Opposing Roles of CD28:B7 and CTLA-4:B7 Pathways in Regulating In Vivo Alloresponses in Murine Recipients of MHC Disparate T Cells

Mice

B10.BR/SgSnJ (H2^k), C57BL/6 (termed B6:H2^b,CD45.2), CD28-deficient B6 (CD28^−/−), C.H2^bm1 (termed bm1; CD45.2), and C.H2^bm12 (termed bm12; CD45.2) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). BALB/c-SCID (H2^d) and B6-CD45.1 congenic mice were purchased from the National Institutes of Health (Bethesda, MD). Mice were housed in a specific pathogen-free facility in microisolator cages. Donors and recipients were used at 8–10 wk of age.

GVHD generation

To determine the effect of CD28/CTLA-4:B7 interactions on the GVHD capacity of CD4⁺ or CD8⁺ T cells, we used a system in which purified T cell subsets are given to MHC disparate sublethally irradiated recipients, as initially described by Sprent (22). This system permits highly accurate quantification of the degree of GVHD responses as related to T cell dose. For this purpose, MHC class II (bm12) or class I (bm1) disparate recipients were irradiated with 6 Gray (Gy) TBI on day 0 from a ¹³⁷Cesium source at a dose rate of 85 cGy/min (20, 22). Four to six hours after TBI, sublethally irradiated bm12 recipients were given purified lymph node (LN) B6CD4⁺ T cells at doses of 0.3–3 × 10⁵ per recipient, while bm1 recipients were injected with CD8⁺ T cells at doses of 0.3, 1, or 3 × 10⁶ per recipient. To purify LN cells, single cell suspensions of axillary, mesenteric, and inguinal LN cells were obtained (as a source of GVHD-causing effector cells), depleted of NK cells, and enriched for CD4⁺ or CD8⁺ T cells by depletions with anti-CD8 (hybridoma 2.43, rat IgG2b, provided by Dr. David Sachs, Charlestown, MA) or anti-CD4 (hybridoma GK1.5, rat IgG2b, provided by Dr. Frank Fitch, Chicago, IL), respectively. Rat mAb-coated T cells were passaged through a goat anti-mouse and goat anti-rat Ig-coated column (Biotex, Edmonton, Canada). The final composition of T cells in the donor graft was determined by flow cytometry and was always found to be ≥94% T cells of the desired phenotype. Hematocrit values were obtained at periodic intervals as an indicator of the possible BM destructive effects of infused T cells (20, 22). For all GVHD and engraftment systems, mice were monitored daily for survival and clinical appearance and weighed twice weekly.

To avoid a host anti-donor response present in the sublethal irradiation systems, we employed two different types of strategies. In the first, mice were lethally irradiated with 8 Gy TBI by x-ray (39 cGy/min) on day −1, followed on day 0 by the i.v. infusion of T cell-depleted (TCD) BM by anti-Thy-1.2 (clone 30-H-12, provided by Dr. David Sachs, Cambridge, MA) + complement treatment. To measure CD4⁺ and CD8⁺ T cell GVHD responses, irradiated B10.BR recipients were infused i.v. with 20 × 10⁶ TCD BM along with 0 or 15 × 10⁶ splenocytes from B6 or CD28^−/− donors. To measure CD4⁺ T cell GVHD responses, irradiated bm12 recipients were infused with 8 × 10⁶ TCD BM along with 0 or 0.3 × 10⁶ purified T cells from B6 or CD28^−/− donors. In a second system designed to analyze the effect of CD28^−/− T cells in the complete absence of tissue damage induced by high dose TBI, BALB/c-SCID recipients were treated with anti-asialo GM1 antisera (25 μl on days −4 and −2) to deplete endogenous NK activity and then infused with 2 × 10⁶ LN T cells from B6 or CD28^−/− donors without further conditioning of the recipient.

Thoracic duct cannulation

For thoracic duct lymphocyte (TDL) isolation, cannulae were inserted in the thoracic duct of recipients at the time of peak proliferation (day 6) post-BMT. TDL were collected over a period of 18 h (20).

GVL studies

Our GVL induction procedure has been described in detail (21). In brief, mice undergoing BMT for GVL assessment were conditioned with 800 cGy irradiation from an x-ray source on day −1 and infused with 8 × 10⁶ B10.BR TCD BM on day 0. Donor B10.BR splenocytes were administered i.v. at a dose of 25 × 10⁶ on day 21 post-BMT. On day 28 post-BMT, mice were challenged with an MHC class I⁺II⁻ acute myeloid leukemia line, C1498 (21). C1498 (American Type Culture Collection, Manassas, VA), originally derived as a spontaneous tumor line from a B6 mouse, was grown in RPMI (Life Technologies, Grand Island, NY) with 10% heat-inactivated FBS (HyClone, Ogden, UT), 2 mmol/L l-glutamine, 50 mmol/L 2-ME, and penicillin/streptomycin. C1498 cells (2 × 10⁵) were given via the i.v. route. All available animals were examined for evidence of GVHD and tumor by necropsy (21).

Engraftment studies

B6-CD45.1 mice were irradiated with 6 Gy TBI by x-ray on day −1 and then given bm1 or bm12 TCD BM (10 × 10⁶) cells (23). Peripheral blood cells were phenotyped for chimerism twice post-BMT. In addition, the presence of donor or host origin cells in individual lineages was analyzed in detail at least once post-BMT.

Anti-CTLA-4 mAb preparation

Clone UC10-4F10-11 (hamster anti-murine CTLA-4) was generously provided by Dr. Jeffrey Bluestone (University of Chicago, Chicago, IL) (9). mAb for injection was generated as ascites fluid and purified by ammonium sulfate precipitation, followed by dialysis. Control hamster IgG was purchased from Rockland Laboratories (Gilbertsville, PA). Cohorts of mice were given anti-CTLA-4 mAb or irrelevant hamster IgG mAb. In initial GVHD studies, anti-CTLA-4 mAb or control mAb was given at a dose of 250 μg/dose on days −1 and 0, and then 100 μg/dose three times per week until day 10. In a subsequent study, a more intensive regimen was used consisting of anti-CTLA-4 mAb or control mAb at 400 μg/dose from days −1 to +5, then 200 μg/dose three times a week until day 14 (for engraftment studies) or day 21 (for GVHD studies) or day 24 after delayed lymphocyte infusion (for GVL studies).

Flow cytometry

The T cell, B cell, and granulocyte/macrophage constituency of splenocytes, peripheral blood cells, or TDL effluent were measured using mAb directed toward CD4 or CD8, CD19, and Mac1, respectively. All fluorochrome-labeled mAb, unless indicated, were obtained from PharMingen (San Diego, CA). For bm1 recpients of B6-CD45.1 CD8⁺ T cells, splenocyte suspensions were monitored with αCD45.2 (clone 104-2, rat IgG2a) and αCD45.1 (clone A20-1.7, rat IgG2a), both provided by Dr. U. Hammerling (New York, NY). For splenic flow cytometry studies, cells were first incubated with 2.4G2 to block Fc receptors, and then incubated with an optimal concentration of fluorochrome-labeled mAb for 45 min at 4°C. Cells were washed three times and resuspended for analysis by two- or three-color flow cytometry using FITC-, PE-, or biotin (along with streptavidin-PerCP)-conjugated mAb purchased from PharMingen or Becton Dickinson (Mountain View, CA). Irrelevant mAb control values were subtracted from values obtained with relevant mAbs. All results were obtained using a FACScalibur (Becton Dickinson). Forward and side scatter settings were gated to exclude red cells and debris, and 1 × 10⁴ cells were analyzed for each determination.

In vitro MLR cultures and CTL assessment

For quantifying MLR responses, CD4⁺ LN T cells or TDL were mixed with irradiated (30 Gy) TCD splenocyte stimulators from host strain mice. Cells were suspended in MLR media consisting of Dulbecco’s MEM (Bio Whittaker, Walkersville, MD), 10% FCS (HyClone, Logan, UT), 2-ME (5 × 10⁻⁵ M) (Sigma, St. Louis, MO), 10 mM HEPES buffer, 1 mM sodium pyruvate (Life Technologies, Grand Island, NY), amino acid supplements (1.5 mM l-glutamine, l-arginine, and l-asparagine) (Sigma), and antibiotics (penicillin, 100 U/ml; streptomycin, 100 μg/ml) (Sigma). A total of 0.3–1 × 10⁵ cell responders and 1 × 10⁵ irradiated stimulators were plated into 96-well round-bottom (Costar) plates and placed at 37°C and 10% C0₂ for 1 to 6 days. In some wells, as indicated, anti-CTLA-4 mAb was added at a final concentration of 50 μg/ml. Δ cpm were calculated by subtracting the syngeneic proliferative response from the allogeneic proliferative response. No scintillant was used to amplify cpm detection. For CTL analysis, splenocytes were activated with Con A (2 μg/ml) for 3 days in MLR media and then labeled overnight with [³H]thymidine (4 μCi/ml) in 24-well plates containing 2 × 10⁶ cells in a 2 ml vol. B6 or CD28^−/− TDL T cells were spun at 500 rpm (Beckman TJ-6) for 5 min with targets (10⁴ targets/well) at E:T ratios of 12.5:1–100:1 in MLR media and then incubated together for a period of 4 h. Cell pellets were harvested and then counted as described above for the MLR culture. The percent specific lysis was based upon the retention of [³H]thymidine in the experimental groups as compared with targets not exposed to effector cells (24).

In situ antisense:sense mRNA hybridization of TDL

TDL cytospins were hybridized with riboprobes specific for IL-2 237–837(237–837), IL-4 1–393(1–393), IL-5 1–1534(1–1534), IFN-γ 1–270(1–270), IL-10 392–937(392–937), granzyme B 239–775(239–775), and perforin 1–1700(1–1700) mRNA. The mRNA:RNA probe hybrids were detected using an alkaline phosphatase α-digoxigenin F(ab′)₂ Ab and BCIP/NBT (5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium) color substrate. The in situ hybridization technique has been described elsewhere (25).

Statistical analyses

Group comparisons of continuous data were made by Student’s t test. Survival data were analyzed by lifetable methods using the Mantel-Peto-Cox summary of χ² (26). Actuarial survival and relapse rates are shown. p values < 0.05 were considered significant.

Role of CD28:B7 interaction on regulating GVHD lethality early post-BMT

Blockade of CD28/CTLA-4:B7 interactions by CTLA4-Ig or anti-B7 mAb has been shown to diminish GVH alloresponses, resulting in a decrease in GVHD-induced mortality in some studies (17, 18, 19, 20, 27, 28, 29, 30). Because CTLA4-Ig and anti-B7 mAb inhibit both the CD28:B7 and CTLA-4:B7 interactions, the analysis of the contributions of each pathway individually in GVHD generation has required alternative approaches, including the use of CD28^−/− mice. To determine whether precluding CD28:B7 interactions would have a beneficial effect on preventing GVH lethality, highly purified wild-type or CD28^−/− T cells were infused into sublethally irradiated wild-type or MHC disparate recipients. In this setting, alloreactive donor T cells cause multiorgan system GVH tissue destruction (20), and because the hemopoietic compartment is a target of alloresponsive donor T cells, recipients frequently experience lethal BM aplasia. A major advantage of this type of system includes the close correlation between T cell dose and lethality rates.

To quantify the effect of CD28:B7 interactions on GVH lethality of CD4⁺ T cells, CD28^−/− CD4⁺ T cells were infused at doses of 0.3–3 × 10⁵ into sublethally irradiated bm12 (MHC class II only disparate) recipients. At all cell doses tested, CD28^−/− CD4⁺ T cells were found to be incapable of causing GVHD lethality (Fig. 1⇓A). In contrast, recipients of as few as 0.3 × 10⁵ wild-type cells had a 70% lethality rate, while recipients of 1 or 3 × 10⁵ wild-type cells uniformly succumbed to GVHD. Weight loss was observed in recipients of wild-type, but not CD28^−/− cells, reflective of survival results. Day 14 mean hematocrit values were ≤17% in recipients of ≥10⁵ wild-type T cells that were significantly lower than the mean values of 25% in recipients of 0.3 × 10⁵ wild-type T cells or ≥32% obtained from recipients of CD28^−/− T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

The GVHD lethality capacity of CD28^−/− T cells. Sublethally irradiated bm12 (A) or bm1 (B) or heavily irradiated B10.BR (C) recipients were given the indicated numbers of highly purified CD4⁺ T cells, CD8⁺ T cells, or nonpurified splenocytes, respectively, from CD28^−/− or wild-type B6 donors on day 0. On the x-axis are the days posttransfer, and on the y-axis is the proportion of mice surviving. The numbers of recipients per group are listed. Data in the sublethal systems were pooled from two replicate experiments with similar results. A significant (p ≤ 0.01) decrease in GVHD-induced lethality with CD28^−/− as compared with wild-type cells given at the same cell dose was noted in the bm12 and bm1 systems and at the highest splenocyte dose in the B10.BR system.

To quantify the effect of CD28:B7 interactions on GVH lethality of CD8⁺ T cells, CD28^−/− CD8⁺ T cells were infused at doses of 0.3 or 1 × 10⁶ into sublethally irradiated bm1 (MHC class I only disparate) recipients. The infusion of CD28^−/− CD8⁺ T cells at a dose of 10⁶ resulted in the mortality of 31% of recipients, while 0.3 × 10⁶ cells were not lethal (Fig. 1⇑B). As few as 0.3 × 10⁶ wild-type cells resulted in 69% lethality and 10⁶ cells were 100% lethal. Weight curves were reflective of survival. Day 14 hematocrit values were significantly lower (≤17%) in recipients of wild-type cells as compared with recipients of CD28^−/− T cells (≥27%). Based upon these data, there was at least a 10-fold reduction in the GVH lethality capacity of CD28^−/− CD4⁺ T cells and >3- but <10-fold reduction in the GVH lethality of CD28^−/− CD8⁺ T cells.

To determine whether CD4⁺ and CD8⁺ T cell-containing inocula from CD28^−/− donors were capable of causing GVHD, we used a different model system in which the B6 (H2^b) donor and B10.BR (H2^k) recipient were mismatched at both MHC class I and class II loci. For this purpose and to simulate a clinical BMT setting in which there is no host anti-donor response as would be present in sublethally irradiated recipients, we used a model system in which the recipient is lethally irradiated and then given 5 or 15 × 10⁶ B6 splenocytes along with TCD BM (Fig. 1⇑C). Recipients of the higher numbers of CD28^−/− had a significantly higher actuarial survival rate as compared with wild-type controls, although both groups ultimately succumbed to GVHD lethality. The degree of GVHD protection was <3-fold since recipients of 15 × 10⁶ CD28^−/− splenocytes had a significantly lower survival rate than recipients of 5 × 10⁶ wild-type cells. Recipients of as few as 5 × 10⁶ CD28^−/− cells had a high survival rate (80% at 100 days), but also had clinical and weight evidence of sublethal GVHD.

The lower degree of GVHD protection could have been due to the use of more irradiation in the B6→B10.BR model than in the sublethal system. To investigate this possibility, we chose a different system in which donor CD4⁺ and CD8⁺ T cells are infused into an MHC class I and class II disparate recipient (BALB/c-SCID) that does not require any TBI conditioning of the recipient to permit donor T cell-mediated GVHD lethality. NK-depleted BALB/c-SCID mice were given 2 × 10⁶ T cells from B6 or CD28^−/− donors. Ninety percent of recipients of either CD28^−/− or wild-type cells succumbed to GVHD mortality within 1 mo after transfer, with both groups averaging >30% loss in mean body weights (data not shown). Together, these data suggest that the reduced GVHD response observed when low numbers of CD28^−/− CD4⁺ or CD8⁺ T cells are infused into sublethally irradiated recipients can be overwhelmed when larger numbers of donor CD28^−/− CD4⁺ and CD8⁺ T cells are coinfused into MHC disparate recipients.

Blockade of CTLA-4:B7 interactions increases donor T cell expansion and GVHD lethality, as well as host-mediated BM graft rejection early post-BMT

The reduced GVHD lethality observed with CD28^−/− T cells in sublethally irradiated recipients could be due to reduced expansion of CD28^−/− T cells due to the absence of a positive regulatory signal or to unperturbed CTLA-4:B7 interactions due to the presence of a negative regulatory signal. In the latter instance, CTLA-4:B7 interactions might be expected to down-regulate donor T cell responses. To directly investigate the latter possibility, anti-CTLA-4 mAb was given in vivo to block CTLA-4:B7 interactions. To address the possibility that anti-CTLA-4 mAb could directly affect donor anti-host T cell alloresponses, in vivo donor T cell expansion and function and GVHD mortality were quantified.

To quantify donor T cell expansion unencumbered by a host anti-donor T cell response, BALB/c-SCID recipients were depleted of NK cells and then given 0.2 × 10⁶ or 2 × 10⁶ T cells from (B6 × 129)F₂ (H2^b) donors. On days 4 and 6 posttransfer, the number of TDL was quantified in recipients of 2 × 10⁶ cells as an indicator of donor T cell expansion in vivo. We chose to isolate TDL rather than study splenic T cells since more donor T cells are collectable in the lymphatics over an 18-h period than present in the spleen and because splenic but not TDL T cells reside in an environment that inhibits the generation of an immune response. At both time periods, recipients of anti-CTLA-4 mAb had a 2-fold increase in TDL T cell number as compared with those that received irrelevant mAb (day 4, 0.295 × 10⁶/ml vs 0.155 × 10⁶/ml; day 6, 8.14 × 10⁶/ml vs 4.17 × 10⁶/ml, respectively) (data not shown). In a different experiment, recipients were given lower numbers of T cells (0.2 × 10⁶) and cannulated on day 9 posttransfer. Once again, the number of TDL was 2-fold higher in anti-CTLA-4 mAb-treated as compared with irrelevant mAb-treated controls (1.09 × 10⁶/ml vs 0.49 × 10⁶/ml, respectively) (data not shown). Phenotypic analysis of days 4, 6, and 9 TDL cells showed no substantial differences in either the proportion of CD4⁺ and CD8⁺ T cells or the expression of a broad panel of activation Ags (data not shown). Proliferative responses to host stimulator cells were measured using day 6 posttransfer TDL from irrelevant or anti-CTLA-4 mAb-treated recipients. In both experiments, peak responses to host stimulator cells were seen after 1 day of culture using TDL obtained from both cohorts as compared with the typical 5 days required for peak response using TDL obtained from non-BMT donor-strain controls. Results from one of these two replicate experiments (Fig. 2⇓) are consistent with an in vivo priming effect. In situ hybridization analysis for cytokine mRNA in fresh day 6 TDL obtained from a pool of four to five mice in each group indicated that anti-CTLA-4 mAb led to a skewed Th2 response, as indicated by an increase in the percentages of TDL cells expressing IL-4 mRNA (1.2% vs 16.4%, respectively), although the percentage of IL-10 mRNA-expressing cells remained similar. The percentage of TDL cells expressing Th1 cytokines (IL-2, IFN-γ) was low in both the control and anti-CTLA-4 mAb groups (<0.3%). The frequency of TDL cells expressing granzyme B was comparable (12.3% vs 11.4%, respectively).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Anti-CTLA-4 mAb infusion increases the expansion of allogeneic donor T cells without affecting the in vitro proliferative responses to alloantigen-bearing stimulators. BALB/c-SCID recipients were depleted of NK cells and given (B6 × 129)F₂ T cells (2 × 10⁶) with either hamster IgG (hIgG) irrelevant mAb or anti-CTLA-4 mAb, as described in Materials and Methods. TDL were isolated from cohorts of recipients (n = 6–8 mice/group) given 2 × 10⁶ donor T cells. On day 6 posttransfer, a mean of 4.17 or 8.14 × 10⁶ TDL T cells/ml was collected from recipients of irrelevant mAb or anti-CTLA-4 mAb, respectively. The in vitro proliferative response of these TDL to host alloantigen-bearing stimulator cells is shown. On the y-axis is mean cpm using a log10 scale, and on the x-axis are days in culture. The error bars shown are 1 SD. cpm for syngeneic responses were <10% of cpm for allogeneic responses. These findings were replicated in a second experiment.

To determine whether the increased T cell expansion observed in anti-CTLA-4 mAb-treated recipients would increase mortality rates, we used our well-studied lethal TBI system to simulate a clinical BMT setting. B10.BR recipients were irradiated, reconstituted with high numbers of B6 donor BM along with donor splenocytes, and given either irrelevant or anti-CTLA-4 mAb. Anti-CTLA-4 mAb-treated recipients had a significantly (p = 0.001) lower actuarial survival rate as compared with controls, with the vast majority of mAb-treated recipients dying by day 9 post-BMT (Fig. 3⇓). Weight curves (data not shown) and clinical appearance of recipients were reflective of survival curves, consistent with a fulminant GVHD process in anti-CTLA-4 mAb-treated recipients. These findings were reproduced in a second experiment.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

The roles of CTLA-4:B7 interaction on GVHD induced lethality in heavily irradiated recipients of allogeneic donor splenocytes. B10.BR recipients were irradiated (8 Gy TBI) and then given B6 TCD BM plus supplemental splenocytes (15 × 10⁶) and either hamster IgG irrelevant mAb or anti-CTLA-4 mAb, as described in Materials and Methods. On the x-axis are the days post-BMT, and on the y-axis are the proportion of mice surviving. Data from two replicate experiments with similar results are pooled.

Our data clearly indicate that anti-CTLA-4 mAb infusion enhances the expansion of donor T cells exposed to host alloantigens in vivo. To determine whether anti-CTLA-4 mAb would increase host anti-donor T cell-mediated BM rejection, we needed to analyze sublethally irradiated recipients reconstituted with allogeneic BM cells. To facilitate graft rejection, BM was depleted of T cells to preclude a graft-enhancing effect of donor T cells. The separate effects of host CD4⁺ and CD8⁺ T cell-mediated graft rejection were analyzed using the same two strain combinations as for quantifying the GVH responses of purified donor CD4⁺ and CD8⁺ T cells.

Donor TCD BM was infused in sublethally irradiated recipients of MHC class I (bm1→B6) or II (bm12→B6) disparate recipients. Peripheral blood chimerism of T, B, and myeloid lineages was assessed at time periods up to 145 days post-BMT. In both strain combinations, recipients given anti-CTLA-4 mAb had significantly lower proportions of donor cells and higher proportions of host cells, with all lineages being affected (Table I⇓). Particularly striking were the increases in the proportion of host CD4⁺ and CD8⁺ T cells in the anti-CTLA-4 mAb-treated recipients, indicating that anti-CTLA-4 mAb stimulates a host anti-donor response culminating in a higher degree of graft resistance. These data provide a likely explanation for the improved survival rate in sublethally irradiated recipients of allogeneic T cells and anti-CTLA-4 mAb (data not shown), which is presumably due to an augmented host anti-donor resistance process, limiting the GVH lethality effect of donor T cells. Together with the GVH results, these data indicate that the selective targeting of CD28:B7 rather than targeting both CD28:B7 and CTLA-4:B7 interactions would be desirable in preventing GVH lethality and preserving alloengraftment early post-BMT.

Anti-CTLA-4 mAb augments the GVL effect of delayed donor splenocyte infusion

The enhanced donor T cell expansion caused by anti-CTLA-4 mAb could be advantageous in augmenting the GVL effect of donor T cells given post-BMT. Because anti-CTLA-4 mAb increases GVHD lethality when given early post-BMT, we elected to study the potential GVL effect of anti-CTLA-4 mAb given later post-BMT when GVHD lethality of donor T cells is markedly reduced. For this purpose, we utilized a model of delayed splenocyte infusion in which lethally irradiated B6 recipients are reconstituted with B10.BR TCD BM for a period of 3 wk before donor splenocyte infusion (21). Donor T cells given at this later time period post-BMT have a markedly lower capacity to cause GVHD lethality than when given on day 0 of BMT (21). One week after donor splenocyte infusion, recipients are challenged with a supralethal dose of AML cells of the B6 strain. In previous studies, we have shown that the GVL effects of delayed splenocyte infusion against AML cells in this model are dependent upon the binding of B7 ligand(s) to T cell counter-receptor(s) (21). Donor CD8⁺ T cells and, to a lesser extent, CD4⁺ T cells present in the donor spleen cells are responsible for the GVL effect of DLI in this setting, while NK cells do not substantially contribute to this response (21).

To discern whether blockade of CTLA-4:B7 binding would enhance the GVL effect against AML cells, anti-CTLA-4 mAb was given to reconstituted recipients of delayed post-BMT donor splenocyte infusion, which are challenged with AML cells. Donor splenocytes provided a GVL effect that delayed the occurrence of leukemia as compared with recipients not given donor splenocytes (Fig. 4⇓A), although all mice treated with donor splenocytes eventually succumbed with leukemia by day 75 post-BMT. In striking contrast, none of the anti-CTLA-4 mAb-treated recipients of donor splenocytes and AML cells examined during the entire 220-day risk period had evidence of leukemia, despite the fact that donor splenocyte infusion alone did not prevent leukemia occurrence in any of the recipients analyzed 75 days post-BMT.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Anti-CTLA-4 mAb infusion augments the GVL capacity of delayed donor splenocyte infusion in recipients challenged with acute myeloid leukemia cells. B6 mice were irradiated with 8 Gy TBI on day −1 and then given B10.BR T and NK cell-depleted BM (8 × 10⁶) on day 0 of BMT. On day 21, cohorts of mice, as indicated, were given B10.BR splenocytes (25 × 10⁶). Anti-CTLA-4 mAb or irrelevant hamster IgG was administered as described in Materials and Methods. On day 28, C1498 acute myeloid leukemia cells were given i.v. at a supralethal dose (2 × 10⁵). Ten mice per group were transplanted. A, Leukemia incidence. All mice dying post-BMT that had received C1498 cells were examined for the presence of tumor. The proportion of mice with identifiable leukemia is shown on the y-axis, and days post-BMT on the x-axis. Eight to nine mice per group were analyzed for relapse. Differences in actuarial rates of leukemia are as follows: C1498 vs spleen/C1498 (p = 0.00094); C1498 vs spleen/C1498/anti-CTLA-4 mAb (p = 0.0015). B, On the x-axis are the days post-BMT, and on the y-axis is the proportion of mice surviving. The actuarial survival is plotted. As compared with control recipients of no spleen/C1498, differences in actuarial survival rates were as follows: spleen/no C1498 (p = 0.0079); spleen/C1498 (p = 0.00069); spleen/C1498/anti-CTLA-4 mAb (p = 0.014). There was no significant survival difference (p = 0.28) between groups of recipients that received spleen/C1498 with or without anti-CTLA-4 mAb.

The overall survival of recipients of donor splenocytes, AML cells, and anti-CTLA-4 mAb was modestly, but not significantly (p = 0.28), decreased, as compared with a group of mice receiving donor splenocytes, no AML cells, and no mAb treatment (Fig. 4⇑B). Anti-CTLA-4 mAb-treated recipients had a tendency toward lower weights than controls that received neither AML cells nor anti-CTLA-4 mAb (data not shown). Of note, the 100-day GVHD mortality risk, presumably the greatest risk period for observing mAb-induced GVHD acceleration of donor splenocytes, was highly comparable with survival rates of 44% or 50% in these two groups, respectively. Taken together, these data indicate that anti-CTLA-4 mAb given later post-BMT in the context of DLI mediated a potent GVL effect, while at most only slightly increasing the GVH side effects of DLI.

Role of CTLA-4:B7 interactions on modulating CD28-dependent and CD28-independent T cell responses

CTLA-4:B7 interaction may serve to regulate CD28:B7 responses. Therefore, experiments were performed to examine the role of CTLA-4:B7 in modifying the responses of CD28^−/− T cells to alloantigens. We first examined CD4⁺ T cell responses obtained from wild-type or CD28^−/− donors on in vitro proliferative responses (Fig. 5⇓A) and in vivo GVH lethality responses (Fig. 5⇓B) to MHC class II disparate bm12 strain mice in the presence or absence of anti-CTLA-4 mAb. In vitro proliferative responses of CD28^−/− CD4⁺ T cells to bm12 stimulator cells were lower than observed using CD4⁺ T cells from wild-type controls beginning on day 2 of culture (Fig. 5⇓A). Anti-CTLA-4 mAb increased proliferation of wild-type but not CD28^−/− T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Anti-CTLA-4 mAb increases the in vitro proliferation and in vivo GVHD capacity of wild-type but not CD28^−/− CD4⁺ T cells responsive to MHC class II disparate stimulator cells. In A, results of proliferation studies, using purified B6 wild-type or CD28^−/− CD4⁺ T cells incubated with bm12 TCD splenic stimulators in the presence of anti-CTLA-4 mAb (50 μg/ml), are shown. On the x-axis are the days of cell culture. The mean cpm are shown on the y-axis using a log10 scale. Error bars represent 1 SD of the mean. cpm for syngeneic responses were <10% of cpm for allogeneic responses. The in vivo GVHD results for 0.3 × 10⁶ (B) or 1 or 3 × 10⁶ wild-type or 1 × 10⁶ CD28^−/− T cells (C), infused along with TCD BM into heavily irradiated recipients, are shown. As indicated, recipients were given either irrelevant or anti-CTLA-4 mAb. On the x-axis are the days post-BMT, and on the y-axis is the proportion of mice surviving. All groups had five mice, except for the groups receiving 10⁶ CD28^−/− CD4⁺ T cells, which had 10 mice per group. In B, recipients of 0.3 × 10⁶ wild-type but not CD28^−/− T cells and anti-CTLA-4 mAb had a significantly (p = 0.001) lower survival than those given irrelevant mAb. In C, recipients given 1 × 10⁶ wild-type or 1 × 10⁶ CD28^−/− T cells and anti-CTLA-4 mAb each had a trend (p = 0.08) toward a lower survival rate than recipients given irrelevant mAb. <226>, p <0.05 vs CD28^+/+ group; #, p <0.05 vs CD28^-/- group that did not receive mAb.

To determine whether the absence of CD28:B7 interactions would affect the accelerated in vivo GVH effects of anti-CTLA-4 mAb infusion, bm12 recipients were given a uniformly sublethal number of B6 wild-type or CD28^−/− CD4⁺ T cells (0.3 × 10⁶) and either irrelevant or anti-CTLA-4 mAb. In recipients of wild-type T cells, survival was 100%, significantly higher than the 0% observed in recipients that were also given anti-CTLA-4 mAb (Fig. 5⇑B). In contrast, recipients given CD4⁺ T cells from CD28^−/− donors had 100% survival regardless as to whether these mice received irrelevant or anti-CTLA-4 mAb. To attempt to uncover a more modest effect of anti-CTLA-4 mAb on accelerating GVHD lethality of CD28^−/− CD4⁺ T cells, the number of donor T cells was increased to 10⁶ per recipient. Under these conditions, acceleration of GVHD by anti-CTLA-4 mAb given to recipients of wild-type CD4⁺ T cells was difficult to assess since these recipients all succumbed by day 10 post-BMT. Anti-CTLA-4 mAb-treated recipients uniformly died by day 9 post-BMT. Recipients (n = 10/group) given 10⁶ CD4⁺ T cells from CD28^−/− donors and irrelevant mAb had an 80% long-term (>2.5-mo post-BMT) survival rate (Fig. 5⇑C). Anti-CTLA-4 mAb-treated recipients had a 50% survival rate that was of borderline (p = 0.08) significance. Because recipients of a 3-fold higher number of CD28^−/− CD4⁺ T cells (3 × 10⁶) had 0% survival, these data would indicate that anti-CTLA-4 mAb infusion resulted in <3-fold increase in CD28^−/− CD4⁺ T cell-mediated lethality.

To determine whether alloreactive CD28^−/− CD4⁺ and CD8⁺ T cells would be stimulated to expand in the presence of anti-CTLA-4 mAb, TDL were obtained from cohorts of lethally irradiated B10.BR recipients of B6 or CD28^−/− donor splenocytes on days 6 and 9 post-BMT. At the time of peak expansion (day 6), donor T cell expansion was significantly higher in recipients of wild-type as compared with CD28^−/− donor cells (Fig. 6⇓A). The increase in TDL T cell number in anti-CTLA-4 mAb-treated BMT recipients was CD28 dependent since no increases in TDL T cells were observed in mAb-treated recipients of CD28^−/− T cells. Thus, under these conditions, the negative regulatory role of CTLA-4:B7 interaction is CD28 dependent. Flow cytometry analysis of TDL obtained on days 6 and 9 did not reveal any obvious differences in activation profiles (data not shown). In recipients of wild-type and CD28^−/− donor splenocytes, donor CD4⁺ T cells comprised a higher proportion of the TDL on days 6 and 9 post-BMT in recipients given anti-CTLA-4 mAb as compared with irrelevant hamster IgG mAb (≥50% increase over baseline percentages at both time points). In situ hybridization analysis for cytokine mRNA was performed using fresh TDL obtained from a pool of four mice per group. In recipients of CD28^−/− T cells, TDL were skewed toward a Th2 phenotype, as evidenced especially by increases in the frequency of IL-5 (5.7-fold) and IL-10 (5.5-fold), with a concomitant decrease in IFN-γ mRNA-expressing cells (Table II⇓). In addition, as observed in the (B6 × 129)F₂→BALB/c-SCID system, anti-CTLA-4 mAb administration led to a skewed Th2 response in TDL obtained from recipients of wild-type cells, and further skewed the Th2 response of CD28^−/− T cells increasing the frequency of TDL cells expressing IL-4, IL-5, and IL-10 (≥2.5-fold), while decreasing the frequency of IFN-γ-expressing cells to frequencies that were 3-fold to 6-fold lower than controls. Also as previously noted, anti-CTLA-4 mAb did not substantially alter the percentage of granzyme B and perforin mRNA-expressing TDL T cells (Table II⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The increased expansion of allogeneic donor T cells induced by anti-CTLA-4 mAb is CD28 dependent. B10.BR recipients were irradiated (8 Gy TBI) and then given B6 or CD28^−/− TCD BM plus supplemental splenocytes (15 × 10⁶), and either hamster IgG irrelevant mAb or anti-CTLA-4 mAb, as described in Materials and Methods. TDL were isolated from each group on days 6 (n = 6/group) and 9 (n = 6–7 per group, except for recipients of wild-type splenocytes and anti-CTLA-4 mAb that had a high early post-BMT mortality rate (n = 3/group)). A, Quantification of TDL cell number. On the y-axis are the number of TDL per ml quantified in individual mice, and on the x-axis are the days post-BMT. Groupwise comparisons were made. B, TDL proliferation. On the y-axis is mean cpm using a log10 scale, and on the x-axis are days in culture. Error bars represent 1 SD. cpm for syngeneic responses were <10% of cpm for allogeneic responses. C, TDL CTL function. On the y-axis is percent specific cytolysis. E:T ratios are indicated on the x-axis. Error bars represent 1 SD. Significant differences between CD28^−/− and wild-type recipients of irrelevant mAb are designated by ∗. Significant differences between recipients that received anti-CTLA-4 mAb as compared with irrelevant mAb are designated by #.

TDL T cells were examined for in vitro proliferative responses to host alloantigen-bearing stimulator cells (Fig. 6⇑B). Peak responses in all groups on day 1 were indicative of in vivo priming. Wild-type TDL had greater proliferation to stimulator cells than CD28^−/− TDL. Proliferative responses of TDL obtained from anti-CTLA-4 mAb-treated recipients were not higher than observed with TDL from irrelevant mAb-treated recipients. Cytolysis against host target cells was higher when using CD28^−/− T cells as effector cells (Fig. 6⇑C). Anti-CTLA-4 mAb did not increase the cytolytic capacity of TDL obtained from either irrelevant or anti-CTLA-4 mAb-treated recipients, despite the fact that the proportion of CD8⁺ T cells was lower in TDL obtained from anti-CTLA-4 mAb versus irrelevant mAb-treated recipients. Therefore, although CTLA-4:B7 regulates the expansion of wild-type but not CD28^−/− T cells, neither the in vitro anti-host proliferative nor cytolytic responses of TDL T cells from recipients of wild-type or CD28^−/− appear to be substantially increased by anti-CTLA-4 mAb infusion.