Fas Ligand-Mediated Lysis of Self Bystander Targets by Human Papillomavirus-Specific CD8⁺ Cytotoxic T Lymphocytes

Mice.

C57BL/6.P⁰ (B6.P⁰) mice derived as described previously (11) were obtained from Gary Kupariah, John Curtin School of Medical Research, Canberra, Australia. C57BL/6 ICAM-1-deficient (B6.ICAM⁰) mice (32) were obtained from The Jackson Laboratories, Bar Harbor, Maine. These strains and B6 β2μ-deficient (B6.β2μ⁰) mice were maintained and bred at the Austin Research Institute Biological Research Laboratories. C3H/HeJ, BALB/c, and B6 mice were purchased from The Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia.

Cells and reagents.

The mouse EL4 (H-2^b) thymoma, RMA (H-2^b) lymphoma, and RMA-S (H-2^b) mutant lymphoma (derived from the Raucher virus-induced murine cell line RBL-5 and defective for peptide loading of MHC class I molecules) (18) cell lines and an E7-transfected clone of RMA, RMA-E7 (43), were grown in RPMI medium supplemented with 10% (vol/vol) fetal calf serum (FCS), 2 mM glutamine, 100 U of penicillin/ml, and 100 μg of streptomycin (Gibco, Grand Island, N.Y.)/ml. d12S, the 12th serial subcloning of the rat × mouse hybridoma PC60 cell line (3), was grown in Dulbecco’s modified Eagle’s medium (Gibco) supplemented with 5% FCS and additives as described above. Phorbol myristate acetate (PMA) (Sigma Chemical Co., St. Louis, Mo.) and ionomycin (ION) (Calbiochem Corp., San Diego, Calif.) were purchased. Soluble recombinant human FasL (srFasL) (produced in COS cell supernatants and used neat), mouse Fas-Fc, and human p80 TNFR-Fc fusion proteins were a kind gift from David Lynch, Immunex Corp., Seattle, Wash. Recombinant human interleukin-2 (IL-2) was a kind gift from Chiron Corp., Emeryville, Calif. Peptides of human papillomavirus type 16 protein E7^49-57 (RAHYNIVTF) and chicken OVA^257-264 (SIINFEKL) were synthesized (>98% pure) on an Applied Biosystems model 430A automated peptide synthesizer (12). d12S cells were incubated for 3 h before the cytotoxicity tests with a mixture of PMA (5 ng/ml) and ION (1 μg/ml). Spleen cells were harvested from B6, B6.ICAM⁰, C3H/HeJ (H-2^k), and BALB/c (H-2^d) mice and cultured in RPMI medium supplemented with the additives described above and ION (1 μg/ml) and PMA (5 ng/ml) for 3 days. Spleen blasts were then washed three times in medium prior to their use as target cells. Spleen blasts were >98% CD3⁺ as determined by flow cytometry.

Peptide immunization and induction of B6 and B6.P⁰ anti-E7^49-57 CTL.

One hundred micrograms (100 μl of 1 mg/ml peptide solution) of peptide was extensively mixed with 100 μl of immunofluorescent antibody and 0.5% bovine serum albumin. The 200-μl mixture was injected subcutaneously into B6 or B6.P⁰ mice, and the procedure was repeated after 2 weeks. Two weeks after the second immunization the mice were sacrificed and spleen cells were cultured overnight on plastic flasks to remove adherent cells. Responder cells were then treated with anti-CD4 (H129.19; rat immunoglobulin G2a [IgG2a]) (Sigma) and complement (1/30 dilution of normal rabbit serum) as previously described (36). Stimulator RMA-S cells were cultured at 25°C for 24 h and incubated with 100 μM E7^49-57 peptide for 2 to 4 h at 33°C. Stimulator cells (10⁶) were extensively washed, irradiated (20,000 rad), and then cultured (25-cm² culture flasks) with responder cells (2 × 10⁷) in RPMI supplemented with 10% FCS and 5 μM peptide. After 5 days, responder lymphocytes were harvested and cultured in 24-well plates at a density of 10⁶ cells per well in the same medium supplemented with 5 U of IL-2/ml. Each well received 10⁶ irradiated RMA-E7 cells as stimulator cells. Responder cells were restimulated weekly under the same conditions; they were >90% CD8⁺ as determined by immunofluorescence assay, and their CTL activity was determined in a 4-h ⁵¹Cr release assay with RMA-E7 target cells.

⁵¹Cr release assays.

The cytotoxicity of responding B6 anti-E7^49-57 or B6.P⁰ anti-E7^49-57 CTL, d12S cells, or srFasL was assessed by ⁵¹Cr release assays for 4 h as described previously (30). Briefly, Ag-bearing target cells (including temperature-induced RMA-S cells that were pulse labeled with immunizing or control peptide, as described above) or Ag-free bystander cells were labeled with 50 μCi of ⁵¹Cr for 1 h at 37°C and then washed three times. ⁵¹Cr-labeled Ag-bearing target cells (10⁴) and unlabeled bystander cells (10⁴) were incubated with different numbers of CTL as indicated, according to the effector/target cell ratio. In other wells, ⁵¹Cr-labeled Ag-free cells and unlabeled Ag-bearing target cells (also at a 1:1 ratio) were added to CTL. The spontaneous release of ⁵¹Cr was determined by incubating the target cells with medium alone, whereas the maximum release was determined by adding sodium dodecyl sulfate to a final concentration of 5%. The percent specific lysis was calculated as follows: 100 × [(experimental release − spontaneous release)/(maximum release − spontaneous release)]. In some experiments, B6 or B6.P⁰ anti-E7^49-57 CTL were incubated for 3 h before the cytotoxicity tests with a mixture of PMA (5 ng/ml) and ION (1 μg/ml). These cells were washed three times prior to the assay. In other experiments, B6 or B6.P⁰ CTL were depleted with monoclonal antibody (MAb) (10 μg/ml; anti-CD8 [1803; rat IgG2a] or control anti-CD4 and complement [1/30 dilution of normal rabbit serum]) prior to the examination of cytolysis against RMA-E7 and bystander syngeneic target cells. Cytotoxicity tests were sometimes performed either in the absence or presence of Fas-Fc or TNFR-Fc (at a final concentration of 5 μg/ml) and/or EGTA (Sigma) (at a final concentration of 2.5 mM). In other tests MAb (anti-LFA-1-α [M17/4; rat IgG2a] [Pharmingen, San Diego, Calif.], anti-CD8 [53-6.7; rat IgG2a] [Sigma], or anti-Thy 1.1 [OX-7; mouse IgG₁]) was added at a final concentration of 20 μg/ml prior to, or 1 h after, the commencement of the cytotoxicity assay.

Generation of E7^49-57-specific B6 and B6.P⁰ CD8⁺ CTL.

Following immunization of B6 and B6.P⁰ mice with E7^49-57 peptide, harvested spleen cells were depleted of CD4⁺ T cells and responding CTL were expanded by several rounds of stimulation with irradiated RMA-E7 cells. Resultant B6 and B6.P⁰ CTL cultures were assayed for specific cytolytic activity towards RMA-E7 or RMA-S cells pulsed with the E7^49-57 (RAHYNIVTF) peptide in a 4-h ⁵¹Cr release assay (Fig. 1). RMA-S cells have a defect affecting peptide loading of MHC class I molecules, and as a consequence, the cell surface expression of these molecules is low but increases when K^b- or D^b-binding peptides are added to the culture medium (18). Both B6 and B6.P⁰ CTL cultures effectively lysed RMA-E7 or RMA-S cells pulsed with E7^49-57 but not RMA-S cells pulsed with OVA^257-264 or RMA-S cells alone. From several independent immunizations and in vitro cultures, peptide-specific lysis by B6 CTL was consistently greater than that by B6.P⁰ CTL. Pretreatment of these cultures with anti-CD8 MAb and complement, but not with control anti-CD4 MAb and complement, completely abrogated B6 anti-E7^49-57 or B6.P⁰ anti-E7^49-57 CTL lysis of RMA-E7 (Fig. 1). Overall, these data indicate that the cytotoxic effects studied were mediated by CD8⁺ CTL and were specific for E7^49-57 peptide.

B6 and B6.P⁰ CD8⁺ CTL specific for E7^49-57 mediate bystander lysis.

Figure 2 demonstrates that E7^49-57-specific CTL from B6 or B6.P⁰ mice could efficiently lyse Ag-free noncognate targets when the latter were bystanders to lysis of Ag-bearing RMA-E7 target cells. In particular, syngeneic (H-2^b) B6 blasts and, to a lesser extent, RMA cells acted as susceptible bystanders, while syngeneic (H-2^b) EL4 cells did not (Fig. 2A). These data correlated with the relative Fas sensitivity of these bystander cells (B6 blasts were more sensitive to FasL than RMA cells, and EL4 cells were insensitive [see Fig. 5]). The levels of bystander lysis mediated by B6 and B6.P⁰ CTL specific for E7^49-57 were approximately equivalent. B6 anti-E7^49-57 or B6.P⁰ anti-E7^49-57 CTL lysed RMA-E7 cells irrespective of the type of bystander cell present (Fig. 2B).

Bystander lysis by E7^49-57-specific B6 or B6.P⁰ CD8⁺ CTL is FasL mediated.

The cytolytic mechanisms employed by B6 and B6.P⁰ CD8⁺ CTL specific for E7^49-57 were determined by the addition of an inhibitory Fas-Fc fusion protein and/or the Ca²⁺ chelator, EGTA (to block granule-mediated lysis). Direct lysis of RMA-E7 target cells by B6 CTL specific for E7^49-57 was not inhibited by Fas-Fc, was significantly inhibited by EGTA, but was completely inhibited by a combination of Fas-Fc and EGTA (Fig. 3A). B6 CTL inhibited from lysing RMA-E7 targets (by the addition of EGTA) still efficiently killed bystander B6 blasts. Thus, B6 CTL specific for E7^49-57 lysed RMA-E7 target cells preferentially via granule exocytosis, but in the absence of granule exocytosis by FasL. By contrast, Fas-Fc alone completely blocked B6.P⁰ anti-E7^49-57-mediated lysis of RMA-E7, and EGTA was without effect (Fig. 3B), suggesting that these P⁰ CTL mediated Ag-specific lysis via FasL. A potential nonspecific inhibition of lysis by the Fas-Fc fusion protein was unlikely given that a control TNFR-Fc fusion protein at the same concentration was ineffective. Bystander lysis of B6 blasts in the presence of RMA-E7, mediated by either B6 (Fig. 3C) or B6.P⁰ (Fig. 3D) CTL specific for E7^49-57, was inhibited by Fas-Fc alone but not by TNFR-Fc or EGTA.

Mode of CTL activation determines quantity of bystander lysis.

A previous report by Kojima et al. highlighted the fact that maximal bystander lysis was mediated by CTL clones that were stimulated by PMA and ION, thus bypassing TCR stimulation (13). In this scenario, interactions between activated CTL and bystanders could be observed in the absence of third-party, Ag-bearing stimulator cells. These experiments were also of particular interest, since PMA-ION-stimulated CTL of H-2^d origin lysed H-2^b bystander cells. By contrast, we had previously examined bystander lysis only of syngeneic (H-2^b) cells by alloreactive CTL (b anti-k) in the presence of H-2^k target cells (34). To further investigate bystander lysis following CTL activation, we stimulated B6.P⁰ anti-E7^49-57 CTL with PMA-ION and/or Ag-bearing RMA-E7 cells and evaluated the ability of these CTL to lyse bystander blasts of H-2^b, H-2^k, and H-2^d origin (Fig. 4A to D). B6.P⁰ anti-E7^49-57 CTL that were activated with RMA-E7 cells lysed labeled syngeneic bystander B6 blasts but did not significantly lyse H-2^k C3H/HeJ blasts, H-2^d BALB/c blasts, B6.β2μ⁰ blasts, or FasL-insensitive EL4 cells (Fig. 4A). None of these bystander cells were lysed by B6.P⁰ anti-E7^49-57 CTL in the absence of RMA-E7 cells (Fig. 4C). By contrast, PMA-ION activation of B6.P⁰ anti-E7^49-57 CTL stimulated equivalent lysis of all blasts of the three H-2 haplotypes, but not EL4, in the absence of RMA-E7 stimulator cells (Fig. 4D). Only bystander lysis of B6 blasts was greater if RMA-E7 stimulator cells were included in the assay (Fig. 4B). This PMA-ION-activated bystander lysis was also completely inhibited by Fas-Fc fusion protein (data not shown). Importantly, blasts from B6, B6.β2μ⁰, BALB/c, and C3H/HeJ mice were equally sensitive to srFasL, suggesting that Fas sensitivity did not explain differential susceptibility to B6.P⁰ anti-E7^49-57 CTL-mediated bystander lysis (Fig. 4E). It should be noted that preliminary data (not shown) also suggest that alloreactive mouse CTL (b anti-k) do not effectively lyse H-2^d BALB/c blasts acting as bystanders. Further experiments with B6 anti-E7^49-57 CTL were performed, and their mode of bystander lysis was identical to that of B6.P⁰ anti-E7^49-57 CTL. In particular, perforin-mediated bystander lysis of allogeneic blasts was not observed, even after pretreatment of B6 anti-E7^49-57 CTL with PMA-ION (data not shown).

Bystander lysis by E7^49-57-specific CTL is not via srFasL.

We considered the possibility that differences in TCR versus PMA-ION stimulation of B6.P⁰ anti-E7^49-57 CTL might be due to bystander killing by soluble FasL (sFasL). CTL supernatants were collected after incubation of B6 anti-E7^49-57 or B6.P⁰ anti-E7^49-57 CTL with either RMA-E7 cells or PMA-ION for 3 h. These supernatants were used as media for incubation with ⁵¹Cr-labeled syngeneic target cells (Fas sensitive or insensitive). Figure 5 shows that no supernatant had lytic activity toward Fas-sensitive B6 blasts or RMA cells, while sFasL significantly lysed B6 blasts and RMA cells but not EL4 cells. These data indicate that there is no evidence that srFasL is responsible for effects of CTL on bystanders irrespective of the CTL stimulus.

CTL LFA-1–bystander ICAM-1 interaction is critical for bystander lysis.

The contact dependence of bystander lysis was investigated with MAbs to surface proteins on the B6.P⁰ anti-E7^49-57 CTL. Previous reports have highlighted the importance of an LFA-1–ICAM-1 interaction for Th-1-mediated B-cell apoptosis (40). MAbs were present throughout the assay or were added after 1 h of preactivation of the CTL with RMA-E7 stimulator cells. As previously demonstrated for many forms of Ag-specific lysis (15, 16), the anti-LFA-1 MAb or the anti-CD8 MAb completely inhibited RMA-E7 cell lysis when present throughout an effector-target cell assay (data not shown). By contrast, after 1 h of incubation between B6.P⁰ anti-E7^49-57 CTL and ⁵¹Cr-labeled RMA-E7 cells, the addition of anti-LFA-1, anti-CD8, or negative control anti-Thy 1.1 MAb did not block E7^49-57-specific lysis (Fig. 6A). These data suggested that the important interactions involving the TCR-CD8 of CTL and class I molecules-peptide of the RMA-E7 target cells were required for activation of CTL-mediated lysis. Not surprisingly therefore, when present throughout the assay, the anti-LFA-1 and anti-CD8 MAbs could block B6.P⁰ anti-E7^49-57 CTL–RMA-E7 interactions, required for TCR-induced FasL expression on CTL, and thus subsequent bystander lysis of labeled B6 blasts in an effector-target-bystander assay (Fig. 6B). However, the simultaneous addition of ⁵¹Cr-labeled bystander B6 blasts and MAb to CTL preactivated by RMA-E7 cells 1 h before the assay made it possible to determine the requirements for the delivery of a lethal hit to bystanders by activated B6.P⁰ anti-E7^49-57 CTL (Fig. 6C). Anti-LFA-1 MAb completely blocked the lysis of bystander B6 blasts when added 1 h after CTL activation (Fig. 6C). The ability of anti-CD8 MAb to inhibit bystander lysis was considerably diminished when it was added 1 h after the exposure of CTL to RMA-E7 cells (Fig. 6C).

Lysis of bystander B6 blasts by PMA-ION-preactivated B6.P⁰ anti-E7^49-57 CTL was also completely inhibited by anti-LFA-1 MAb but not by anti-CD8 MAb, further confirming a role for LFA-1 in interactions between activated B6.P⁰ anti-E7^49-57 CTL and bystander B6 blasts (data not shown). Pretreatment of PMA-ION-preactivated B6.P⁰ anti-E7^49-57 CTL with anti-LFA-1 MAb resulted in a complete inhibition of bystander lysis, while pretreatment of bystanders with anti-LFA-1 MAb had no effect (data not shown). Thus, the inhibitory effects of anti-LFA-1 MAb on bystander lysis were not due to the inhibition of CTL–RMA-E7 cell interactions, required for CTL FasL upregulation (38) or LFA-1 activation (7, 21). Only preactivated CTL, as opposed to resting CTL, used their surface LFA-1 molecules to facilitate the effective formation of CTL-bystander conjugates, and after activation, CTL no longer required the signals from Ag-bearing target cells to kill bystander cells. Importantly, experiments with the anti-CD8 MAb demonstrate that the CD8-TCR-dependent activation of CTL is critical for bystander lysis. These data may indicate important changes in the affinity of CTL surface LFA-1 (7, 21) that enabled conjugation with Ag-free bystander cells. Thus, while the expression of FasL and LFA-1 on TCR-activated CTL was necessary, neither alone was sufficient for bystander lysis.

Confirmation that bystander cell ICAM-1 was interacting with CTL LFA-1 was obtained by using ⁵¹Cr-labeled blasts from B6.ICAM-1⁰ mice (Fig. 6D). ICAM-1-deficient B6 blasts were not susceptible bystanders, while control ICAM-1-expressing B6 blasts were efficiently lysed by B6.P⁰ anti-E7^49-57 CTL. Importantly, B6 and B6.ICAM-1⁰ blasts were equally susceptible to d12S effector cells that kill in a non-MHC-restricted FasL-dependent manner (Fig. 6D). Furthermore, ICAM-1 expression on B6 blasts but not on B6.ICAM⁰ blasts was confirmed by flow cytometry (data not shown). These results and the data shown in Fig. 6A to C strongly support a critical interaction between activated CTL LFA-1 and bystander ICAM-1. This critical interaction between CTL LFA-1 and bystander ICAM-1 was also observed for B6 anti-E7^49-57 CTL (data not shown).

Conclusions.

This efficient and rapid lysis of Fas-sensitive bystanders by E7^49-57 peptide-specific CD8⁺ CTL extends previously published observations of bystander lysis (2, 6, 13, 34, 41). However, the issue of the variable sensitivity of bystander cells to activated Ag-specific CTL remains somewhat unresolved. In our experimental system we have demonstrated that Ag-specific CTL will lyse syngeneic bystander cells preferentially to third-party allogeneic bystander cells, providing the stimulus is TCR mediated and the bystander expresses class I molecules. If PMA-ION is used to activate Ag-specific CTL they will lyse Fas-sensitive bystanders irrespective of MHC class I molecule expression. Our findings are consistent with an earlier study by Duke that defined self-recognition of bystander target cells by using allospecific CTL (6). However, in contrast, Kojima et al. (13) have shown that allospecific CTL of H-2^b origin and specific for H-2^k can lyse L1210-Fas bystander cells (H-2^d). The discrepancy between the data of Kojima et al. and ours is not simply explained by CTL activation, since these allospecific CTL were not apparently stimulated with PMA-ION (13). Nor is it explained by the lack of ICAM-1 expression on our blast populations, as spleen blasts from B6, BALB/c, and C3H/HeJ mice all expressed similar levels of ICAM-1 (data not shown). The ability of CTL to lyse bystander cells is dependent on important cell-cell interactions and the signals these interactions mediate, and thus the most important interactions between activated CTL and bystander cells of different lineages and activation states may need to be defined to resolve these differences.

In the future, activation of Ag-specific CTL and/or T-cell bystander cells by other T-cell stimuli, such as lectins (e.g., con A), anti-CD2 or anti-CD3 cross-linking, cytokines (e.g., IL-2), or superantigens, will be of interest. Maintenance of clonotype specificity in Fas-mediated apoptosis of mature T cells has been hypothesized to result from important TCR sensitization signals at the time of Fas engagement (9). In these studies, neither resting nor con A-stimulated T-cell blasts acted as bystanders to Ag-induced T-cell death, suggesting that the sensitization event necessary for bystander T-cell death was very specific. By contrast, in our assays, it appeared that the sensitization was provided to bystander T cells after prestimulation with PMA-ION. Arguably, such a powerful stimulus to bystander T cells is not physiologically relevant, but if it is, then these data raise the possibility that T cells of different clonotypes may be destroyed in vigorous or chronic Ag-specific CTL responses. In particular, cytokines such as IL-2 and gamma interferon are important in predisposing T cells to TCR-induced death (4, 17) or creating an environment favorable for Fas-mediated apoptosis (2, 4, 17, 23, 25). Thus, these cytokines may play an important general role in regulating bystander lysis and, consequently, immune responses and the resolution of disease or tumor burden.

The role that CTL-mediated bystander lysis plays in outcomes following virus infection remains to be determined; however, it is becoming evident that many viruses initially enhance the susceptibility of newly infected cells to Fas-mediated apoptosis (31), and thus CTL-mediated bystander lysis of Ag-free infected cells might play an important protective role in this regard. Alternatively, bystander lysis could play a potentially damaging role in chronic infections that involve tissues that can express high levels of Fas (e.g., liver and heart tissues) (2, 24, 29). It is now of paramount importance that bystander lysis be demonstrated convincingly in vivo. In this light, murine models of viral infection or tumor rejection that assess the role of CTL FasL in direct versus bystander cytotoxicity must be developed.