-SRBC response (day 15), the mean primary IgG anti-SRBC Ab responses were suppressed by >30-fold in treated animals vs the control group (Fig. 2 a; p < 0.05).
In addition to assessing the influence of anti-CD40 treatment on primary Ab responses, we evaluated the impact of the single treatment on secondary responses to the same Ag. On day 149 after drug administration, after the serum levels of Chi220 had fallen below putatively immunosuppressive levels (<∼10 μg/ml) and the number of peripheral blood B cells had returned to pretreatment levels in all groups, the animals were immunized a second time with SRBCs. As expected, control animals mounted strong secondary IgG Ab responses to SRBCs (Fig. 2,b). Upon rechallenge, animals that had previously been treated with the lower dose (10 mg/kg) of Chi220 mounted primary IgM and IgG Ab responses to SRBCs that were generally comparable to the primary Ab response (150 days earlier) in control (PBS-treated) animals (data not shown and Fig. 2,b). Interestingly, the Ab response to SRBCs was still partially suppressed in the 40 mg/kg group and was substantially inhibited in the 100 mg/kg cohort (∼90% suppressed compared with the mean primary anti-SRBC Ab response of the control animals). Despite demonstrating a blunted response to the original immunogen (SRBCs), the treated animals were capable of generating a productive Ab response against a novel Ag (KLH) comparable in magnitude to that in control animals when challenged at the same time (Fig. 2 c). These experiments confirmed that all animals were immunocompetent and suggested that higher doses of Chi220 may facilitate Ag-specific hyporesponsiveness of T cell-dependent Ab responses.
Anti-CD40 treatment combined with LEA29Y promotes islet allograft acceptance
Recent outcomes in islet transplantation have dramatically improved with changes in isolation techniques and immunosuppressive agents. Unfortunately, the side effects associated with the immunosuppressive regimens limit the application to a select patient population. Given the efficacy of Chi220 to inhibit humoral immune responses in vivo, we sought to evaluate it in a nonhuman primate model of pancreatic islet transplantation alone and in combination with the mutant CTLA4-Ig molecule, LEA29Y.
The dose and schedule of administration were selected based on prior pharmacokinetic studies. Using these data, we constructed a regimen designed to maintain serum trough LEA29Y concentrations of ≥10 μg/ml for the first 3 mo and ≥2 μg/ml for the next 3 mo. The chimeric anti-CD40 mAb was given as an induction agent with animals receiving six doses in the first 2 wk (Fig. 3 a). Diabetes was induced by surgical pancreatectomy of recipient animals and was confirmed by a pretransplant i.v. glucose tolerance test. Donor-recipient pairings were defined based on molecular typing using a panel of previously defined MHC alleles (18, 19, 20). Pairings maximized disparity at both class I and II loci. Rejection was defined as two consecutive fasting blood glucoses >130 mg/dl on subsequent days. Intraportal infusion of allogeneic islets (>10,000 IEQ/kg) resulted in initial restoration of euglycemia and insulin independence in diabetic monkeys.
Treatment of pancreatectomized macaques with a 2-wk course of Chi220 alone only moderately prolonged allograft survival (14, 16, and 84 days; Fig. 3,b) relative to historic controls (mean survival time, 7 days). Similarly, administration of LEA29Y alone prolonged islet allograft survival, but recipients returned to hyperglycemia while receiving therapy (58 and 60 days; Fig. 3,c). Despite clear evidence in rodents that simultaneous blockade of both the CD40-CD154 and CD28-CD80/86 pathways results in synergistic immunosuppressive effects, a survival benefit has not been observed when tested in nonhuman primates (5, 25, 26). Given these data, we evaluated the effect of combined costimulation blockade in the nonhuman primate using Chi220 and LEA29Y. In contrast to the individual therapies, the combination of Chi220 and LEA29Y dramatically enhanced allograft survival (237, 237, 220, >185, and 172 days; Fig. 3 d).
Interestingly, despite the early loss of glycemic control in two of the three animals from the anti-CD40 alone group, these single-treated animals continued to have detectable allograft function (stimulated C peptide; data not shown) and required minimal or intermittently no exogenous insulin (data not shown). Histological analysis of these animals confirmed the presence of donor islets, but with a surrounding cellular infiltrate (Fig. 4,b) In contrast, examination of the liver from a representative animal treated with the combination of LEA29Y and Chi220 at 6 mo demonstrated healthy islets without any evidence of a surrounding infiltrate (Fig. 4,a). Upon further examination using immunohistochemistry, the majority of infiltrating cells in the anti-CD40 group appeared to be positive for CD8 (Fig. 4,d). Recipients treated with LEA29Y alone showed no residual function after rejection, and no remaining islets were found after extensive histological analysis (Fig. 4, c and e).
The development of anti-donor Abs after transplantation generally denotes a poor prognosis, with a higher proportion of these individuals progressing on to chronic rejection. When we evaluated the experimental groups, none of the animals treated with LEA29Y, either alone or in combination with Chi220, ever developed a measurable anti-donor Ab response (data not shown). Similarly, animals treated with anti-CD40 alone failed to develop an Ab response to the donor at the time of islet rejection (data not shown).
Effects of anti-CD40 are not mediated solely through B cell depletion
CD40 is expressed on numerous cell types, including B cells, lymphoid dendritic cells, and perhaps a subset of T cells (27). We have previously reported that treatment with Chi220 results in a transient depletion of peripheral B cells (28). To compare the effects of treatment with LEA29Y alone, Chi220 alone, or their combination, we monitored the various leukocyte lineages by flow cytometry. Although the absolute numbers of CD3+, CD4+, and CD8+ cells did not differ over time between the treated groups, animals treated with anti-CD40 alone or in combination with LEA29Y demonstrated an almost complete depletion of peripheral CD20+ B cells (Fig. 5, a–c). The depletion was rapid, but transient, because CD20+ cells could be found as early as 50 days post-transplant.
To determine whether the effects of the anti-CD40 treatment were mediated via depletion of B cells, we treated an additional group of animals with LEA29Y and a depleting anti-CD20 mAb (rituximab). The anti-CD20 treatment, which is used clinically for the treatment of B cell lymphomas, induced rapid depletion of CD19+ cells from the peripheral blood in a similar fashion to the anti-CD40 treatment (data not shown). Despite the profound depletion of B cells, animals treated with anti-CD20 and LEA29Y quickly rejected their islet allografts (survival time, 14 and 16 days; Fig. 5 d). This was in stark contrast to the group receiving the combination of LEA29Y and Chi220, which had a median survival time of >200 days.
Combined costimulatory blockade does not impair maintenance of humoral immunity to CMV
The role of the CD40/CD154 pathway in the generation of an effective Ab response has been known for over a decade (29). In this report we have shown that treatment with anti-CD40 can dramatically inhibit a primary Ab response to a neo-Ag (SRBCs). The impact of this pathway on the maintenance of humoral immunity, however, is an important question that requires additional study. Because CD40 is expressed on B cells, but presumably not on terminally differentiated plasma cells, the subtype primarily responsible for the maintenance of serum Ab titers, we hypothesized that treatment with Chi220 would not interfere with the maintenance of humoral immunity despite causing a transient depletion of B cells. To test whether Ag-specific Ab titers were affected by significant B cell depletion, we sampled and analyzed the CMV-specific Ab response before, during, and after treatment with the various protocols. CMV was selected because of its relevance to transplantation and convenience, because all animals were seropositive before the initiation of the study.
Not surprisingly, treatment with LEA29Y alone did not significantly alter the mean titer of anti-CMV Ab. Interestingly, despite effectively depleting B cells, administration of anti-CD40 alone or in combination with LEA29Y did not significantly diminish anti-CMV Ab levels (Fig. 6,a). In addition to measuring the virus-specific Ab titers, we monitored CMV viral load from recipient blood samples using a quantitative PCR assay (Fig. 6 b). Similar to the results for the Ab titers, none of the animals in any of the treatment groups developed a significant amount of viral reactivation. One animal in each of the experimental groups developed a transient, detectable viral load, but all subsequently controlled the virus without further intervention. Importantly, no animal developed any signs of CMV-related disease at any time. Included for comparison are blood samples from a control animal from a separate, unrelated experimental protocol, which was diagnosed clinically with CMV disease (diarrhea, weight loss, etc.) and subsequently recovered after treatment with antiviral therapy.
In recent years many of the pathways and molecules that are critical to T cell activation and function have been characterized. Initial observations included the description of cell surface proteins involved in providing accessory signaling during T cell activation. There are now dozens of molecules with reported costimulatory potential. It has become clear that their relative importance can vary dependent upon the model, type of response (naive or memory), or timing within the immune response (early activation and proliferation vs late effector function, etc.). Although disruption of any individual pathway may temper an immune response, it is now known that simultaneous blockade of multiple pathways results in synergistic suppression. For example, when either CTLA4-Ig or anti-CD154 is administered alone in attempts to block the CD28-CD80/86 or CD40-CD154 pathway, respectively, improved survival was noted in rodent models of transplantation. Notably when the same two pathways were interrupted simultaneously, skin allograft survival, a rigorous immunological challenge, was dramatically prolonged (5). Interestingly, this observation of synergy has yet to be replicated in primate transplant models, perhaps due to the limited demonstrated efficacy of CTLA4-Ig in these models (25, 28, 30).
Although extremely effective in models of autoimmunity and transplantation, the exact mechanism by which anti-CD154 exerts its immunomodulatory effects is still in question. The traditional dogma suggests that interruption of the CD40-CD154 interaction by administering blocking Ab would deprive the T cell or APC of critical costimulatory signals that are required for the generation of an effective immune response. Recent rodent data suggest that the mechanism of action of anti-CD154 may not be through blockade of costimulation, but, rather, through depletion of activated T cells expressing the ligand (12). Although this is an intriguing hypothesis, it does not fully explain why anti-CD154 is so uniquely effective relative to other depleting Abs targeting T cell Ags (ex anti-IL-2R) or why memory T cells, which purportedly express CD154, but are refractory to anti-CD154 treatment, would be resistant to a depletional mechanism.
In contrast to Abs targeting CD154, Abs against CD40 are conventionally thought of as immunostimulatory and are commonly used to promote immune responses both in vitro and in vivo (14, 15). More recently, however, it has been shown in a rodent model of autoimmunity that agonistic Abs against CD40 can also have therapeutic potential by inhibiting self-directed responses (16). In the current study we have developed and evaluated a chimeric anti-human CD40 mAb with potent immunosuppressive properties. We show that anti-CD40 treatment can effectively inhibit primary humoral responses and at high doses facilitate Ag-specific hyporesponsiveness of T-dependent Ab responses upon rechallenge. Interestingly, treatment with the Ab results in transient depletion of peripheral B cells, which express CD40. One possible mechanism of action could be direct depletion of cells expressing CD40, including APCs. Given that Chi220 consists of a human IgG1 tail, an isotype known to fix complement and facilitate Ab-dependent cellular cytotoxicity via its interactions with FcRs, it would not be surprising if one mechanism is cellular depletion. Although there is a limited ability to dissect the role of depletion in nonhuman primates, we do show that the depletion of B cells alone is not solely responsible for Chi220’s immunosuppressive effects. In addition to APCs, there are data to suggest that activated T cells may express CD40. These studies suggest that CD8 T cells receive CD4 help directly through CD40 and that this is necessary for the effective generation of CD8 memory (31). Given these data, it may be that the anti-CD40 Ab works to deplete activated T cells as well as APCs. However, other studies fail to support a role for CD40 on CD8+ T cells. Moreover, the immunomodulatory effects of a short course of Chi220 persist much longer than the period of cell depletion. More likely is the potential for a combined mechanism of action, where depletion plays a critical role, but is not sufficient alone.
A second mechanism of action could involve activation-induced cell death of CD40-bearing cells. For example, anti-CD40 therapy has been shown to promote apoptosis of human CD40-bearing multiple myeloma cells (32). This effect appears not to be due to depletion, but, rather, to ligation, which increases the susceptibility of the tumor cells to activation-induced cell death when the Ab treatment is combined with additional chemotherapy. We have shown that not only does Chi220 block CD154 binding, but it also acts as a partial agonist to deliver weak proliferative signals in vitro. Although fully agonistic Abs targeting CD40 augment immune responses, weak partial signals may not only limit the response, but may actually promote the cell’s susceptibility to anergy, death, or control by other immune mechanisms. These partial signals through CD40 may occur very early as the immune response is initiated and could influence the resulting cascade of immune activation even if the cells are eventually depleted. This concept is supported by a recent study demonstrating the reciprocal effects of differential signaling through CD40 in a mouse model of leishmaniasis (33). Thus mechanisms underlying the immunosuppressive effects of the anti-CD40 tested in this study may differ from those in prior studies of nondepleting and antagonistic anti-CD40 mAbs (30, 34).
The implementation of novel immunosuppressive strategies in clinical practice is the principal focus of transplantation research. The major limitations of current immunosuppression are the concomitant nonimmune side effects and inability to prevent chronic rejection. For example, calcineurin inhibitors, the mainstay of immunotherapy in transplantation, are associated with a host of side effects, including hypertension, diabetes mellitus, dyslipidemia, accelerated cardiovascular disease, etc. Not only are these effects deleterious to patient health, but they also limit the population that may benefit from life-saving therapy. The development of reagents that are both immunospecific and nontoxic is essential as the field of transplantation moves forward. Unfortunately, one of the most promising reagents, anti-CD154, has been hampered by clinical complications. These same sequelae have now been observed in primate models as well (35, 36). Importantly, a comprehensive gross and histologic assessment in both toxicity and transplant studies using Chi220 in rhesus and cynomolgus macaques showed no evidence of thromboembolism. In addition, animals in the toxicity studies were free of nephrotoxicity, hypertension, and glucose intolerance, side effects commonly associated with calcineurin inhibitor therapy (data not shown). This is not surprising because the targets of both LEA29Y and Chi220 are immunospecific and would be expected to avoid the limitations of currently used maintenance immunosuppressants, which target pathways used in many organs (i.e., calcineurin, mammalian target of rapamycin, and steroid receptors).
In this report we demonstrate that a protocol using combined CD40/CD28 blockade effectively promotes the survival of allogeneic islets. In contrast to previous reports, when the addition of CTLA4-Ig to either anti-CD40 or anti-CD154 failed to provide a survival benefit in primate studies, the combination of LEA29Y, a high affinity variant of CTLA4-Ig, and Chi220 acted synergistically to prolong allograft survival. The use of a partial agonist Ab lends potential insight into the mechanism of CD154-CD40 pathway manipulation. Our data suggest that the original paradigm of costimulation blockade using antagonist-blocking Abs that solely inhibit ligand-receptor interactions may need to be modified. A newer model might also include the use of Abs or molecules that induce weak signaling through critical pathways or promote short term depletion cellular subsets, thereby promoting active manipulation of the immune response toward a desired outcome. In this study we show that short term treatment with a partial agonist Ab against CD40 has powerful synergy with CD28 blockade. The targeting of CD40 may circumvent the major barrier to the clinical translation of a reagent that blocks the CD40-CD154, namely, thromboembolic events. Our results in nonhuman primates provide a strong rationale for the implementation of clinical trials to test these strategies in human islet transplantation.
We are indebted to the excellent support from the Yerkes veterinary staff, in particular, Jack Orkin, Dan Anderson, and Rachel Fest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants U19-AI44644 and U19-AI51731 with additional support from Juvenile Diabetes Research Foundation Center Grant 4-2001-922, the Engineering Research Center Program of the National Science Foundation under Award EEC-9731643, the Carlos and Marguerite Mason Trust, and the Livingston Foundation.
Abbreviations used in this paper: KLH, keyhole limpet hemocyanin; IEQ, islet equivalent.