Analysis of the Requirements for the Induction of CD4⁺ T Cell Alloantigen Hyporesponsiveness by Ex Vivo Anti-CD40 Ligand Antibody¹

Bone marrow transplantation (BMT)³ is being utilized as the standard of care for the treatment of a number of malignant and nonmalignant disorders of both hematologic and nonhematologic origin. Graft-vs-host disease (GVHD) remains a major cause of morbidity and mortality after sibling, matched unrelated, haploidentical, or mismatched bone marrow transplantion (1). Although GVHD can be prevented by rigorous T cell depletion ex vivo of the donor graft or prolonged, global immunosuppression in vivo in the recipient, these strategies have associated risks of increased rates of leukemic relapse, infectious complications, or graft failure (1). A strategy that would selectively target only that small fraction of donor T cells capable of mediating GVHD but preserve those T cells that might mediate antitumor, anti-infectious effects and promote engraftment would be desirable. An in vitro blockade strategy that would not involve global in vivo immune suppression or nonspecific multiorgan toxicity also would be advantageous.

Productive T cell activation and proliferation require two signaling events. The first signal is the engagement of the TCR with the MHC-peptide ligand complex on the surface of the APC. Additional costimulatory signals are required for the full activation of the intracellular signaling cascade, IL-2 production, and ultimate T cell proliferation (2, 3, 4, 5, 6, 7). Multiple pathways capable of costimulation have been described, including CD40L:CD40, CD28/CTLA4:B7-1/B7-2, OX40:OX40L, 41BB:41BBL, CD30:CD30L, LFA-1/ICAM-1, CD27:CD70, and CD2:CD48. CD40L is a member of the TNF family and is transiently expressed on activated CD4⁺ T cells (7, 8). The counter-receptor, CD40, is a member of the TNFR family and is found on APCs, including B cells, bone marrow-dendritic cells, follicular dendritic cells, and activated macrophages (7, 8). The CD40L:CD40 pathway was initially described for its importance in B and T cell interactions, B cell activation, Ig production, and isotype switching. Subsequently, CD40L:CD40 interactions were found to be essential for the initiation and activation of Ag-specific T cell effector functions and the activation of costimulatory activity (including the up-regulation of B7 molecules) on APCs, including B cells, macrophages, and dendritic cells (7). CD40L is expressed early in activation subsequent to the engagement of the TCR with the MHC-peptide ligand complex on the surface of the APC. CD40L transduces a signal to its counter-receptor, CD40, constitutively expressed on APCs, resulting in the up-regulation of additional molecules involved in further T cell costimulation (7, 9).

Administration of anti-CD40L mAb in vivo has been effective in preventing collagen-induced arthritis and ameliorating GVHD in murine models (10, 11). The in vivo administration of B cells from CD40-deficient mice or the simultaneous administration of donor splenocytes with anti-CD40L mAb has conferred alloantigen-specific tolerance to recipient mice (12, 13). In primates, administration of anti-CD40L mAb in vivo has been shown to prevent acute renal and intrahepatic islet allograft rejection (14, 15, 16). Although the effect of in vivo administration of anti-CD40L mAb has been impressive in preventing the rejection of solid organ allografts, we found that the ex vivo tolerization of CD4⁺ T cells led to a more profound reduction in GVHD lethality than did the in vivo administration of anti-CD40L mAb in the same setting (11, 17). In vivo tolerance may be more difficult to achieve after BMT due to the 1) induction of proinflammatory cytokines and increased expression of costimulatory molecules on APCs as a result of the intense conditioning protocols used for BMT (18, 19), and 2) difficulty of achieving complete in vivo blockade at all possible sites of allorecognition. As an undesired side effect, the in vivo blockade of costimulation could result in the induction of tolerance to tumor Ags as well as alloantigens, preventing an effective graft vs tumor effect.

We have previously described an ex vivo approach in which the blockade of CD40L:CD40 interactions during a 10-day culture induces donor CD4⁺ T cells to become tolerant to host alloantigens, resulting in a ≥30-fold reduction in GVHD lethality with no additional in vivo immunosuppression (17). Additionally, the tolerized bulk T cell population retained intact responses to Ags not present during tolerization. Tolerance was long-lived and not readily reversible in vivo. In these studies, we sought to determine the prerequisites for tolerance induction in this ex vivo culture system.

B6.C-H2^bm12/KhEg (termed bm12), B10.D2 (H2^d), and inbred C57BL/6 IL-10^−/− mice were purchased from The Jackson Laboratory (Bar Harbor, ME). C57BL/6 (termed B6) were purchased from the National Institutes of Health (Bethesda, MD). (B6 × 129SV)F₁ IL-4^−/− mice, backcrossed six generations onto a B6 background, were obtained from Dr. Manfred Kopf (Basel, Switzerland). bm12 and B6 (both H2^b) mice differ at three amino acids due to mutations in the IA region. Mice were used at 9–10 wk of age. All mice were housed in a specific pathogen-free facility in microisolator cages.

To purify B6 CD4⁺ T cells, axillary, mesenteric, and inguinal lymph nodes were mashed, and single cell suspensions were passed through a wire mesh and collected into PBS containing 2% FBS. Cell preparations were depleted of NK cells (hybridoma PK136, rat IgG2a, provided by Dr. Gloria Koo, Rahway, NJ) and CD8⁺ T cells (hybridoma 2.43, rat IgG2b, provided by Dr. David Sachs, Charlestown, MA) by coating with mAb, followed by passage through a goat anti-mouse and goat anti-rat Ig-coated column (Biotex, Edmonton, Canada). The final composition of purified T cells was determined by flow-cytometric analysis to be ≥94% CD4⁺ T cells. Responder B6 CD4⁺ T cells were mixed with irradiated (30 Gy), anti-Thy-1.2 mAb (hybridoma 30H-12, rat IgG2b, provided by Dr. David Sachs), and anti-NK1.1 mAb plus baby rabbit complement (Nieffenegger, Woodland, CA)-depleted bm12 splenic stimulators. Responder and stimulator cells were suspended at a final concentration of 0.5 × 10⁶/ml in 24-well plates (Costar, Acton, MA) containing DMEM (BioWhittaker, Walkersville, MD) with 10% FBS (HyClone, Logan, UT), 50 mM 2-ME (Sigma, St. Louis, MO), 10 mM HEPES buffer, 1 mM sodium pyruvate (Life Technologies, Grand Island, NY), and amino acid supplements (1.5 mM l-glutamine, l-arginine, and l-asparagine) (Sigma) and antibiotics (penicillin, 100 U/ml; streptomycin, 100 mg/ml) (Sigma). Anti-CD40L mAb (hybridoma MR1, hamster IgG) was obtained by culturing the hybridoma in 10% FBS/DMEM in a hollow fiber bioreactor (Accucyst^Jr, Cellex Biosciences, Minneapolis, MN). Supernatant was purified by ammonium sulfate precipitation. Anti-CD40L mAb was added at a final concentration of 50 μg/ml to primary MLR cultures. Plates were incubated at 37°C and 10% CO₂ for 4, 7, or 10 days. On day 5, the culture was fed 1:1 with new media including mAb. To monitor primary MLR proliferation, 96-well round-bottom microtiter plates (Costar) were set up to contain 10⁵ responders and 10⁵ stimulators cells per well in the presence or absence of exogenous IL-2 (50 IU/ml) (Amgen, Thousand Oaks, CA). In some experiments, murine IL-4 (Schering-Plough, Kenilworth, NJ) (sp. act., 1.5 × 10⁷ U/mg) was added at the indicated doses at the initiation of culture and readded at the day 5 refeed.

To monitor secondary MLR proliferation, 3 × 10⁴ washed, adjusted responders and 10⁵ irradiated (30 Gy) non-T cell-depleted stimulators were plated in the presence or absence of IL-2 (50 IU/ml). Anti-CD40L mAb was not present in the secondary MLR. Microtiter wells were pulsed with tritiated thymidine (1 μCi/well; Amersham Life Science, Buckinghamshire, U.K.) on the indicated days for 16–18 h before harvesting and counted in the absence of scintillation fluid on a β plate reader (Packard Instrument Company, Meriden, CT). Six wells were analyzed for each data point.

Proliferating T lymphocyte precursor frequency analysis was accomplished by setting up eight 3-fold serial dilutions of responder B6 CD4⁺ T cells at 30 replicates in 96-well round-bottom plates and incubating for 7 days with irradiated (30 Gy), non-T cell-depleted bm12 splenic stimulators in the absence of IL-2. Anti-CD40L mAb was added to the primary LDA at a final concentration of 50 μg/ml. Secondary LDA were established with washed, 10-day cultured control-primed or anti-CD40L-tolerized responder cells and bm12 or third-party B10.D2 splenic stimulators. Anti-CD40L mAb was not present in the secondary LDA. Wells were pulsed with tritiated thymidine for 18 h before harvesting and counted in the absence of scintillation fluid on a β plate reader. Wells were scored positive if their cpm exceeded the average cpm plus 3 SDs of the stimulators plated without responders. Using Poisson distribution statistics according to the method of Taswell and with the aid of a computer program, the likelihood of a single hit was confirmed and a frequency estimate calculated.

Washed 10-day cultured control-primed or anti-CD40L-tolerized cells were incubated 4 h with 10⁴ tritiated thymidine-labeled syngeneic or class II-disparate Con A (Sigma)-stimulated splenocyte blasts at various E:T ratios according to the JAM CTL assay (20). Wells, plated in triplicate, were harvested and counted in the absence of scintillation fluid on a β plate reader. Percent lysis was calculated according to the equation: {[(cpm of targets in absence of killers) − (cpm of targets in presence of killers)]/(cpm of targets in absence of killers)} × 100.

bm12 recipients were sublethally irradiated by exposing mice to 6 Gy total body irradiation from a ¹³⁷Cesium source at a dose rate of 85 cGy/min. Day 4, 7, or 10 MLR-cultured cells were injected i.v. at the doses indicated. Peripheral blood was obtained by retroorbital venipuncture for measurement of day 14 and 28 hematocrit (HCT) values as an indicator of the bone marrow-destructive effects of infused T cells.

Freshly purified and day 4, 7, and 10 MLR-cultured cells were assessed for evidence of activation by forward scatter (FSC) and side scatter (SSC) profiles and the coexpression of CD4 and activation Ags, including CD25, L-selectin (CD62L), and CD40L. All studies were performed with two-color flow cytometry using fluorescein- and PE-conjugated mAb (PharMingen, San Diego, CA). All results were obtained using a FACSCalibur (Becton Dickinson, San Jose, CA). FSC and SSC settings were gated to exclude debris. A total of 10,000 cells were analyzed for each determination.

Murine cytokine levels in the supernatant of MLR cultures were quantitated by ELISA (R&D Systems, Minneapolis, MN). Sensitivity of the assays was between 1 and 10 pg/ml for each assay. A standard curve using recombinant protein was generated with each assay.

Survival data were analyzed by life-table methods, and actuarial survival rates are shown. Group comparisons were made by log-rank test statistics. For other data, group comparisons were made by Student’s t test. Values of p ≤ 0.05 were considered significant.

Studies by other investigators have indicated that the induction of anergy in mouse T cell clones, human T cell clones, and primary T cells requires from 16 h up to 10 days of primary culture (5, 21, 22, 23, 24, 25, 26). We have previously reported a 10-day ex vivo tolerization approach of primary murine CD4⁺ T cells, which results in profound secondary in vitro hyporesponsiveness and a ≥30-fold protection from GVHD mortality (17). In the studies presented in this work, we sought to determine the minimum culture duration required for the induction of alloantigen hyporesponsiveness in vitro and inhibition of GVHD in vivo. Purified B6 CD4⁺ lymph node T cells were mixed with irradiated, T cell-depleted, MHC class II-disparate bm12 splenic stimulators in the presence or absence of anti-CD40L mAb. On days 4, 7, and 10, aliquots of cells were washed to remove Ab and cytokines, and cells were either infused in vivo into irradiated bm12 recipients or reexposed in vitro to irradiated bm12 splenic stimulators in the absence of anti-CD40L mAb. Fig. 1,A illustrates the proliferative responses in the primary culture. The addition of anti-CD40L mAb to the cultures progressively reduced proliferation by 21% (day 2), 91% (day 4), and 99% (day 6) as compared with the control cultures. The addition of exogenous IL-2 (50 U/ml) to the cultures prevented inhibition of proliferation by anti-CD40L mAb, with both groups having equivalent proliferative responses (Fig. 1,B). Fig. 2,A–C illustrates the secondary proliferative responses of the cultures established after 4, 7, or 10 days in primary culture in the presence of anti-CD40L mAb. All secondary cultures were established in the absence of anti-CD40L mAb and were derived from the same primary culture. Because our unpublished data have indicated that a 1- to 3-day primary culture has an impaired capacity to mediate lethal GVHD in vivo (P. A. Taylor and B. R. Blazar, unpublished observation), we chose 4 days as the shortest culture period to study. Upon reexposure to alloantigen restimulation in the secondary MLR, the day 4 control culture peaked on day 3 instead of day 6 as in the primary culture indicative of priming (Fig. 2,A). Responses of cells exposed to anti-CD40L mAb in the primary culture for 4 days were inhibited by 45–74% during the first 3 days of the secondary culture. In secondary MLRs established after 7 days of primary culture, the control group had an early vigorous response upon reexposure to alloantigen, peaking on day 1 and gradually declining (Fig. 2,B). This rapid secondary response by day 7 control-primed cells has been reproducibly seen in these cultures. Responses of cells exposed to anti-CD40L mAb in the primary culture for 7 days were inhibited by 71–81% during the first 3 days of the secondary culture. Seven-day anti-CD40L cultures were more hyporesponsive upon alloantigen restimulation than 4-day anti-CD40L cultures. The peak magnitude of the 7-day anti-CD40L-cultured cells was 5400 cpm compared with 10,000 cpm for the 4-day anti-CD40L-cultured cells. As compared with cultures established from 4 or 7 days of primary culture, secondary MLR cultures established from 10-day anti-CD40L-cultured cells were more profoundly hyporesponsive to alloantigen restimulation. At the time of peak secondary response, the proliferative response of the 10-day anti-CD40L-cultured cells was inhibited by 95% as compared with the control-primed cells (Fig. 2 C). The peak magnitude of the 10-day anti-CD40L-cultured cells was 2700 cpm. These data indicate that hyporesponsiveness upon alloantigen restimulation becomes more profound with prolonged ex vivo blockade of the CD40:CD40L pathway. As in the primary MLR, the addition of IL-2 fully restored secondary responses to alloantigen restimulation (data not shown).

As other measures of in vitro function, we determined the frequency of B6 CD4⁺ cells proliferating in response to bm12 alloantigen and the CTL killing of bm12 targets after 10 days of primary culture with anti-CD40L mAb. Analysis of a 7-day LDA established on day 0 of primary culture in the absence of IL-2 showed that the frequency of B6 CD4⁺ cells responding to bm12 alloantigen was 1:5,887 in the absence of mAb and 1:22,352 in the presence of mAb, a 4-fold reduction. When LDAs were established with washed, 10-day cultured cells, the frequency increased about 15-fold in the control group to 1:404. In contrast, the frequency in the anti-CD40L-tolerized cells remained the same as in the primary culture (1:24,602). Although the frequency of anti-CD40L mAb-tolerized B6 CD4⁺ responding to bm12 alloantigen restimulation was decreased about 60-fold compared with the control-primed cells, the frequency of third party B10.D2 allo-responses was only decreased by less than 2-fold, indicating a high degree of specificity (data not shown).

Because other investigators have reported a state of split anergy in which anergic murine T cell clones can be rendered hyporesponsive to Ag reexposure in terms of proliferation and cytokine production but have intact CTL (27, 28), we placed washed 10-day control-primed cells and anti-CD40L-tolerized cells in a 4-h CTL assay against bm12 Con A splenocyte blasts at various E:T ratios. The percent specific lysis of bm12 targets by the 10-day anti-CD40L-tolerized cells was approximately one-fourth that of the control-primed cells (Fig. 3). Consistent with these data, there was an 8–10-fold reduction in the percentage of anti-CD40L-tolerized cells staining positive for granzymes A and B mRNA (data not shown). Granzymes, serine proteases stored in secretory CTL granules, induce DNA fragmentation and CTL lysis of target cells (29, 30). Therefore, a prolonged culture with anti-CD40L mAb prevented the generation of optimal CTL effector function.

To determine whether anti-CD40L mAb treatment prevented T cell activation, control- and anti-CD40L-cultured cells were evaluated by flow cytometry for their degree of blastogenesis and activation (Fig. 4). Although there was up-regulation of CD40L in the control cultures by day 4, there was very little increase in cell size (FSC) or internal complexity (SSC) at this time. Additionally, CD25 expression was low in both groups. L-selectin (CD62L), a homing receptor lost during activation (31), was marginally down-regulated in only 10% of cells in the control cultures on day 4. By day 7, both primed and tolerized cells had substantially increased FSC, SSC (although values were higher in the primed cells as compared with the tolerized cells), and CD25 expression (more than 50% positive in both groups), indicating a greater degree of activation than day 4 MLR cells (Fig. 4). Primed cells had substantially down-regulated or lost expression of L-selectin by day 7. CD25 expression remained high on day 10 in both groups, with 62% positive in the anti-CD40L-cultured cells compared with 32% positive in the control-primed cells. Notably, L-selectin remained high in 85% of the tolerized cells, while more than one-half of the primed cells lost expression (Fig. 4). We conclude from these data that although both control- and anti-CD40L-cultured cells exhibit marked evidence of blastogenesis and activation by day 7, the tolerized cells do not become fully Ag experienced based upon high L-selectin expression.

To determine the length of time in culture required to induce tolerance induction in vivo, aliquots on days 4, 7, and 10 of the same primary cultures used to establish the above secondary MLRs were washed and infused into sublethally irradiated recipients bearing the same alloantigen (bm12) used as stimulator cells in the primary MLR culture. Data from our laboratory suggest cells cultured for shorter periods (1–4 days) have a reduced capacity to mediate GVHD lethality (P. A. Taylor and B. R. Blazar, unpublished observation). Therefore, we infused 10⁶, a higher number of cells, or 10⁵ cells, which is typically lethal to all recipients. In contrast to our previous data demonstrating no GVHD lethality with the infusion of 10⁶ 10-day anti-CD40L mAb-tolerized cells (17), a 4-day culture period was insufficient to confer uniform GVHD protection. All recipients of 10⁶ control-primed or anti-CD40L-cultured cells died of GVHD-induced BM aplasia by day 25 after infusion of cells (Fig. 5,A). Sixty percent of mice receiving 10⁵ anti-CD40L-cultured cells vs 20% of mice receiving 10⁵ control-primed cells survived the 65-day observation period (p = 0.078 vs control). As an indicator of donor CD4⁺ T cell-mediated GVHD-induced bone marrow aplasia, HCT values were assessed in all mice on day 14 posttransfer of cells (Table I). There were no statistically signifi- cant differences in HCT values between groups at either cell dose from 4-day cultured cells. Collectively, these data indicate that 4 days of primary culture with anti-CD40L mAb are insufficient to confer a high level of tolerance in vitro. Additionally, the relatively moderate degree of hyporesponsiveness seen in secondary MLR established from 4-day anti-CD40L-cultured cells did not translate to protection from GVHD lethality in vivo.

In contrast, a 7-day MLR culture in the presence of anti-CD40L mAb was sufficient to confer GVHD protection at the cell doses used in these studies. All recipients of both cell doses of control-cultured cells (10⁵ or 10⁶) died of GVHD-induced BM aplasia by day 23 (Fig. 5,B). With the exception of one unusually late death at the higher cell dose, all recipients of anti-CD40L-tolerized cells survived the observation period (Fig. 5,B) and had normal HCT values at the time of elective sacrifice on day 65 (data not shown). As expected, all recipients of 10-day anti-CD40L-tolerized cells survived long term (Fig. 5,C) and had normal HCT values at the time of elective sacrifice (data not shown), while all recipients of control-cultured cells died by day 16 (Fig. 5 C). Although there were no significant differences in GVHD mortality at the cell doses used in this study between a 7- and a 10-day primary culture with anti-CD40L mAb, the time to death and day 14 HCT values suggest that 10-day control-primed cells may have a greater GVHD capacity than those cultured for 7 days. Mice receiving 10⁵ 10-day control cells succumbed to GVHD mortality 1 wk earlier than those receiving 10⁵ 7-day control cells (p = 0.050). Additionally, the day 14 HCT values in mice receiving 10⁵ control cells suggest a greater degree of aplasia with increasing culture duration (4-day MLR HCT, 30.9; 7-day MLR HCT, 23.5; 10-day MLR HCT, 15.3). These data suggest that although a 7-day ex vivo tolerization period confers protection from GVHD lethality, a 10-day ex vivo tolerization period may confer even greater protection.

Fowler et al. (32) and Krenger et al. (40) have found that in vitro generated, alloreactive Th2 (IL-4, IL-5, and IL-10) populations have reduced GVHD capacity. To determine whether the GVHD protection of the anti-CD40L-cultured cells was mediated by a skewing toward CD4⁺ Th2 cells, supernatants were taken from the primary (days 2, 4, 7, and 10) and secondary cultures (days 1, 3, and 5) and evaluated for the presence of Th1 (IL-2, IFN-γ) and Th2 cytokines. Both control-cultured and anti-CD40L-cultured cells produced IL-2 by day 2 of the primary MLR (Table II). IL-2 continued to accumulate in the control culture, peaking at 312 pg/ml on day 7. Cultures containing anti-CD40L mAb produced 12-fold less IL-2 (27 pg/ml vs 312 pg/ml) during the primary culture than the control. Peak amounts of IFN-γ were detected at the end of the primary culture in the control cells, with very low accumulation in the anti-CD40L cultures. The Th2 cytokines, IL-4 and IL-10, were not detected in the supernatants of either culture in the primary MLR.

Secondary cultures established from both groups of 4-day primary cultures produced large amounts of IL-2 and IFN-γ as early as day 1 of culture (Table III). On a per cell basis, the cells exposed to anti-CD40L mAb in the primary culture produced 10-fold more IL-2 (915 pg/0.15 × 10⁶ cells vs 312 pg/0.5 × 10⁶ cells) and 6-fold more IFN-γ (523 pg/0.15 × 10⁶ cells vs 281 pg/0.5 × 10⁶ cells) in the secondary MLR than the control group produced in the primary MLR, suggesting significant Ag priming. In contrast to the primary MLR, the Th2 cytokines, IL-4 and IL-10, were produced by the 4-day control-primed cells upon in vitro alloantigen restimulation, indicating that the control bulk culture consisted of a mixed population of Th1- and Th2-primed cells (Table III). Although in secondary MLRs, Th2 cytokines by the anti-CD40L-cultured cells were reduced in comparison with the control-primed cells, there was still substantial production. These data indicated that although a 4- day culture in the presence of anti-CD40L mAb reduced cytokine production upon reexposure to alloantigen in comparison with the control cultures, there was still a vigorous Th1 and Th2 cytokine response to alloantigen restimulation. Given the high levels of IL-2 production by the 4-day anti-CD40L-cultured cells, it is perhaps not surprising that they did mediate GVHD lethality in vivo (Fig. 5 A).

IL-2 and IFN-γ production in secondary cultures established from 7-day anti-CD40L cultures was reduced 3-fold (286 pg/ml vs 915 pg/ml) and 11-fold (48 pg/ml vs 523 pg/ml), respectively, as compared with that established from 4-day anti-CD40L cultures. It is interesting to note that although the IL-2 production was reduced 75% as compared with the 7-day control group, there was more IL-2 produced on a per cell basis by the 7-day anti-CD40L cultures in secondary MLR than in the primary MLR control group. Surprisingly, in spite of this substantial IL-2 production in vitro upon Ag restimulation by the anti-CD40L-cultured cells, these cells did not mediate GVHD in vivo (Fig. 5,B). IFN-γ production by the 7-day anti-CD40L cultures was reduced by 99% compared with the control-primed cultures. As with the Th1 cytokines, there were lower levels of Th2 cytokines produced in secondary MLRs by 7-day anti-CD40L-cultured cells than by 4-day cultures. Th1 and Th2 cytokine production in secondary cultures established from 10-day anti-CD40L cultures were the most profoundly suppressed (Table III). Most notably, IL-2 production in secondary MLR established from 10-day anti-CD40L cultures was 9-fold (33 pg/ml vs 286 pg/ml) lower than that from 7-day anti-CD40L cultures. Overall, we found a pronounced and progressive dimunition of Th1 and Th2 cytokine production in the secondary MLRs of the anti-CD40L cultures with increasing primary culture duration.

Because ELISA measures the total amount of immunologically reactive protein present in the supernatants, but doesn’t take into account cytokine consumption, Th2 skewing may have been present, but not at a level detectable by ELISA. As additional evidence that GVHD protection was not due to a skewing toward a Th2 phenotype, we used CD4⁺ cells from IL-4 or IL-10 knockout mice as responders in our 10-day ex vivo tolerization MLR cultures. Ex vivo blockade of the CD40L:CD40 pathway via anti-CD40L mAb in IL-4 knockout CD4⁺ cells resulted in decreased proliferation in the primary MLR (Fig. 6,A) and profound hyporesponsiveness upon alloantigen restimulation in vitro (data not shown). Mice receiving 10⁵ control-primed IL-4 knockout CD4⁺ cells died of BM aplasia by 3 wk. Mice receiving 10⁵ anti-CD40L-cultured IL-4 knockout CD4⁺ cells survived the 65-day observation period (Fig. 6,B) and had normal HCT at time of elective sacrifice (data not shown). Similar results were found with IL-10 knockout CD4⁺ cells. Proliferation in primary and secondary MLRs was profoundly reduced in anti-CD40L-cultured IL-10 knockout CD4⁺ cells (data not shown). Mice receiving 10⁵ anti-CD40-cultured IL-10 knockout CD4⁺ cells survived the observation period (Fig. 7). In contrast, mice receiving 10⁵ control-primed IL-10 knockout CD4⁺ cells died of BM aplasia by 2 wk posttransfer. These data indicate that production of the Th2 cytokines, IL-4 or IL-10, was not required for anergy induction or GVHD protection.

IL-4 has been reported to preclude anergy induction by signaling through the common γ-chain of the IL-2R, IL-4R, IL-7R, and IL-15R (21, 33, 34). IL-4 has also been found to prevent cell death without cell division and proliferation by preventing the decay of bcl-2 in activated T cells (35). Because cell recovery in the 10-day anti-CD40L cultures is relatively low (≤10% of starting cell number), we examined the effect of a day 0 addition of a low concentration IL-4 as a survival factor to determine whether we could improve recovery without precluding the induction of anergy. The addition of 0.1 ng/ml of IL-4 did not reproducibly improve recovery. However, the addition of 1 ng/ml or 10 ng/ml of IL-4 added to the cultures on day 0 did reproducibly and favorably increase recovery from <10% to 32–39% of the starting cell number (data not shown). The addition of 1 ng/ml of IL-4 added to cultures on day 0 did not preclude the inhibition of proliferation by anti-CD40L mAb in the primary MLR (Fig. 8,A). Peak proliferation in the primary culture was inhibited by anti-CD40L mAb by 96% even in the presence of 1 ng/ml IL-4. Cell recovery on day 10 was 8% in the anti-CD40L cultures in the absence of IL-4 and 48% in the presence of IL-4. Despite a high degree of blockade of proliferation in the primary MLR cultures containing both anti-CD40L mAb and IL-4, these cells were only modestly hyporesponsive upon in vitro alloantigen restimulation in the secondary MLR (Fig. 8,B). This modest reduction (<40%) in secondary response as compared with the control-primed cells did not translate to GVHD protection in vivo. All recipients of either 10⁵ control-primed/IL-4 or anti-CD40L/IL-4-cultured cells died of GVHD-induced BM aplasia by day 18 after cell transfer (Fig. 8 C). In contrast, all recipients of 10⁵ anti-CD40L-cultured cells in the absence of supplemental IL-4 survived the 2-mo observation period. These data indicate that IL-4 (1 ng/ml) added as a survival factor on day 0 precluded the induction of anergy.

In this study, we have demonstrated that anergy induction via in vitro blockade of the CD40:CD40L costimulatory pathway became more profound with increasing duration in primary culture as measured by in vitro secondary MLR proliferative responses, cytokine production, and in vivo GVHD protection. Th1 and Th2 cytokine production were both reduced in the primary and secondary MLRs of the anti-CD40L-cultured cells, indicating that GVHD was not mediated by a skewing toward a Th2 phenotype. Additionally, IL-4 and IL-10 knockout CD4⁺ cells were susceptible to the induction of alloantigen hyporesponsiveness in our model, further indicating that Th2 cytokines were not obligatory. IL-4 added to the cultures at doses that improved cell recovery precluded the induction of anergy, indicating that IL-4 does not mediate but rather can prevent the induction of tolerance.

We have previously demonstrated a profound reduction in GVHD lethality with a 10-day ex vivo tolerization culture duration (17). Because studies by others have indicated that the time to optimal tolerance induction can vary between 16 h and 10 days, we wished to determine whether this relatively prolonged duration was essential for GVHD protection (5, 21, 22, 23, 24, 25, 26). Our data indicated a 4-day MLR culture in the presence of anti-CD40L mAb was insufficient to induce tolerance. Although CD40L expression was up-regulated by day 4 of control cultures, cells had not substantially increased in size and internal complexity, CD25 expression was not increased, and L-selectin had not yet significantly down-regulated. Together these data may suggest that T cells require a certain level of activation for tolerance induction to occur.

In contrast to the partial inhibitory effect of 4-day anti-CD40L MLR cultures, 7 days of primary culture with mAb were sufficient to induce hyporesponsiveness in vitro and tolerance in vivo, while 10 days of culture resulted in a more profoundly tolerized population of cells. Although the cell doses used for in vivo GVHD generation did not allow us to conclude that 10 days of primary culture led to greater in vivo GVHD prevention than a 7-day primary culture, the secondary MLR and cytokine data indicated that the secondary in vitro hyporesponsiveness was more profound with a 10-day primary tolerization culture duration. In all experiments, there was a small burst of proliferation on day 1 of secondary MLR by these cultures. This slight degree of proliferation in the hyporesponsive cells in the secondary MLR was accompanied by substantial IL-2 production especially from secondary cultures established from 7-day tolerized MLR cultures. Nine-fold more IL-2 was produced in response to alloantigen restimulation by 7-day tolerized cultures as compared with 10-day tolerized cultures. Based upon the high levels of IL-2 detected in the supernatants of secondary cultures of tolerized cells established from 7-day MLR cultures, one might have predicted greater GVHD mortality. However, this early proliferation and IL-2 production by the tolerized population were not sustained and rapidly declined. Although we do not yet know whether the Ag-specific cells in the tolerized cultures were deleted, we hypothesize that deletion is probably not the sole mechanism involved, as the bulk population had a partial early proliferation and produced IL-2 in response to alloantigen restimulation.

The degree of hyporesponsiveness seen in the secondary MLR upon reexposure to alloantigen correlated well with GVHD protection in vivo, suggesting that the in vitro assays could potentially be used as a monitoring device, which might be predictive of the clinical outcome. The primary MLR may be less predictive of successful tolerance induction than the secondary MLR (see Fig. 8 A). Although the degree of blockade of proliferation was very high by anti-CD40L mAb plus IL-4 in the primary MLR, there was only modest secondary hyporesponsiveness and no GVHD protection.

The requirement for a relatively long primary culture duration to induce alloantigen hyporesponsiveness is in contrast to studies by other investigators. Taub et al. found that a 3-day incubation of the phosphatidylinositol-3-kinase inhibitor, wortmanin, was sufficient to induce alloantigen-specific tolerance in a murine GVHD model (22). However, as the degree of GVHD protection was not quantified, it is unknown whether a longer incubation would have resulted in a more profound resistance to GVHD induction. Although it is possible that tolerance induction via small intracellular signaling inhibitors may occur faster than that via Ab blockade of cell surface costimulatory molecules, not all surface-mediated tolerance-inducing events require prolonged ex vivo culture. For example, Jenkins and Schwartz found that only a 16-h incubation of pigeon cytochrome c-specific T cell clones with Ag and chemically modified splenocytes was sufficient to induce unresponsiveness to restimulation with normal APC and Ag (5). Although these studies used T cell clones and a peptide Ag and not bulk CD4⁺ lymph node and alloantigen as ours do, it is tempting to speculate that the relatively short culture duration needed to tolerize these cells may be due to the fact that APCs in their culture system were completely incapable of up-regulating the cell surface expression of all potential costimulatory molecules. In our system, costimulatory molecules other than those entirely dependent upon induction via CD40L:CD40 interaction potentially could be induced to be expressed and therefore may have transduced some costimulatory signal. Alternatively, their model may involve a higher frequency of responding T cells and a more rapid response requiring a shorter culture duration to induce anergy. Optimal anergy induction via blockade of costimulation may require a culture duration of sufficient length to ensure that all potentially alloreactive cells have had maximal opportunity for TCR engagement, initial activation and up-regulation, and subsequent blockade of CD40L. Although CD40L was up-regulated by day 2, peak proliferative responses occurred on day 5 or 6 of primary culture. Additionally, maximal up-regulation of the activation Ag, CD25, was not seen until day 7 of primary culture, with very low expression by day 4. Blockade of costimulation as a strategy to induce anergy induction may be required throughout the duration of these events. The minimum culture duration required to obtain optimal anergy may depend as much on the kinetics of the response as on the strategy used to induce anergy.

In a model more analogous to ours, Gribben et al. found that maximal hyporesponsiveness of human alloreactive T cells to alloantigen restimulation was achieved with 36 h via blockade of the CD28:B7-1/B7-2 pathway (21). Other studies have indicated that a longer culture duration is necessary or at least beneficial. Two studies found that a minimum of 5- to 7-day culture of human T cells with stimulator cells in the presence of anti-B7-1 and cyclosporine, followed by 1- to 2-day rest was needed to yield a hyporesponsive state (23, 26). In a system in which human alloreactive T cells were rendered hyporesponsive by IL-10, T cells were noted to become progressively more hyporesponsive during a 3- to 10-day culture period (25). These data and ours suggest that hyporesponsiveness becomes increasingly profound with increasing culture duration in bulk cultures consisting of naive, potentially alloreactive T cells.

We have previously demonstrated that OVA-specific T cells exposed to anti-CD40L mAb during tolerization to alloantigen have intact OVA responses at optimal Ag concentrations (17). Consistent with these data, although the frequency of responding T cells to relevant alloantigen was decreased by 60-fold in anti-CD40L-cultured cells as compared with the control-cultured cells, the frequency of the tolerized cells to third party alloantigen was decreased only 2-fold. This suggests that a bulk T cell population that is tolerized ex vivo to alloantigen may still be capable of responding to viral or tumor Ags encountered in vivo.

In contrast to other groups reporting split anergy in which proliferation and cytokine production were diminished, but cytotoxicity was not impaired (27, 28), we found CTL killing by the anti-CD40L-cultured cells was approximately one-fourth that of the control-primed cells. Perhaps more importantly, anti-CD40L-cultured cells failed to induce mortality in mice, while mice receiving control-primed cells died of BM aplasia. Consistent with this, IFN-γ required for CTL generation was profoundly reduced in both the primary and secondary MLR of the anti-CD40L-cultured cells. We and others have found that an intact CD40L:CD40 pathway is required for optimal in vivo CTL generation, and that the administration of anti-CD40L mAb abrogates a GVL effect (36, 37, 38, 39) (P. A. Taylor and B. R. Blazar, unpublished observation). Collectively, our data indicate that the anti-CD40L-cultured cells are hypofunctional in terms of proliferation, cytokine production, and cytotoxicity.

Although Th2 cells have been reported in some (32, 40) but not all (41) studies to result in reduced GVHD as compared with Th1 cells, our hyporesponsive cells were not skewed toward a Th2 phenotype. The anti-CD40L-cultured cells made far less of the Th2 cytokines, IL-4 and IL-10, upon alloantigen restimulation than did the control-primed cells. As well as being a Th2 cytokine, IL-10 is an immunosuppressive cytokine that has been shown to induce tolerance in human alloreactive T cells (25). In another study, endogenously produced IL-10 and TGF-β were found to mediate superantigen-induced tolerance in a murine in vivo model (42). However, both IL-4 and IL-10 knockout CD4⁺ cells were susceptible to anergy induction in our model, indicating there is not an absolute requirement for either of these cytokines for hyporesponsiveness in vitro and GVHD protection in vivo. Additionally, we did not find free TGF-β, another immunosuppressive cytokine, in the supernatants of either culture group in either the primary or the secondary MLR (data not shown). We did find the chemokines, RANTES and macrophage-inflammatory protein 1-α, present in the supernatants of the anti-CD40L cultures in both the primary and the secondary MLR, indicating that the tolerized cells are not globally, metabolically disabled (data not shown).

Because cell recovery is typically low (≤10%) in the anergic cultures, we examined the effect of the addition of IL-4 as a survival factor to the primary cultures in an attempt to increase our cell recovery. We did not uncover a concentration of IL-4 that both increased recovery and permitted the induction of secondary hyporesponsiveness. As IL-7 also has been reported to be involved with T cell survival in vivo and in vitro (35), we examined the addition of IL-7 on cell recovery and induction of secondary hyporesponsiveness and discovered that a dose or schedule that improved cell recovery also precluded anergy induction (P. A. Taylor and B. R. Blazar, unpublished data). We surmise that the addition of IL-4 or IL-7 at concentrations high enough to increase recovery results in sufficient signaling through the IL-2R common γ-chain to preclude the induction of anergy.

These studies indicate that ex vivo anti-CD40L tolerization becomes more profound with increasing primary culture duration, highlighting the importance of kinetics studies in tolerance-induction protocols. Our data indicate that a combination of flow cytometry and in vitro secondary proliferative, cytotoxic, and cytokine responses may be useful indicators of the efficacy of a tolerizing procedure in preventing GVHD generation in vivo. Anti-CD40L mAb ex vivo tolerization warrants consideration as a potential therapeutic modality for the prevention of GVHD.

We thank Dr. Arlene Sharpe for critical review of the manuscript.