Ionizing Radiation and Chemotherapeutic Drugs Induce Apoptosis in Lymphocytes in the Absence of FAS or Fadd/Mort1 Signaling: Implications for Cancer Therapy

Fas (also known as CD95 or APO-1) belongs to a subset of the TNFR family, whose members have a cytoplasmic ‘death domain.’ When Fas is engaged by multimeric Fas ligand (L), its death domain binds to the death domain in the cytoplasmic adaptor protein FADD/MORT1 (Fas-associated death domain protein). A further homotypic interaction occurs between the ‘death effector domain’ in FADD/MORT1 and a death effector domain in procaspase-8. The close proximity of caspase-8 zymogens facilitates their autocatalytic activation, and this appears to initiate the proteolytic cascade that leads to apoptosis 1. It has been proposed that Fas signaling plays a critical role in drug-induced apoptosis, stemming from the observation that neutralization of FasL prevented doxorubicin-induced apoptosis in T cell leukemia lines 2. Subsequent reports have differed as to whether impaired Fas signaling in tumor cell lines blocks 3 ,4 ,5 or has no effect 6 ,7 ,8 ,9 ,10 ,11 ,12 ,13 ,14 on drug-induced apoptosis. As mutations affecting the cell death machinery can contribute to tumor formation 15 ,16 ,17 ,18, these discrepancies may have arisen because the transformed cell lines studied harbor unknown mutations in genes that regulate cell death. To avoid such complications, we sought evidence of a role for Fas and FADD/MORT1 in gamma radiation– and drug-induced apoptosis using nontransformed cells. We used lymphocyte populations from genetically modified mice to determine if nontransformed cells that are resistant to FasL-induced apoptosis are also refractory to cytotoxic anticancer treatments.

All experiments with mice were approved by the Royal Melbourne Hospital Research Foundation Animal Ethics Committee. lck^pr-FADD-DN (dominant negative) strain 64 transgenic mice 19, Eμ-bcl-2-36 and Eμ-bcl-2-25 transgenic mice 20, Fas-deficient mutant lpr mice 21, and p53^−/− mice 22 have been described previously. All animals had a C57BL/6J genetic background. bcl-2 transgenic T cells were isolated from Eμ-bcl-2-25 mice, whereas bcl-2 transgenic B cells were isolated from Eμ-bcl-2-36 mice. Spleen B cells and lymph node T cells were purified by negative cell sorting in a FACSII™ or FACStar^PLUS™ sorter (Becton Dickinson). FITC-negative, propidium iodide (PI)-negative cells were selected after surface staining with a cocktail of FITC-conjugated mAbs. MI/70 anti–Mac-1, 8C5 anti–Gr-1, Ter119 antierythroid cell surface marker, RA3-6B2 anti-B220, 5.1 anti-IgM, and 11-26C anti-IgD antibodies were used to isolate T cells. MI/70 anti–Mac-1, 8C5 anti–Gr-1, Ter119 antierythroid cell surface marker, H129.19.6.8 anti-CD4, YTS 169 anti-CD8, and T24.31.2 anti–Thy-1 antibodies were used to isolate B cells.

Lymphocytes were cultured in the high-glucose version of DME supplemented with 13 μM folic acid, 250 μM l-asparagine, 50 μM 2-ME, and 10% FCS (Biosciences; starting concentration 10⁵ cells/ml). B cells were stimulated with 20 μg/ml LPS, 20 μg/ml F(ab′)₂ goat anti–mouse IgM (Jackson ImmunoResearch), 100 U/ml IL-2, and 100 U/ml IL-4 plus 100 U/ml IL-5. T cells were stimulated with immobilized KT3 rat anti–mouse CD3 and 37N51 hamster anti–mouse CD28 antibodies plus 100 U/ml IL-2. Tissue culture plates were coated in PBS that contained 10 μg/ml of each antibody. Resting and activated cells were treated with gamma radiation (2.5–10 Gy from a ⁶⁰Co source), 5–500 ng/ml doxorubicin (David Bull Labs.), 0.1–10 μg/ml etoposide, and 0.1–10 μg/ml cis-platin (Della West Pty Ltd.) or were cultured on Neuro-2A.FasL or Neuro-2A.neo cell monolayers 23. Activated cells were maintained in cytokines during treatment to minimize death due to growth factor withdrawal. Cell viability was determined by flow cytometric analysis of PI-stained cells in a FACScan™ (Becton Dickinson). To correct for doxorubicin fluorescence at higher doses, living and dead cells were also discriminated on the basis of their forward and side light-scattering properties.

To determine clonogenic survival, purified T cells were incubated for 8 h at 37°C in medium that contained 0, 0.1, 1, or 5 μg/ml cis-platin or etoposide. Cells were then washed free of drug and plated at limiting dilution with 2 ng/ml PMA and 0.1 μg/ml ionomycin plus 100 U/ml IL-2 in flat-bottomed 96-well plates. Three dilutions (1, 5, 10, 50, or 100 cells/well) were prepared for each treatment, and 96 wells were prepared for each dilution. Wells were scored microscopically as positive or negative for T cell growth after 7-d incubation at 37°C. The surviving fraction (S) after each treatment was calculated according to the equation:

where x is the fraction of wells negative for cell growth and y is the number of cells initially plated per well for treated (i) and untreated (u) cells 24. In other words, comparisons were made of the slopes of graphs plotting ln (fraction of wells negative for growth) versus the number of cells plated per well.

We have previously shown that T cells from Fas-deficient lpr mice 21 and FADD-DN transgenic mice that express a dominant interfering mutant of FADD/MORT1 19 are highly resistant to apoptosis induced by cross-linked soluble human FasL 19. Resistance was also observed when we cultured the cells with Neuro-2A neuroblastoma cells stably transfected with a mouse FasL expression construct ( Fig. 1A and Fig. B). These Neuro-2A.FasL cells express membrane-anchored FasL because they lack the matrix metalloproteinase responsible for FasL cleavage from the cell surface 23. In contrast, wild-type thymocytes and mature T cells underwent extensive apoptosis when cultured with Neuro-2A.FasL cells ( Fig. 1A and Fig. B). FasL triggered this apoptosis, because it was not observed in lymphocytes cultured with Neuro-2A.neo cells transfected with vector DNA (data not shown). Similarly, wild-type T cells activated in culture for 3 d with cross-linking antibodies to CD3 and CD28 or wild-type B cells activated with LPS plus anti-IgM antibody for 3 d were sensitive to FasL killing, whereas their lpr counterparts were resistant ( Fig. 1D and Fig. E). Wild-type resting B cells were more resistant than activated B cells to FasL killing ( Fig. 1 C), presumably due to very low surface expression of Fas (data not shown). The effect of FADD-DN on the sensitivity of B cells to killing by FasL could not be examined, because FADD-DN expression is limited to T cells in our transgenic mice. The proliferation defect of FADD-DN T cells prevented examination of activated FADD-DN T cells 19. As the anti-apoptotic protein Bcl-2 and the tumor suppressor p53 are known modulators of gamma radiation– and drug-induced apoptosis 22 ,24 ,25 ,26 ,27, we included in our analysis lymphocytes from bcl-2 transgenic mice 20 and p53-deficient mice 22. Neither p53 deficiency nor constitutive expression of Bcl-2 protected lymphocytes from apoptosis induced by FasL ( Fig. 1A–E).

To test sensitivity to a range of anticancer treatments, lymphocytes were exposed in culture to ionizing radiation (2.5–10 Gy) or the cytotoxic drugs cis-platin (0.1–10 μg/ml), doxorubicin (50–500 ng/ml), and etoposide (0.1–10 μg/ml). Cell survival was determined after 24 and 48 h. Apoptosis in cultures of wild-type, lpr, or FADD-DN transgenic lymphocytes occurred in a similar dose-dependent manner ( Fig. 1), and a comparison of the kinetics of cell killing revealed no difference between the three genotypes ( Fig. 2). In contrast, thymocytes and resting mature B and T cells from p53^−/− or bcl-2 transgenic mice were resistant to these apoptotic stimuli. For example, 2 d after a 10-Gy dose of gamma radiation or continuous culture in 1 μg/ml etoposide, >70% of all wild-type, lpr, and FADD-DN transgenic thymocytes were dead, whereas apoptosis in cultures of similarly treated p53^−/− or bcl-2 transgenic thymocytes did not exceed 20% ( Fig. 2 A). Interestingly, enforced Bcl-2 expression or p53 deficiency had relatively little impact on the death of gamma-irradiated or drug-treated cycling cells ( Fig. 1D–E, and Fig. 2D–E).

Short-term survival assays may only reflect differences in the rate of cell killing. For cancer treatment, the critical determinant of success is whether or not cells retain the capacity to proliferate in the long term. Therefore, we examined the capacity of drug-treated T cells for mitogen-induced proliferation. Resting mature T cells from wild-type, lpr, bcl-2 transgenic, and p53^−/− mice were cultured in varying concentrations of etoposide or cis-platin (0, 0.1, 1, and 5 μg/ml) for 8 h and then plated at limiting dilution in 96-well microtiter plates containing PMA and ionomycin plus recombinant IL-2. Plates were scored microscopically for T cell proliferation 7 d later. The results of these experiments are shown in Fig. 3. Clonogenic survival of wild-type cells decreased as the dose of drug increased. There was on average a 260-fold drop in the frequency of clonogenic cells after culture in 5 μg/ml cis-platin ( Fig. 3 A) and a 12-fold drop after culture in 5 μg/ml etoposide ( Fig. 3 B). lpr T cells treated with cis-platin or etoposide exhibited a similar drop in clonogenic survival ( Fig. 3). In comparison, enforced Bcl-2 expression or loss of p53 markedly enhanced clonogenic survival after exposure to cis-platin and etoposide. After treatment with 5 μg/ml cis-platin, clonogenic survival of bcl-2 transgenic and p53^−/− T cells was reduced 16- and 8-fold, respectively ( Fig. 3 A), and after treatment with 5 μg/ml etoposide, cells of both genotypes showed only a 1.8-fold reduction in clonogenic survival ( Fig. 3 B).

We conclude from our findings that signals transduced by Fas and/or FADD/MORT1 are dispensable for ionizing radiation– and cytotoxic drug–induced apoptosis in lymphocytes. Our results show that in resting lymphocytes, these anticancer treatments trigger apoptosis signaling pathways that can be regulated by Bcl-2 and p53. Significantly, neither of these proteins influenced Fas death signaling ( Fig. 1). These results and those of others 27 ,28 directly contradict a recent study suggesting that Bcl-2 confers drug resistance via inhibition of nuclear factor of activated T lymphocytes (NFAT)-induced FasL expression 3. Increased Fas expression in response to anticancer drugs was reported to be dependent on p53, but the issue of whether induction of the Fas gene is essential for p53-induced apoptosis was not addressed 29. Our data from lymphocytes, together with the finding that FADD/MORT1- 30 or caspase-8–deficient 31 mouse embryo fibroblasts are resistant to killing by FasL but normally sensitive to drug-induced apoptosis, indicates that Fas is not a common mediator of cell killing by cytotoxic agents. By contrast, certain mouse cell types lacking Apaf-1 or caspase-9 are resistant to both p53-induced apoptosis 32 and killing by ionizing radiation or cytotoxic drugs 33 ,34 ,35 ,36. These observations support a model in which ionizing radiation and cytotoxic drugs trigger apoptosis via p53 and Apaf-1–mediated activation of caspase-9, whereas apoptosis signaling by death receptors such as Fas is dispensable ( Fig. 4). Low levels of apoptosis in p53^−/− resting lymphocytes after drug treatment may be due to p53-related proteins such as p73. Recent studies implicate phosphorylation of p73 by the tyrosine kinase c-Abl as a crucial event in apoptosis induced by gamma radiation or cis-platin 37 ,38 ,39. However, the mechanism by which p73 triggers apoptosis remains to be established. Failure of enforced Bcl-2 expression or loss of p53 to perturb the death of activated B and T cells treated with ionizing radiation or cytotoxic drugs ( Fig. 1 and Fig. 2) may reflect reproductive/necrotic cell death 40. Thus, even though apoptosis signaling is disabled in these cells, their capacity for continued cell growth is limited by irreparable DNA damage.

We thank T. Jacks, A. Harris, S. Cory, and J. Adams for gifts of transgenic and knockout mice, K. Gray for animal husbandry, F. Battye, D. Kaminaris, V. Lapatis, and J. Parker for operating the cell sorters, and J. Birtles for editorial assistance. We are grateful to A. Harris, D. Huang, L. O'Connor, and D. Vaux for critical review of the manuscript.

K. Newton is supported by a Melbourne University Ph.D. scholarship, and A. Strasser is a Scholar of the Leukemia Society of America and a recipient of a Clinical Investigator Award from the Cancer Research Institute (New York). This work was supported by grants from the National Health and Medical Research Council (Canberra, Australia), the Dr. Josef Steiner Cancer Foundation (Bern, Switzerland), and the Anti-Cancer Council of Victoria (Melbourne, Australia).