Alloreactive CD8 T cells rescued from apoptosis during co-stimulation blockade by Toll-like receptor stimulation remain susceptible to Fas-induced cell death

Authors


Correspondence: Dr Michael A. Brehm, Program in Molecular Medicine, University of Massachusetts Medical School, 373 Plantation Street, Biotech 2, Suite 218, Worcester, MA-01605, USA. Email: michael.brehm@umassmed.edu

Senior author: Michael A. Brehm

Summary

Blockade of co-stimulatory signals to T cells is extremely effective for the induction of transplantation tolerance in immunologically naive rodents. However, infections and inflammation compromise the efficacy of co-stimulation blockade regimens for the induction of tolerance, thereby stimulating the rejection of allografts. Previous studies have shown that stimulation of innate immunity abrogates tolerance induction by preventing the deletion of alloreactive CD8+ T cells that normally occurs during co-stimulation blockade. Although inflammation prevents the deletion of alloreactive T cells during co-stimulation blockade, it is not known if this resistance to cell death is the result of a mechanism intrinsic to the T cell. Here, we used syngeneic bone marrow chimeric mice that contain a trace population of T-cell receptor transgenic alloreactive CD8+ T cells to investigate the early apoptotic signature and activation status of alloreactive T cells following exposure to inflammatory signals during co-stimulation blockade with an antibody specific for CD154. Our findings revealed that the presence of bacterial lipopolysaccharide during co-stimulation blockade enhanced the early activation of alloreactive CD8+ T cells, as indicated by the up-regulation of CD25 and CD69, suppressed Fas ligand expression, and prevented apoptotic cell death. However, alloreactive CD8+ T cells from lipopolysaccharide-treated mice remained sensitive to Fas-mediated apoptosis in vitro. These findings suggest that alloreactive T cells rescued from deletion during co-stimulation blockade by inflammation are still sensitive to pro-apoptotic signals and that stimulating these apoptotic pathways during co-stimulation blockade may augment the induction of tolerance.

Abbreviations
DES

desire

DST

donor-specific transfusion

LPS

lipopolysaccharide

PCD

passive cell death

Introduction

Blockade of co-stimulatory pathways during T-cell activation is an effective strategy to induce peripheral tolerance to transplanted allogeneic tissues.[1, 2] For example, blockade of the CD28 pathway with CTLA-4-immunoglobulin alone or in combination with blockade of additional co-stimulatory pathways such as ICOS, CD134, 4-1BB and CD27 promote allograft survival.[1, 3, 4] As an alternative to targeting CD28–B7 interactions, blockade of the CD40–CD154 pathway induces tolerance to allogeneic skin, islets and bone marrow[5-7] and has been used successfully to induce tolerance to allografts in non-human primates.[8, 9] Two mechanisms by which co-stimulation blockade is thought to induce tolerance are by the deletion of activated alloreactive T cells[6, 10, 11] and by promoting the emergence of regulatory T cells that are essential for the maintenance of tolerance.[12]

Inflammation resulting from the activation of pattern-recognition receptors, such as Toll-like receptors (TLR), nucleotide oligomerization domain-like receptors and retinoic acid-inducible gene-I-like receptors,[13] by either surgical procedures or viral and bacterial infections reduces the efficacy of co-stimulation blockade and results in graft rejection.[6, 14-18] Exposure to TLR agonists, such as poly I:C or lipopolysaccharide (LPS), stimulates a vigorous inflammatory reaction and prevents deletion of alloreactive T cells through a type-I interferon (IFN) -dependent mechanism.[6, 7] Recent studies have also shown that CpG, a TLR9 agonist, abrogates tolerance induced by CD40–CD154 blockade by impairing the function of regulatory T cells, enabling T helper differentiation and promoting the development of T helper type 17 responses.[19, 20]

The balance between the frequency of activated alloreactive effector T cells and regulatory T cells determines the fate of allografts during tolerance induction.[21, 22] Consequently, the lack of alloreactive T-cell deletion in the presence of inflammation may be crucial for shifting the balance towards effector T-cell responses, abrogating tolerance and allograft rejection.[23] We therefore questioned if alloreactive T cells are completely refractory to deletion in the presence of inflammation during co-stimulation blockade. To determine this, we first investigated the expression of cell death-associated genes by alloreactive CD8 T cells during the induction of tolerance by co-stimulation blockade in the absence or presence of the TLR4 agonist LPS. Using a bone marrow chimera model[24], we show that T-cell receptor transgenic alloreactive CD8+ T cells increased expression of Fas ligand (FasL), a pro-apoptotic molecule involved in activation-induced cell death, during co-stimulation blockade with an antibody specific for CD154 (MR1).[25] However, the presence of LPS during co-stimulation blockade diminished the expression of FasL and enhanced the expression of the Fas death receptor by alloreactive CD8 T cells. Provision of exogenous Fas agonist in vitro induced apoptosis in these alloreactive T cells. This finding suggests that alloreactive T cells prevented from undergoing deletion by LPS treatment during co-stimulation blockade are not completely refractory to the apoptotic signals and that intervention during this early time period may be a strategy to bypass the anti-apoptotic effects of inflammation on alloreactive T cells and thereby allow for the survival of the allograft.

Materials and methods

Mice

Male C57BL/6J (H2b), BALB/CJ (H2d) and CBA/J (H2k) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). (CBA/J × KB5.CBA/J) F1 CD8+ T-cell male and female mice that express a transgenic T-cell receptor that recognizes H2Kb as alloantigen[24, 26, 27] were bred at the University of Massachusetts Medical School (UMMS) Department of Animal Medicine. KB5 synchimeric mice were generated as described previously.[24] The B6.MRL-Fas lpr/J (Fas-deficient) mice were bred at the UMMS Department of Animal Medicine and B6 Smn.C3-Faslgld/J mice (FasL-deficient) were purchased from the Jackson Laboratory. Lpr and gld mice were matched for age and weight and used when between 6 and 7 weeks of age before they developed lymphoproliferative disorders. All experiments were carried out in compliance with the institutional guidelines as approved by the Institutional Animal Care and Use Committee of UMMS.

Tolerance induction regimen

Recipient mice were given donor-specific transfusion (DST) and MR1 antibody as described previously.[6] To determine the effect of TLR activation at the time of tolerance induction, some mice were also injected intraperitoneally with 100 μg LPS (Ultra Pure LPS from Escherichia coli 0111:B4 strain, Invivogen San Diego, CA) on day –7.

Flow cytometry

Monoclonal antibodies were purchased from BD Biosciences. For tracking alloreactive KB5 transgenic CD8+ T cells in KB5 synchimeric mice, primary DES clonotypic antibody was used as previously described.[24] Samples were analysed using a BD LSRII Flow Cytometer (BD Biosciences, San Jose, CA) and FlowJo software (Tree star Inc., Ashland, OR).

Real time PCR array

Alloreactive DES+  CD8β+ T cells were recovered from KB5 synchimeric mice at the indicated time-points following the initiation of tolerance induction and then purified using the MoFlo™ XDP cell sorter (Beckman Coulter, Inc., Brea, CA) to > 98% purity. Total RNA was isolated using an RNA isolation kit (Qiagen Valencia, CA), and genomic DNA was removed using an RNase-free DNase kit (Qiagen). RNA was then reverse-transcribed into cDNA using RT2 PCR Array First Strand Kit (Superarray, Valencia, CA). Real-time PCR was performed on the synthesized cDNA using the RT2 Real-Time™ SYBR Green/ROX PCR master mix according to the manufacturer's protocol (Superarray). The fold regulation was calculated using the ΔΔCt method (Superarray). The positive fold changes > 1 indicated the fold up-regulation and fold changes < 1 were represented as negative inverse of fold change, which indicated fold down-regulation.

Annexin-V staining

Spleens from KB5 synchimeric mice were harvested at the indicated time-points, and pre-incubated for 4 hr in vitro at either 37° or at 4°. After the incubation, cells were stained for surface markers. This was followed by incubation with annexin-V (BD Biosciences) and 7-aminoactinomycin-D (BD Biosciences) in 1 × annexin-V buffer at room temperature in the dark. The cells were washed with annexin-V buffer and immediately analysed on an LSR2 flow cytometer.

FasL staining

The protocol for FasL surface staining was performed as previously described.[28] Stained samples cells were immediately analysed on the LSR2.

Use of Fas agonistic antibodies during in vitro cultures

Splenocytes (106) isolated from KB5 synchimeric mice (after the indicated treatments) were pre-incubated in 48-well flat bottom plates for 4 hr in vitro either at 37° or 4°. To determine the effect of Fas agonistic antibody on the apoptotic profile of alloreactive T cells, splenocytes were incubated with 5 μg/ml Fas (Jo2) antibody (BD Biosciences) for 4 hr in vitro[29] followed by annexin-V staining.

In vivo cytotoxicity assay

The in vivo cytotoxicity assay was performed as previously described.[30] The specific lysis of allogeneic targets was calculated according to the following formula: 100 × {[(% allogeneic target population in experimental/% syngeneic target population in experimental) ÷ (% allogeneic target population in natural killer-depleted DST + MR1 control/% syngeneic target population in natural killer-depleted DST + MR1 control)] × 100}.[31, 32]

Serum cytokine analysis

The levels of interleukin-6 (IL-6), IL-10, monocyte chemoattractant protein-1, IFN-γ, tumour necrosis factor (TNF), and IL-12p70 (IL12p70) were determined in the serum from the indicated mice using the BD™ Cytometric Bead Array.

Statistics

Sample analyses were performed using Graph Pad Prism (Graph Pad Software). Samples were analysed by unpaired t-test wherever specified, and P values are indicated. A one-way analysis of variance with a Tukey post-test was used to compare multiple samples, with a P-value of < 0·05 considered significant.

Results

Treatment with LPS at the time of co-stimulation blockade inhibits the early deletion in alloreactive KB5 transgenic CD8+ T cells

Previously, Thornley et al. showed that the frequencies of alloreactive CD8 T cells were significantly reduced by 24 hr after co-stimulation blockade.[6] To determine the kinetics of alloreactive T-cell deletion before 24 hr, we quantified the percentages and the absolute number of KB5 alloreactive CD8+ T cells in the spleens of KB5 synchimeric mice treated with co-stimulation blockade regimen at 6, 11 and 15 hr (or as indicated). The blockade regimen consisted of treating mice with an injection of donor splenocytes (donor-specific transfusion or DST) and injection of an antibody-specific for CD154 (MR1). The percentages and absolute numbers of alloreactive KB5 transgenic CD8+ T cells (KB5 Tg CD8+ T cells) decreased significantly by 11 hr post-treatment (DST + MR1) relative to the 6-hr time-point and were further decreased by 14–15 hr (Fig. 1a,b). The percentages in untreated mice were unaffected at any time-point tested (Fig. 1a).

Figure 1.

Lipopolysaccharide (LPS) treatment at the time of co-stimulation blockade prevents the decline in alloreactive KB5 transgenic (Tg) CD8+ T cells. KB5 CD8+ synchimeric mice were left untreated, given donor-specific transfusion (DST) in the form of B6 splenocytes along with MR1 antibody in the absence or presence of LPS. At the indicated time-points, recipient spleens were harvested and stained for transgenic alloreactive T cells using KB5-DES clonotypic antibody and anti-CD8β. The percentages (a) and the absolute numbers (b) of KB5 Tg CD8+ T cells are shown at the indicated time-points. (c) Absolute numbers of KB5 Tg CD8+ T cells at the indicated time-points. (d) Percentages of annexin-V-positive cells in three different treated groups (columns) at three time-points (rows). (e) Mean fluorescence intensity (MFI) of annexin-V of KB5 Tg CD8+ T cells in three groups at 8–9 hr post treatment. Each dot represents an individual mouse.

As shown above, the deletion of alloreactive T cells was first evident by 12 hr after initiation of co-stimulation blockade, and we predicted that the effects of LPS on alloreactive T-cell survival would be detectable at these early time-points. Both the absolute numbers (Fig. 1c) and the percentage (data not shown) of alloreactive KB5 Tg CD8+ T cells in the DST + MR1 + LPS-treated groups were significantly higher at 12 hr than in the DST + MR1-treated groups. Moreover, the percentages of annexin-V-positive alloreactive KB5 Tg CD8+ T cells in the LPS-treated mice were significantly lower compared with mice not treated with LPS during the early time-points tested (Fig. 1d). The reduction in the percentage of annexin-V-positive alloreactive KB5 Tg CD8+ T cells correlated with a lower mean fluorescent intensity (MFI) for annexin-V during DST + MR1 + LPS treatment (Fig. 1e). Together, these results indicate that LPS prevents KB5 Tg CD8+ T cells from undergoing apoptotic cell death early during co-stimulation blockade.

LPS treatment during co-stimulation blockade enhances the expression of CD25 on activated alloreactive KB5 Tg CD8+ T cells

Co-stimulation blockade does not prevent the activation and division of alloreactive T cells as measured by the up-regulation of CD44 and Ki67 staining.[6] To determine if exposure to LPS alters this activation profile, we examined changes in the expression of early activation markers, including CD44, CD25, CD62L, CD69 and CD127 (IL-7R) in the KB5 Tg CD8+ T cells (Fig. 2). The percentages of cells that down-regulated CD62L and CD127 expression were similar between the DST + MR1 and DST + MR1 + LPS groups (Fig. 2a–c). Treatment with DST alone or DST + MR1 increased the percentage of cells expressing CD25 (25·4 ± 5·7 for DST and 19·5 ± 3·7 for DST + MR1, P > 0·05) and CD69 (86·4 ± 3·3 for DST and 85·7 ± 7·4 for DST + MR1 P > 0·05) to a similar level, and this was further enhanced following simultaneous exposure to LPS. CD69 was also up-regulated in the endogenous non-transgenic T cells upon LPS treatment (Fig. 2b), consistent with previous studies.[33] In addition, LPS treatment significantly increased the expression of inflammatory cytokines, including IL-6, TNF, IFN-γ, IL-12 and monocyte chemoattractant protein-1 (see Supplementary material, Fig. S1). Together these results indicate that there was no overall defect in the early activation of alloreactive KB5 Tg CD8+ T cells during co-stimulation blockade and that exposure to LPS enhances the activation of alloreactive T cells.

Figure 2.

Co-stimulation blockade does not prevent the early activation of alloreactive KB5 transgenic (Tg) CD8+ T cells. Groups of KB5 synchimeric mice were either treated with donor-specific transfusion (DST) + MR1 or DST + MR1 along with lipopolysaccharide (LPS) for 9 hr as described in the Materials and methods. (a, b) Activation profiles of the indicated markers during DST, DST + MR1 and DST + MR1 + LPS treatment in KB5 Tg CD8+ T cells and endogenous non-transgenic CD8+ T cells. (c, d) Average percentages of activated alloreactive T cells with respect to each of the activation markers and their corresponding changes in the mean fluorescence intensity (MFI) of each of the markers in KB5 Tg CD8+ T cells.

LPS exposure alters the apoptotic signature of activated alloreactive T cells during co-stimulation blockade

Next we examined the intrinsic apoptotic gene profile of alloreactive T cells from KB5 synchimeric mice treated with co-stimulation blockade and evaluated how the expression pattern of these same genes is altered in the presence of LPS. To do this, we used a real-time PCR array consisting of a focused set of genes involved in apoptosis, and we analysed the transcriptional signatures of these apoptotic genes in alloreactive KB5 Tg CD8+ T cells 12 hr after treatment with DST + MR1 or DST + MR1 + LPS. We focused on genes that were differentially regulated by at least fivefold between DST + MR1 + LPS-treated groups and DST + MR1-treated groups (see Supplementary material, Figs S2 and S3). The fold induction values of various genes with respect to the untreated group are shown in Table 1 (average of two experiments).

Table 1. Transcriptional differences in the expression of pro-apoptotic and anti-apoptotic genes in alloreactive KB5 Tg CD8+ T cells during lipopolysaccharide (LPS)-mediated abrogation of tolerance induction
Gene familyDST + MR1DST + MR1 + LPS
PositionGene symbolExperiment 1Experiment 2Experiment 1Experiment 2
  1. Positive values indicate the fold up-regulation and negative values indicate fold down-regulation of genes with respect to the untreated groups in two independent experiments.

Bcl-2 family
A05Bad−9·77−14·52−9·65−9·46
A09Bax3·041·394·612·31
A11Bcl-2−6·67−9·32−3·84−2·19
A12Bcl2 l12·35−1·031·39−1·05
B03Bid1·351·091·17−1·02
Card family
C08Casp2−2·18−2·62−3·61−3·54
C09Casp32·82−1·491·57−1·20
C10Casp41·8−2·533·031·70
C12Casp7−2·19−3·56−1·59−1·11
D02Casp9−10·08−14·93−12·39−1·92
Death effector domain family
D01Casp8−7·88−4·29−4·03−2·03
D12Fadd−1·65−2·19−2·211·22
TNF ligand family
E02FasL30·0031·561·832·77
F07TNF8·479·0010·2226·50
F12TRAIL−20·94−16·56−7·14−5·91
G02CD40L−4·46−1·88−6·35−4·63
TNF receptor family
E01Fas−2·91−4·08−1·85−1·98
F11CD40−7·15−25·81−1·33−1·62
Anti-apoptotic family
E04IL10−3·41−3·23122·2858·40

The expression of FasL, a member of the TNF superfamily, was an average of 14·5-fold lower in mice treated with DST + MR1 + LPS relative to mice treated with only DST + MR1 (see Supplementary material, Fig. S3). This change was attributed to a nearly 30-fold up-regulation of FasL in the DST + MR1-treated groups, in contrast to only a twofold up-regulation in the DST + MR1 + LPS-treated groups when normalized to the untreated group (Table 1). Changes in FasL mRNA levels were validated by examining the expression of FasL protein in alloreactive KB5 Tg CD8+ T cells in KB5 synchimeric mice that were either untreated or treated with DST + MR1 and DST + MR1 + LPS for 9 hr and then stained for surface expression of FasL (Fig. 3a,b). These findings suggest that LPS treatment inhibited FasL up-regulation in alloreactive T cells that occurred during co-stimulation blockade.

Figure 3.

Detection of Fas and Fas ligand (FasL) protein on the surface of KB5 transgenic (Tg) CD8+ T cells during donor-specific transfusion (DST), DST + MR1 and DST + MR1 + lipopolysaccharide (LPS) treatment. Splenocytes of KB5 synchimeric mice treated with DST + MR1 and DST + MR1 + LPS for 9 hr were stained for surface FasL expression as described in the Materials and methods section. (a) Percentages of cells positive for FasL (red histograms) in alloreactive KB5 CD8+ transgenic T cells and the endogenous non-transgenic CD8+ T cells. The black histograms indicate the FasL isotype staining in each of the groups. (b) Comparison of averages of FasL MFI profile in KB5 Tg CD8+ T cells from two experiments with P < 0·05. (c, d) Percentages of cells positive for Fas receptor and their respective mean fluorescence intensities (MFI).

The Fas–FasL pathway is not required for co-stimulation blockade-induced tolerance

To test the role of the Fas–FasL pathway in the deletion of alloreactive T cells during co-stimulation blockade, we used lpr (Fas mutant) and gld (FasL mutant) mice as hosts in an in vivo cytotoxicity assay (Fig. 4). This assay is a reliable test for activated alloreactive T cells, because it determines their ability to lyse allogeneic target cells.[31, 34] B6 mice injected with BALB/c DST efficiently killed BALB/c allogeneic target cells in vivo, and this killing was prevented by treatment with MR1 (Fig. 4). Both lpr (Fig. 4a,b) and gld (Fig. 4c,d) mice injected with BALB/c DST efficiently killed BALB/c allogeneic target cells, confirming that these mice are able to generate alloreactive T-cell responses. In contrast, lpr and gld mice treated with DST + MR1 were unable to kill BALB/c target cells, indicating that the alloreactive T-cell response was tolerized. Together, these results suggest that Fas and FasL are not critical for co-stimulation blockade-induced tolerance induction.

Figure 4.

Fas and FasL expression in the recipient is not necessary for the induction of co-stimulation-blockade tolerance. Wild-type (WT) B6 (H2b), lpr (H2b) and gld (H2b) mice were subjected to co-stimulation blockade regimen with allogeneic BALB/c splenocytes (H-2d) as donor-specific transfusion (DST) and three injections of MR1 (αCD154) as described in the Materials and methods. On day 1, mice were given NK1.1 depleting antibody. The following day (day 0), mice were injected with syngeneic and allogeneic splenocyte populations that were labelled with the indicated concentrations of CFSE. On day + 1, recipient spleens (WT, lpr, gld) were harvested and the lysis of allogeneic target cells was determined as described in the Materials and methods. (a) Representative histograms of allogeneic (H-2d) and syngeneic (H-2b) target cells before transfer and after the indicated treatments in WT and lpr recipients. (b) Comparison of per cent lysis of allogeneic target cells with respect to the syngeneic populations normalized to WT B6 DST + MR1-treated group between WT and lpr hosts. (c) Representative histograms of allogeneic (H-2d) and syngeneic (H-2b) target cells before transfer and after the indicated treatments in WT and gld recipients. (d) Comparison of per cent lysis of allogeneic target cells with respect to the syngeneic populations normalized to the WT B6 DST+MR1-treated group between WT and gld hosts.

In vitro activation of Fas receptor renders alloreactive T cells from the LPS-treated mice susceptible to apoptosis

Although FasL expression was suppressed in KB5 Tg CD8+ T cells from mice treated with DST + MR1 + LPS, Fas receptor expression was up-regulated in these cells (Fig. 3c,d). Given that the Fas receptor expression on alloreactive T cells is increased in the presence of LPS, we hypothesized that alloreactive T cells emerging from a pro-inflammatory environment (co-stimulation blockade + LPS) may have enhanced sensitivity to Fas-mediated apoptosis in vitro. Therefore we tested whether the engagement of Fas on the surface of alloreactive T cells from DST + MR1 + LPS-treated mice would induce their apoptosis in vitro. Splenocytes were recovered from KB5 synchimeric mice 9 hr after treatment with DST + MR1 and DST + MR1 + LPS and incubated with Fas agonistic or isotype-control antibodies for 4 hr in vitro and then examined by annexin-V binding assay. Incubation with Fas agonistic antibody increased the percentages of annexin-V-positive cells and also increased the MFI of the annexin-V stain in the LPS-treated groups (Fig. 5a–c). The endogenous non-transgenic CD8+ and CD4+ T cells did not show any changes in annexin-V profile in the presence of the Fas agonistic antibody. Together, these results suggest that alloreactive T cells that are prevented from deletion in the presence of LPS remain sensitive to Fas-mediated apoptosis.

Figure 5.

Engagement of Fas in vitro selectively induces apoptosis in alloreactive KB5 Tg CD8+ T cells from donor-specific transfusion (DST) +MR1 and lipopolysaccharide (LPS) -treated group. Splenocytes of KB5 synchimeric mice treated with DST, DST + MR1 and DST + MR1 + LPS for 9–10 hr were incubated for 4 hr before staining for annexin-V. Splenocytes from the DST + MR1 + LPS group were incubated in the presence of either Fas agonist antibody or isotype control (5 μg/ml) during the 4 hr incubation period followed by annexin-V staining. (a) Representative percentages of cells that are annexin-V-positive, (b, c) average percentages of annexin-V-positive cells and the annexin-V mean fluorescence intensity (MFI), respectively. These results are representative of four individual experiments from a total of 17 mice.

Discussion

The early deletion of alloreactive T cells during co-stimulation blockade protocols is considered as an important mechanism for the induction of peripheral transplantation tolerance and for allograft acceptance.[6, 12] Exposure to TLR agonists during tolerance induction prevents this early deletion in a type-I IFN-dependent manner.[6, 7] However, it was not known if exposure to inflammatory stimuli during co-stimulation blockade would induce intrinsic alterations in the susceptibility of alloreactive T cells to apoptosis. Here, we show that LPS strikingly blocks the co-stimulation blockade-induced up-regulation of FasL, a major mediator of T-cell apoptosis, while concurrently inhibiting apoptosis in this cell population. Although the up-regulation of FasL expression by alloreactive T cells was suppressed in mice treated with DST + MR1 + LPS, these cells were still susceptible to Fas-mediated apoptosis in vitro. These findings suggest that alloreactive T cells prevented from undergoing deletion during co-stimulation blockade by exposure to inflammation remain susceptible to apoptotic stimuli.

Although the Fas–FasL pathway plays a vital role in mediating the process of activation-induced cell death, previous studies in lpr and gld mice[35-37] and our findings (Fig. 4) suggest that the Fas–FasL apoptotic pathway is not crucial for the deletion of alloreactive T cells during tolerance induction. Studies on immune privileged sites such as the cornea have revealed the involvement of both FasL and TRAIL in mediating tolerance.[38, 39] Recently, both Fas and B-cell lymphoma 2 interacting mediator of cell death (BIM) have been shown to have a synergistic role in maintaining lymphocyte homeostasis, as defects in both the molecules led to a rapid onset of autoimmune diseases in mice compared with mice deficient in either molecule alone.[40-42] In addition, we observed a modest increase in BIM expression (1·5-fold increase in MFI) in alloreactive T cells during co-stimulation blockade (data not shown). These studies suggest that compensatory or redundant apoptotic mechanisms can drive the deletion of alloreactive T cells in the absence of Fas–FasL signalling, which can lead to tolerance induction.

Lipopolysaccharide inhibited the up-regulation of FasL expression selectively in alloreactive T cells but induced a modest up-regulation of Fas protein on the surface of all T cells (endogenous non-transgenic CD8+ T cell and CD4+ T cells). Fas agonistic antibody treatment in vitro caused selective apoptosis of alloreactive T cells from DST + MR1 + LPS-treated mice (Fig. 5), suggesting that alloreactive T cells in this scenario are still susceptible to Fas-mediated apoptosis. The increased levels of FasL and Fas by alloreactive T cells during co-stimulation may render this population susceptible to a “fratricide” mechanism[43, 44], although this does not appear to be essential for the deletion of alloreactive T cells. We speculate that activation of pro-apoptotic signalling pathways (including the Fas–FasL pathway) in alloreactive T cells along with blockade of co-stimulatory signals may ensure their robust deletion and overcome the effects of inflammation during the early period after co-stimulation blockade treatment. Recently, several alternative strategies have been developed to target the Fas–FasL pathway in vivo without inducing toxicity[45], including the co-transplantation of syngeneic myoblasts engineered to express FasL, the use of CTLA-4-immunoglobulin fused with FasL and the use of DST engineered to display chimeric SA-FasL that is non-cleavable.[46-48]

Humans treated with anti-CD40L therapy unexpectedly showed complications of platelet aggregation and thromboembolism that was later attributed to the expression of CD154 on human platelets[49, 50] leading to the withdrawal of this promising therapy from clinical trials. Additionally, the use of anti-CD40L therapy in humans as a tolerance induction regimen raises concerns regarding its efficacy in the context of inflammation. Moreover, the induction of tolerance to transplantation antigens is also abrogated by the presence of memory T cells that recognize alloantigens and by T cells that have undergone homeostatic proliferation.[51, 52] Given these current challenges, understanding the downstream apoptotic molecular pathways mediating alloreactive T-cell death during blockade of the CD40–CD40L pathway is critical to identify novel and alternative targets for induction of peripheral tolerance. The results of our study suggest that activation of apoptotic pathways during co-stimulation blockade may circumvent the detrimental effects of inflammation on tolerance induction.

Acknowledgements

We thank the University of Massachusetts Medical School Flow Cytometry Core Laboratory for performing cell sorting. We thank the late Dr Nancy Philips, Dr Rabinarayan Mishra, Jean Leif, Carey Zammitti and Keith Daniels for helpful discussions and excellent technical assistance and Linda Paquin for expert mouse management. This work was supported in part by National Institutes of Health Grants AI017672 and AI081675 to R.M.W, AI46629 to D.L.G, AI083911 to M.A.B, an institutional Center for AIDS Research (CFAR) grant AI042845 and grants from the Juvenile Diabetes Research Foundation. R.M.W and D.L.G are members of the UMASS Diabetes Endocrinology Research Center (DERC) grant DK32520.

Disclosures

M.A.B. and D.L.G. are consultants for The Jackson Laboratories and for Perkin Elmer. All other authors declare no financial or commercial interests.

Ancillary