Introduction

The nucleoside analogues cladribine and fludarabine are useful drugs for the treatment of lymphoid malignancies, including hairy cell and chronic lymphocytic leukemia and low-grade lymphomas. Upon cellular uptake, these drugs are metabolized into their triphosphate form and induce DNA strand breaks and apoptosis.

Two major death pathways have been identified so far. The death receptor signaling by CD95/Fas and homologous receptors such as the TRAIL receptors DR4 and DR5 is characterized by recruitment of the FADD adapter protein and caspase-8 to the activated death receptor resulting in formation of the death-inducing signaling complex (DISC) and release of activated caspase-8 at the DISC. The second pathway that is regulated by the Bcl-2 family proteins involves activation of the mitochondria. Proapoptotic signals leading to activation, conformational change and oligomerization of the Bcl-2 homologues Bax and Bak (Hemmati et al., 2002; von Haefen et al., 2002; Gillissen et al., 2003) induce cytochrome c release from the intermembrane space of the mitochondria into the cytosol (reviewed in Daniel et al., 2003). Binding of cytochrome c to the cytosolic adapter protein APAF-1 initiates apoptosome formation and recruitment of caspase-9 in a dATP/ATP-dependent fashion. Caspase-9 is activated in the apoptosome complex, at least in part through induced proximity, and mediates processing and activation of executioner caspases including caspase-3.

Previous data have suggested that cytotoxic drugs induce cell death by activation of the CD95/Fas pathway, for instance in doxorubicin-induced apoptosis of leukemic T cells (Fulda et al., 2000) and in cladribine-induced cell death (Nomura et al., 2000). Other groups, however, could not confirm these findings and established a CD95/Fas and caspase-8-independent mechanism upon activation of a nuclear stress program by genotoxic agents (Eischen et al., 1997; Engels et al., 2000; Newton and Strasser, 2000; Wieder et al., 2001; von Haefen et al., 2003).

In the case of nucleoside analogues, an additional mechanism was suggested that results in direct activation of the mitochondrial pathway. It was demonstrated that endogenous dATP can be replaced by the triphosphate metabolite of cladribine in binding to APAF-1 and formation of the apoptosome complex (Leoni et al., 1998). Moreover, a mechanism where cladribine but not fludarabine directly affects the mitochondria was proposed (Genini et al., 2000a). Another study found that cladribine mediates apoptosis induction through caspase-dependent and -independent activation of the mitochondria (Marzo et al., 2001). It is therefore still controversial how nucleoside analogues initiate and execute apoptosis.

In the present study, we demonstrate the independence of cladribine- and fludarabine-induced apoptosis from death receptor signaling by the use of FADD or caspase-8-deficient cells as well as by interference with CD95/Fas receptor/ligand interaction. In contrast, our data demonstrate that nucleoside analogues trigger the mitochondrial cell death machinery that is mediated by a Bcl-2-sensitive pathway rather than by a previously suggested effect of nucleotide metabolites on apoptosome formation.

Results

Nucleoside analogue-induced apoptosis is independent from caspase-8

Previous reports suggested an involvement of death receptor signaling and caspase-8 (Nomura et al., 2000) or a direct activation of the mitochondria or the APAF-1/apoptosome complex in nucleoside analogue-induced apoptosis (Genini et al., 2000a, 2000b). To functionally dissect the mechanism of apoptosis induced by the nucleoside analogues cladribine and fludarabine, we employed Jurkat T cells lacking caspase-8, FADD or overexpressing either Bcl-2 or a dominant-negative mutant of caspase-9. Cladribine was used in a range from 2.5 to 20 M and fludarabine in a range from 10 to 80 M, because cladribine has a lower LD₅₀ as compared to fludarabine (Genini et al., 2000a). Apoptosis was determined on the single cell level by flow cytometric measurement of DNA fragmentation. Dose–response curves confirmed an approximately 10-fold lower ED50 dose of cladribine to induce apoptosis as compared to fludarabine (Figure 1a). Both, caspase-8 -/- and wild-type Jurkat cells displayed significant DNA fragmentation after treatment with the drugs after 72 h (Figure 1a). This indicates that cell death induction by nucleoside analogues does not depend on caspase-8. To verify that apoptosis occurs in a caspase-8-independent manner, we employed zIETD-fmk, an inhibitor of caspase-8 (Figure 1b). This cell permeable peptide specifically inhibits caspase-8-like enzyme activities by irreversibly binding to the active site of caspase-8 and related caspases including caspase-10. In a control experiment, we first determined the inhibitory efficacy of zIETD-fmk (Figure 1c). Wild-type Jurkat cells were treated with an agonistic anti-CD95 antibody that crosslinks the CD95/Fas receptor and induces cell death via a caspase-8-dependent mechanism. In this setting, DNA fragmentation was almost completely inhibited when zIETD-fmk was added at a concentration of 20 M. In contrast, zIETD-fmk did not prevent DNA fragmentation when Jurkat cells were treated with one of the nucleoside analogues (Figure 1b). This confirms that cladribine- and fludarabine-induced apoptosis occurs in a caspase-8-independent manner.

Figure 1.

Cladribine- and fludarabine-induced apoptosis is independent from caspase-8. (a) Jurkat caspase-8-deficient (-/-) cells (filled circles) and control cells (open circles) were cultured in medium alone or in the presence of cladribine or fludarabine. (b) Caspase-8-proficient Jurkat cells were cultured in the presence (filled circles) or in the absence (open circles) of the caspase-8-like inhibitor (zIETD-fmk) at a final concentration of 20 M or DMSO at 0.25% as control. The inhibitor was added 1 h before treatment with nucleoside analogues. (c) As functional control for the inhibitor zIETD-fmk, cells were either incubated in control medium (control) or medium containing an agonistic anti-CD95 antibody (0.1 g/ml) plus F(ab')₂ fragment goat anti-mouse IgG (H+L) in the absence (minus zIETD-fmk) or in the presence of zIETD-fmk (20 M) (plus z-IETD-fmk). In the absence of zIETD-fmk, the medium contained a solvent control (0.25% DMSO). Apoptosis was determined by flow cytometric analysis of the genomic DNA fragmentation 72 h after treatment. Meanss.d. of hypodiploid (apoptotic) cells from triplicates are shown

Full figure and legend (38K)

Nucleoside analogue treatment of caspase-8 -/- and wild-type Jurkat cells led to significant breakdown of the mitochondrial membrane potential _m after 48 h (Figure 2a). Notably, caspase-8 deficiency did not prevent mitochondrial permeability transition following treatment with cladribine or fludarabine. This indicates that nucleoside analogue-induced cell death proceeds via the mitochondrial apoptosis pathway. To further substantiate these findings, we determined the influence of the caspase-8 inhibitor zIETD-fmk on _m loss in wild-type Jurkat cells after drug treatment (Figure 2b). Jurkat cells cultured in the presence of zIETD-fmk showed the same extent of _m breakdown as the cells cultured in the absence of zIETD-fmk. In contrast, Jurkat cells treated with agonistic anti-CD95 antibody as a positive control for zIETD-fmk function demonstrated a 3.5-fold decreased level of cells with low _m in the presence of zIETD-fmk (Figure 2c). This shows that activation of the mitochondria by the CD95/Fas death receptor pathway is efficiently inhibited by zIETD-fmk, whereas caspase-8 appears to be dispensable for apoptosis induction by nucleoside analogues.

Figure 2.

Breakdown of the mitochondrial membrane potential (_m) is independent from caspase-8 in cladribine- and fludarabine-induced apoptosis. (a) Induction of _m breakdown in caspase-8 -/- (filled circles) or caspase-8-proficient Jurkat cells (open circles) by cladribine or fludarabine. (b) Caspase-8-proficient Jurkat cells were treated with cladribine or fludarabine in the presence of zIETD-fmk at 20 M (filled circles) or a solvent control (0.25% DMSO, open circles). The inhibitor was added 1 h before treatment. (c) As functional control for the inhibitor zIETD-fmk, cells were either incubated in control medium (control) or medium containing an agonistic anti-CD95 antibody (0.1 g/ml) plus F(ab')₂ fragment goat anti-mouse IgG (H+L) in the absence (minus zIETD-fmk) or in the presence of zIETD-fmk (20 M) (plus z-IETD-fmk). In the absence of zIETD-fmk the medium contained a solvent control (0.25% DMSO). Loss of _m was determined 48 h after treatment with the nucleoside analogues by the use of the cationic dye JC-1 that shows loss of red fluorescence upon breakdown of the _m. Meanss.d. of cells with low _m from triplicates are shown

Full figure and legend (37K)

To further investigate the cell death induction by nucleoside analogues, cleavage of caspase-3, -9 and -8 was determined by Western blot analysis in caspase-8 -/- cells and wild-type Jurkat cells. Furthermore, release of cytochrome c from the mitochondria was studied. As shown in Figure 3, the 16/17 kDa active subunit of caspase-3 was detected in Jurkat caspase-8 -/- cells and in wild-type cells as early as 24 h after treatment with 5 M cladribine or 30 M fludarabine. Moreover, the appearance of the 37 kDa subunit of caspase-9 was observed after 24 h in both, caspase-8 -/- and in wild-type cells. Processing of caspase-8 was also detected in wild-type cells after 24 h. Cytochrome c was released after 24 h in wild type as well as in caspase-8-deficient cells and was slightly less pronounced in the caspase-8 -/- cells. This may indicate that caspase-8 takes part in a mitochondrial feedback amplification loop as observed previously in Taxol-induced apoptosis. In contrast to the Western blot data, loss or inhibition of caspase-8 did not inhibit breakdown of the mitochondrial membrane potential at all. Moreover, the slight differences in cytochrome c release are not reflected by the apoptosis data. There, no inhibition of genomic DNA fragmentation was observed in nucleoside analogue-treated cells that either lack caspase-8 or were treated with the caspase-8 inhibitor zIETD-fmk. Also in fludarabine-treated cells, processing of caspase-9 and -3 as well as cytochrome c release was detectable in both caspase-8 -/- and wild-type Jurkat cells.

Figure 3.

Western blot analysis of caspase-8 -/- versus caspase-8 wild-type Jurkat cells treated with cladribine or fludarabine. Caspase-8 wild-type and caspase-8 -/- Jurkat cells were incubated with 5 M cladribine (a) or 30 M fludarabine (b). Cells were harvested 24, 36 and 48 h after treatment. Controls (c) were incubated for 48 h in control medium. Arrows indicate the positions of the procaspases caspase-3, -9 and -8. p37 is a cleavage product of caspase-9, p20, p18, p16 are subunits of caspase-3 and p43/41 are cleavage products of caspase-8. In addition, release of cytochrome c from the mitochondria is shown. Loading was confirmed by reprobing with an antibody against actin. The asterisk denotes a nonspecific protein band

Full figure and legend (81K)

To confirm the data generated in Jurkat cells, we employed another acute lymphoblastic leukemia T-cell line MOLT-3. Measurement of the CD95 expression in MOLT-3 cells shows a very low expression level (Figure 4b). Furthermore, MOLT-3 cells did not show significant DNA fragmentation after incubation with a high dose (0.5 g/ml) of an agonistic anti-CD95 antibody (Figure 4a). Thus, MOLT-3 cells show limited sensitivity to CD95/Fas. Apoptosis signaling through either CD95 or caspase-8 does not appear to be relevant. To exclude that caspase-8 is functionally relevant to induce apoptosis by nucleoside analogues, the cells were incubated with the caspase-8 inhibitor zIETD-fmk prior to treatment. In absence and in presence of the inhibitor zIETD-fmk the cells displayed significant DNA fragmentation (Figure 4c). Moreover, breakdown of the mitochondrial membrane potential was not blocked by the inhibitor (Figure 4d). Virtually identical results were obtained in MOLT-4. These T-ALL cells showed low sensitivity to apoptosis induced upon CD95/Fas ligation and underwent identical rates of apoptosis and mitochondrial membrane potential breakdown in either the presence or the absence of zIETD-fmk when exposed to cladribine or fludarabine (data not shown).

Figure 4.

Apoptosis and breakdown of the mitochondrial membrane potential is independent from death receptor signaling in MOLT-3 cells after nucleoside analogue treatment. (a) CD95/Fas receptor expression was quantified by flow cytometry following staining of the cells with PE-conjugated anti-CD95. (b) To functionally characterize MOLT-3 cells, apoptosis was induced by the use of an agonistic anti-CD95 antibody (0.1–0.5 g/ml) plus goat anti-mouse IgG (H+L) F(ab')₂ fragments as a crosslinker. Apoptosis was determined by flow cytometric analysis of genomic DNA fragmentation 72 h after treatment. (c) MOLT-3 cells were cultured in the presence (filled circles) or in the absence (open circles) of zIETD-fmk at 20 M or DMSO at 0.25% as control. The inhibitor was added 1 h prior to treatment with nucleoside analogues. Apoptosis was determined by flow cytometric analysis of genomic DNA fragmentation 72 h after treatment. Meanss.d. of hypodiploid (apoptotic) cells from triplicates are shown. (d) Loss of _m was determined by the use of the cationic dye JC-1 48 h after treatment with the nucleoside analogues. JC-1 shows loss of red fluorescence upon breakdown of the _m. Meanss.d. from triplicates of cells with low _m are shown

Full figure and legend (44K)

Mechanism of caspase-8 processing in nucleoside analogue-induced apoptosis

These experiments established that induction of the apoptotic machinery by nucleoside analogues is functionally independent from caspase-8. Nevertheless, induction of apoptosis by nucleoside analogues in wild-type Jurkat cells led to cleavage of caspase-8. To further address the mechanism of caspase-8 processing, we performed experiments in FADD -/- Jurkat cells. As shown in Figure 5a, the lack of FADD did not affect the sensitivity to drug-induced apoptosis as determined by analysis of genomic DNA fragmentation. Both cladribine and fludarabine triggered apoptotic DNA fragmentation in 40–50% of the cells. Furthermore, breakdown of the mitochondrial membrane potential occurred to the same extent in wild-type and FADD -/- Jurkat cells following exposure to either cladribine or fludarabine (Figure 5b). Both, wild-type and FADD -/- Jurkat cells exhibited breakdown of the mitochondrial membrane potential in 60% of the cells exposed to the nucleoside analogues.

Figure 5.

Apoptosis and breakdown of the mitochondrial membrane potential (_m) induced by cladribine and fludarabine is independent from FADD. Jurkat cells deficient for FADD (filled circles) and wild-type Jurkat cells (open circles) were incubated in the presence or absence of cladribine or fludarabine. (a) Cells were analysed for DNA fragmentation 72 h after treatment with nucleoside analogues. The genomic DNA content was determined by flow cytometry. Meanss.d. from triplicates of hypodiploid (apoptotic) cells are shown. (b) Loss of _m was determined 48 h after treatment by the use of the cationic dye JC-1. Meanss.d. from triplicates of cells with low _m are shown

Full figure and legend (33K)

Western blot analyses revealed that caspase-8 processing (onset after 36 h) occurred secondary to processing of caspase-9 and -3 (onset after 24 h) in cladribine-treated Jurkat FADD -/- cells (Figure 6a). In this setting, cytochrome c release into the cytosol was still weak after 24 h, increased after 36 h and showed the highest extent at 48 h. Processing of caspase-8 occurred after 24 h in wild-type cells and was slightly more pronounced as compared to the FADD -/- cells. Furthermore, Western blot analysis of fludarabine-treated cells showed cleavage of caspase-9, -3 and -8 in both FADD -/- cells as well as in wild-type cells (Figure 6b). Independence from FADD of nucleoside analogue-induced apoptosis was confirmed in BJAB cells expressing a dominant-negative FADD mutant (FADDdn) lacking the DED death effector domain. Both, BJAB FADDdn and mock-transfected cells displayed comparable apoptosis induction upon treatment with the drugs (Figure 7).

Figure 6.

Western blot analyses of Jurkat FADD -/- cells versus wild-type cells treated with cladribine or fludarabine. FADD deficient (-/-) and wild-type Jurkat cells were treated with 5 M cladribine (a) or 30 M fludarabine (b). Cells were harvested 24, 36 and 48 h after treatment. Controls (c) were incubated 48 h in control medium. Arrows indicate the positions of the procaspases caspase-3, -9 and -8. p37 is the active subunit of caspase-9, p20, p18, p16 are subunits of caspase-3 and p43/41 are products of caspase-8. In addition, release of cytochrome c from the mitochondria is shown. Loading was controlled by reprobing with an antibody against actin

Full figure and legend (99K)

Figure 7.

Apoptosis following nucleoside analogue treatment does not depend on FADD in BJAB cells. BJAB FADDdn cells (filled circles) and control cells (open circles) were cultured in medium alone or in the presence of cladribine or fludarabine. Apoptosis was determined by flow cytometric analysis of genomic DNA fragmentation 72 h after treatment. Meanss.d. from triplicates of hypodiploid (apoptotic) cells are shown

Full figure and legend (15K)

In summary, these results together with the functional experiments described above establish that cell death induced by nucleoside analogues proceeds in a manner that is independent from critical components of the death receptor pathway, FADD and caspase-8.

To substantiate that caspase-8 processing occurs secondary to caspase-9 and -3 activation, we inhibited caspase-3-like caspases by the use of the cell permeable caspase-3 inhibitor zDEVD-fmk. Western blot analyses revealed inhibition of caspase-8 processing when zDEVD-fmk was added to fludarabine- and cladribine-treated cells (Figure 8a, b). Interestingly, cleavage of caspase-3 was not completely blocked by zDEVD-fmk and presence of the inhibitor led to the accumulation of the p20 and p18 products of caspase-3 and interfered with the formation of the p16/17 cleavage product. In contrast, FADDdn abrogates sensitivity to CD95/Fas death receptor-induced apoptosis (Wieder et al., 2001; von Haefen et al., 2003).

Figure 8.

Processing of caspase-8 after treatment with fludarabine or cladribine depends on caspase-3 activation FADD deficient (-/-) and wild-type (wt) Jurkat cells were treated with 5 M cladribine (a) or 30 M fludarabine (b) in the presence or absence of the cell permeable caspase-3 inhibitor zDEVD-fmk at a final concentration of 20 M or DMSO at 0.25% as control. The inhibitor was added 1 h prior to addition of nucleoside analogues. Cells were harvested 24, 36 and 48 h after treatment. Controls (c) were incubated for 48 h in medium alone. Arrows indicate the positions of the procaspases caspase-3 and -8. p20, p18, p16 are subunits of caspase-3 and p43/41 are products of caspase-8. Loading was controlled by reprobing with an antibody against actin

Full figure and legend (111K)

To confirm the data showing independence from death receptor signaling, we inhibited the CD95/Fas receptor/ligand interaction by the use of a blocking anti-CD95 antibody in Jurkat wild-type cells (Figure 9). To confirm the efficacy of the blocking antibody, we performed a control experiment and treated cells in the presence or the absence of the agonistic anti-CD95 antibody (Figure 9c). DNA fragmentation induced by CD95 ligation was reduced 10-fold when a blocking anti-CD95 antibody was added. In contrast, apoptotic cell death induced by exposure to nucleoside analogues could not be inhibited when CD95/Fas receptor/ligand interaction was blocked. No differences in DNA fragmentation were found between cells cultured in the presence or absence of the blocking anti-CD95 antibody (Figure 9a, b).

Figure 9.

Blocking of CD95/Fas receptor/ligand interaction does not interfere with apoptosis induced by nucleoside analogues. Jurkat cells were incubated in control medium (open circles) or with a 5 g/ml blocking anti-CD95 antibody (filled circles). Cells were treated with (a) cladribine or (b) fludarabine. The blocking anti-CD95 antibody was added 1 h prior to addition of the drugs. (c) Functional control for activity of the blocking anti-CD95 antibody (clone SM1/23, IgG_2B). Cells were incubated either in control medium (control), medium containing an agonistic anti-CD95 antibody (clone anti-APO-1 IgG3 at 0.1 g/ml) plus F(ab')₂ fragment goat anti-mouse IgG (H+L) in the absence (-Fas block) or the presence of the blocking anti-CD95 antibody (+Fas block). DNA fragmentation was determined 72 h after treatment by flow cytometric measurement of the cellular DNA content. Meanss.d. from triplicates of hypodiploid (apoptotic) cells are shown

Full figure and legend (24K)

Inhibition of nucleoside analogue-induced apoptosis by Bcl-2

We and others previously observed that caspase-8 cleavage occurs as a downstream event in the mitochondrial pathway in drug-induced apoptosis (Belka et al., 2000; Engels et al., 2000; Wieder et al., 2001). To establish whether caspase-8 processing occurs through a similar mechanism in nucleoside-induced apoptosis, that is, in a death receptor-independent manner via the mitochondrial pathway, we performed analyses in Bcl-2 overexpressing Jurkat cells. Induction of apoptosis was determined by measurement of DNA fragmentation and compared with mock-transfected cells 72 h after drug treatment. In Bcl-2 overexpressing cells, the number of hypodiploid cells was reduced by approximately 59% in cladribine-treated cells and by 50% in fludarabine-treated cells as compared to the mock-transfected cells (Figure 10). Bcl-2 inhibits the release of proapoptotic proteins from mitochondria and interferes with ion fluxes induced by proapoptotic stimuli. We therefore analysed mitochondrial membrane potential breakdown 48 h after exposure to the drugs. Figure 11 shows that Bcl-2 completely blocked mitochondrial permeability transition in fludarabine as well as in cladribine-treated Jurkat cells. In contrast, mock-transfected cells exhibited a loss of _m in up to 25% of the cladribine-treated cells and 35% of the fludarabine-treated cells.

Figure 10.

Bcl-2 overexpression inhibits apoptosis after treatment with cladribine or fludarabine. Jurkat cells overexpressing Bcl-2 or mock-transfected cells were cultured in medium alone (medium) or in the presence of cladribine or fludarabine. Apoptosis was determined after 72 h by flow cytometric measurement of cells with hypodiploid (sub-G1) DNA content. (a) Representative experiment; percentage of apoptotic cells displaying a sub-G1 DNA content is given between markers. (b) Dose response. Filled circles: Bcl-2 overexpressing cells, open circles: mock-transfected cells. Meanss.d. from triplicates are given

Full figure and legend (34K)

Figure 11.

Breakdown of the mitochondrial membrane potential after treatment with nucleoside analogues is prevented by Bcl-2 overexpression in Jurkat cells. Jurkat cells overexpressing Bcl-2 or mock-transfected cells were cultured in medium alone (medium) or in the presence of cladribine or fludarabine. Breakdown of the mitochondrial membrane potential (_m) was determined after 48 h by the use of the cationic dye JC-1. (a) Representative experiment; percentage of cells with loss of _m is given between markers. (b) Dose response. Filled circles: Bcl-2 overexpressing cells, open circles: mock-transfected cells. The experiments were repeated twice and yielded similar results

Full figure and legend (35K)

Western blot analysis showed complete inhibition of cytochrome c release in the Bcl-2 overexpressing cells upon cladribine exposure and significant reduction of cytochrome c release from the mitochondria after fludarabine treatment when compared to mock-transfected cells (Figure 12). Interestingly, caspase activation was not completely inhibited by Bcl-2 overexpression. Processing of caspase-9 was delayed in Bcl-2 overexpressing cells compared to mock-transfected cells upon treatment with cladribine (Figure 12a) or fludarabine (Figure 12b). Processing of caspase-3 to the active p16/17 kDa subunit was prevented by Bcl-2. Finally, caspase-8 activation was delayed in cladribine-treated Bcl-2 overexpressing cells, as well as in fludarabine-treated Bcl-2 overexpressing cells when compared to the mock-transfected cells. The vector control cells, but not the Bcl-2 overexpressing Jurkat cells showed a strong decrease of the caspase-8 proform at 48 h after drug exposure (Figure 12a, b). Therefore, cleavage of procaspase-8 appears to occur downstream of the mitochondria and secondary to activation of other caspases in a death receptor-independent manner.

Figure 12.

Western blot analyses in Jurkat cells overexpressing Bcl-2 versus mock-transfected cells treated with cladribine or fludarabine. Jurkat cells overexpressing Bcl-2 or mock-transfected cells were treated with 5 M cladribine (a) or 30 M fludarabine (b). Cells were harvested after 24, 36 and 48 h. Controls (c) were incubated 48 h in control medium. Arrows indicate the positions of the procaspases caspase-3, -9 and -8. p37 is the active subunit of caspase-9, p20, p18, p16 are subunits of caspase-3 and p43/41 are cleavage products of caspase-8. In addition, release of cytochrome c from the mitochondria into the cytosol is shown. Loading was confirmed by reprobing with an antibody against actin

Full figure and legend (92K)

Western blot analyses show caspase activation not only in mock-transfected cells but also in Bcl-2 overexpressing cells (Figure 12). Therefore, caspase activation was measured by the use of FAM-fluorescence-labeled peptide substrates. These fluorescent fluoromethyl ketone (fmk)-modified peptides bind covalently to active caspases. Caspase activation was determined in both, Bcl-2 overexpressing cells and mock-transfected Jurkat cells after treatment. In the mock cells, activation of caspase-8-like caspases was observed in 76.7% of the cladribine-treated cells and in 73.4% of the fludarabine-treated cells. In contrast, only 25.9% of the Bcl-2 overexpressing cells displayed caspase-8 activation upon treatment with cladribine and 11.2% of the cells upon treatment with fludarabine. Similar data were obtained for caspase-9 and -3-like enzyme activities (Figure 13). This is in accordance with the Western blot data for processing of caspase-8 and other caspases (Figure 12).

Figure 13.

Caspase activation in Bcl-2 overexpressing Jurkat cells. Jurkat cells overexpressing Bcl-2 or mock-transfected cells were treated with 5 M cladribine (a) or 30 M fludarabine (b). Caspase activation was determined 48 h after treatment by the use of the fluorescence-labeled peptide substrates. Filled bars: Bcl-2 overexpressing cells, open bars: mock-transfected cells. Meanss.d. from triplicates are given

Full figure and legend (47K)

Discussion

The adenine deoxynucleosides fludarabine and cladribine induce DNA damage and are capable of killing quiescent lymphoid cells (Bromidge et al., 1995; Huang et al., 1995). A controversial discussion about how these compounds induce apoptosis was raised and it is still a matter of debate whether death receptor signaling is functionally involved cell death initiated upon drug treatment. Involvement of death receptor signals and caspase-8 processing was demonstrated (Nomura et al., 2000), whereas other groups showed CD95-independent apoptosis after exposure of malignant T cells to nucleoside analogues (Genini et al., 2000a). Additionally, the 5'-triphosphate metabolite of cladribine, which is similar to dATP, was shown to directly bind to APAF-1 and cytochrome c resulting in caspase activation in cytoplasmic extracts (Leoni et al., 1998). Another study demonstrates that 5'-triphosphate metabolites of nucleosides are able to promote apoptosome activation in CLL extracts and in a reconstituted cell-free system (Genini et al., 2000b). Moreover, cladribine was demonstrated to disrupt the integrity of mitochondria leading to AIF release, whereas fludarabine did not affect mitochondrial function (Genini et al., 2000a).

In the present study, we investigated the mechanism of nucleoside analogue-induced death by employing various mutants and transfectants of Jurkat cells that exhibit integrity or disruption of key signaling events in apoptosis pathways. We report that apoptosis and activation of the mitochondria by fludarabine and cladribine is functionally independent of death receptor signaling. Thus, cells lacking caspase-8 or FADD showed no defective apoptosis response upon exposure to fludarabine or cladribine. Similar data were observed in BJAB cells expressing a dominant-negative FADD mutant. By the use of an inhibitor of caspase-8-like caspases, zIETD-fmk, we blocked CD95/Fas receptor-induced cell death, but not nucleoside-induced mitochondrial membrane potential breakdown and DNA fragmentation in Jurkat, MOLT-3 and MOLT-4 cells. Moreover, blocking of CD95/Fas receptor/ligand binding did not interfere with cladribine- and fludarabine-induced apoptosis in Jurkat cells. Notably, not only the CD95/Fas, but also TRAIL receptors DR4 and DR5 and the p75 TNF receptor type I utilize the FADD/caspase-8 pathway for the initiation of apoptotic cell death. Thus, nucleoside analogue-induced apoptosis proceeds in a manner that is entirely independent from death receptor signaling and FADD or caspase-8. This is corroborated by the fact that processing of caspase-9 and -3 occurred to a comparable extent in cells lacking either caspase-8 or FADD. This establishes that induction of downstream apoptosis signaling upon treatment with nucleoside analogues does not depend on the extrinsic pathway, that is, occurs in death receptor-independent fashion. Furthermore, our data demonstrate that caspase-8 processing occurs independently from the adapter protein FADD that is critically involved in DISC complex formation. Western blot analyses in FADD -/- cells showed cleavage of caspase-8 upon exposure to cladribine and fludarabine. Since caspase-8 processing was delayed and occurred secondary to cleavage of caspase-3 to active species, caspase-8 appears to be processed via a caspase-3-dependent mechanism downstream of the mitochondria, as described earlier for paclitaxel (von Haefen et al., 2003) and epirubicin-induced cell death in B-lymphoid cells (Wieder et al., 2001). Here, this was observed also for nucleoside analogues when caspase-8 processing was blocked by zDEVD-fmk, an inhibitor of caspase-3-like caspases.

In contrast, nucleoside analogue-induced apoptosis proceeds via the intrinsic, mitochondrial pathway, as Bcl-2 overexpression strongly protected cells after treatment with either fludarabine or cladribine. Likewise, breakdown of the mitochondrial membrane potential and cytochrome c release were abrogated by Bcl-2. This complete inhibition of the mitochondrial apoptosis pathway by Bcl-2 argues against an earlier assumption that nucleoside analogues could induce apoptosis by a direct damage of mitochondria (Genini et al., 2000a). Thus, it was suggested that cladribine, but not fludarabine would exert its activity through the direct activation of mitochondrial apoptosis, for instance by binding of the triphosphate metabolite of cladribine to the F₀–F₁ ATPase, which could promote the mitochondrial permeability transition. In addition, the strong inhibition of cell death by Bcl-2 raises doubts whether a direct binding of nucleoside metabolites to APAF-1 and subsequent apoptosome formation, which has been previously observed in a cell-free system (Leoni et al., 1998) and which should be insensitive to Bcl-2, is a relevant mechanism for apoptosis induction. Thus, our findings challenge these alternative mechanisms of a direct toxicity for mitochondria or a direct effect of the nucleoside metabolites on apoptosome assembly, but rather delineate that the mitochondrial pathway and release of proapoptotic factors including cytochrome c is triggered by upstream signals that can be inhibited by Bcl-2.

Interestingly, Bcl-2 overexpression abrogated activation of the mitochondrial apoptosis signal transduction machinery, but only partially inhibited caspase processing. Decreased levels of processed caspase-9, -3 and -8 were observed, whereas release of cytochrome c was almost not present in the Bcl-2 overexpressing cells as compared with the mock transfectants. This may indicate that caspase processing could be additionally initiated through a Bcl-2-insensitive pathway. There, the question arises, how these caspases are activated in Bcl-2 overexpressing cells, while mitochondria are completely protected by Bcl-2. Several explanations are possible for caspase activation in Bcl-2 overexpressing cells. First, caspases may function upstream of the mitochondria. In fact, such caspase activities being initiated upstream of the mitochondria were reported recently (Marsden et al., 2002). A similar function has been attributed to caspase-2 during nuclear stress (Lassus et al., 2002). Another possibility for the failure of Bcl-2 to completely inhibit caspase processing could reside in the existence of an alternative pathway. In fact, a third apoptosis pathway was described recently in drug-induced apoptosis that involves the endoplasmic reticulum (ER). Nevertheless, our data obtained in the Bcl-2 transfectants preclude a role for the ER in this putative alternative pathway because Bcl-2 also inhibits the apoptotic function of the ER (Rudner et al., 2002).

Furthermore, data obtained by the use of Jurkat cells expressing a dominant-negative caspase-9 provide additional support that apoptosis signaling induced by nucleoside analogues strongly depends on the mitochondrial pathway and activation of the APAF-1/procaspase-9 apoptosome (data not shown). Nevertheless, higher fludarabine concentrations (above 100 M) partially circumvented inhibition by the dominant-negative caspase-9 mutant. Again, these data might indicate the existence of an alternative pathway that triggers cell death in a caspase-9-independent manner. Moreover, these data argue against the previously suggested direct activation of the APAF-1/procaspase-9 apoptosome by adenine nucleoside analogues in intact cells.

Altogether, our findings provide novel insights into the mechanisms of apoptosis induced by adenine nucleoside analogues. In detail, we excluded that cladribine or fludarabine initiate a FADD/caspase-8-dependent extrinsic death pathway. In contrast, induction of apoptosis by adenine nucleoside analogues can be inhibited by overexpression of Bcl-2 and a dominant-negative mutant of caspase-9 and is therefore mediated by caspase-9, that is, the intrinsic pathway. The differential effect of Bcl-2 on mitochondrial membrane permeability transition and caspase activation also suggest that in particular high concentration of the nucleoside analogues could trigger a pathway that does not fully rely on the APAF-1/caspase-9 apoptosome. Such a pathway would well fit to our previous observation that fludarabine is capable of inducing cell death in B-CLL cells with loss of Bax (Bosanquet et al., 2002). A model for these pathways of cell death induction by nucleoside analogues is shown in Figure 14. The further elucidation of these signaling events is warranted, especially in view of the impressive clinical activity of these compounds and the deleterious consequences when resistance to nucleoside analogues arises (Bosanquet and Bell, 1996; Bosanquet et al., 1999).

Figure 14.

Model of apoptosis induction by nucleoside analogues. Apoptotic cell death induced by adenine nucleotide analogues proceeds through a mechanism that is independent from signaling via the CD95/Fas death receptor. Likewise, loss of caspase-8 or FADD has no impact on execution of drug-induced apoptosis. Cell death proceeds through the intrinsic, mitochondrial apoptosis pathway that is susceptible to inhibition by Bcl-2. Activation of the mitochondrial pathway results in release of cytochrome c, processing of the initiator caspase-9 and the executioner caspase-3 and ultimately results in demise of the damaged cell by apoptosis. Notably, Bcl-2 abrogated activation of the mitochondrial pathway but only partially inhibited caspase processing. This implies that nucleoside analogues might trigger apoptosis through a separate, death receptor-independent and Bcl-2 insensitive pathway. Caspase-8 is irrelevant in this pathway and does not, unlike to paclitaxel-induced apoptosis, mediate a mitochondrial feedback amplification loop. A direct activation of the APAF-1/procaspase-9 apoptosome was suggested earlier based on biochemical analyses in cell free systems. High doses of fludarabine or cladribine appear, however, to trigger apoptosis in a caspase-9-independent manner

Full figure and legend (62K)

Materials and methods

Cell culture

BJAB cells (Burkitt-like lymphoma) stably transfected with either control vector-(pcDNA3-mock-transfected) or pcDNA3-FADDdn (a dominant-negative FADD mutant lacking the N-terminal death effector domain), the acute lymphoblastic leukemic T cells, MOLT-3, MOLT-4 and Jurkat were grown in RPMI-1640 medium supplemented with 10% fetal calf serum (FCS), 0.56 g/l L-glutamine, 100 U/ml penicillin and 0.1 g/ml streptomycin (all from Invitrogen, Karlsruhe, Germany) at 37°C with 5% CO₂ in fully humidified atmosphere. Bcl-2 overexpressing and wild-type Jurkat cells were described previously (Rudner et al., 2001). FADD and caspase-8-deficient Jurkat cells as well as the corresponding parental cell line A3 were kindly provided by J Blenis (Harvard Medical School, Boston, MA, USA). Caspase-9 dominant-negative (C9DN) Jurkat cells were generated by overexpression of a dominant-negative mutant of caspase-9 (Engels et al., 2000). The caspase-9 cysteinyl mutant carries a T7 tag and Western blot analyses confirmed expression of the transgene by the use of an anti-T7 tag antibody as previously described for a caspase-8 cysteinyl mutant (Belka et al., 1999).

Antibodies and reagents

Mouse monoclonal anticytochrome c antibody (clone 7H8.2C12, IgG_2B) was from BD-Pharmingen (Heidelberg, Germany), rabbit anti-actin antibody was from Sigma-Aldrich (Taufkirchen, Germany). Both were used at a dilution of 1 : 500. Goat polyclonal antibodies against caspase-3 and -9 were from R&D Systems (Wiesbaden, Germany) and used at a dilution of 1 : 2000. Monoclonal mouse anti-caspase-8 antibody (clone 1C12, IgG₁) was purchased from Cell Signaling Technology (Beverly, MA, USA) and used at a dilution of 1 : 500. Secondary goat-anti-mouse, goat-anti-rabbit IgG and donkey-anti-goat IgG antiserum coupled to horseradish peroxidase (HRP) were from Southern Biotechnology Associates (Birmingham, AL, USA) and used at 1 : 5000, 1 : 10 000 and 1 : 2500, respectively. All antibodies used for Western blot analysis were diluted in PBS supplemented with 0.05% Tween-20 (PBS-T) and 3% nonfat dry milk. Agonistic, monoclonal anti-CD95 antibody (anti-APO-1 IgG₃) (Dhein et al., 1992) was diluted in growth medium to give a final concentration of 0.1 g/ml. F(ab')₂ fragment goat anti-mouse IgG (H+L) was from Jackson ImmunoResearch (West Grove, PA, USA). The Fas receptor blocking monoclonal antibody anti-Apo-1/Fas (clone SM1/23, IgG_2B) was purchased from Bender MedSystems (Vienna, Austria). Caspase-8 inhibitor zIETD-fmk and caspase-3 inhibitor zDEVD-fmk were from Calbiochem (Bad Soden, Germany) and dissolved in dimethyl sulfoxide (DMSO) to give a 10 mM stock solution. RNase A was from Roth (Karlsruhe, Germany). Fludarabine and cladribine were purchased from Sigma-Aldrich. PE-conjugated antibody to human CD95 was purchased from BD PharMingen (Heidelberg, Germany).

Measurement of apoptotic cell death by flow cytometry

Apoptosis was determined on the single cell level by measuring the DNA content of individual cells by flow cytometry as described (Essmann et al., 2000). Cellular DNA content was measured with a logarithmic amplification in the FL-3 channel of a FACScan flow cytometer (Becton Dickinson; Heidelberg, Germany) equipped with the CellQuest software. Data are given in percent hypoploidy (i.e. the percentage of cells with a sub-G1 DNA content), which reflects the percentage of apoptotic cells with fragmented genomic DNA.

Measurement of the mitochondrial permeability transition

Breakdown of the mitochondrial membrane potential was determined by staining of the cells with 5,5',6,6'-tetrachloro-1,1',3,3'-tetraethyl-benzimidazolylcarbocyanin iodide (JC-1; Molecular Probes, Leiden, The Netherlands) as described (Radetzki et al., 2002). Mitochondrial permeability transition, that is, mitochondria displaying a lower membrane potential (_m), was subsequently quantified by flow cytometric determination of cells with decreased red fluorescence in the FL-2 channel. Data are given in % cells with low _m.

Measurement of caspase activation

Cells were harvested and washed once in PBS 48 h after drug treatment. Caspase activation was determined by staining the cells with FAM-fluorescence-labeled fluoromethyl ketone conjugated peptides (FAM-DEVD-fmk binds to caspase-3 and -7; FAM-LETD-fmk (caspase-8); FAM-LEHD-fmk (caspase-9)) (Serotec, Oxford, UK). Equal numbers of cells (2 10⁵) were incubated with one of the conjugated peptides in 300 l PBS for 1 h at 37°C under 5% CO₂. Cells were then centrifuged at 300 g, 4°C for 5 min, washed twice with washing buffer and finally resuspended in 200 l washing buffer. Caspase activity was subsequently quantified by flow cytometric determination of cells with increased green fluorescence in the FL-1 channel. Data are given as percentages of green fluorescent cells.

Immunoblotting

Western blot analyses were performed as described (Essmann et al., 2000). Protein bands were visualized using the enhanced chemiluminescence (ECL) system and exposed to Hyperfilm (both Amersham-Buchler; Braunschweig, Germany). For control of equal protein loading, membranes were stripped in 62.5 mM Tris-HCl, pH 6.8 supplemented with 100 mM -mercaptoethanol and 2% SDS at 55°C for 5 min, washed three times with PBS-T and reprobed with an antibody against actin. For measurement of cytochrome c release, cells were collected at indicated time points by centrifugation at 300 g, 4°C for 5 min and washed with PBS. The pellet was resuspended and incubated for 5 min on ice in hypotonic lysis buffer (20 mM HEPES pH 7.4, 10 mM KCl, 2 mM MgCl₂, 1 mM EDTA). Thereafter, debris was pelleted and the supernatants were collected. The membrane fraction was removed by washing twice at 10 000 g at 4°C for 10 min and the supernant of the second centrifugation was used as cytosolic extract. After determination of the protein concentration, Western blot analyses for cytochrome c content were performed by the use of 20 g cytosolic protein per lane as described above.

References

1. Belka C, Marini P, Lepple-Wienhues A, Budach W, Jekle A, Los M, Lang F, Schulze-Osthoff K, Gulbins E and Bamberg M. (1999). Oncogene, 18, 4983–4992. | Article | PubMed | ISI | ChemPort |
2. Belka C, Rudner J, Wesselborg S, Stepczynska A, Marini P, Lepple-Wienhues A, Faltin H, Bamberg M, Budach W and Schulze-Osthoff K. (2000). Oncogene, 19, 1181–1190. | Article | PubMed | ISI | ChemPort |
3. Bosanquet AG and Bell PB. (1996). Blood, 87, 1962–1971. | PubMed | ISI | ChemPort |
4. Bosanquet AG, Johnson SA and Richards SM. (1999). Br. J. Haematol., 106, 71–77. | Article | PubMed | ISI | ChemPort |
5. Bosanquet AG, Sturm I, Wieder T, Essmann F, Bosanquet MI, Head DJ, Dorken B and Daniel PT. (2002). Leukemia, 16, 1035–1044. | Article | PubMed | ISI | ChemPort |
6. Bromidge TJ, Howe DJ, Johnson SA and Phillips MJ. (1995). Cytometry, 20, 257–260. | PubMed |
7. Daniel PT, Schulze-Osthoff K, Belka C and Guner D. (2003). Essays Biochem., 39, 73–88. | PubMed | ISI | ChemPort |
8. Dhein J, Daniel PT, Trauth BC, Oehm A, Moller P and Krammer PH. (1992). J. Immunol., 149, 3166–3173. | PubMed | ISI | ChemPort |
9. Eischen CM, Kottke TJ, Martins LM, Basi GS, Tung JS, Earnshaw WC, Leibson PJ and Kaufmann SH. (1997). Blood, 90, 935–943. | PubMed | ISI | ChemPort |
10. Engels IH, Stepczynska A, Stroh C, Lauber K, Berg C, Schwenzer R, Wajant H, Janicke RU, Porter AG, Belka C, Gregor M, Schulze-Osthoff K and Wesselborg S. (2000). Oncogene, 19, 4563–4573. | Article | PubMed | ISI | ChemPort |
11. Essmann F, Wieder T, Otto A, Muller EC, Dorken B and Daniel PT. (2000). Biochem. J., 346 (Part 3), 777–783. | Article | PubMed | ISI | ChemPort |
12. Fulda S, Strauss G, Meyer E and Debatin KM. (2000). Blood, 95, 301–308. | PubMed | ISI | ChemPort |
13. Genini D, Adachi S, Chao Q, Rose DW, Carrera CJ, Cottam HB, Carson DA and Leoni LM. (2000a). Blood, 96, 3537–3543. | PubMed | ISI | ChemPort |
14. Genini D, Budihardjo I, Plunkett W, Wang X, Carrera CJ, Cottam HB, Carson DA and Leoni LM. (2000b). J. Biol. Chem., 275, 29–34. | Article | PubMed | ChemPort |
15. Gillissen B, Essmann F, Graupner V, Starck L, Radetzki S, Dorken B, Schulze-Osthoff K and Daniel PT. (2003). EMBO J., 22, 3580–3590. | Article | PubMed | ISI | ChemPort |
16. Hemmati PG, Gillissen B, von Haefen C, Wendt J, Starck L, Guner D, Dorken B and Daniel PT. (2002). Oncogene, 21, 3149–3161. | Article | PubMed | ISI | ChemPort |
17. Huang P, Robertson LE, Wright S and Plunkett W. (1995). Clin. Cancer Res., 1, 1005–1013. | PubMed |
18. Lassus P, Opitz-Araya X and Lazebnik Y. (2002). Science, 297, 1352–1354. | Article | PubMed | ISI | ChemPort |
19. Leoni LM, Chao Q, Cottam HB, Genini D, Rosenbach M, Carrera CJ, Budihardjo I, Wang X and Carson DA. (1998). Proc. Natl. Acad. Sci. USA, 95, 9567–9571. | Article | PubMed | ChemPort |
20. Marsden VS, O'Connor L, O'Reilly LA, Silke J, Metcalf D, Ekert PG, Huang DC, Cecconi F, Kuida K, Tomaselli KJ, Roy S, Nicholson DW, Vaux DL, Bouillet P, Adams JM and Strasser A. (2002). Nature, 419, 634–637. | Article | PubMed | ISI | ChemPort |
21. Marzo I, Perez-Galan P, Giraldo P, Rubio-Felix D, Anel A and Naval J. (2001). Biochem. J., 359, 537–546. | Article | PubMed |
22. Newton K and Strasser A. (2000). J. Exp. Med., 191, 195–200. | Article | PubMed | ISI | ChemPort |
23. Nomura Y, Inanami O, Takahashi K, Matsuda A and Kuwabara M. (2000). Leukemia, 14, 299–306. | Article | PubMed |
24. Radetzki S, Kohne CH, von Haefen C, Gillissen B, Sturm I, Dorken B and Daniel PT. (2002). Oncogene, 21, 227–238. | Article | PubMed | ISI | ChemPort |
25. Rudner J, Jendrossek V and Belka C. (2002). Apoptosis, 7, 441–447. | Article | PubMed | ISI | ChemPort |
26. Rudner J, Lepple-Wienhues A, Budach W, Berschauer J, Friedrich B, Wesselborg S, Schulze-Osthoff K and Belka C. (2001). J. Cell Sci., 114, 4161–4172. | PubMed | ISI | ChemPort |
27. von Haefen C, Wieder T, Essmann F, Schulze-Osthoff K, Dorken B and Daniel PT. (2003). Oncogene, 22, 2236–2247. | Article | PubMed | ISI | ChemPort |
28. von Haefen C, Wieder T, Gillissen B, Starck L, Graupner V, Dorken B and Daniel PT. (2002). Oncogene, 21, 4009–4019. | Article | PubMed | ISI | ChemPort |
29. Wieder T, Essmann F, Prokop A, Schmelz K, Schulze-Osthoff K, Beyaert R, Dorken B and Daniel PT. (2001). Blood, 97, 1378–1387. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We wish to thank Mrs Antje and Anja Richter for their expert technical assistance. We are grateful to Dr John Blenis (Department of Cell Biology, Harvard Medical School, Boston, Massachusetts, USA) for kindly providing FADD -/- and Caspase-8 -/- Jurkat cells. AM was a recipient of a fellowship by the Deutsche Forschungsgemeinschaft. The Deutsche Krebshilfe, the Deutsche José Carreras Leukämiestiftung and the Deutsche Forschungsgemeinschaft supported this work.