Restimulation of previously activated T cells via the T-cell receptor (TCR) leads to activation-induced cell death (AICD), which is, at least in part, dependent on the death receptor CD95 (APO-1, FAS) and its natural ligand (CD95L). Here, we characterize cutaneous T-cell lymphoma (CTCL) cells (CTCL tumor cell lines and primary CTCL tumor cells from CTCL patients) as AICD resistant. We show that CTCL cells have elevated levels of the CD95-inhibitory protein cFLIP. However, cFLIP is not responsible for CTCL AICD resistance. Instead, our data suggest that reduced TCR-proximal signaling in CTCL cells is responsible for the observed AICD resistance. CTCL cells exhibit no PLC-γ1 activity, resulting in an impaired Ca2+release and reduced generation of reactive oxygen species upon TCR stimulation. Ca2+ and ROS production are crucial for up-regulation of CD95L and reconstitution of both signals resulted in AICD sensitivity of CTCL cells. In accordance with these data, CTCL tumor cells from patients with Sézary syndrome do not up-regulate CD95L upon TCR-stimulation and are therefore resistant to AICD. These results show a novel mechanism of AICD resistance in CTCL that could have future therapeutic implications to overcome apoptosis resistance in CTCL patients. [Cancer Res 2009;69(10):4175–83]
- cutaneous T-cell lymphoma (CTCL)
- activation-induced cell death (AICD)
- apoptosis resistance
- CD95 ligand (CD95L)
Cutaneous T-cell lymphomas (CTCL) are a heterogeneous group of slowly progressive diseases characterized by skin homing clonal CD4+ T lymphocytes. Sézary syndrome is the leukemic variant of CTCL ( 1, 2). Apoptosis resistance rather than proliferation is implicated in the pathogenesis ( 3, 4).
Apoptosis resistance plays a major role in the pathogenesis and therapy of many cancers ( 5). T-cell receptor (TCR) restimulation induces activation-induced cell death (AICD) in T cells ( 6). Removal of activated and expanded T cells by AICD in vivo can be mimicked by an in vitro model system using activated T cells cultured in vitro with interleukin (IL)-2 and restimulated through the TCR or by phorbol 12-myristate 13-acetate (PMA) and ionomycin ( 7). Stimulation of the TCR-complex by antigen or agonistic antibodies leads to production of death receptor ligands including the CD95 ligand (CD95L; ref. 8). After ligation of CD95 (APO-1/Fas) with CD95L, the death-inducing signaling complex is formed. Within the death-inducing signaling complex, caspase-8 is cleaved and activates effector caspases such as caspase-3, leading to apoptosis that can be blocked by the inhibitory protein cFLIP. Short-term activated primary T cells are resistant to CD95-mediated AICD due to up-regulated cFLIP levels. Long-term activated T cells, in contrast, down-regulate cFLIP and are sensitized to AICD ( 9).
Here, we show that CTCL tumor cells are highly resistant to AICD. We found absent PLCγ-1 activation and subsequent failure to release intracellular Ca2+ and reactive oxygen species (ROS) in CTCL tumor cells. Both signals are required for CD95L production ( 10). Consistent with these data, we could show that the majority of CTCL cell lines and primary tumor cells from Sézary patients lack TCR-induced production of CD95L. We suggest that the reduced TCR signaling in CTCL tumor cells is sufficient to up-regulate cFLIP but fails to induce CD95L expression, which is causative for AICD resistance in CTCL. Thus, this newly identified defect of AICD in CTCL cells may offer new perspectives for the treatment of CTCL.
Materials and Methods
Cell lines and preparation of primary T cells from healthy donors and Sézary patients. CTCL cell lines HH, HUT78, MyLa, and SeAx and the T-cell line Jurkat J16-145 were cultured as previously described ( 11, 12). The ethics committee II of the Ruprecht-Karls-University of Heidelberg approved all described studies. Experiments were performed according to the Declaration of Helsinki guidelines. CD4+ CTCL cells were isolated by magnetic beads. Isolation, activation, and cultivation of primary human T cells were described previously ( 11).
Antibodies and reagents. The antibodies used were as follows: α-CD3 (OKT3), α-CD4, α-CD28, isotype control antibody (all BD Biosciences), mouse monoclonal antibodies α-CD95 (α-Apo-1), and recombinant CD95L as previously described ( 9). Annexin V was obtained from Immunotools; PMA and Ionomycin were purchased from Sigma; Streptavidin-APC and biotinylated α-CD95L (NOK1) were obtained from Becton-Dickenson; the APC-specific FASER kit and magnetic beads (T-cell isolation kit II) from Miltenyi; TRIzol from Invitrogen; the MuLV reverse transcriptase from Applied Biosystems; and antibodies for Western blot analysis of p-c-Jun-NH2-kinase (JNK), JNK, p-extracellular signal-regulated kinase (ERK), ERK, p-PLCγ1, and PLCγ1 were purchased from Cell Signaling; antibodies to cFLIP (NF6), Actin (Sigma), and Tubulin (Santa Cruz) were used as previously described ( 11). Flow cytometry was performed with a FACSCanto II or a FACScan (Becton-Dickenson), and the data were analyzed with FACSDiva- and CellQuest-Software (Becton-Dickenson).
Apoptosis assays. For TCR-mediated apoptosis induction, cells were seeded in triplicates on culture plates precoated with the indicated concentrations of α-CD3. Or apoptosis was induced by addition of PMA/Ionomycin, α-CD95, or CD95L to the cells. Cell death of unstained cells was quantified by FSC/SSC analysis ( 13) or by staining with 2.5 μg/mL propidium iodide and/or Annexin V and analyzed by flow cytometry. Specific apoptosis was calculated as (percentage of induced apoptosis − percentage of spontaneous apoptosis)/(100 − percentage of spontaneous apoptosis) × 100.
Flow cytometry. Flow cytometry was performed on a FACSCanto II or a FACScan with at least 10.000 cells counted. Fluorescence-activated cell sorting (FACS) data were analyzed with FACSDiva- and CellQuest-Software. For CD3 surface expression measurement, cells were incubated with cychrome-labeled α-CD3 for 15 min. For CD95 surface expression, cells were either incubated for 15 min with FITC-labeled or biotinylated α-CD95 (α-APO1, own production) or isotype control antibody. After washing, cells previously incubated with biotinylated α-CD95 were incubated for 30 min with Streptavidin-APC. For CD95L surface expression measurement, cells were either unstimulated or incubated with 30μg/mL plate-bound α-CD3 for 20 h. Cells were incubated for 1 h with biotinylated α-CD95L or isotype control antibody. After washing, cells were incubated for 30 min with Streptavidin-APC. After washing, cells were incubated at 4°C with the APC-specific FASER kit according to the manufacturer's protocol. Measurements of Ca2+ release and ROS production were performed as previously described ( 10). Briefly, to detect Ca2+, cells were stained with 1 μmol/L Fluo-4-AM and cross-linked with 1μg/mL goat α-mouse antibody for 30 min. Thereafter, cells were stimulated and Ca2+_influx into the cytosol was measured by flow cytometry. To determine ROS levels, cells were stained with DCFDA (5 μmol/L) for 30 min. Experimental data are presented as “Increase in Mean Fluorescence Intensity (MFI)” (%), calculated according to the formula: “Increase in MFI” (%) = (MFI (αCD3 or PMA-stimulated cells ± Inhibitor) − MFI (untreated cells or ± Inhibitor)). One hundred percent value was defined as a difference between MFI (αCD3 or PMA-stimulated cells) and MFI (untreated cells).
Reverse transcription PCR and quantitative reverse transcription-PCR for CD95L and cFLIP. T cells were stimulated with 30μg/mL plate-bound α-CD3 for the indicated time. Cells (4 × 106) were lysed in TRIzol, and RNA was isolated according to the manufacturer's instructions. Fifty nanograms of RNA were reverse transcribed into cDNA using MuLV reverse transcriptase followed by quantitative PCR for cFLIP and CD95L as previously described ( 14).
Stable siRNA expression. We used stable expression of siRNA against c-FLIP as recently published ( 15– 17). To obtain optimal knockdown efficiencies, pRS-derived retroviruses were produced using a high titer–producing FLYRD18 producer cell line ( 18). Retrovirus containing supernatant was then used to infect HUT78 or Jurkat cells. Infected cells were selected by treatment with puromycin. Infection efficiency and knockdown of cFLIP was monitored by GFP expression as well as Western blotting.
Immunoprecipitation and Western blotting. Immunoprecipitation and Western blotting have been described previously ( 11).
CTCL cells are resistant to AICD. We have used an ex vivo model to investigate the status of TCR-induced AICD of CTCL cells. Upon activation and subsequent culture of T cells in IL-2–containing medium, cells switch from an AICD-resistant (d1) to an AICD-sensitive phenotype (d6; ref. 9). The CTCL cell lines HH, HUT78, MyLa, and SeAx were derived from CTCL patients and are used as a CTCL model system. CTCL cell lines showed varying degrees of TCR expression ( Fig. 1A ). However, all CTCL cell lines were resistant to TCR-induced AICD ( Fig. 1B). To obtain CTCL tumor cells without antibodies binding to the TCR (e.g., Vβ chain antibodies, which could lead to stimulation), we isolated CD4+ lymphocytes from Sézary patients (Supplementary Table S1) by negative magnetic bead isolation. T cells from patients and healthy donors showed similar TCR expression levels as established by CD3 surface staining ( Fig. 1C). Interestingly, despite similar expression levels of CD3, CD4+ cells from Sézary patients showed reduced sensitivity to AICD on day 6 in contrast to healthy controls ( Fig. 1D).
CTCL cells are sensitive to CD95-induced apoptosis. After TCR stimulation, AICD is induced, e.g., via the CD95-dependent pathway ( 6). Blocking of the CD95L leads to AICD resistance in T cells ( 19). To exclude a defect in the CD95-induced apoptosis pathway, we triggered CD95 on CTCL cells by exogenous addition of LZ-CD95L. All four investigated CTCL cell lines as well as primary tumor cells showed high expression levels of CD95 ( Fig. 2A and C ). Treating these cells with LZ-CD95L induced apoptosis in all CTCL cell lines tested ( Fig. 2B). Similar levels of LZ-CD95L– or α-APO-1–induced specific apoptosis were seen in primary T cells from Sézary patients as well as with cells from healthy controls ( Fig. 2D). In contrast to TCR-induced apoptosis, these experiments show that CTCL cells are sensitive to CD95-mediated apoptosis when this pathway is triggered by exogenous CD95L or α-APO-1 and exclude a defect within the CD95 signaling cascade.
cFLIP up-regulation is not responsible for AICD resistance of CTCL cells. Next, we studied the role of potential inhibitors of CD95-mediated apoptosis, which might explain the observed AICD resistance of CTCL cells. Members of the cFLIP protein family were shown to inhibit CD95-mediated apoptosis by interfering with caspase-8 activation ( 9). It is known that cFLIP is an important regulator of AICD as its degree of TCR-induced up-regulation determines AICD resistance versus sensitivity ( 9). In line with the AICD-resistant phenotype, we detected increased basal levels of cFLIP protein in CTCL cell lines and primary CTCL cells ( Fig. 3A and B ). However, comparing T cells from Sézary patients to control T cells and the CTCL cell line HUT78 to control Jurkat cells ( Fig. 3C), we found that TCR-induced up-regulation of cFLIP was comparable. To further analyze a functional role of cFLIP in AICD, we performed a stable knockdown of cFLIP by shRNA expression and tested these cells for AICD sensitivity. However, knockdown of cFLIP did not sensitize HUT78 to TCR-induced AICD ( Fig. 3D). Therefore, we conclude that other pathways different from cFLIP are involved in mediating AICD resistance in CTCL.
TCR signaling is decreased in CTCL tumor cells. AICD is triggered by TCR stimulation. Functional TCR signaling is crucial for the transmission and execution of the death signal. Therefore, we investigated the TCR signaling pathway in CTCL cells. FACS analysis of CTCL cell lines revealed heterogeneity regarding TCR expression levels. Only a minority of HH and MyLa cells showed TCR expression. High TCR levels were seen on SeAx cells and on HUT78 cells ( Fig. 1A). To investigate the TCR signaling capacity, we stimulated these cells with α-CD3/α-CD28 antibodies and analyzed downstream signaling events. In contrast to Jurkat control cells, we found decreased but clearly evident TCR-induced ERK signaling in all CTCL cell lines investigated ( Fig. 4A ). Despite of different TCR expression levels, all CTCL cell lines showed ERK phosphorylation at 5 minutes after TCR stimulation. In contrast to Jurkat cells, the ERK-phosphorylation signal in CTCL cell lines was not detectable after prolonged TCR stimulation. Phosphorylation of JNK was completely diverse when CTCL cell lines were compared with Jurkat cells. MyLa cells showed normal JNK signaling, whereas in HUT78 cells activation of JNK was completely abrogated, and SeAx cells showed a constitutive activation of JNK ( Fig. 4B). Because of the diversity in TCR-induced JNK activation in different CTCL cell lines, it is highly unlikely that the JNK pathway plays any role in mediating the AICD resistant state. In summary, all tested CTCL cell lines showed rapid activation of ERK, but the phosphorylation signal faded much faster when compared with control Jurkat cells. This points to a general defect of CTCL cell lines in prolonged ERK activation.
CTCL tumor cells fail to induce PLC-γ1 activation, Ca2+ influx, and ROS production after TCR stimulation. TCR stimulation activates ERK via two downstream pathways. ERK can be activated quickly via LAT, GRB2, and RAS. PLC-γ1 is crucial for a much slower ERK activation. It converts PIP2 into IP3 and diacylglycerol (DAG). IP3 leads to an intracellular Ca2+ release from the endoplasmic reticulum into the cytosol, and DAG recruits RAS-GRP1, which then activates RAS and ERK. Because of the defects in prolonged ERK phosphorylation in CTCL cell lines, we investigated the status of PLC-γ1 activation in these cells. Interestingly, none of the CTCL tumor cell lines showed any PLC-γ1 activation after TCR stimulation ( Fig. 5A ). This is consistent with our findings concerning the ERK signaling pathway and could explain the missing ERK phosphorylation upon prolonged TCR stimulation ( Fig. 4A). Beyond the effect on ERK PLC-γ1 activation mediates intracellular Ca2+ release. In line with the abrogated PLC-γ1 activation, CTCL tumor cells did not show an increase in intracellular Ca2+ after TCR stimulation ( Fig. 5B). In primary CTCL tumor cells, a significantly different kinetic with delayed Ca2+ influx depending on tumor cell burden was shown ( Fig. 5B). Furthermore, it is known that DAG generation via PLC-γ1 is essential for TCR-induced ROS production from mitochondria ( 20). It was not possible to induce similar amounts of ROS in CTCL tumor cells by TCR-stimulation compared with Jurkat control cells ( Fig. 5C). In contrast to that, we could not observe a significant reduction of ROS production in clinical samples ( Fig. 5C). This points into the direction that impairment of the Ca2+−dependent pathway might be more important for primary CTCL tumor cell survival. However, treatment of CTCL tumor cells with the Ca2+ ionophore ionomycin induced Ca2+ influx ( Fig. 5B) and PMA induced ROS production ( Fig. 5C) by bypassing the proximal TCR signaling machinery including PLC-γ1. Finally, we treated CTCL tumor cells with PMA and ionomycin to artificially generate Ca2+ and ROS signals and measured cell death after 24 hours of incubation. Interestingly, treatment with PMA and ionomycin induced apoptosis in all tested CTCL cell lines and primary CTCL tumor cells ( Fig. 5D). We conclude that Ca2+ and ROS signals are needed to break AICD resistance of CTCL tumor cells.
TCR stimulation does not induce CD95L in CTCL cells. Previously, we have shown that Ca2+ and ROS signaling are crucial for up-regulation of CD95L and the execution of AICD ( 10). Thus far, our findings concerning abrogated Ca2+ release and reduced ROS production point into the direction that CTCL cells might be unable to induce CD95L expression upon TCR stimulation. To investigate CD95L up-regulation in CTCL tumor cells, we quantified CD95L mRNA after TCR triggering in CTCL tumor cells from Sézary patients. In contrast to T cells from healthy donors, CTCL tumor cells failed to up-regulate CD95L mRNA after TCR stimulation ( Fig. 6A ). Furthermore, restoration of Ca2+ and ROS signaling by treatment with PMA and ionomycin induced CD95L mRNA expression in all CTCL cell lines (data not shown; ref. 10). The failure to induce CD95L mRNA is also evident on the protein level. Primary CTCL tumor cells showed decreased levels of TCR-induced CD95L surface expression as determined by flow cytometry ( Fig. 6B). These results clearly show a defect in TCR-induced CD95L-induction.
In summary, we have shown that CTCL tumor cells exhibit no PLC-γ1 activity that results in decreased ROS production and abrogated Ca2+ signaling. Both signals are needed for CD95L up-regulation in AICD. Consistent with this, CTCL tumor cells from Sézary patients show reduced CD95L surface expression upon TCR stimulation, which is causative for the resistance of these cells to AICD ( Fig. 6C).
In the present study, we investigated the underlying molecular mechanism determining AICD resistance of CTCL cells. AICD is induced by TCR stimulation, leading to CD95L expression that triggers CD95 to execute AICD. A decrease in CD95L induction is associated with reduced AICD ( 19). cFLIP is a potent inhibitor of death receptor signals and is involved in the regulation of AICD ( 9). Besides its function in the immune system, cFLIP is implicated in the pathogenesis of anaplastic large cell lymphoma ( 21) and was correlated with a poor clinical outcome in patients with Burkitt's lymphoma ( 22). We showed higher levels of cFLIP protein expression in CTCL cell lines and tumor cells from Sézary patients in comparison with lymphoma cell lines and T cells from healthy donors. Increased expression of cFLIP protein was also seen by Braun and colleagues ( 23) in CTCL cell lines and skin and blood samples from CTCL patients. Contassot and colleagues ( 24) recently attributed resistance of CD95-expressing CTCL tumor cells to CD95-mediated apoptosis to increased levels of cFLIP mRNA in a number of Sézary patients. Both studies suggest c-FLIP–mediated apoptosis resistance at least in CTCL patients with CD95-expressing tumor cells. However, here, we showed a similar degree of TCR-induced cFLIP mRNA induction in CTCL cells and control cells. However, knockdown of c-FLIP by shRNA in CTCL cells did not determine resistance to AICD. cFLIP can be induced by nuclear factor-κB (NF-κB; ref. 25) and is able to activate NF-κB by itself ( 26). We speculate that cFLIP is either induced by or is responsible for the known constitutive activation of NF-κB found in CTCL cell lines and CTCL cells from Sézary patients ( 27). These studies and our own data indicate that the overall amount of cFLIP is likely to be significantly higher in CTCL tumor cells even if the relative expression levels of cFLIP are similar between healthy controls and CTCL tumor cells.
We have shown reduced but functional TCR signaling regarding phosporylation of mitogen-activated protein kinases in CTCL cells. Further downstream of the TCR, we identified a loss of PLCγ-1 activation that leads to reduced ROS production and abrogated intracellular Ca2+ release specifically in CTCL cells. In line with our findings, Nikolova and colleagues ( 28) suggested that the reduction of TCR signaling by inhibitory receptors (CD158k/KIT3DL2 and CD85j/ILT-2R) may lead to AICD resistance as seen in ILT-2R–expressing T cells. The finding of reduced TCR signaling in CTCL cells was also reported by Fargnoli and colleagues ( 29) who showed decreased activities of Zap70, Syk and membrane-associated Csk after TCR stimulation. These studies support our results of abrogated PLC-γ1 activation in CTCL cells after TCR stimulation.
AICD is dependent on Ca2+ release and ROS production to initiate transcription of CD95L, its excretion and subsequent stimulation of CD95. Artificial restoration of Ca2+ influx and ROS production sensitized CTCL cells to AICD underlining the significance of our suggested pathway. Consistent with the abrogated Ca2+ release and the reduced ROS production, we have shown that the majority of CTCL cell lines and primary CTCL cells fail to induce CD95L mRNA and protein after TCR stimulation. Thus, in the panel of patients examined, AICD resistance of CTCL cells is caused by reduced TCR-induced CD95L induction. This finding represents a novel mechanism of apoptosis resistance in CTCL and, therefore, has an effect on future therapeutic regimens aiming at induction of CD95L in CTCL cells independent of TCR signaling.
Resistance to apoptosis may play a significant role in the pathogenesis of many cancers and has implications for their treatment ( 5). In the pathogenesis of CTCL, an important role of apoptosis resistance rather than proliferation is implicated ( 30). Different splice variants of the death receptor CD95 were identified in CTCL, which might be responsible for apoptosis resistance ( 31). The analysis of skin biopsies from CTCL patients revealed high expression rates of CD95 in early stages of the disease and a decrease of CD95 levels in more advanced disease ( 32). This is also seen in a number of Sézary patients that showed a loss of CD95 and subsequent resistance to CD95-mediated apoptosis ( 24, 33). However, these observations do not seem to be the sole mechanism in CTCL patients because we and others ( 23) could show high expression rates of CD95 on CTCL cells paralleled by increased sensitivity to CD95-induced apoptosis. Taken together, these data point to the fact that the balance between CD95 expression and intracellular inhibition such as cFLIP may determine a required CD95 signaling strength to induce apoptosis.
CTCL cells might be protected from AICD by increased levels of Bcl-2. Bcl-2 is also known to be induced by NF-κB ( 34), and increased levels of Bcl-2 could be shown in CTCL skin lesions ( 35). Furthermore, cytokines such as IL-7 were found to maintain survival of CTCL tumor cells by increased expression of Bcl-2 ( 36, 37). In contrast, another group found two CTCL tumor cell lines that lost Bcl-2 expression, suggesting that Bcl-2 increases apoptosis resistance while the loss of Bcl-2 can be compensated ( 38). Furthermore, the signal transducer and activator of transcription-3 (STAT3) can mediate apoptosis resistance in CTCL by activating antiapoptotic genes BCL-XL, MCL1, and Survivin and proliferation-inducing genes MYC and cyclin D1 ( 39). STAT3 can be induced in CTCL cells by IL-2 and IL-15 ( 40) and gain of STAT3/STAT5 genes ( 41), which is of functional relevance in CTCL patients being resistant to treatment with the histone deacetylase inhibitor vorinostat ( 42). STAT3 was shown to be activated by TCR stimulation with α-CD3 and Staphylococcal enterotoxin ( 43). All these factors might contribute to AICD resistance in CTCL.
The discussed studies and our own results show resistance to apoptosis as a major feature of CTCL tumor cells. However, the identified apoptosis defects are heterogenic in different patients and likely reflect the different clinical courses observed in Sézary syndrome. Here, we show that CTCL tumor cells are resistant to AICD despite an intact CD95 pathway. We found that PLC-γ1 activity is abrogated leading to a lack of Ca2+ influx and reduced ROS production in CTCL upon TCR stimulation. These defects are responsible for the low CD95L expression resulting in AICD resistance in CTCL. Therefore it might be possible to break AICD resistance in CTCL by inducing CD95L independent of TCR ligation, pointing to potential future therapeutic options.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: Deutsche Forschungsgemeinschaft (SFB 405; to C.D. Klemke and P.H. Krammer), Intramurales Förderprogramm DKFZ (D. Brenner), Wilhelm-Sander-Stiftung (2000.092.2, 2008.072.1), Deutsche Krebshilfe (106849) and Deutsche Forschungsgemeinschaft (Le 953/5-1; M. Leverkus), Wilhelm-Sander-Stiftung (2004.064.2, 2007.126.1) and Landesstiftung Baden-Württemberg (Network Aging Research; K. Gülow), and Deutsche Krebshilfe and Helmholtz Alliance of Immunotherapy (P.H. Krammer).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank all patients and healthy donors who supported our study; Dr. Nina Booken for taking patient's blood; Cosima Kretz for providing the biotinylated α-APO1 (α-CD95); PD Dr. Elisabeth Suri-Payer and Michael Kieβling for critical reading of the manuscript; Dr. Alexander Golks for discussion; Dorothee Süss, Manuel Scheuermann, and Klaus Hexel for technical assistance; and Miguel Martins and Marion MacFarlane for the puromycine-resistant variant of the pRS retroviral construct.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
C.D. Klemke and D. Brenner contributed equally to this work.
- Received December 5, 2008.
- Revision received February 10, 2009.
- Accepted March 20, 2009.
- ©2009 American Association for Cancer Research.