Cutting Edge: Early Attrition of Memory T Cells during Inflammation and Costimulation Blockade Is Regulated Concurrently by Proapoptotic Proteins Fas and Bim

Apoptosis is a critical mechanism regulating T cell homeostasis and is essential for T cell development, for suppression of autoreactive T cells, and for the contraction phase of an Ag-specific T cell response (1, 2). The well-described attrition of T cells occurring early after both viral infection (3) and costimulation blockade (CoB) therapy to extend allograft survival (4) also involves apoptosis of T cells. However, specific cell death pathways regulating the early apoptosis of T cells after viral infection or CoB are not well defined. Two distinct pathways regulate T cell apoptosis: the intrinsic and extrinsic pathways (5). The intrinsic pathway is regulated by the members of the B-cell lymphoma 2 family and includes prosurvival proteins and proapoptotic BH3 proteins and proapoptotic pore formers (6). The extrinsic pathway is activated by the binding of death ligands such as Fas ligand (FasL) to cognate death receptors (Fas) and results in the formation of the death-inducing signaling complex and activation of the initiator caspase 8 (1).

The immune response to viral infections involves two distinct stages of T cell apoptosis, which have been studied during acute infection with lymphocytic choriomeningitis virus (LCMV) in mice (3). The first wave of apoptosis occurs early after LCMV infection (2 to 4 days postinfection), and the second wave occurs after Ag is cleared during the contraction phase (3, 7). During the early T cell attrition phase, memory phenotype (CD44^hi) CD8⁺ T cells are more susceptible to deletion than naive (CD44^lo) CD8⁺ T cells (3). The viral dsRNA mimetic poly(I:C) simulates the early apoptosis observed during LCMV infection, and cell death is dependent on type I IFNs IFN-α/β (3, 8). FasL-deficient gld mice are not resistant to early CD8⁺ T cell deletion induced by poly(I:C), suggesting that the extrinsic death receptor pathway regulated by Fas-FasL interactions is not necessary for early T cell apoptosis (3). Mice lacking the proapoptotic protein Bim show a partial resistance to early deletion of CD44^hi CD8⁺ T cells postinfection with LCMV, suggesting a partial role for the intrinsic apoptosis pathway during this early T cell death (9).

Strategies to prolong allograft survival that target the CD28-B7 and CD40-CD154 pathways have been tested extensively in animal models (10). Death of alloreactive T cells is an important component for prolonged allograft survival during blockade of costimulation pathways (11). CoB using the reagents CTLA4-Ig and anti-CD154 mAb induce tolerance to skin, islets, heart, and kidney allografts in mice (12). CoB (CTLA4-Ig + anti-CD154) prolongs survival of allografts in FasL-deficient gld mice and in Fas-deficient lpr mice, suggesting that the Fas-FasL pathway is not necessary for tolerance induction (13–15). Moreover, Bim-deficient mice are sensitive to tolerance induction by CoB (CTLA4-Fc + anti-CD154), indicating that the intrinsic cell death pathway regulated by Bim is dispensable for peripheral tolerance induction (16).

In the current study, we investigated the effects of a combined deficiency of Fas and Bim on T cell apoptosis during early stages of viral infection and CoB-induced prolonged survival of skin allografts by generating mice lacking Bim and harboring the lpr mutation in Fas (Bcl2l11^−/−Fas^lpr/lpr). Mice lacking Bim and bearing the lpr mutation have a block in the contraction of Ag-specific T cells in chronic and certain acute viral infections, and they have dysregulated homeostatic proliferation and develop lymphadenopathy and autoimmunity (2, 7, 17, 18). Our studies show that Bcl2l11^−/−Fas^lpr/lpr mice are resistant to the early T cell attrition resulting from LCMV infection, and this resistance was due to inhibition of T cell apoptosis. Moreover, Bcl2l11^−/−Fas^lpr/lpr mice were resistant to prolonged survival of skin allografts induced by a CoB therapy consisting of donor-specific transfusion (DST)/anti-CD154 mAb. These results indicate that Fas and Bim both function to regulate T cell apoptosis in early stages of viral infections and during CoB.

C57BL/6J (H2^b), Bcl2l11^−/− (H2^b), Fas^lpr/lpr (H2^b), and BALB/c (H2^d) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Bcl2l11^−/− and Fas^lpr/lpr mice were back-crossed (10 generations) to the C57BL/6J strain in the animal facility of the University of Massachusetts Medical School and then interbred to generate the Bcl2l11^−/−Fas^lpr/lpr double knockout mouse line. Mice were given autoclaved food and housed in microisolator cages in a specific pathogen-free facility accredited by the American Association for Laboratory Animal Care, and all studies were approved by the University of Massachusetts Medical School Institutional Animal Care and Use Committee.

Spleens were recovered from mice and single cell suspensions were prepared. To remove contaminating erythrocytes, leukocyte preparations were treated with 0.84% ammonium chloride.

LCMV (strain Armstrong) was propagated in BHK cells as previously described (19). Mice were inoculated i.p. with 5 × 10⁴ PFU of LCMV.

Leukocytes were stained for CD8β (YTS156.7.7; BioLegend), CD4 (RM4-5; BD Pharmingen), CD44 (IM7; eBioscience), and IFN-γ (XMG1.2; BD Biosciences). For intracellular cytokine staining, splenocytes from recipient mice were incubated with LPS-matured, irradiated syngeneic (H2^b) or allogeneic (H2^d) stimulator cells and assessed for intracellular IFN-γ production as described previously (20). Samples were analyzed on an LSR II flow cytometer (Becton Dickinson).

For TUNEL staining, 1 × 10⁶ leukocytes were incubated in 48-well plates at 37°C directly ex vivo for 5 h in culture media. TUNEL assay (APO-DIRECT Kit; BD Biosciences) staining was carried out according to the manufacturer’s protocol.

Recipient mice of the specified strain were transplanted with complete MHC-mismatched skin and treated with a tolerizing regimen of donor-derived splenocytes (DST) and anti-CD154 mAb, as described previously (21). A DST of 1 × 10⁷ donor splenocytes was injected i.v. on day −7 relative to skin transplantation on day 0. Anti-CD154 mAb (0.5 mg, clone MR1; Bio X Cell) was given i.p. on days −7, −4, 0, and +4 relative to skin transplantation. Skin graft rejection was defined as the first day when the entire graft was necrotic.

Statistical analyses were performed using GraphPad PRISM software. Data significance (p values) was calculated using an unpaired Student t test. To compare three or more means, one-way ANOVA was used. All error bars represent the SEM. Allograft survival curves were plotted by the Kaplan–Meier method, and the incidence of graft rejection between groups was compared by χ² analysis (22). For all statistical analyses significance is defined as p < 0.05 and indicated as NS.

CD44^hi CD8 T cells undergo apoptotic attrition in the early phase of acute LCMV infection (3, 8). However, the apoptosis pathways mediating this early attrition have not been fully defined. Previous studies have shown that the absence of Fas signaling does not protect CD8 T cells from cell death after inflammation and that deletion of Bim is only partially protective (3, 9). To elucidate the apoptotic pathways involved in this deletion, wild type (WT) C57BL/6J and Bcl2l11^−/−Fas^lpr/lpr mice were infected with LCMV. CD44^hi CD8⁺ T cells underwent a significant reduction in percentages and numbers in the spleens of WT mice 3 d postinfection (Fig. 1A, 1B). Consistent with previous reports, untreated Bcl2l11^−/−Fas^lpr/lpr mice developed splenomegaly (data not shown) and had higher levels of memory phenotype CD44^hi CD8⁺ T cells in the spleen compared with untreated WT mice (7, 17, 18). However, these cells did not undergo attrition in response to LCMV infection (Fig. 1A, 1B). To determine whether this block in attrition of memory phenotype CD8 T cells was due to inhibition of apoptosis, we performed the TUNEL assay to measure DNA fragmentation. There was a significant increase in the frequency of TUNEL⁺ CD44^hi CD8⁺ T cells following LCMV infection of WT mice compared with the untreated controls, but not after LCMV infection of Bcl2l11^−/−Fas^lpr/lpr mice (Fig. 1C). These data indicated that CD44^hi CD8⁺ T cells from Bcl2l11^−/−Fas^lpr/lpr mice are resistant to apoptotic cell death.

Memory phenotype CD44^hi CD4⁺ T cells from WT mice showed a decline in percentages and numbers 3 d post-LCMV infection, although the decline was not as pronounced for CD4 T cells as for CD8 T cells (Fig. 2A, 2B). CD44^hi CD4⁺ T cells showed a high percentage of apoptotic TUNEL⁺ cells (Fig. 2C). CD44^hi CD4⁺ T cells from Bcl2l11^−/−Fas^lpr/lpr mice were resistant to apoptotic cell death and attrition in the spleen (Fig. 2A–C). Together, these data indicate that memory phenotype T cells depend on both Fas and Bim to undergo apoptotic deletion in the early phase of acute viral infection.

We have previously published that T cells deficient in Fas-FasL interactions or lacking Bim (Bcl2l11^−/− mice) have no or only partial defects in memory T cell attrition following LCMV infection or treatment with poly(I:C) (3, 8, 9, 23). These results indicate that Fas and Bim function concurrently to regulate T cell apoptosis in early stages of viral infections. One explanation for the importance of both the intrinsic and extrinsic cell death pathways in this system is that the early T cell deletion in viral infection relies predominantly on the mitochondrial pathway activated by Bim, with a smaller contribution by the death receptor pathway activated by Fas. In Fas- or FasL-deficient mice, the functional mitochondrial pathway activated by Bim in response to viral infection would delete T cells by apoptosis (24). In Bim-deficient mice, the mitochondrial pathway that is primarily responsible for the early T cell deletion is inactivated. However, cell death would occur to a limited extent by the death receptor pathway, and hence T cell attrition is partially blocked. Combined deficiency of Fas and Bim would inactivate both pathways of apoptosis, thereby completely inhibiting T cell deletion.

CoB including the combination of DST and anti-CD154 mAb significantly prolongs survival of fully mismatched skin, heart, bone marrow, and islet grafts in mice and nonhuman primates (25, 26). In addition, we have demonstrated that DST/anti-CD154 mAb treatment results in early deletion of alloreactive CD8 T cells (21). Previous studies have shown that CoB will significantly extend allograft survival in mice deficient in Bim or Fas (13, 16). We have also evaluated gene expression profiles of alloreactive CD8 T cells during DST/anti-CD154 mAb treatment using RNA-sequencing (data not shown). No change in levels of Bim mRNA were detected between the untreated and mice treated with DST/anti-CD154 mAb. However, levels of FasL mRNA increased significantly in the DST/anti-CD154 mAb (10.7-fold) treated groups as compared with untreated mice, consistent with our previous observations and suggesting a role of Fas-FasL interactions during apoptosis of T cells during CoB (27). RNA isolation, sequencing, and analysis were performed as previously described (28). The raw data for this study were deposited in the Gene Expression Omnibus (accession number GSE89030, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE89030).

To determine the apoptotic pathways involved in allograft survival induced by CoB, WT C57BL/6J, Bcl2l11^−/−, and Bcl2l11^−/−Fas^lpr/lpr mice were transplanted with MHC-mismatched BALB/c skins and treated with DST/anti-CD154 mAb. WT mice that were given no treatment rejected skin grafts rapidly, with a median survival time (MST) of 12 d (Fig. 3A), and DST/anti-CD154 mAb treatment showed prolonged allograft survival (Fig. 3A) (29). Bcl2l11^−/− mice that received no treatment rejected grafts with kinetics similar to those of WT mice (MST = 12 d). Interestingly, 10 out of 17 Bcl2l11^−/− mice that received DST/anti-CD154 mAb rejected grafts (MST = 13 d), suggesting that Bim is partially necessary for prolonged allograft survival. Skin allografts were rejected in Bcl2l11^−/−Fas^lpr/lpr mice that received no treatment, as expected (MST = 14 d) (Fig. 3A). Notably, the majority (10 out of 11 mice) of Bcl2l11^−/−Fas^lpr/lpr mice that were treated with DST/anti-CD154 mAb rejected their grafts with an MST of 15 d (Fig. 3A). Skin allograft survival in Bcl2l11^−/− mice that were treated with DST/anti-CD154 mAb was significantly increased as compared with the Bcl2l11^−/−Fas^lpr/lpr mice (p = 0.04), indicating that mice deficient in Bim alone are partially sensitive to CoB. To confirm previous observations that mice lacking Fas signaling are still sensitive to CoB-induced survival of allografts (13, 16), WT C57BL/6J and Fas^lpr/lpr mice were transplanted with MHC-mismatched BALB/c skin grafts and treated with DST/anti-CD154 mAb (Fig 3B). Both untreated WT and Fas^lpr/lpr mice rapidly rejected the skin allografts. Treatment with DST/anti-CD154 mAb significantly prolonged allograft survival in both WT and Fas^lpr/lpr mice (Fig 3B), demonstrating that elimination of Fas signaling alone does not abrogate survival of skin allografts following CoB. These results indicate that loss of Fas function in addition to that of Bim completely prevents the increased survival of skin allografts in mice treated with CoB.

We next investigated the ability of alloreactive CD8⁺ T cells in each group of mice to produce IFN-γ in response to syngeneic (C57BL/6J) or allogeneic (BALB/c) stimulation. Consistent with our previous observation, DST/anti-CD154 mAb CoB treatment significantly reduced the frequency of donor-reactive IFN-γ–producing effector/memory CD8⁺ T cells in WT mice, correlating with increased survival of skin allografts in these mice (Fig. 3C) (20). Donor-reactive IFN-γ–producing CD8⁺ T cell proportions were lowered in Bcl2l11^−/− mice treated with DST/anti-CD154 mAb despite the observation that 10 out of 17 of these mice rejected skin allografts. However, the total number of IFN-γ–producing CD8 T cells in the Bcl2l11^−/− mice treated with DST/anti-CD154 mAb was significantly higher as compared with the numbers in WT C57BL/6J mice (5.6 ± 0.6 × 10⁴ for Bcl2l11^−/− mice and 2.5 ± 0.5 × 10⁴ for WT mice, p = 0.003), which suggests that CD8 T cells in Bcl2l11^−/− mice do maintain a low level of functionality. High proportions of IFN-γ–producing CD8⁺ T cells were detectable in Bcl2l11^−/−Fas^lpr/lpr mice treated with DST/anti-CD154 mAb (Fig. 3C). Together, these data indicate that the deletion of alloreactive T cells by CoB requires both Fas and Bim.

Our data show that in a stringent BALB/c to B6 skin allograft model, Fas^lpr/lpr mice are sensitive to CoB-induced allograft survival, and Bcl2l11^−/− mice are partially resistant. However Bcl2l11^−/−Fas^lpr/lpr mice are completely resistant to prolonged allograft survival. We propose that the passive cell death pathway is the primary mechanism for the prolonged allograft survival, with a small contribution by the death receptor pathway. Recent studies have shown that the anti-CD154 mAb inhibits dendritic cell expression of inflammatory cytokines that are required for productive T cell activation and expansion (30). In Fas- or FasL-deficient mice, the passive death pathway activated by the lack of cytokines resulting from CD154 antagonism would mediate apoptosis of alloreactive T cells. In Bim-deficient mice, cell death would occur by activation-induced cell death or mitochondrial apoptosis by an alternate BH3 protein such as Puma, Noxa, or Bid, resulting in partial susceptibility to CoB. However, in mice lacking Fas and Bim, the inhibition of both extrinsic and intrinsic apoptosis completely prevents prolonged survival of allografts.

Overall, our results show that apoptosis of memory T cells in response to acute LCMV infection and death of alloantigen-specific T cells in response to alloantigen and CoB are mediated by both Fas and Bim pathways. Previous studies have shown that Fas and Bim have redundant roles in regulating cell death pathways that are important for controlling the development of autoimmunity and the contraction of virus-specific memory CD8 T cells (7, 17, 18). The results of our studies demonstrate that the Fas and Bim redundancy is also critical for cell death observed in the context of inflammation-induced attrition of memory T cells and the extension of allograft survival by CoB. Our results highlight the concurrent function of multiple apoptotic pathways in regulating immune responses during early viral infection and during CoB. Together, these findings suggest that collaboration between Bim and Fas is essential to regulate T cell homeostasis and cell death at several stages during the T cell life cycle (31, 32).

Michael Brehm is a consultant for The Jackson Laboratory. The other authors have no financial conflicts of interest.