Introduction

Ewing's sarcoma (ET), a childhood bone tumor, and the less frequent peripheral neuroepithelioma, are tumors of neuroectodermal origin, characterized by a highly specific recurrent balanced chromosomal translocation t(11; 22) (q24; q12) (Aurias et al., 1983; Horowitz et al., 1997). This translocation results in the fusion, in the der (22) chromosome, of the 5′ half end of a resident EWS gene (Ewing's sarcoma gene) of unknown function, with the 3′ portion of the chromosome 11-derived Fli-1 gene (Friend Leukemia Integrator-1 gene) carrying the Ets domain with a specific DNA-binding motif (Delattre et al., 1992; Bailly et al., 1993). In biological studies, the resultant EWS/Fli-1 fused protein initiates cellular transformation and has been suggested to function as an aberrant transcription factor that may modulate the expression of a different set of target genes (May et al., 1993; Ouchida et al., 1995; Kovar et al., 1996; Tanaka et al., 1997; Jaishanker et al., 1999).

Although 50–60% of patients with Ewing's sarcoma can be cured using a combination of local surgery and/or radiation and systemic chemotherapy, current treatment protocols in patients with metastatic development rarely result in complete remission or cure. Innovative approaches are therefore under investigation to increase disease-free survival (Horowitz et al., 1997; Fizazi et al., 1998). The hypothesis that interferons (IFNs) may reduce carcinogenesis has been supported by the results of numerous studies (Gutterman, 1994; Tanneberger and Harelia 1996; Borden, 1998). Type I IFNs were demonstrated to depress cell growth and proliferation, and to modify certain complex functions inhibiting the synthesis of several proteins and enzymes induced by growth factors (Gutterman, 1994; Grander et al., 1997). IFNs were recently introduced into combination protocols for cancer prevention together with chemotherapeutic agents (Dorr, 1993; Gutterman, 1994; Borden, 1998). As yet however, very little is known about their action on Ewing's sarcoma.

IFNs consist of three main protein families (-α, -β, and -γ), whose production can be induced in most cells by different stimuli. IFNs exert their characteristic biological actions by binding to high-affinity cell surface receptors. IFN-α and IFN-β (type I IFNs) share a common receptor (Uzé et al., 1990; Novick et al., 1994; Domanski et al., 1996); IFN-γ (type II IFN) binds to a different cell surface receptor (Aguet et al., 1988). Recent reports showed that the subunits of its receptor engage IFN-β in a set of interactions distinct from those of IFN-α (Croze et al., 1996; Domanski et al., 1998; Runkel et al., 1998). Early events in type I IFN signaling are tyrosine phosphorylation of the type I IFN receptor subunits (IFNAR1 and IFNAR2), and the activation of the receptor-associated Tyk-2 and Jak-1 Janus kinases. Engagement of these kinases regulates tyrosine phosphorylation of Stat proteins, and the activation of signaling cascades downstream of these proteins. Type I IFNs activate the transcription of Interferon Stimulated Genes (ISGs) through the assembly and translocation from the cytoplasm to the nucleus of Interferon Stimulated Gene Factor 3 (ISGF3), a multisubunit transcription factor which interacts with the Interferon Stimulated Response Elements (ISRE), which are located upstream of the ISGs promoters. ISGF-3α contains Stat-1/Stat-2 heterodimers that are activated by phosphorylation on tyrosine of the two component polypeptides of the complex. These heterodimers associate with ISGF3γ, a 48 kDa DNA-binding subunit, to form ISGF3, a complex which displays high-affinity binding to ISRE and is competent to activate transcription of ISRE-containing promoters (Velasquez et al., 1992; Darnell et al., 1994; Colamonici et al., 1995; Ihle, 1996; Meraz et al., 1996). The crucial role of the p38 MAPkinase for the serine phosphorylation of Stat-1 and transcriptional changes induced by IFNs has recently been demonstrated (Goh et al., 1999; Uddin et al., 1999).

IRFs (Interferon Responsive Factors) are a family of transcription factors that, like the Stats, mediate IFN signaling. However, unlike the Stats which are activated within minutes of ligand binding to the receptor, IRFs constitute a secondary wave of response to the IFN signal (Nguyen et al., 1997a). IRF-1 and IRF-2 are also key transcription factors in the regulation of cell growth, cell cycle and apoptosis. IRF-1 functions as a transcriptional activator, whereas IRF-2 represses IRF-1 functions (Harada et al., 1993, 1994, 1998). Ectopic overexpression of IRF-1 results in strong inhibition of cell growth, and deletion of the IRF-1 gene has been demonstrated in a number of human leukemias and myelodysplasias (Harada et al., 1998). Several ISGs have been shown to be involved in regulating cell proliferation and are target genes of ISGF3 and IRFs. The two best characterized ISGs, 2′5′oligo(A) synthetase and double-stranded RNA-dependent protein kinase, participate in the regulation of cell growth and are targets for regulation by ISGF3 and IRF-1 (Stark et al., 1998).

Within the framework of the regulation of apoptosis by IRF-1, some candidate genes, such as p21WAF-1/Cip-1, have been identified (Tanaka et al., 1996); furthermore, IRF-1 can cooperate with p53 to transactivate the p21WAF-1/Cip-1 promoter (Tanaka et al., 1996; Gartel et al., 1999). Published data suggest that IFN-α and IFN-β induce growth arrest and differentiation in a variety of cell lines by increasing the levels of p21WAF-1/Cip-1 mRNA, but the mechanisms responsible for these effects have not been established. (Chin et al., 1996; Hobeika et al., 1997; Giandomenico et al., 1998). On the other hand, a recent study described different cellular locations of p21WAF-1/Cip-1 resulting in anti-apoptotic properties (Asada et al., 1999).

In various cell systems, overexpression of the oncoprotein Blc2 can delay or prevent apoptosis by various death-promoting signals (Reed, 1998). Bcl-2 governs a cell-death commitment step upstream of caspase activation, and caspase-family cell death proteases are the ultimate effectors of apoptosis (Salvesen et al., 1997). There are two major executioner caspases, caspase-3 and caspase-7, which have different substrate specificities. The exact order of the executioners, and the place of other caspases in the pathway, are still controversial, but it has been acknowledged that the signaling of death is transmitted in part by sequential caspase activation (Salvesen et al., 1997; Wolf et al., 1999). The results of several studies suggest that caspase-7, but not caspase-3, is the executioner caspase in human prostatic carcinoma cell lines. More importantly, overexpression of caspase-7 has the ability to bypass the antiapoptotic effect of the oncoprotein Bcl-2, a putative mediator of resistance to apoptosis in androgen-independent prostate cancer (De Marco et al., 1995; Roklin et al., 1996; Marcelli et al., 1998).

Although type I IFNs share not only the same receptor but also a common set of signaling events that ultimately regulate the expression of common biological activities, evidence is accumulating that significant differences exist between the effects of the different type I IFNs (Aguet et al., 1988; Uzé et al., 1990; Novick et al., 1994; De Marco et al., 1995; Croze et al., 1996; Domanski et al., 1996, 1998; Runkel et al., 1998; Grumbach et al., 1999). Although their mechanisms of action are not completely understood, IFNs have been studied for their therapeutic efficacy in a number of pathologies including leukemia (Michalevicz et al., 1988; Gutterman, 1994; De Marco et al., 1995; Borden, 1998; Sacchi et al., 1998; Grumbach et al., 1999). Kaposi's sarcoma in acquired immunodeficiency syndrome (Hadida et al., 1999), type C viral hepatitis (Salmeron et al., 1999), and multiple sclerosis (Paty et al., 1993).

Because little is known about the efficacy of the action of IFNs on solid tumors of childhood, this study was designed to examine the action of type I IFNs, i.e. IFN-α2a and IFN-β, on the proliferation of tumoral cells derived from Ewing's sarcoma (ET cells), and their behavior during IFN treatment. Our results demonstrate that IFN-β exerts a more powerful anti-proliferative effect than IFN-α in Ewing's sarcoma cells, regardless of their p53 status (Kovar et al., 1993; Harmelin et al., 1994). In contrast to IFN-α, IFN-β treatment led to rapid activation of p38 MAPkinase and induced Stat-1-Ser727 phosphorylation which correlated with the strong activation of ISGF3 and the 2′5′oligo(A)synthetase gene. Furthermore, IFN-β also induced IRF-1-mediated cell death in wild-type p53 cell lines, i.e. EW-7 and COH. Ectopic expression of IRF-1 confirmed the role of IRF-1 in triggering programmed cell death involving the induction and activation of caspase-7.

Results

Comparative antiproliferative effect of IFN-α and -β on cells derived from Ewing's sarcoma

The antiproliferative effects of IFN-α and IFN-β (type I IFNs) were tested in four ET cell lines derived either from primary tumors (EW-7 and EW-1 cells), or metastatic secondary tumors (COH and ORS cells). Exponentially growing cell cultures were exposed to IFNs (500 U/ml) and cells were counted 24 h later (Figure 1a). In all four cell lines, IFN-α exhibited 20–30% inhibition of cell proliferation, whereas treatment with IFN-β inhibited growth by more than 60% with, for instance, 70% growth inhibition in EW-7 cells. The effect of various concentrations of IFNs on the proliferation of Ewing's sarcoma cells was explored with the EW-7 cell line, as it displays maximal susceptibility to the antiproliferative action of IFNs (Figure 1b). Treatment of EW-7 cells with IFN-α for 24 h resulted in a limited, concentration-dependent inhibition of cell growth of 10–36%. On the other hand, IFN-β induced a dose-dependent inhibition of cell growth of 40% with a minimal IFN-β concentration of 50 U/ml. IFN-β at 500 U/ml inhibited cell growth by 70% and a maximal inhibition of cell growth of 80% was obtained with 2000 U/ml of IFN-β. Analysis of type I IFN receptors on EW-7 cells revealed the presence of about 1500 homogeneous binding sites, with identical equilibrium constants for the two IFNs (kd 1×10−10 M). Unlabeled IFN-α was effective in displacing radiolabeled IFN-β from the receptor with equivalent concentration as unlabeled IFN-β (data not shown). These results indicate that the differences in cell growth inhibition by IFN-α and IFN-β were not due to differences in IFN receptor binding.

Figure 1
figure 1

Growth inhibition of Ewing's sarcoma cell lines after IFN-α and IFN-β treatment. Exponential growing cells (5×104) were seeded in 25 cm2 culture flasks. (a) After 24 h, medium was withdrawn and the four ET cell lines, EW-7, EW-1, COH and ORS, were treated with 500 U/ml of IFN-α2a or IFN-β. Cells were harvested after 24 h of stimulation and were counted with a ZM Coulter Counter equipped with a Coultronic 256 channelizer. The percentage of inhibition of cell proliferation was calculated taking cell growth in untreated cultures as 100% proliferation. Values are means±s.e.m. from five separate experiments. (b) Dose response of IFN-α and IFN-β in EW-7 cells. Cells were incubated with 0, 50, 500, 2000 or 5000 U/ml IFN-α or IFN-β for 24 h, as described in a. Harvested cells were counted as described in a. Values are means±s.e.m. from three separate experiments

Full size image

Cell signaling in response to IFN-α or IFN-β treatment of ET cells

To determine whether IFN-α and IFN-β lead to similar Jak-1 and Tyk-2 phosphorylation, as well as to similar activation and nuclear translocation of Stat-1/Stat-2-containing complexes, nuclear protein extracts were prepared from both untreated EW-7 cells and cells stimulated for 15 min with either IFN-α or IFN-β and subjected to immunoprecipitation with specific Tyk-2 or Jak-1 antibodies, followed by immunoblot analysis with anti-phosphotyrosine specific antibody. As shown in Figure 2a, tyrosine phosphorylation of Tyk-2 and Jak-1 was similar after IFN-α and IFN-β treatment. Hybridization of membranes with specific Tyk-2 or Jak-1 antibodies resulted in similar yields of Tyk-2 and Jak-1 proteins in untreated cells and IFN-treated cells. The effects of IFN-α and -β treatment on Stat-1 and Stat-2 activation were then evaluated using EW-7 cell extracts stimulated for 15 min with IFN-α or IFN-β and prepared under non-denaturing conditions. Co-immunoprecipitation with antibodies directed against Stat-1 followed by immunoblot analysis (Figure 2b) showed similar yields of Stat-1 protein in untreated and IFN-treated cells. Rehybridization with an antibody directed against Stat-2 showed that Stat-2 coprecipitated with Stat-1 in IFN-stimulated but not in unstimulated cells (Figure 2b). Consecutive probing of the blot with antibodies directed against phosphotyrosine and phosphoserine demonstrated the presence of tyrosine-phosphorylated Stat-1 and Stat-2 after IFN-α and IFN-β treatment (Figure 2c). In unstimulated cell extracts, no tyrosine-phosphorylated Stat-2 was observed, but a faint band was detected corresponding to constitutive tyrosine-phosphorylated Stat-1. Antibody directed against phosphoserine only recognized a Stat-1 phosphorylated protein in the extract of IFN-β-treated cells. Similar results were also observed for the three other ET cells (data not shown). Overall, these results indicate effective activation of Stat proteins by both IFN-α and IFN-β. However, phosphorylation of Stat-1 on serine was only induced by IFN-β in the ET cell lines. Interestingly, antibodies directed against phosphorylated Ser727-Stat-1 revealed a rapid phosphorylation of Stat-1, after IFN-β treatment. Ser727-Stat-1 was still detectable after 60 min of IFN-β stimulation (Figure 2d). Recently, it has been shown that p38 MAPkinase is required for Stat-1 serine phosphorylation and transcriptional activation induced by interferons (Goh et al., 1999; Uddin et al., 1999). In agreement with these data, we observed a correlation between the early activation of p38 MAP kinase (5 min) – shown by the increase of phosphorylated p38 yield (p-p38) – and the serine phosphorylation of Stat-1 (p-S727Stat-1) after IFN-β treatment (Figure 2d). Stimulation of EW-7 cells by IFN-α resulted in a modest activation of p38 MAPkinase after 5 min which decreased thereafter (Figure 2d). The activation of p38 MAPkinase was blocked by treatment of cells with the specific p38 inhibitor SB203580, impairing the induction of serine phosphorylation of Stat-1. The phosphorylation of Stat-1 on tyrosine residue was not modified by SB203580 treatment (Figure 2e).

Figure 2
figure 2

IFN-α and IFN-β-dependent phosphorylation of Tyk-2 and Jak-1, and of Stat-1 and Stat-2 proteins in the EW-7 cell line. (a) EW-7 cells were treated for 15 min with IFN-α (α) or IFN-β (β) (500 U/ml). Cell lysates were immunoprecipitated with antibodies directed against Tyk-2 or Jak-1, and analysed by Western blot. The blots were probed with antibodies directed against phosphotyrosine. After dehybridization, blots were probed with specific Tyk-2 or Jak-1 antibodies. (b) Untreated EW-7 cells, or cells treated with IFN-α (α) or IFN-β (β) for 15 min, were lysed under non-denaturing conditions as described in Materials and methods. The total extracts were submitted to immunoprecipitation using specific antibodies against Stat-1, followed by Western blot analysis. Blots were successively hybridized with monoclonal antibodies against Stat-2 and Stat-1 proteins. (c) The same blot (described in b) was stripped and reprobed with antibodies against phosphotyrosine. After another dehybridization step, the same blot was reprobed with antibodies against phosphoserine. (d) Induction of pS727-Stat-1 and activation of MAPkinase p38 by IFN-β in EW-7 cells. Cells were treated for different times with IFN-α or IFN-β, before total lysates were subjected to SDS–PAGE and Western blot analysis. Membranes were probed with pS727-Stat-1 or p-p38 antibodies before stripping and reprobing with Stat-1 or p38 mAb, respectively. (e) EW-7 cells were incubated in RPMI medium without serum for 2 h and then treated with the indicated IFNs for 30 min, in the presence or absence of 10 μM SB203580, which was added 30 min prior to IFN treatment. The cells were then lysed as described in Materials and methods and submitted to Western blot analysis. Membranes were probed with pS727-Stat-1, pY701-Stat-1 or p-p38 antibodies before stripping and reprobing with Stat-1 or p38 antibodies

Full size image

Induction of ISGF3 binding activity in ET cells by IFN-α and IFN-β treatment. Differential induction of the 2′5′oligo(A)synthetase gene promoter

Nuclear extracts from the four ET cell lines were stimulated for 2 h with IFN-α or IFN-β, and then tested for ISGF3 binding activity, using a radiolabeled ISRE probe spanning the ISRE motif of the 2′5′oligo(A)synthetase gene (Figure 3a). The pattern of the protein-DNA complexes, revealed by EMSA was very similar for the four ET cell lines studied, and showed an inducible complex detectable after either IFN-α or IFN-β stimulation, albeit in greater amounts after IFN-β treatment. The specificity of this protein-DNA complex was confirmed by the complete competition of the binding activity by unlabeled ISRE (not shown). Addition of antibodies against Stat-1 or against Stat-2 supershifted the protein-DNA inducible complex (Figure 3b) and antibodies directed against ISGF3γ/p48 impaired this complex formation, thus demonstrating that the protein-DNA inducible complex was ISGF3. The phosphorylation of Stat-1 in both tyrosine and serine residues was shown by the supershifting of the ISGF3 complex by the antibodies directed against phosphotyrosine and by the inhibition of ISGF3 complex formation by antibodies directed against pSer727-Stat-1 (Figure 3c). The protein-DNA inducible complex, identified as an ISGF3 complex, was not affected by either IRF-2 or IRF-1 antibodies. An inducible protein-DNA complex with a high electrophoretic mobility was detectable after 6 h of IFN stimulation. This complex was sensitive to IRF-1 antibodies (data not shown).

Figure 3
figure 3

IFN-α- and IFN-β-induced binding of specific nuclear proteins to Interferon Stimulating Response Element (ISRE) of the 2′5′oligo(A)synthetase gene. (a) Nuclear extracts of the EW-7, EW-1, COH and ORS cell lines treated for 2 h with IFN-α or IFN-β, were tested for their binding activity, using radiolabeled ISRE, by EMSA. The upper arrow showed specific ISGF3 complex. NS indicates non-specific binding activity. (b) Nuclear extracts from EW-7 cells treated for 2 h with IFN-β were used to determine the specificity of the protein-DNA inducible complex. Before adding the radiolabeled ISRE probe, nuclear extracts were preincubated with antibodies against Stat-1, Stat-2 and ISGF2γ-p48 or with non-immune serum. (c) Nuclear extracts from EW-7 cells treated for 2 h with IFN-β or untreated were used to determine the state of the phosphorylation of ISGF3 complex formation. Antibodies directed against the phosphorylated serine residue 727 of Stat-1 (S727-Stat-1) or against the phosphotyrosine (P-tyrosine) were added before the radiolabeled ISRE probe

Full size image

ISGF3 is a major mediator of the signaling pathway of type I IFNs, which triggered an antiproliferative effect in cells by inducing several endogenous emzymes that antagonize cell growth. One such enzyme is the 2′5′oligo(A)synthetase (2-5A) whose promoter contains the classical IFN-stimulated response element (ISRE) recognized by ISGF3 (Benech et al., 1987). Functional analysis of the 2-5A gene promoter linked to a pCAT-basic reporter plasmid was investigated in transiently transfected EW-7 cells stimulated with either IFN-α or IFN-β and tested for CAT activity after 18 or 32 h (Figure 4a). Unstimulated cells displayed no significant response, whereas IFN-α induced a 30-fold increase in reporter gene activity. IFN-β is more effective than IFN-α in inducing 2-5A gene transcription since a 120-fold increase in reporter gene activity was induced in the same conditions. These results are in agreement with the greater yield of ISGF3 containing pS727-Stat-1 induced by IFN-β. The differential induction of 2-5A promoter correlated with the mRNA transcript levels of the endogenous 2-5A gene induced by IFN-α and IFN-β (Figure 4b). The stronger induction of 2-5A gene by IFN-β was observed in all four ET cell lines. For comparison, no difference in the level of induced 2′5′oligo(A)synthetase mRNA was observed in monocytic THP-1 cells after IFN-α or IFN-β treatment (Figure 4b).

Figure 4
figure 4

Functional analysis of ISRE of 2′5′oligo(A)synthetase gene: relative CAT activity. (a) To assess basal ISRE activity and its responsiveness to IFN-α and IFN-β, EW-7 cells were transfected as described in Materials and methods. CAT activity was determined after 6, 18, and 32 h. The 2′5′oligo(A)synthetase promoter response (-fold induction) is expressed by the ratio of CAT activity in stimulated cells to that in unstimulated cells, which is defined as 1.0. Values are means of three independent experiments with a standard deviation of less than 10%. (b) 2′5′oligo(A)synthetase gene expression in the four ET cell lines and in monocytic THP-1 cell. Total RNAs (20 μg) of ET cell lines treated with IFN-α or IFN-β for 18 h were analysed by Northern blot hybridization with the 2′5′oligo(A)synthetase-radiolabeled probe. After dehybridization, membrane transfers were hybridized with GAPDH probe to validate the amounts of RNA

Full size image

Effect of IFN-β on the expression of IRF-1 in Ewing's sarcoma derived cell lines. Relationship with apoptosis

It should be interesting to imply that IRF-1 could take part in the transcriptional activation of the same set of genes by possibly cooperating with ISGF3 function in the activation of the transcription. Thus, experiments were done to investigate the regulation of IRF-1 by IFN-α and IFN-β in ET cells. Immunoblot analysis of IRF-1 and β-actin proteins in cell extracts stimulated for 18 h with IFN-α and IFN-β showed that IFN-β treatment induced a higher level of IRF-1 protein expression than IFN-α in all four cell lines (Figure 5). However, the levels of IRF-1 protein expression differed between the four cell lines used; in EW-7 and COH cells, the IRF-1 expression induced by IFN-β was stronger than that induced in EW-1 and ORS cells.

Figure 5
figure 5

Modulation of IRF-1 gene expression and IRF-1 protein induction: correlation with p21WAF-1/CIP-1 and Bcl2 protein levels. Total cellular extracts (50 μg) of the four ET cell lines, either untreated, or treated with IFN-α or IFN-β for 18 h, were submitted to Western blot analysis. Specific, proteins were successively revealed with antibodies against IRF-1, p21WAF-1/CIP-1 or Bcl2 proteins. Antibodies directed against β-actin were used to indicate the amounts of protein loaded onto the gels

Full size image

IRF-1 has a crucial role in many aspects of host defense, and also accumulating evidence has suggested that IRF-1 regulates DNA damage and induced cell cycle arrest in cooperation with the tumor suppressor p53 through transcriptional activation of the p21WAF1/CIP1 gene. In this study, we observed that after stimulation of EW-7 and COH cells for up to 48 h with IFN-β, subsequent cell detachment and cell death were observed. IFN-α did not induce such an effect. The EW-1 and ORS cells did not show this IFN-β-dependent cell death effect. We were particularly interested in determining whether stimulation of these ET cell lines by IFN-β provoked apoptosis, and the possible relationship between apoptosis and IRF-1 gene expression. For this purpose, nuclear fragmentation was evaluated by Hoechst staining, and quantified by counting the number of apoptotic cells in the total populations of untreated cells or cells treated with IFN-α and IFN-β for 48 or 72 h (Figure 6). No significant apoptotic effect was observed after IFN-α treatment whatever the cell lines considered. EW-7 and COH cells exhibited a time-dependent increase in apoptotic cells after IFN-β treatment, whereas EW-1 and ORS cells underwent no significant apoptosis during either period of IFN-β treatment. The differences in sensitivity of the four ET cell lines to IFN-β-dependent apoptosis, and the correlation between p21WAF-1/Cip-1 and Bcl-2 protein levels were analysed by immunoblot analysis after IFN-β treatment. Downregulation of total cellular Bcl-2 protein correlated with the upregulation of p21WAF-1/Cip-1 and IRF-1 expression in EW-7 and COH cells, whereas no change in Bcl-2 protein expression was observed in EW-1 or ORS cells. The p21WAF-1/Cip-1 protein was not detectable in EW-1 or ORS cells, whether or not they were treated with IFNs.

Figure 6
figure 6

Quantification of apoptotic cells in cell lines derived from Ewing's sarcoma. The four ET cell lines were stimulated with IFN-α and IFN-β. After 48 and 72 h, the fragmented nuclei were detected using Hoechst 33258 fluorescent dye on harvested cells. Nuclear fragmentation was evaluated by counting the apoptotic cells in the total populations of untreated and IFN-treated cells respectively. The percentages of apoptotic cells are the means of three independent experiments, after two readings by two experimenters, for each of the four cell lines

Full size image

Because induction of apoptosis by cytokines has been shown to involve the expression of caspase genes, we explored the possible enhancement of caspase gene expression in ET cell lines by IFN-β. For this purpose, RNase protection assays were performed for the caspase family using the hApo1 kit (Pharmingen), after stimulation of ET cells with IFN-α or IFN-β for 18 h. Expression of caspase-7 mRNA rose significantly in the EW-7 cells stimulated with IFN-β (Figure 7a), and to a lesser extent the mRNA level of caspase-3 and -8; the mRNA levels of caspases-1, -6 and -9 remained unchanged. IFN-α did not affect the expression of any caspase mRNAs. Similar results were observed in COH cells. In EW-1 and ORS cells, the mRNA levels of caspases remained unchanged.

Figure 7
figure 7

Effects of IFN-α and IFN-β on caspase gene expression and protein regulation. (a) IFN-β increased caspase-7 mRNA expression. The ET cell lines were stimulated with IFN-α or IFN-β. After 18 h, the total RNAs were isolated and 2 μg of each RNA was submitted to an RNase protection assay using the hApo-1b multi-probe set, to detect caspases-1, -2, -3, -6, -7, -8 and -9, and the housekeeping genes L32 and GAPDH which indicate RNA loading on the gel. The gel was dried and exposed overnight to a PhosphoImager screen. Note that each probe band (hApo-1b multiprobe) migrates slower than its protected band. This is due to the presence in the probe of flanking sequences that are not protected by mRNA. (b) Caspase-1 and caspase-7 activation by IFN-α and IFN-β in ET cell lines. EW-7, EW-1, COH and ORS cells were treated for 18 h with IFN-α and IFN-β. Total cell extracts were submitted to Western blot analysis. Caspase-1 and caspase-7 were successively revealed with specific antibodies. The proteolyzed 20 kDa caspase-1 was first revealed after probing with antibodies against the active form of caspase-1. After stripping the blot, the 35 kDa and 18 kDa forms of caspase-7 were detected with the specific antibodies against caspase-7

Full size image

In view of the RNase protection results, we investigated the possibility that IFN-β enhances proteolysis leading to the formation of active caspase subunits. Immunoblot analysis with antibodies against caspase-7 revealed that IFN-β not only increased the amount of procaspase-7 protein but also generated the proteolyzed active subunit of 18 kDa (Figure 7b). These processes were only observed in the wild-type p53 cells EW-7 and COH tumor cells. An increase in the 20 kDa caspase-1 subunit was also observed in EW-7 and COH cells after IFN-α or IFN-β treatment, but was not detectable in EW-1 or ORS cells (Figure 7b). Since we did not detect apoptotic cells after IFN-α treatment, the involvement of caspase-1 as a direct effector of apoptosis is unlikely. Both the increase in the gene expression of caspase-7 and the cleavage of pro-caspase-7 detected after IFN-β treatment suggest that in cells derived from Ewing's sarcoma, IFN-β triggers cell death by selectively upregulating activated caspase-7.

Ectopic expression of IRF-1

To demonstrate the role of IRF-1 in apoptosis in Ewing's sarcoma cells, we attempted to generate cells expressing IRF-1. ET cell lines constitutively expressing IRF-1 have been difficult to establish because of the strong cytotoxic effects of endogenous IRF-1. Thus, an inducible form of IRF-1 was used to determine whether the induction of IRF-1 enhanced the death of the Ewing's tumor cells that are normally resistant to IFN-β-mediated cell death. Consequently, we used a tetracycline-inducible system utilizing the reverse tTA activator (rtTA) which permits doxycycline (Dox) inducible expression of IRF-1 (Nguyen et al., 1997b). Cell clones from rtTA-IRF-1 of all four ET cell lines were expanded individually and screened for IRF-1 expression by Western blot analysis, following 48 h of growth in medium with or without Dox (Figure 8). Dox stimulation led to induction of IRF-1 protein in the four ET-TA/IRF-1 cell lines, but not in the control ET-rtTA cell lines (Figure 8, left part). To determine whether IRF-1 induction was associated with nuclear fragmentation, cells were stained with Hoechst fluorescent dye after Dox treatment of the transfected cells for 24 or 48 h (Figure 9a,b). All ET-TA/IRF-1 cell lines, regardless of their p53 status, underwent time-dependent apoptosis after stimulation by Dox. Forty to 50% of the apoptotic cells were detected at 48 h of Dox stimulation, and almost all the cells had undergone apoptosis by 96 h after stimulation (not shown). Western blot analysis of total extracts of ET-TA/IRF-1 cells showed a correlation between IRF-1 induction by Dox and p21WAF-1/Cip-1 upregulation with concomitant downregulation of Bcl 2 protein (Figure 8, right part) but not in the control ET-rtTA cells (Figure 8, left part). An increasing level of the proteolyzed active subunit caspase-7 (Figure 8, right part) was also observed in the four ET-TA/IRF-1 cell lines treated with Dox for 48 h. Taken together, these results suggest that caspase-7 is involved in the apoptosis mediated by IRF-1, in Ewing's sarcoma cells.

Figure 8
figure 8

Inducible expression of IRF-1: correlation of IRF-1 expression and the activation of caspase-7. Whole cellular extracts of the four control ET-rtTA and ET-rt-TA/IRF-1 cell lines, either untreated, or treated with 2 μg/ml of doxycycline (Dox) for 48 h, were submitted to Western blot analysis. First, IRF-1 protein induction was revealed with specific antibodies directed against human IRF-1. Then, after successive dehybridizations, the blots were reprobed with antibodies directed against p21WAF-1/CIP-1 or against Bcl-2, antibodies against the two forms of caspase-7 (35 kDa and 18 kDa). β-Actin bodies were used to verify the amounts of protein loaded onto the gels

Full size image
Figure 9
figure 9

Inducible IRF-1 induced apoptosis in the four ET-rt-TA/IRF-1 cell lines. (a) Percentage apoptosis induced in the rt-TA/IRF-1 expressing cells. Transfected ET-rt-TA/IRF-1 cell lines were cultured in the presence of 1 μg/ml or 2 μg/ml doxycycline (Dox) for 24 or 48 h. Harvested cells were submitted to Hoechst staining as described in the legend to Figure 6. The percentages of apoptotic cells are means of five independent experiments for 24 h of Dox induction, and of four independent experiments for 48 h of Dox induction, after reading by two different experimenters, for each of the four transfected cell lines. (b) Nuclear morphology of ET-rt-TA/IRF-1 cell lines cultured for 48 h without doxycycline (−Dox) or with 2 μg/ml of Dox, after Hoechst staining. Apoptotic nuclei display fragmentation in Dox-treated cells. (Cells were photographed at ×630 magnification)

Full size image

Discussion

We have examined the antiproliferative and apoptotic effects of type I IFNs, i.e. recombinant interferon-α2a and recombinant interferon-β, on human cell lines established from Ewing's sarcoma (ET) at different stages of the disease. Although these cell lines exhibited a similar phenotype, the expression and the status of various genes such as the p53 gene, differed. Two of the four ET cell lines studied, EW-7 and COH, expressed wild-type p53, while EW-1 and ORS cells expressed mutated p53 gene. Our results showed that the four ET cell lines were more responsive to the antiproliferative effect of IFN-β than of IFN-α. Previous reports also showed that IFN-α had a weak antiproliferative effect of 20 to 30% on ET cell lines (Van Valen et al., 1993; Rosolen et al., 1997).

Binding studies in Ewing's sarcoma cells showed that IFN-α and IFN-β displayed similar properties in relation to their common receptors. Therefore, differences in sensitivity to the antiproliferative effect of IFN-α and IFN-β were not due to the alterations in binding properties to their common type I IFN receptors expressed by ET cell lines. These results are in agreement with published data describing the presence of specific functional receptors for IFN-α in a series of pediatric tumors, including Ewing's sarcoma, and in cell lines derived from these tumors (Rosolen et al., 1997). Despite the fact that all type I IFNs bind to the same receptor and activate a common set of signaling elements, there is now accumulating evidence of differences in their signaling pathways (Croze et al., 1996; Domanski et al., 1998; Runkel et al., 1998). For instance, selective Jak-1 phosphorylation induced by IFN-β in human myocardial fibroblasts was recently reported (Grumbach et al., 1999). Velasquez et al. (1992) showed that cells with Tyk-2 kinase deficiency exhibit residual sensitivity to IFN-β but not to IFN-α. Also, IFN-β was shown to induce the association of the phosphorylated forms of the IFNAR-I and IFNAR-2 subunits of type I IFN receptors, whereas IFN-α did not (Croze et al., 1996; Domanski et al., 1998; Runkel et al., 1998).

To try to clarify the mechanism responsible for the difference in ET cell sensitivity to both IFN-α and IFN-β, we analysed the triggering of the early signals involved in type I IFN-signal transduction. Our results showed that both IFNs induced tyrosine phosphorylation of the Jak-1 and Tyk-2 kinases and of the transcription factors Stat-1 and Stat-2, with similar efficacy. However, after IFN-β treatment, additional phosphorylation of Stat-1 was observed on serine 727, which correlated with the activation of the p38 MAP kinase. The fact that inhibition of p38 activation by SB203580 results in abrogation of serine phosphorylation of Stat-1 without change of tyrosine phosphorylation of Stat-1 suggests that p38 plays a role in IFN-β signaling in Ewing's cells. Our data are in line with recent findings showing an essential role of p38 MAPkinase in regulating the serine phosphorylation and transcriptional activity of Stat-1 and ISGF3 (Goh et al., 1999; Uddin et al., 1999). Although in other cell lines of different origin, both IFN-α and IFN-β activate p38 MAPkinase and induce phosphorylation of Stat-1 on serine 727 in ET cells, IFN-β, but not IFN-α, was able to trigger these effects. Thus, in ET cells, IFN-β-dependent gene transcription via ISRE involved not only the Jak/Stat pathway but also p38 MAPkinase activation and Stat-1 serine phosphorylation. It has been shown that tyrosine phosphorylation is critical for translocation and binding to DNA, and serine/threonine phosphorylation is often crucial for maximal transcriptional activation (Pestka et al., 1987; Eilers et al., 1995; Wen et al., 1995; Horvath et al., 1997; Goh et al., 1999; Uddin et al., 1999). The strong ISGF3 complex activation involving pS727-Stat-1 may explain the greater efficacy of IFN-β in inducing the promoter of the 2-5A gene, which has been clearly demonstrated to mediate the antiproliferative effects of IFNs (Pestka et al., 1987; Gutterman, 1994; Stark et al., 1998). These findings explain, at least in part, the greater biological effect observed for IFN-β compared to IFN-α in ET cells. Serine phosphorylation was also shown to be required for the formation of stable Stat-3 DNA complexes during IL-6 activation in EW-1 cells. However, this seems to be restricted to cells of lymphoid and neuronal origin (Zhang et al., 1995).

The transcription factor ISGF3 initiated the first wave of transcription induced by IFN-α and IFN-β. IRF-1 protein may prolong IFN-induced gene expression, either by cooperating with the ISGF3, or by functioning independently after the dissociation of ISGF3 due to the deactivation of its components (Kimura et al., 1996). Experiments conducted with Stat or IRF family transcription factor gene-targeted mice have clearly shown that ISGF3 and IRFs factors have specific, nonredundant roles in cytokine signaling (Pellegrini et al., 1989).

The 72 h treatment of ET cells with IFN-β led to the delivery of an apoptotic signal in EW-7 and COH cells, which are wild-type p53 cell lines. Although the EW-1 and ORS cell lines were sensitive to the antiproliferative effect of IFNs, they did not undergo cell death. Our results suggest that their resistance to IFN-β-induced apoptosis might be due to the failure to maintain sustained expression of IRF-1 in these cell lines. Accumulating evidence suggests that IRF-1 controls tumor development (Harada et al., 1998), regulates DNA damage, and induces cell cycle arrest in collaboration with the tumor suppressor p53 through the transcriptional activation of the p21WAF1/CIP1 gene, a cell cycle inhibitor (Tanaka et al., 1996; Gartel et al., 1999).

In our experiments, IRF-1 expression after IFN-β treatment only correlated with an increase in the p21WAF1/CIP1 protein level, in the two wild-type p53 EW-7 and COH cell lines. However, p21WAF1/CIP1 expression did not correlate with the growth inhibition induced by IFNs, because we did not detect any p21WAF1/CIP1 protein in the p53-mutated EW-1 and ORS cells. In a recent report, p21WAF1/CIP1 was described as an inhibitor of the cell cycle or of apoptosis, as defined by its subcellular localization (Asada et al., 1999). As regards the concomitant regulation of the anti-apoptotic protein Bcl-2, we observed a decrease in its level after IFN-β treatment in the p53-wild-type EW-7 and COH cell lines, which correlated with the increase in the level of p21WAF1/CIP1 protein and in the number of apoptotic cells. These results are in line with those reported in various studies showing that suppression of Bcl-2 expression induces or accelerates cell death (Arriola et al., 1999; Hotchkiss et al., 1999).

The caspase family cell death proteases play a critical role in the biochemical events governing apoptosis. In this connection, it is noteworthy that IRF-1 has been reported to induce the activation of caspase-1 (Tamura et al., 1995). In ET cell lines, both the increase in caspase-7 gene expression and the cleavage of caspase-1 and -7, demonstrated here by the appearance of activation fragments, were detected after IFN-β treatment in the EW-7 cells. The fact that both IFN-α and IFN-β upregulated the level of proteolyzed active 20 kDa caspase-1, but that only IFN-β increased the gene expression of caspase-7 in EW-7 cells, suggests that caspase-7 is probably the major effector protease involved in IFN-β-induced apoptosis.

In many reports, uncertainty was expressed about the role of caspase-1 in apoptosis, but caspase-7, described as an apoptotic executioner, was shown to play a critical role in the biochemical events governing this process (Salvesen et al., 1997). Although the role of IRF-1 as a tumor suppressor has been supported by the results of several studies, the mechanisms by which IRF-1 exerts its apoptotic effects are not clear. In the present work, we showed that IRF-1 mediated IFN-β-induced apoptosis in cells derived from Ewing's sarcoma, and that this apoptosis involved caspase-7 induction and activation.

To clarify the putative role of IRF-1 in IFN-β-mediated apoptosis, the four ET cell lines carrying an inducible IRF-1 vector were generated. Ectopic expression of IRF-1 induced cell apoptosis in the four ET-rtTA/IRF-1 cell lines, whatever their p53 status. In addition, ectopic expression of IRF-1 correlated with the increase in the p21WAF1/CIP1 protein level and in the active form of 18 kDa caspase-7 released from the precursor by proteolysis. These results indicate that although the IRF-1 expression induced by IFN-β correlated with the expression of wild-type p53, the consequent downstream events involving IRF-1 and leading to apoptosis seem to be independent of the p53 status. Our results are in agreement with those reported in a number of recent studies concerning the apoptotic properties of IRF-1 (Tamura et al., 1995; Tanaka et al., 1996; Harada et al., 1998; Stark et al., 1998; Hotchkiss et al., 1999).

The EWS-Fli-1 fusion protein may function as an antiapoptotic factor in Ewing's sarcoma (Ouchida et al., 1995; Kovar et al., 1996). However, in our system, the cell death induced by IFN-β treatment via IRF-1 induction did not correlate with the downregulation of this protein (data not shown). The observations concerning ectopic IRF-1 expression confirmed the results obtained in parental ET cell lines treated with IFN-β, as this expression correlated with the induction of caspase-7, which is known to be involved in the effector phase of apoptosis.

Many authors have established that caspases are essential mediators of apoptosis, and that disordered apoptosis can promote human disease. Insufficient apoptosis, resulting from caspase inactivation, may promote oncogenesis by allowing cell accumulation (Salvesen et al., 1997; Wolf et al., 1999; Roklin et al., 1996; Marcelli et al., 1998). Manipulation of the apoptotic pathway for the treatment of human diseases is becoming an important field of investigation, because many conditions associated with excessive or impaired programmed cell death have been described. Caspase activity has already been blocked or enhanced for therapeutic purposes in animal disease models (Hara et al., 1997). Recently, overexpression of caspase-7 was shown to induce apoptosis in a human prostatic carcinoma cell line (LNCaP) and caspase-7 expression was suggested as a potential gene therapy designed to force prostate cancer cells to undergo programmed cell death (Marcelli et al., 1998).

Although extensive clinical experiments have been performed with IFN-α, IFN-β is now used in human therapy, and an increasing number of recent clinical studies focus on the use of IFN-β against human carcinomas (Michalevicz et al., 1988; Gutterman, 1994; Borden, 1998; Grumbach et al., 1999). Since the possibility of using IFN-β, both alone and combined with chemotherapeutic agents, has already been examined in several clinical trials, it might be of interest to explore the possibility of using IFN-β alone, or with these agents, in solid tumors (Gutterman, 1994; Borden, 1998). In this report, our results concerning the general antiproliferative effect of IFN-β on Ewing's sarcoma cells provide a rational foundation for a promising therapeutic approach to Ewing's sarcoma. Furthermore, the sensitivity of Ewing's sarcoma cells to IFN-β-mediated apoptosis could be monitored by following the expression and proteolyzed activation of caspase-7 in surgical samples of Ewing's tumors.

Materials and methods

Cell culture

Four cell lines derived from Ewing's sarcoma were used in this study: (1) wild-type p53 EW-7 cells, primary tumor localized on scapula; (2) p53 mutated EW-1 cells, primary tumor localized on rib; (3) wild-type p53 COH cells, metastatic tumor localized on femur; (4) p53 mutated ORS cells from unknown localization (Kovar et al., 1993; Hamelin et al., 1994). ET cell lines were grown (7% CO2) in RPMI 1640 (Gibco-BRL) supplemented with 10% heat-inactivated FCS (Myoclone Plus, Gibco-BRL).

Reagents

rHuIFN-α2a (sp. act. 2×108 U/mg protein) was provided by Hoffmann-La Roche (Basel, Switzerland); rHuIFN-β-1a (sp. act., 4×108 U/mg protein) was a gift from Ares-Serono International (Geneva, Switzerland). The biological activity of both IFNs was verified by antiviral protection against Vesicular Stomotitis Virus on human Wish cells (ATCC, CCL-25); as expected, IFN-α2a and IFN-β showed a similar antiviral activity on these cells. Polyclonal antibodies against human Tyk-2, Jak-1, Stat-1 (p91-p83), ISGF3γ (p48), IRF-1, Bcl-2, caspase-1 (20 kDa), p38 and p-p38 proteins were from Santa Cruz Biotechnology, Inc. (Tebu, France). Monoclonal antibodies against Stat-1/p91, Stat-2/p113, Cip1/WAF1/p21, phosphotyrosine (PY20) were from Transduction Laboratories (Becton Dickinson). Monoclonal antibodies against pS727-STAT-1, pY701-Stat-1 were from Upstate Biotechnology. Monoclonal antibodies against phosphoserine was from Sigma. Purified mouse antibodies against Caspase-7 was from Pharmingen (Becton Dickinson). Monoclonal mouse anti-actin, clone C4 was from ICN. Renaissance Enhanced Luminol Reagent was from NENTM Life Science Products. P38 MAPkinase inhibitor SB203580 was from Alexis Biochemicals.

Transient transfections

2×107 ET cells mixed gently with 15 μg of dried supercoiled 2′5′oligo(A)synthetase promoter-plasmid DNA (Benech et al., 1987) and 5 μg of pSVβ-galactosidase control plasmid as an internal reference (Promega) and were electroporated as previously described (Sancéau et al., 1995). Cells were stimulated 16 h after electroporation with IFN-α2a (500 U/ml) or with IFN-β (500 U/ml), in fresh RPMI 1640/10% FCS. After stimulation of 18 or 36 h, cells were analysed for chloramphenicol acetyl transferase (CAT) and β-galactosidase activities, essentially as previously (Sancéau et al., 1995). Plasmids DNA cloned in DH5α competent cells (Gibco-BRL) were prepared with the Quantum prep. Kit (Bio Rad).

Nuclear extracts and electrophoretic mobility shift assay (EMSA)

Nuclear extracts and EMSA were performed essentially as previously described (Sancéau et al., 1995). Nuclear extracts, prepared by lysis 40×106 cells, were submitted to EMSA experiments (Sancéau et al., 1995). Nuclear proteins (10 μg) was incubated with radiolabeled probe (20 000 c.p.m.). After resolving of the nucleoprotein complexes by a non-denaturing electrophoresis, the gel was dried and exposed over-night to a PhosphoImager screen (Molecular Dynamics, Sunnyvale, CA, USA). For competition experiments, a 400-fold molar excess of the unlabeled oligonucleotides was added 15 min before incubation of nuclear extracts with the end-labeled oligonucleotide, while antisera were mixed directly with nuclear extracts and binding buffer 30 min before adding radiolabeled probe. The synthetic oligo nucleotides covering the IFN-stimulated-response-element (ISRE) of 2′5′oligoAsynthetase 5′-ctagaGATCCATGCCTCGGGAAAGGGAAACCGAAACTGAAGCCt-3′ was used (Benech et al., 1987; Sancéau et al., 1995).

Western blot analysis and immunoprecipitation analysis

After stimulation with IFN-α (500 U/ml) or with IFN-β (500 U/ml) for indicating times, cells (10×106 cells) were lysed in 100 μl of RIPA buffer (150 mM NaCl, 50 mM Tris pH 7.4, 1 mM EDTA, 0.1% Np40 (v/v), 0.5% Na deoxycholate (w/v), 0.1% SDS (w/v), 20 mM of anti-phosphatase mixture (sodium orthovanadate, NaF, β-glycerophosphate) 0.2 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg/ml of protease inhibitors (pepstatin, leupeptin, aprotinin). After successive drawing up through 26G needles, extracts were centrifuged for 10 min at 12 000 r.p.m. Proteins content in the supernatant was determined using Bradford's method. Protein samples (50 μg) were mixed with 20 μl of 2% SDS, 30% glycerol, 150 mM KCl, 10 mM Tris-HCl pH 6.7, 200 mM β mercapto-ethanol (10 min at 95°C), and submitted to Western blot analysis. For protein detection, blots were first blocked in TBS-T (125 mM NaCl, 25 mM Tris pH 8, 0.1% Tween 20) containing 5% nonfat-milk. Milk saturated blots were incubated with primary antibodies over-night at 4°C (1 : 1000 dilution, in TBS-T containing 5% nonfat-milk), washed with TSB-T, and then incubated with secondary antibodies peroxidase-conjugated goat anti-mouse (or anti-rabbit) immunoglobulins (Dako, Denmark). Enhanced chemiluminescence protein detection (Renaissance Enhanced Luminol Reagent, NEN) was performed according to the manufacturer's procedures. Immunoprecipitation experiments were performed as previously described (Sancéau et al., 1995).

For coimmunoprecipitation experiments, 40×106 ET cells were lysed in 400 μl of 10 mM NaCl, 10 mM Tris pH 7.4, 1 mM EDTA, 20 mM of anti-phosphatase mixture, 0.2 mM PMSF, 10 μg/ml of protease inhibitors, 0.5% Brij35 (Sigma). After incubation of 30 min on ice, cellular extracts were centrifuged for 15 min at 12 000 r.p.m. (4°C). Protein samples (800 μg) were diluted for 1 ml of 150 mM NaCl, 10 mM Tris pH 7.4, 0.5 mM EGTA, 1 mM EDTA, 0.5 mM PMSF, 20 mM of anti-phosphatase mixture and 10 μg/ml protease inhibitors. After incubation with antibodies against Stat-1 (6 h at 4°C), 8 mg of Protein-G-Sepharose (Pharmacia) was added and gently rocked overnight at 4°C. The protein-G-Sepharose immuno-complexes were successively washed five times with 150 mM NaCl, 10 mM Tris pH 7.4, 0.5 mM EGTA, 1 mM EDTA. All washes were performed with buffers containing the anti-proteases and anti-phosphatases mixtures. Proteins were eluted in 50 μl of 2% SDS, 30% glycerol, 150 mM KCl, 10 mM Tris-HCl pH 6.7, 200 mM β mercapto-ethanol (10 min at 95°C), and specific immuno-protein-complexes were separated by SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane (0.22 μM, BA85 Schleicher & Schuell) in refrigerated conditions (200 mA, 3 h). Protein detection was performed as described above. Membrane was incubated with primary monoclonal antibodies against Stat-1 (p91) overnight at 4°C, (1 : 1000 dilution), in TBS-T containing 5% nonfat-milk. After successive dehybridizations, the membrane was reprobed with antibodies against Stat-2, against phosphotyrosine and finally against phosphoserine.

Generation of IRF-1 cell lines

Construction of CMVt-rtTA/IRF-1 plasmid was already described (Nguyen et al., 1997b). Plasmid CMVt-rtTA was introduced into the four ET cell lines by electroporation as described above. After 48 h, cells were selected in RPMI 1640 supplemented with 10% FCS and containing 2.5 μg/ml puromycin (Sigma) for 14 days. Resistant cells carrying the CMVt-rtTA plasmid (rtTA-cells) were transfected with CMVt-IRF-1 plasmid. After 48 h cells were selected with 2.5 μg/ml puromycin and 1.5 mg/ml G418 (Life Technologies, Inc.) for 18 days. To characterize rtTA/IRF-1 expressing cells (ET-TA/IRF-1), some cellular clones were cultured in the presence of 2 μg/ml doxycycline (Dox-Sigma). Cells were harvested after 24, 48 or 72 h of Dox induction. Total cell extracts were performed as described above, and IRF-1 protein was evaluated by Western blot analysis using polyclonal antibodies directed against human IRF-1. The cellular clones of each ET cell line expressing the higher level of IRF-1 protein were selected for subsequent experiments.

Northern blot analysis, and cDNA probe

Total RNA (20 μg) were submitted to Northern blot analysis as previously described (Sancéau et al., 1995). The 2′5′oligo(A)synthetase cDNA (Benech et al., 1987) was labeled using the Redi-prime random primer labeling kit (Amersham) using α-32P-dCTP (DuPont NEN).

RNase protection assay

Total RNA samples (2 μg) were analysed for the presence of transcripts of mRNAs related to apoptosis. An hApo-1 Multi-Probe Template Set including probes for caspases-1, -2, -3, -5, -6, -7, -8, and -9, and granzyme B was purchased from Pharmingen. L32 and GAPDH were included as internal controls. RNA protection assays were performed with the RiboQuant TM Ribonuclease Protection Assay (RPA) Kit (Pharmingen) according to the manufacturer's recommendations. Protected transcripts were separated by denaturing polyacrylamide gels. The gels were dried and exposed over-night to a PhosphoImager screen (Molecular Dynamics, Sunnyvale, CA, USA).

Analysis for apoptosis

For quantification of apoptotic cells, the membrane-permenant bisbenzimide dye Hoechst 33342 (Sigma) was used to stain nuclei. A concentrated stock solution (2 mg/ml) was made up in water. After incubation either with IFNs or with Doxycyclin (1 μg/ml or 2 μg/ml) for indicated times, cells (1×105) were washed with phosphate buffer, and were fixed in 100 μl 1% formaldehyde, 0.2% glutaraldehyde in a phosphate buffer for 30 min with gentle mixing. After centrifugation (5 min, 100 g), cells were washed twice with phosphate buffer. Nuclear fragmentation was evaluated by Hoechst staining (60 min incubation in dark, 30 μl Hoechst 2 μg/ml) and quantified by counting the apoptotic cells in the total populations of untreated and treated cells respectively by luminescence microscopy. Cells were photographed at ×630 magnitude.