Abstract
Activation of the c-Jun NH2-terminal kinase type 1 (JNK1)signaling pathway is often associated with apoptosis. In this report,we elucidated the role of this kinase in the programmed cell death induced by the nucleoside analogue 9-β-d-arabinosyl-2-fluoroadenine (F-ara-A). Treatment of ML-1 cells with 3 or 10 μm F-ara-A specifically killed cells in the S-phase of the population. Incorporation of F-ara-ATP, the nucleoside triphosphate of F-ara-A, into DNA resulted in the activation of JNK1 in a time- and dose-dependent fashion. Activation of JNK1 temporally preceded DNA fragmentation. When incorporation of F-ara-A into DNA was blocked by pretreatment of the cells with aphidicolin to inhibit DNA synthesis, neither JNK1 signaling nor apoptosis was evident. Furthermore, inhibition of JNK1 by treatment of the cells with forskolin or by pretreatment with an antisense oligonucleotide directed against JNK1 mRNA resulted in a decrease in F-ara-A-induced apoptosis. Finally, the JNK1 signaling pathway appeared to be upstream to that of the effector caspases in nucleoside analogue-induced apoptosis. Thus,our data strongly suggest that JNK1 is involved in transduction of F-ara-A-induced distress signals into an apoptotic response.
INTRODUCTION
The JNK13is a component of a sequential protein kinase cascade that is activated in response to stress (1, 2, 3). Several lines of evidence support a role for the JNK1 cascade as a signaling intermediary involved in converting cellular stress stimuli into an apoptotic response. For instance, overexpression of the constitutively activated forms of Rac1, cdc42, or mitogen-activated protein kinase kinase 1,which are upstream elements in the JNK1 pathway, correlated with an increase in apoptosis (4, 5, 6, 7). Conversely, apoptosis was inhibited by overexpression of the dominant-negative counterparts of these molecules as well as stress-activated protein kinase/Erk kinase 1 SEK1, the protein kinase immediately upstream of JNK1 (4, 5, 7, 8, 9). Other studies demonstrated that treatment of cells with antisense oligonucleotides down-regulated cellular levels of JNK1 protein and that was associated with resistance to cell death induced by topoisomerase inhibitors (10, 11). In addition, direct activation of c-jun, a major downstream target of JNK1, corresponded with an increase in cell death (12), whereas apoptosis induced by a variety of stimuli was inhibited by overexpression of the inactive dominant-negative form of c-jun (5, 13, 14, 15, 16). Thus, the JNK1 cascade appears to function as a critical effector of the cellular response to stress.
JNK1 is also activated by DNA-damaging agents such as UV and ionizing radiation (17, 18, 19, 20), whereas cell types that are resistant to the lethal effects of radiation do not exhibit JNK1 activation(18, 19). Similar activation of JNK1 has been demonstrated in cells responding to DNA adducts after treatment with the alkylating agents mitomycin C and cisplatin (21, 22) and to DNA strand breaks associated with the topoisomerase inhibitors camptothecin and etoposide (10, 21). In addition, JNK1 and p38 are activated in response to inhibition of DNA replication caused by the nucleoside analogue ara-C (23, 24, 25). Thus, diverse types of DNA damage can serve as initial signals for the activation of JNK1.
Nucleoside analogues are agents that induce selective DNA damage in that they act only upon incorporation into DNA and therefore are specific for S-phase cells (26). F-ara-A, the nucleoside of the therapeutic agent fludarabine, specifically interferes with DNA replication (27, 28) and repair (29, 30). The most extensively characterized action of fludarabine involves the misincorporation of its triphosphate F-ara-ATP into DNA, resulting in termination of DNA strand elongation (27). Furthermore,F-ara-A-terminated DNA fragments resist excision repair by the 3′-5′exonuclease activity associated with DNA polymerases (31),and ligation to adjacent DNA strands is inhibited (32),thus creating a situation recognized by the cell as DNA damage.
Although it has been demonstrated that the c-abl tyrosine kinase functions upstream to JNK1 in the cellular response to ara-C and ionizing radiation (17, 23), the precise cascade of kinases involved in the DNA damage-induced activation of JNK1 is unclear. Furthermore, the consequences of this activation on nucleoside analogue-induced cell death remain unknown. Therefore, the objective of this investigation was to assess the importance of the JNK1 signaling pathway in F-ara-A-induced apoptosis. Specifically, we sought to determine: (a) whether the JNK1 signaling cascade was activated in response to distress stimuli initiated by incorporation of F-ara-ATP into DNA; and (b) whether the JNK1 signaling pathway facilitated F-ara-A-initiated DNA damage signals into a cell death response.
MATERIALS AND METHODS
Cell Culture and Chemicals.
The ML-1 myeloid leukemia cell line was a gift from Dr. M. J. Kastan(Memphis, TN). CCRF-CEM and U937 leukemia cells were obtained from the American Type Culture Collection (Manassas, VA). All cells were cultured in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum and maintained at 37°C in 5% CO2in a fully humidified incubator. The cell doubling time was 24 h under these conditions. F-ara-A was produced by alkaline phosphatase treatment of fludarabine (Berlex Laboratories, Richmond, CA). 8-[3H]F-ara-A (specific activity, 11 Ci/mmol)was prepared by and obtained from Movarek Biochemicals, Inc. (Brea,CA). Anti-JNK1 antibodies (sc-474) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), anti-PARP antibodies from Oncogene Research Products, (Cambridge, MA), phospho-c-jun antibodies from Cell Signaling Technology (Beverly, MA), and anti-actin antibodies from Sigma Chemical Co. (St. Louis, MO). FK009 was purchased from Enzyme System Products (Livermore, CA), sense and antisense oligonucleotides to JNK1 were from Life Technologies, Inc.(Gaithersburg, MD), and the APO-DIRECT kit for TUNEL assay was from PharMingen (San Diego, CA). All other chemicals were reagent grade.
Quantitation of Intracellular F-ara-ATP and Its Incorporation into DNA.
Quantitation of cellular F-ara-ATP and its incorporation into DNA was done as described previously (27). Briefly, cells were incubated with 3 or 10 μm F-ara-A for various times,washed with cold PBS, and treated with 0.4 n perchloric acid to extract nucleotides. The intracellular F-ara-ATP was quantitated by HPLC analysis. To determine the incorporation of F-ara-A into DNA, ML-1 cells were incubated with 3 and 10 μm[3H]F-ara-A for the indicated times and washed twice with ice-cold PBS. Cells were then lysed in 10 mmTris-HCl (pH 7.4), 100 mm NaCl, 25 mm EDTA, and 0.5% SDS containing 2 mg/ml proteinase K for 12 h at 50°C,extracted with phenol and chloroform, and precipitated in ethanol. The precipitated nucleic acids were then dissolved in a final volume of 1 ml of TE buffer [10 mm Tris-HCl (pH 7.8), 1 mmEDTA]. An aliquot of the total nucleic acid sample was treated with RNase A (50 mg/ml) at 37°C for 2 h and then extracted with phenol and chloroform. DNA was precipitated twice with three volumes of ethanol, dissolved in water, and quantitated in terms of UV absorbance at 260 nm. The [3H]F-ara-AMP content in DNA was measured by liquid scintillation counting. These data were used to calculate the pmol of F-ara-AMP/mg of DNA. Furthermore, using the equation 1 × 1012 pmol = 6.02 × 1011 molecules, we calculated the number of molecules of F-ara-AMP incorporated into the DNA under experimental conditions. The percentage of cells in S-phase was determined by flow cytometry (∼40%) and then used to calculate the number of molecules of F-ara-AMP/S-phase cell in each experiment.
Pulsed-Field Gel Electrophoresis.
To detect high molecular weight DNA fragmentation, 1–2 × 106 cells were incubated as described(26) with 3 or 10 μm F-ara-A for the indicated times and then washed once with PBS (pH 7.0). The cells were embedded in 0.6% agarose plugs containing 75 mm NaCl, 5 mm EDTA, and 5 mm Tris-HCl (pH 7.8). The plugs were allowed to solidify at 4°C for 20 min and then incubated in lysis buffer containing 1% Sarkosyl, 50 mm EDTA, 50 mm Tris-HCl (pH 7.8), and 0.2 mg/ml proteinase K at 45°C for 16 h. The plugs were analyzed by pulsed-field gel electrophoresis (CHEF-DR II; Bio-Rad Laboratories, Richmond, CA) at 200 V with a switch time of 50 s for 16 h at 7°C in an electrophoresis buffer containing 50 mm Tris-borate (pH 8.2) and 1 mm EDTA. After electrophoresis, the gel was stained with ethidium bromide and photographed. Large DNA fragments were quantitated by densitometric scanning of the negative films.
TUNEL Assay.
Cells were incubated with 3 or 10 μm F-ara-A for various times, washed twice with cold PBS, fixed in 1% paraformaldehyde for 20 min on ice, and stored in 70% ethanol. The cells were analyzed for apoptosis using the APO-DIRECT system. Briefly, cells were stained to simultaneously assess DNA nicks by TUNEL assay and DNA content by propidium iodide uptake according to the manufacturer’s instructions and then analyzed by flow cytometry using an Epics Profile II flow cytometer (Coulter Electronics, Miami, FL). The percentages of TUNEL-positive and -negative cells were obtained by standard analysis techniques using Elite software (Coulter Electronics). The cell cycle profiles were obtained using Multicycle software (Phoenix Flow Systems,San Diego, CA). The TUNEL-negative regions were further divided into analysis regions to quantitate the percentage of cells in G1, S, and G2-M phases of the cell cycle.
Inhibition of JNK1 by Forskolin and Examination of Apoptotic Cell Morphology.
Forskolin has been demonstrated previously to inhibit JNK1 in a time-and dose-dependent fashion (33, 34). ML-1 cells were pretreated with 10 μm forskolin for 2 h and then treated with 10 μm F-ara-A for an additional 3.5 h. The cells were then centrifuged onto glass slides, fixed in methanol,and stained with Wright-Giemsa stain. Using a Nikon HFX-II microscope,cell morphology was examined, and the numbers of cells undergoing apoptotic changes such as nuclear condensation, fragmentation of nuclei, and rupture into debris were quantified.
Inhibition of JNK1 by Antisense Oligonucleotides and Examination of Apoptotic Morphology.
JNK1 sense [SnJNK1 (sense oligonucleotide to JNK1 mRNA),5-ATCATGAGCAGAAGCAAGCGTGAC-3] and antisense [AsnJNK1 (antisense oligonucleotide to JNK1 mRNA), 5-GTCACGCTTGCTTCTGCTCATGAT-3]oligonucleotides were synthesized under phosphorothioate-modified conditions and purified by HPLC (Life Technologies, Inc., Gaithersburg,MD). These sequences represent the amino acid codons 1 to 7 of JNK1(10). The oligonucleotides were dissolved in 30 mm HEPES (pH 7.0) and added into culture media. Continued exposure to either oligonucleotides proved nonspecifically toxic to ML-1 cells; therefore, U937 myeloid leukemia cells were pretreated with 100 μm SnJNK1 or 100 μm AsnJNK1 oligonucleotides for 72 h and then exposed to 10 μmF-ara-A for 6 h and centrifuged onto glass slides. Cells showing apoptotic morphology were scored as described above.
Protein Kinase Assays.
Cells (1 × 107) were harvested,and nuclear and cytosolic fractions were collected as described(17, 19). Nuclear JNK1 activities were immunoprecipitated by anti-JNK1 antibodies. The precipitates were washed twice with lysis buffer, twice with LiCl buffer [500 mm LiCl, 100 mm Tris-Cl (pH 7.6)] and twice with kinase buffer [20 mm 4-morpholinepropanesulfonic acid (pH 7.2), 2 mm EGTA, 10 μm MgCl2, 1 mm DTT, 0.1% Triton X-100, and 0.1 mmNa3VO4]. The pellets were mixed with 2 μg of GST-c-jun-1–79(1–79), 15 μm cold ATP,and 5 μCi of [γ-32P]ATP in 30 μl of kinase buffer for 20 min at 30°C. The reaction was terminated with an equal volume of Laemmli sample buffer, and the products were resolved by 10% SDS-PAGE and autoradiography. The relative kinase activities were normalized to the amounts of immunoprecipitated JNK1 assayed by immunoblotting and visualized by chemiluminescence. Normalized JNK1 activity in control cultures was set at a value of 1.
Immunoblot Assays.
Cell lysates (30 μg protein) were resolved by 10% SDS-PAGE and then electrophoretically transferred onto Immobilon P nitrocellulose membranes (Millipore). After blocking with 4% nonfat dry milk in PBS-T(PBS with 0.05% Tween 20) for 1 h, the membranes were probed variously with antibodies specific to JNK1, actin, phospho-c-jun or PARP for 1 h. The membranes were washed in TBS-T and incubated with horseradish peroxidase-conjugated goat antirabbit for 1 h before visualization using an enhanced chemiluminescence detection system (Amersham International). The relative expression of JNK1 was quantitated using a densitometer and normalized to the value obtained for actin within the same samples.
Analysis of DNA Fragmentation.
Cells (5 × 106) were harvested,washed, and incubated in 50 μl of 50 mm Tris-HCl (pH 8.0), 10 mm EDTA, 0.5% SDS, and 0.5 mg/ml proteinase K(Sigma) for 6 h at 50°C. The samples were then incubated with 50μl of 10 mm EDTA (pH 8.0) containing 2% (w/v)low-melting-point agarose and 40% sucrose for 10 min at 70°C. The DNA was separated in 2% agarose gels. After treatment with RNase, the gels were visualized by UV illumination.
Statistical Analyses.
Quantitative data were averaged and expressed as the mean ± SE from at least three separate experiments. Differences among groups were statistically analyzed first by one-way ANOVA and then by Bonferroni’s post hoc test. Comparisons between two experimental groups were based on the two-tailed t test. P < 0.01 was considered significant.
RESULTS
Incorporation of F-ara-AMP into DNA Results in the Apoptosis of S-phase ML-1 Cells.
Cells treated with 3 μm[3H]F-ara-A accumulated up to 20μ m [3H]F-ara-ATP over 4 h,whereas those cells incubated with 10 μm[3H]F-ara-A accumulated >35 μm[3H]F-ara-ATP over the same time (Fig. 1,A). Thus, the phosphorylation of F-ara-A to its triphosphate occurred in a time- as well as dose-dependent manner in ML-1 cells. In parallel, DNA was extracted from similarly treated cells to quantitate the amounts of F-ara-AMP incorporated into DNA and to calculate the number of analogue molecules per S-phase cell. The incorporation of F-ara-AMP into DNA was time and concentration dependent during the first 2 h (Fig. 1 B). Subsequently, similar levels were achieved in each culture, suggesting that F-ara-AMP incorporation was self-limiting.
DNA fragmentation associated with apoptosis results in part from endonuclease-mediated events that cleave DNA into nucleosomal-sized fragments of 200 bp (35) and is different from the DNA fragments that result because of incorporation of a chain-terminating analogue into DNA. The former represents a systematic destruction of cellular DNA as part of the apoptotic process, whereas the latter causes relatively few DNA breaks, which presumably serve to signal initial DNA damage. ML-1 cells were treated with 3 or 10μ m F-ara-A for various times and analyzed with the TUNEL method to detect nucleosomal length DNA fragments as a measure of apoptosis. Propidium iodide staining was used to assess the cell cycle distribution of drug-treated cells. In cultures exposed to 10μ m F-ara-A, the S-phase cells that became TUNEL positive dramatically increased after 3 h (Fig. 1,C) and approached 50% of the total population by 4 h. In contrast, it took 4 h of exposure to 3 μm F-ara-A for TUNEL-positive cells to appear. The relative constancy of the G1 and G2 populations over this relatively brief time course (data not shown) is consistent with the conclusion that S-phase cells were specifically affected by F-ara-A treatment (Fig. 1 C). Incorporation of between 3 and 6 × 105 molecules of F-ara-AMP into DNA appeared to be essential for triggering apoptosis. Thus,F-ara-A-induced damage to the DNA of actively replicating cells appeared to initiate distress signals that finally resulted in the apoptotic death specifically in the S-phase population.
Activation of JNK1 in F-ara-A-induced Apoptosis.
The effect of F-ara-A on the kinase activity associated with JNK1 was quantitated to determine the action of the analogue on this signaling pathway. There was a visible change in JNK1 phosphorylation status(Fig. 2,A, lower blot) in cells treated with 10 or 30μ m F-ara-A that was accompanied by robust activation of its kinase activity (Fig. 2, A, upper blot and B). In contrast, cells incubated with 1 or 3μ m F-ara-A showed no discernable change in electrophoretic mobility and correspondingly exhibited much less JNK1 activation at 2 h. Examination of the time course of JNK1 activation demonstrated that in cells treated with 10μ m F-ara-A (Fig. 3,A, upper blot) JNK1 activity was initiated by 1.5 h, was nearly maximal by 2 h, and was sustained thereafter. In contrast, 3 μm F-ara-A elicited initial JNK1 activation at 2.5 h, which continued to increase over time. Again, there was a visible change in the JNK1 phosphorylation status that coincided with activation of the protein kinase (Fig. 3,A, middle blot). This indicated that the temporal activation of JNK1 was commensurate with the amounts of analogue incorporated into DNA at that time. Once activated, JNK1 phosphorylates c-jun, among other targets. c-jun is phosphorylated on multiple sites, notably on Ser-63 and Ser-73, which increases its transactivating potential and DNA binding activity (36). Therefore, we assessed the levels of endogenous c-jun that was phosphorylated in response to activation of JNK1. Exponentially growing ML-1 cells demonstrated a low basal level of c-jun phosphorylation. Exposure to 3 μm F-ara-A resulted in a 2-fold increase in the levels of p-Ser-63-c-jun after 2 h, whereas cells exposed to 10 μm F-ara-A demonstrated this increase starting at 1.5 h (Fig. 3A, lower blot).
The use of high molecular weight DNA fragmentation as a measure of F-ara-A-induced apoptosis has been documented previously(26). Activation of JNK1 preceded the appearance of high molecular weight DNA fragments because the lag time between the initial activation of JNK1 and the first appearance of such fragments in cells exposed to either 3 or 10 μm F-ara-A was ∼1 h (Fig. 3 B).
Activation of JNK1 Is a Specific Signaling Response to Incorporation of F-ara-AMP into DNA.
To determine whether nucleotide analogue incorporation is required for the initiation of signaling to apoptosis, exponentially growing ML-1 cells were pretreated for 24 h with the DNA synthesis inhibitor aphidicolin prior to incubation with 10 μm F-ara-A for 2.5 h. Exposure of cells to aphidicolin inhibited DNA synthesis by>98% (data not shown). This allowed the analogue triphosphate to be formed but prevented the nucleotide analogue from being incorporated into the DNA. Consequently, the aphidicolin-pretreated cells did not activate JNK1 (Fig. 4,A, upper blot); in contrast, ML-1 cells treated with 10 μm F-ara-A alone showed a 50-fold increase in the levels of activated JNK1 (Fig. 4,A, upper blot). Blocking F-ara-A incorporation into DNA with aphidicolin also inhibited the induction of apoptosis (Fig. 4,B); in contrast, cells treated with 10μ m F-ara-A alone for 2.5 h exhibited significant levels of high molecular weight DNA fragmentation (Fig. 4 B). Thus, JNK1 activation is a specific signaling response induced by incorporation of F-ara-A into DNA.
The Role of JNK1 in Nucleoside Analogue-induced Cell Death.
Two approaches were used to determine whether activation of JNK1 is essential to F-ara-A-induced cell death. The use of forskolin to inhibit JNK1 has been demonstrated previously (33, 34).
In our first approach, ML-1 cells were pretreated with 10μ m forskolin for 2 h before incubation with 10 μm F-ara-A for an additional 2 h. Consequently, cells treated with forskolin alone and cells pretreated with forskolin before treatment with F-ara-A did not activate JNK1(Fig. 5,A). In contrast, cells exposed to F-ara-A alone showed a 35-fold increase in the levels of JNK1 activity (Fig. 5,A). Blocking F-ara-A-induced activation of JNK1 with forskolin also had a profound effect on apoptosis, as demonstrated by several parameters. As DNA fragmentation assays demonstrated, cells treated with forskolin prior to treatment with F-ara-A and cells treated with forskolin alone did not undergo DNA fragmentation (Fig. 5,B), whereas cells treated with 10 μm F-ara-A exhibited internucleosomal sized DNA fragments (Fig. 5 B).
Immunoblots from similarly treated cells revealed that cells incubated with forskolin prior to treatment with F-ara-A as well as those exposed to forskolin alone exhibited the uncleaved Mr 115,000 form of PARP (Fig. 5,C). In contrast, cells treated with F-ara-A alone exhibited the cleaved Mr 85,000 form of PARP(Fig. 5 C).
On microscopic examination of the cells for apoptotic morphology such as nuclear condensation, fragmentation of nuclei, and presence of cellular debris, 25 ± 2% of the F-ara-A-treated cells showed signs of apoptosis (Fig. 5,D) versus only 5 ± 0% of the cells pretreated with forskolin and then subsequently treated with F-ara-A (P < 0.01;Fig. 5 D).
The second approach aimed at depleting JNK1 protein levels by targeting its mRNA with antisense oligonucleotides. U937 myeloid leukemia cells were pretreated for 72 h with 100 μm sense(SnJNK1; as a nonspecific control) or antisense oligonucleotides(AsnJNK1) directed specifically against JNK1 mRNA. As immunoblot assays revealed, the total level of JNK1 protein in the cells treated with AsnJNK1 decreased by 80% when compared with the levels in lysates from either control cells or those treated with SnJNK1 (Fig. 6,A, upper gel). The levels of actin in these cells were also compared to confirm the specificity of the antisense oligonucleotide against JNK1 (Fig. 6,A, lower gel). Furthermore, immunoprecipitates from cells treated with either F-ara-A alone or with SnJNK1 and then with F-ara-A showed a 12–14-fold increase in the kinase activity of JNK1. In contrast, there was no demonstrable JNK1 activity in immunoprecipitates from cells pretreated with AsnJNK1 and then treated with 10μ m F-ara-A (data not shown). The consequences of inhibiting JNK1 activation on the morphological appearance of F-ara-A-induced apoptosis were then examined. Changes in morphology typical of cells undergoing apoptosis were seen in 21 ± 4% of the cultures treated with F-ara-A alone and in 14 ± 3% of the cells pretreated with SnJNK1 before being exposed to F-ara-A. In contrast, pretreatment of cells with AsnJNK1 reduced the incidence of F-ara-A-induced apoptosis to only 2 ± 2%(Fig. 6,B). Thus, pretreatment with AsnJNK1 inhibited F-ara-A-induced apoptosis by 88% when compared with treatment with F-ara-A alone (P < 0.01) and by 85% when compared with pretreatment with SnJNK1 and then treatment with F-ara-A(P < 0.01). Thus, the data presented in Figs. 5 and 6 indicate that inhibition of JNK1 activation caused cells to become resistant to the cell death process induced by the incorporation of F-ara-A into DNA.
Activation of JNK1 Occurs Prior to Activation of Caspases in the Apoptotic Program.
The ability of FK009, a tetrapeptide inhibitor of multiple caspases, to inhibit apoptotic death has been demonstrated in a variety of cells(37). To determine the relationship of F-ara-A-induced JNK1 activation to that of the caspases in the apoptotic cascade, ML-1 cells were pretreated with varying concentrations of FK009 for 1 h, followed and then challenged with F-ara-A for 2.5 h. As pulsed-field gel electrophoresis demonstrated, quenching of high molecular weight DNA fragments increased with the increasing concentrations of FK009 (Fig. 7,A). Under the same conditions, nuclear lysates were tested for their ability to phosphorylate GST-jun and showed no quenching of JNK1 activation (Fig. 7,B). Taken together with the finding that cells compromised in their ability to activate JNK1 did not exhibit PARP cleavage or DNA fragmentation (see Fig. 5, C and B), events that usually occur downstream of caspase activation, these results indicated that the JNK1 signaling cascade was upstream to that of the caspases in the apoptotic program.
DISCUSSION
In this report, we show that activation of JNK1 occurs in a time-and dose-dependent fashion after incorporation of a chain-terminating analogue into DNA. Treatment of cells with 10 μm F-ara-A resulted in the incorporation of 4.6 × 105 molecules of F-ara-AMP into the DNA of S-phase cells by 1 h (Fig. 1,B), and JNK1 activation followed 30 min later (Fig. 3,A). Treatment of cells with 3μ m F-ara-A produced a proportionally slower accumulation of F-ara-ATP (Fig. 1,A) and incorporation into DNA (Fig. 1,B). However, once incorporation had reached levels similar to those in cells treated with 10μ m analogue (3–6 × 105 molecules/S phase cell), JNK1 was also activated within 30 min (Fig. 3 A). Thus, there appears to be a temporal linkage between this level of analogue incorporation and the initiation of JNK1 signaling. This in turn suggests that a certain amount of drug has to be incorporated into DNA before JNK1 can be activated. Whether incorporation of the drug into specific replicons has a role in initiating the cascade of DNA damage signals that result in the activation of JNK1 remains to be determined. The time between F-ara-A addition and JNK1 activation reflects the time required to accumulate F-ara-ATP, its incorporation into DNA, and the recognition of DNA damage. Although the nature of the DNA damage is likely to involve termination of nascent DNA chains by the nucleotide analogue(27), further work will be required to identify the sensor molecules involved in the initial damage recognition process.
JNK1 is classically activated by cellular stresses that depend upon membrane-initiated signaling cascades (4, 13, 14, 38). However, because F-ara-A-induced apoptotic stimuli are generated within the nucleus (26), it was important to establish whether JNK1 activation by F-ara-A was a specific signaling response to F-ara-AMP incorporation into DNA or a generic cellular stress response. Consistent with previous reports using various nucleoside analogues in other cell lines (26), our experiments with F-ara-A demonstrated that incorporation of drug into DNA was the critical event required to initiate apoptosis. Inhibition of DNA synthesis by aphidicolin, an agent that is not incorporated into DNA and does not affect F-ara-ATP accumulation, did not activate JNK1 or induce high molecular weight DNA fragmentation. When incorporation of F-ara-A into DNA was inhibited by pretreating ML-1 leukemia cells with aphidicolin,subsequent treatment with F-ara-A failed to either activate JNK1 or induce high molecular weight DNA fragmentation (Fig. 4). Thus,incorporation of F-ara-AMP into DNA was essential for JNK1 activation.
Two lines of evidence support the conclusion that JNK1 activation is required as a facilitator during F-ara-A-induced apoptosis:
(a) we used forskolin to inhibit JNK1 in ML-1 cells. Subsequent treatment of these cells with F-ara-A failed to activate JNK1 (Fig. 5,A) or to initiate apoptosis as measured by DNA fragmentation (Fig. 5,B), by cleavage of the downstream caspase-3 target PARP (Fig. 5,C), and by scoring for apoptotic morphology (Fig. 5 D). These results support the conclusion that, after F-ara-A-induced DNA damage, a functional JNK1 signaling pathway appears to be necessary to translate distress signals into cell death. These results differ from a previous report in which cells transfected with a dominant-negative mutant of c-jun, one of several downstream targets of JNK1, did not acquire resistance to ara-C-mediated toxicity (39).
(b) The levels of JNK1 protein significantly decreased when we treated U937 cells with antisense oligonucleotides directed against JNK1 mRNA as opposed to sense JNK1 oligonucleotides (Fig. 6,A). Although lysates from the cells treated with AsnJNK1 had a residual level of JNK1 protein, subsequent treatment of these cells with F-ara-A caused no demonstrable JNK1 activation, as measured by in vitro kinase assays (data not shown). Concordantly,when the cells were challenged with F-ara-A after antisense treatment,there was a significant inhibition of F-ara-A-induced apoptosis (Fig. 6 B). In contrast, the percentage of apoptotic cells in cultures treated with sense oligonucleotides prior to the addition of F-ara-A did not differ significantly from cultures treated with F-ara-A alone (P > 0.01). Furthermore, MCF-7 cells coexpressing green fluorescent protein (GFP) as well as dominant-inactive mutant of JNK1 exhibited resistance to F-ara-A-induced apoptosis in comparison to cells expressing green fluorescent protein and the vector alone.4
Although the exact transcriptional targets of the JNK1 pathway in nucleoside-analogue induced apoptosis remain unknown, apoptosis induced by certain types of DNA damage in T-cell lines depends on the JNK1-mediated transcriptional up-regulation of the Fas ligand(40). However, in our studies, ML-1 cells treated with anti-Fas antibody did not undergo cell death (data not shown),indicating that the Fas-mediated apoptotic pathway is not active in this cell line. In addition, the susceptibility of U937 leukemia cells to certain cytotoxic drugs was shown to be independent of the Fas receptor/ligand (3). Therefore, up-regulation of Fas ligand expression is not a likely target of the JNK1 pathway during F-ara-A-induced apoptosis in these myeloid leukemia cell lines.
Caspases have roles in initiating as well as executing cellular disassembly during apoptosis (41, 42). For instance, they regulate apoptosis by specifically cleaving a variety of substrates,some of which help protect cells from apoptosis. Such substrates include the antiapoptotic proteins bcl-2 or bcl-XL (43), an inhibitor of caspase activated DNA fragmentation factor ICAD (44) and DNA-PK(45). Caspases also directly contribute to apoptosis by cleaving proteins required for the maintenance of subcellular integrity(46). In our experiments with the pan-caspase inhibitor FK009, cells treated with this agent did not exhibit high molecular weight DNA fragmentation in response to F-ara-A (Fig. 7,A). This is consistent with earlier published reports that demonstrate that DNA fragmentation factor (47, 48), a protein species that functions downstream to caspase-3, is responsible for the DNA fragmentation seen during apoptosis. However, cells pretreated with z-VAD-fmk continued to activate JNK1 in response to F-ara-A, although DNA fragmentation was blocked (Fig. 7,B). This suggests that the JNK1 cascade functions upstream of the caspases. This conclusion is supported by our finding that inhibition of JNK1 led to an inhibition of PARP cleavage in ML-1 cells (Fig. 5 C), indicating a lack of activation of caspase-dependent protein degradation.
In summary, we have shown that: (a) F-ara-A incorporation into DNA is associated with the time- and dose-dependent activation of JNK1; (b) despite its effects on deoxynucleotide triphosphate pools and RNA synthesis, incorporation of F-ara-A into DNA is essential to activate JNK1 as well as trigger apoptosis;(c) the JNK1 signaling pathway acts upstream of the caspases in F-ara-A-induced apoptosis; and (d) activation of the JNK1 signaling pathway facilitates DNA damage-initiated distress stimuli into apoptotic cell death.
Accumulation of fludarabine triphosphate, its incorporation into DNA over time, and the induction of apoptosis in ML-1 cells exposed to 3 or 10 μm F-ara-A. A, at the indicated times, F-ara-ATP was extracted from exponentially growing cells exposed to [3H]F-ara-A. Extract from 2 × 106 cells was fractionated by HPLC, and the cellular concentration of F-ara-ATP was calculated; bars, SD. B, cells were incubated with[3H]F-ara-A for the indicated times; then, the number of molecules of F-ara-AMP incorporated per S-phase cell was determined as described in “Materials and Methods.” Each point represents the mean of three independent experiments; bars, SD. C, ML-1 cells were treated with F-ara-A. Propidium iodide staining was used to distinguish cell cycle distribution. The appearance of TUNEL-positive cells was quantitated by flow cytometry. Each point represents the mean of two independent experiments; bars, SD.
Accumulation of fludarabine triphosphate, its incorporation into DNA over time, and the induction of apoptosis in ML-1 cells exposed to 3 or 10 μm F-ara-A. A, at the indicated times, F-ara-ATP was extracted from exponentially growing cells exposed to [3H]F-ara-A. Extract from 2 × 106 cells was fractionated by HPLC, and the cellular concentration of F-ara-ATP was calculated; bars, SD. B, cells were incubated with[3H]F-ara-A for the indicated times; then, the number of molecules of F-ara-AMP incorporated per S-phase cell was determined as described in “Materials and Methods.” Each point represents the mean of three independent experiments; bars, SD. C, ML-1 cells were treated with F-ara-A. Propidium iodide staining was used to distinguish cell cycle distribution. The appearance of TUNEL-positive cells was quantitated by flow cytometry. Each point represents the mean of two independent experiments; bars, SD.
Effect of varying concentrations of F-ara-A on JNK1 activity. Cells were treated with 0, 1.0, 3.0, 10, and 30μ m F-ara-A for 2 h. A, JNK1 activity was assayed by immune complex kinase assays (upper blot)using GST-jun as a substrate. The relative kinase activities obtained for each sample were quantitated using a densitometer and normalized to the densitometric value of immunoblots of JNK1. The immunoblot is shown to demonstrate comparable amounts of the immunoprecipitated kinase(JNK1; lower blot). B, fold increase in GST-jun phosphorylation. Data shown are the means of two independent experiments; bars, SD.
Effect of varying concentrations of F-ara-A on JNK1 activity. Cells were treated with 0, 1.0, 3.0, 10, and 30μ m F-ara-A for 2 h. A, JNK1 activity was assayed by immune complex kinase assays (upper blot)using GST-jun as a substrate. The relative kinase activities obtained for each sample were quantitated using a densitometer and normalized to the densitometric value of immunoblots of JNK1. The immunoblot is shown to demonstrate comparable amounts of the immunoprecipitated kinase(JNK1; lower blot). B, fold increase in GST-jun phosphorylation. Data shown are the means of two independent experiments; bars, SD.
Time course of JNK1 activation by F-ara-A. A, cells were treated with 3 and 10 μmF-ara-A and harvested at the indicated times. JNK1 activity was assayed by immune complex kinase assays using GST-jun as a substrate(upper blot), and the JNK1 protein was measured by immunoblotting (middle blot). The lower blot represents cells treated with 3 and 10 μmF-ara-A and harvested at 0, 1, 1.5, 2.0, 2.5, and 3 h. Cell lysates were immunoblotted for the levels of endogenous c-jun p-Ser(63). B, the temporal relationship between DNA damage-induced JNK1 activation and DNA fragmentation. Cells were treated with 3 μm (left panel) and 10μ m (right panel) F-ara-A for various times. The kinetics of JNK1 induction and the kinetics of the induction of high molecular weight DNA fragmentation were compared. Data shown are the means of three separate experiments; bars, SD.
Time course of JNK1 activation by F-ara-A. A, cells were treated with 3 and 10 μmF-ara-A and harvested at the indicated times. JNK1 activity was assayed by immune complex kinase assays using GST-jun as a substrate(upper blot), and the JNK1 protein was measured by immunoblotting (middle blot). The lower blot represents cells treated with 3 and 10 μmF-ara-A and harvested at 0, 1, 1.5, 2.0, 2.5, and 3 h. Cell lysates were immunoblotted for the levels of endogenous c-jun p-Ser(63). B, the temporal relationship between DNA damage-induced JNK1 activation and DNA fragmentation. Cells were treated with 3 μm (left panel) and 10μ m (right panel) F-ara-A for various times. The kinetics of JNK1 induction and the kinetics of the induction of high molecular weight DNA fragmentation were compared. Data shown are the means of three separate experiments; bars, SD.
Effect of inhibition of incorporation of F-ara-A into DNA on JNK1 activation. A, ML-1 cells were pretreated with 1 μm aphidicolin for 24 h to inhibit DNA synthesis prior to the addition of 10 μm F-ara-A for 2.5 h. Cells were then harvested and assayed for JNK1 activation. The relative kinase activities obtained for each sample were quantitated using a densitometer and normalized to the densitometric value of immunoblots of JNK1. The immunoblot is shown to demonstrate comparable amounts of the immunoprecipitated kinase (JNK1). The upper panel shows the results of typical kinase and immunoblot assays, whereas the lower panel shows quantitation from three independent experiments. Bars,SD. B, aliquots of ML-1 cells from the same experiment as above were treated as indicated and harvested for pulsed-field gel electrophoresis. The fold induction of DNA fragmentation was calculated by densitometric analysis from photographic negatives of ethidium bromide-stained gels. The untreated control values were defined as 1.
Effect of inhibition of incorporation of F-ara-A into DNA on JNK1 activation. A, ML-1 cells were pretreated with 1 μm aphidicolin for 24 h to inhibit DNA synthesis prior to the addition of 10 μm F-ara-A for 2.5 h. Cells were then harvested and assayed for JNK1 activation. The relative kinase activities obtained for each sample were quantitated using a densitometer and normalized to the densitometric value of immunoblots of JNK1. The immunoblot is shown to demonstrate comparable amounts of the immunoprecipitated kinase (JNK1). The upper panel shows the results of typical kinase and immunoblot assays, whereas the lower panel shows quantitation from three independent experiments. Bars,SD. B, aliquots of ML-1 cells from the same experiment as above were treated as indicated and harvested for pulsed-field gel electrophoresis. The fold induction of DNA fragmentation was calculated by densitometric analysis from photographic negatives of ethidium bromide-stained gels. The untreated control values were defined as 1.
Effect of inhibition of JNK1 on F-ara-A-induced apoptosis in ML-1 cells. A, cells were pretreated with 10μ m forskolin for 2 h, incubated with 10μ m F-ara-A for an additional 2 h, harvested, and assayed for JNK1 activation. B–D, cells were pretreated with 10 μm forskolin for 2 h, treated with 10μ m F-ara-A for another 3.5 h, harvested, and assayed for DNA fragmentation (B), PARP cleavage(C), and apoptotic morphology (D).
Effect of inhibition of JNK1 on F-ara-A-induced apoptosis in ML-1 cells. A, cells were pretreated with 10μ m forskolin for 2 h, incubated with 10μ m F-ara-A for an additional 2 h, harvested, and assayed for JNK1 activation. B–D, cells were pretreated with 10 μm forskolin for 2 h, treated with 10μ m F-ara-A for another 3.5 h, harvested, and assayed for DNA fragmentation (B), PARP cleavage(C), and apoptotic morphology (D).
Effect of antisense JNK1 oligonucleotide on JNK1 protein levels and on F-ara-A-induced apoptosis in U937 cells. A, cells were treated with 10 μm F-ara-A for 6 h or pretreated with 100 μm SnJNK1 or 100μ m AsnJNK1 for 72 h before being exposed to 10μ m F-ara-A for 6 h. The upper blotshows immunoblots of JNK1 demonstrating the effect of SnJNK1 and AsnJNK1 oligonucleotides on the absolute levels of JNK1 protein. The lower blot shows the levels of actin in these same lysates. B, cells were treated with 10 μmF-ara-A for 6 h, pretreated with 100 μm SnJNK1 or 100 μm AsnJNK1 alone for 72 h, or pretreated with 100 μm SnJNK1 or 100 μm AsnJNK1 for 72 h, followed by 10 μm F-ara-A for 6 h and then spread on slides by cytocentrifuge techniques and analyzed for apoptotic morphology after staining with Giemsa-Wright stain. Data shown represent the SE (bars) of 1000 cells counted in three independent experiments.
Effect of antisense JNK1 oligonucleotide on JNK1 protein levels and on F-ara-A-induced apoptosis in U937 cells. A, cells were treated with 10 μm F-ara-A for 6 h or pretreated with 100 μm SnJNK1 or 100μ m AsnJNK1 for 72 h before being exposed to 10μ m F-ara-A for 6 h. The upper blotshows immunoblots of JNK1 demonstrating the effect of SnJNK1 and AsnJNK1 oligonucleotides on the absolute levels of JNK1 protein. The lower blot shows the levels of actin in these same lysates. B, cells were treated with 10 μmF-ara-A for 6 h, pretreated with 100 μm SnJNK1 or 100 μm AsnJNK1 alone for 72 h, or pretreated with 100 μm SnJNK1 or 100 μm AsnJNK1 for 72 h, followed by 10 μm F-ara-A for 6 h and then spread on slides by cytocentrifuge techniques and analyzed for apoptotic morphology after staining with Giemsa-Wright stain. Data shown represent the SE (bars) of 1000 cells counted in three independent experiments.
Effect of inhibition of caspases on JNK1 activation and induction of high molecular weight DNA fragmentation associated with apoptosis in ML-1 cells exposed to F-ara-A. A, ML-1 cells were pretreated with 10, 30, or 100 μm FK009,followed by treatment with 10 μm F-ara-A for 2.5 h. The cells were then prepared for pulsed-field gel electrophoresis, and the fold induction of DNA fragmentation was calculated by densitometric analysis from photographic negatives of ethidium-stained gels. The untreated control values were defined as 1. B, ML-1 cells were pretreated with 100 μm FK009, followed by cotreatment with F-ara-A for 2 or 2.5 h. The relative kinase activities obtained for each sample were quantitated using a densitometer and normalized to the densitometric value of immunoblots of JNK1. Data shown are representative of two independent experiments; bars, SD.
Effect of inhibition of caspases on JNK1 activation and induction of high molecular weight DNA fragmentation associated with apoptosis in ML-1 cells exposed to F-ara-A. A, ML-1 cells were pretreated with 10, 30, or 100 μm FK009,followed by treatment with 10 μm F-ara-A for 2.5 h. The cells were then prepared for pulsed-field gel electrophoresis, and the fold induction of DNA fragmentation was calculated by densitometric analysis from photographic negatives of ethidium-stained gels. The untreated control values were defined as 1. B, ML-1 cells were pretreated with 100 μm FK009, followed by cotreatment with F-ara-A for 2 or 2.5 h. The relative kinase activities obtained for each sample were quantitated using a densitometer and normalized to the densitometric value of immunoblots of JNK1. Data shown are representative of two independent experiments; bars, SD.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Supported in part by Grant CA28596 from the Department of Health and Human Services, Grant DHP-1 from the American Cancer Society, and a research grant from Schering AG, Berlin,Germany.
The abbreviations used are: JNK1, c-Jun NH2-terminal kinase 1; ara-C,1-β-d-arabinofuranosylcytosine; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; F-ara-A,9-β-d-arabinosyl-2-fluoroadenine; FK009, z-VAD-fmk; PARP,poly(ADP-ribose) polymerase; HPLC, high-performance liquid chromatography.
D. Sampath and W. Plunkett, unpublished observations.
Acknowledgments
We are grateful to Drs. Xianjun Fang and Peng Huang for advice on experimental procedures and comments on the manuscript.