Report

Requirement of JNK for Stress- Induced Activation of the Cytochrome c-Mediated Death Pathway

See allHide authors and affiliations

Science  05 May 2000:
Vol. 288, Issue 5467, pp. 870-874
DOI: 10.1126/science.288.5467.870

Abstract

The c-Jun NH2-terminal kinase (JNK) is activated when cells are exposed to ultraviolet (UV) radiation. However, the functional consequence of JNK activation in UV-irradiated cells has not been established. It is shown here that JNK is required for UV-induced apoptosis in primary murine embryonic fibroblasts. Fibroblasts with simultaneous targeted disruptions of all the functional Jnkgenes were protected against UV-stimulated apoptosis. The absence of JNK caused a defect in the mitochondrial death signaling pathway, including the failure to release cytochrome c. These data indicate that mitochondria are influenced by proapoptotic signal transduction through the JNK pathway.

The development and maintenance of healthy tissues involves apoptosis, a program of physiologically regulated cell death (1). Dysregulated apoptosis contributes to many pathologies, including tumor promotion, autoimmune and immunodeficiency diseases, and neurodegenerative disorders (2). Therefore, signaling pathways that trigger apoptosis are of intense interest. The c-Jun NH2-terminal kinase (JNK) signaling pathway is essential for neuronal apoptosis in response to excitotoxic stress (3). However, the role of JNK in the apoptotic responses of other cell types is unclear. The goal of this study was to define the requirement for JNK in apoptosis of primary murine embryonic fibroblasts (MEF).

MEF were prepared from wild-type (WT) embryos and mutant embryos in which the Jnk genes were disrupted (3, 4). WT MEF expressed large amounts of 46-kD JNK1 and 55-kD JNK2 isoforms and small amounts of 55-kD JNK1 and 46-kD JNK2 isoforms (Fig. 1A). The neuronal JNK isoform (JNK3) was not detected (5). The compound mutantJnk1 /− Jnk2−/− MEF did not express JNK1 or JNK2. These data indicated thatJnk1−/− Jnk2−/− MEF may lack a functional JNK signal transduction pathway. This conclusion was confirmed by analysis of JNK activity in MEF exposed to ultraviolet (UV) radiation (Fig. 1B) or treated with fetal bovine serum (Fig. 1C). Disruption of the Jnk1 and Jnk2 genes did not alter expression of related mitogen-activated protein kinases (MAPK), ERK and p38, or the JNK activators MKK4 and MKK7 (Fig. 1A). The MKK4 and MKK7 activity inJnk1−/−Jnk2−/− MEF after UV treatment was 66% and 77%, respectively, of that detected in WT MEF (5). UV-activated p38 MAPK (Fig. 1B) and serum-activated ERK (Fig. 1C) were detected in JNK-deficient cells. ERK activation was comparable in WT andJnk1−/−Jnk2−/− MEF, but a slight reduction of p38 MAPK activation was detected inJnk1−/−Jnk2−/− MEF. These data establish that Jnk1−/−Jnk2−/− MEF possess no JNK and represent a useful model for studies of the JNK signaling pathway.

Figure 1

Isolation of JNK-deficient MEF. (A) Extracts were prepared from WT,Jnk1 −/−, Jnk2 −/−, andJnk1 /− Jnk2 / MEF (25). Expression of JNK, p38, ERK, MKK4, and MKK7 was examined by protein immunoblot analysis (26). (B) Activation of the JNK and p38 MAPK in response to UV radiation (UV-C, 60 J/m2) and incubation for the indicated times were measured by in vitro kinase assays with c-Jun and ATF2 as the substrates, respectively (27). Phosphorylated c-Jun and ATF2 were detected after SDS–polyacrylamide gel electrophoresis (PAGE) by autoradiography (upper), quantitated by PhosphorImager analysis (Molecular Dynamics), and are presented in arbitrary units (lower). (C) Activation of the JNK and ERK MAPK in response to treatment with fetal bovine serum (10%) and incubation for the indicated times were measured by in vitro kinase assays with c-Jun and c-Myc as the substrates, respectively (27). Phosphorylated c-Jun and c-Myc were detected after SDS-PAGE by autoradiography (upper), quantitated by PhosphorImager analysis (Molecular Dynamics), and presented as above (lower). (D) Proliferation of MEF was examined by crystal violet staining [mean optical density (OD) at 590 nm (±SD); n = 3] after addition of 1 × 104 cells to 20-mm tissue culture dishes and culture in medium supplemented with 10% fetal calf serum. (E) Saturation growth density of MEF in different concentrations of serum was examined by crystal violet staining [mean OD at 590 nm (±SD);n = 3]. Relative cell numbers were measured at day 0 (D = 0) and after culture for 9 days (D= 9).

Jnk2−/− MEF proliferated slightly more rapidly in culture than did WT MEF (Fig. 1D). However, the saturation density of WT and Jnk2−/− MEF was similar. In contrast, Jnk1−/− MEF and compound mutantJnk1−/− Jnk2−/− MEF proliferated more slowly than did WT cells and they reached a lower saturation density (Fig. 1D). The lower saturation density of these mutant MEF was observed at both low and high serum concentrations (Fig. 1E). These data indicated that JNK was required for normal proliferation of MEF and that JNK1 may be more important than JNK2 for proliferation. A greater role for JNK1 is consistent with the observation thatJnk1 gene disruption caused a larger decrease in JNK activity than that caused by disruption of the Jnk2 gene (Fig. 1, B and C). The reduced proliferation potential of MEF without JNK is similar to MEF without c-Jun (6) and MEF with a mutation in c-Jun that eliminated the JNK phosphorylation sites (7). This phenotypic similarity indicates that defects in c-Jun phosphorylation may contribute to the reduced proliferative potential of Jnk1−/−Jnk2−/− MEF.

The JNK signaling pathway has been implicated in the apoptotic response of cells exposed to stress (8). JNK is required for stress-induced neuronal apoptosis (3, 7, 9). However, the role of JNK in apoptosis of other cell types is controversial and the conclusions of several studies indicate that JNK may not mediate apoptotic signaling (10). Therefore we examined apoptosis of JNK-deficient MEF. WT and Jnk2 −/− MEF exposed to UV exhibited decreased viability (Fig. 2A) and increased fragmentation of genomic DNA (Fig. 2B). Partial protection from UV-induced apoptosis was observed for Jnk1 −/− MEF.Jnk1−/−Jnk2−/− MEF were nearly completely protected from the effects of UV on survival (Fig. 2A), DNA fragmentation (Fig. 2B), and sub-G1 DNA content (Fig. 2C). These data indicate that JNK is required for the normal apoptotic response of fibroblasts to UV. The partial protection observed forJnk1 −/− MEF was consistent with the lower amount of JNK activity detected in these cells after exposure to UV (Fig. 1B). The protection against UV-induced apoptosis caused by JNK deficiency was observed a long time after UV exposure (Fig. 2, A and B) and also in response to high doses of UV radiation (Fig. 2D). To confirm that the protection was caused by JNK deficiency, we examined the effect of Jnk gene dosage on UV-induced apoptosis. The amount of UV-induced apoptosis of Jnk1 −/− MEF correlated with the amount of JNK expression (Fig. 2E). Treatment of MEF with a general caspase inhibitor (zVAD) prevented cell death caused by UV (Fig. 2F). In contrast, inhibitors of protein synthesis (cycloheximide) and mRNA synthesis (actinomycin D) did not inhibit UV-induced apoptosis. These data indicated that caspase activation, but not new gene expression, was required for UV-induced apoptosis. The absence of a requirement for new gene expression distinguishes the JNK-dependent UV-induced apoptosis of MEF (Fig. 2F) from JNK-dependent excitotoxic stress-induced apoptosis of neurons, which requires c-Jun phosphorylation and new gene expression (3, 7,9). We conclude that the JNK signaling pathway is required for the apoptotic response of fibroblasts to UV radiation.

Figure 2

Selective resistance of MEF to apoptotic stimuli. (A) Increased survival ofJnk1 / Jnk2 / MEF after exposure to UV-C radiation. MEF were treated without or with UV-C (60 J/m2) and incubated in medium with serum for the indicated times. Percentage of surviving cells was measured by crystal violet staining [mean OD at 590 nm (±SD); n = 3]. (B) Defective apoptotic response ofJnk1 / Jnk2 / MEF to UV-C radiation. MEF were treated without or with UV-C (60 J/m2) and incubated in medium with serum for the indicated times. Apoptosis was measured (OD) by examination of DNA fragmentation (27) (mean ± SD; n = 3). (C) Apoptotic response of MEF to treatment with UV-C (60 J/m2) was measured after incubation in medium with serum for 20 hours by flow cytometric analysis of propidium iodide staining (27). The sub-G1 population of cells was identified and is expressed as a percentage of the total number of cells (mean ± SD; n = 3). (D) Resistance ofJnk1 / Jnk2 / MEF to both low and high levels of UV-C radiation. MEF were treated without or with the indicated dose of UV-C radiation and incubated in medium with serum for 16 hours. Apoptosis was measured (OD) by examination of DNA fragmentation (mean ± SD; n = 3). (E) Apoptotic defect ofJnk1 / Jnk2 / cells is reduced by JNK expression. The effect of increasing amounts of JNK2 expression on apoptosis caused by UV-C (60 J/m2) was examined after incubation in medium with serum for 16 hours by measurement of DNA fragmentation (mean ± SD; n = 3). (F) Requirement of caspases, but not new gene expression, for UV-C-induced apoptosis. MEF were untreated (Control) or were treated with UV-C (60 J/m2; UV) or with anti-Fas (Jo2 antibody at 1 μg/ml plus cycloheximide at 0.3 μg/ml; Fas). The effect of pretreatment (2 hours) with zVAD (40 μM), zIETD (40 μM), cycloheximide (CHX; 0.3 μg/ml), or actinomycin D (ActD; 0.1 μg/ml) was examined. Cells were incubated in medium with serum for 16 hours and the amount of apoptosis was measured by analysis of DNA fragmentation (mean OD; n = 3). (G)Jnk1 / Jnk2 / MEF are selectively resistant to proapoptotic stimuli. MEF were treated (16 hours) without and with UV-C (60 J/m2; UV), anisomycin (5 μg/ml; ANISO), MMS (0.5 μM; MMS), or anti-Fas (Jo2 antibody at 1 μg/ml plus cycloheximide at 0.3 μg/ml; FAS). Apoptosis was measured (mean OD; n = 3) by examining DNA fragmentation. Amount of apoptosis measured for WT MEF exposed to each stimulus was normalized to a relative OD value of 100 (mean ± SD; n = 3).

The Jnk1−/−Jnk2−/− MEF were also defective in the apoptotic response to the genotoxin methyl methanesulfonate (MMS) (Fig. 2G). Treatment with MMS caused apoptosis of WT cells. In contrast,Jnk1−/−Jnk2−/− cells continued to proliferate after a period of growth arrest when exposed to MMS (5). TheJnk1−/−Jnk2−/− MEF were also resistant to apoptosis caused by the drug anisomycin. In contrast, normal apoptosis of both WT andJnk1 /− Jnk2−/− MEF was observed in response to activation of the Fas death signaling pathway (Fig. 2G). These data indicate that JNK deficiency caused a selective defect in the apoptotic response of MEF.

The apoptotic defect in response to UV may be the consequence of increased survival signaling inJnk1−/−Jnk2−/− MEF. Therefore we examined activation of the transcription factor NF-κB and the protein kinase Akt (PKB), two important components of signaling pathways that promote cell survival (11). UV caused a similar increase in NF-κB DNA binding activity in WT andJnk1−/−Jnk2−/− MEF (5). In contrast, UV did not increase Akt activation [monitored with an antibody that recognizes phosphoserine-473 in Akt (12)] in WT orJnk1−/−Jnk2−/− MEF (5). These data indicate that increased survival signaling may not account for the resistance of Jnk1−/−Jnk2−/− MEF to UV-induced apoptosis.

The p53 tumor suppressor is negatively regulated by c-Jun (6) and JNK (13). Changes in p53 could contribute to the UV resistance and reduced proliferation ofJnk1−/− Jnk2−/− MEF because p53 is implicated in the cellular response to genotoxic stress (14). Therefore we examined the effect of UV on p53 in WT and Jnk1−/−Jnk2−/− MEF. Exposure of WT andJnk1−/−Jnk2−/− MEF to UV caused reduced entry into S phase (Fig. 3A). Because radiation-induced growth arrest involves p53 (14), this observation indicates that p53 may function in Jnk1−/−Jnk2−/− MEF. UV increased the amount of p53 (but not p73) in WT andJnk1−/−Jnk2−/− MEF (Fig. 3B). Immunofluorescence analysis demonstrated nuclear accumulation of p53 in WT and Jnk1−/−Jnk2−/− MEF exposed to UV (5). TheJnk1−/−Jnk2−/− MEF expressed slightly more p53 than did WT MEF and correlated with increased expression of p19ARF and the cell cycle inhibitor p21 (Fig. 3B). Because p19ARF stabilizes p53 by inactivating Mdm2 (15), the increased expression of p19ARF may explain the increased amount of p53 and p53 target gene expression (p21) detected in Jnk1−/−Jnk2−/− MEF. However, increased expression of Bax, another p53 target gene, was not observed (Fig. 3B). Thus, the reduced proliferation potential ofJnk1−/− Jnk2−/− MEF may be related, in part, to changes in p19ARF, p53, and p21.

Figure 3

Similar p53 response induced by UV radiation in WT and JNK-deficient MEF. (A) UV-C radiation causes growth arrest of WT andJnk1 / Jnk2 / MEF. MEF were treated without and with UV-C radiation (60 J/m2) and incubated in medium with serum for 19 hours. Cells were pulse-labeled (1 hour) with BrdU, harvested, and examined by flow cytometry (27). The S-phase population of cells (labeled with BrdU) is presented as the percentage of total cells (mean ± SD; n = 3). (B) WT andJnk1 / Jnk2 / MEF were either untreated (0 hours) or exposed to UV radiation (60 J/m2) and incubated in culture medium for 12 or 20 hours. Expression of p73, p53, ARF, p21, Bcl2, Bax, and actin was examined by protein immunoblot analysis (26).

The caspase group of proteases is required for apoptosis (16). Therefore we examined whether JNK was required for caspase activation in response to UV. Gene knockout studies demonstrate that the initiator caspase required for Fas-induced death (caspase 8) is not required for UV-induced apoptosis (17). Thus, the caspase 8 inhibitor zIETD prevented Fas-induced, but not UV-induced, apoptosis (Fig. 2F). UV-induced cell death requires Apaf-1, the initiator caspase 9, and the effector caspase 3 (18–20). This pathway is initiated by the release of cytochrome c from the mitochondria (21). Therefore we examined the cytochrome c/Apaf-1/caspase 9/caspase 3 pathway in WT and JNK-deficient MEF. Caspase 3 activity was increased after exposure of WT MEF, but not Jnk1−/−Jnk2−/− MEF, to UV (Fig. 4A). Protein immunoblot analysis confirmed that caspase 3 was processed from the proenzyme (32 kD) to the active form (18 kD) when WT MEF, but notJnk1−/−Jnk2−/− MEF, were exposed to UV (Fig. 4B). We also examined the release of mitochondrial cytochrome c. Exposure of WT MEF to UV caused a large increase in cytoplasmic cytochrome c (Fig. 4B) and decreased mitochondrial membrane potential (Fig. 4C). In contrast, UV did not cause cytochrome c release or mitochondrial depolarization inJnk1−/−Jnk2−/− MEF (Fig. 4, B and C). These defects inJnk1−/−Jnk2−/− MEF were selective for exposure to UV because Fas-induced cytochrome c release and caspase 3 processing were normal in JNK-deficient cells (Fig. 4B). Microinjection studies demonstrated that cytoplasmic cytochrome c caused apoptosis of both WT and JNK-deficient MEF (Fig. 4D). Selective resistance of Jnk1−/−Jnk2−/− MEF to UV therefore reflects the failure of cytochrome c release from the mitochondria and caspase activation.

Figure 4

Requirement of JNK for UV-C radiation-stimulated cytochrome c release and caspase activation. (A) Caspase 3 activity is increased after exposure of WT MEF, but not Jnk1−/−Jnk2−/− MEF, to UV-C radiation. WT (red) andJnk1−/−Jnk2−/− (green) MEF were exposed to UV radiation (60 J/m2) and incubated in medium with serum for 15 hours. Caspase 3 activity was measured by incubating cells (1 hour) with the fluorogenic substrate PhiPhiLux-G1D2 and analyzing by flow cytometry (27). (B) Mitochondrial death pathway is activated by UV-C radiation in WT MEF but not inJnk1−/ −Jnk2 / MEF. Cells were exposed to UV-C radiation (60 J/m2) and incubated in medium with serum for 0, 12, and 20 hours. Effect of UV-C was compared with treatment (16 hours) with anti-Fas (Jo2 antibody at 1 μg/ml plus cycloheximide at 0.3 μg/ml; Fas). Expression of cytoplasmic cytochrome c (Cyto.C) and Bid was examined by protein immunoblot analysis. Caspase 3 was detected by protein immunoblot analysis of total cell extracts with a caspase 3 antibody that binds both procaspase 3 (p32) and activated caspase 3 (p18) (upper) and an antibody that specifically binds activated caspase 3 (p18) (lower). (C) UV-C radiation causes mitochondrial depolarization in WT MEF but not inJnk1 / Jnk2 / MEF. WT (red) andJnk1 / Jnk2 / (green) MEF were exposed to UV-C radiation (60 J/m2) and incubated in medium with serum for 20 hours. Mitochondrial membrane potential was examined with the potentiometric dye 3,3′-dihexyloxacarbocyanine iodide [DiOC6(3)] and analysis by flow cytometry (27). (D) Cytochrome c induces apoptosis in both WT MEF andJnk1 / Jnk2 / MEF. Cells were microinjected with fluorescein-conjugated dextran (3 mg/ml) and the indicated concentrations of cytochrome c (28). Apoptosis was scored 2 hours after injection. Averages of at least 150 cells from three independent determinations are shown. Error bars represent SD. Control experiments demonstrated that treatment of cells with the caspase inhibitor zVAD (40 μM) for 2 hours before microinjection reduced the cytochrome c–induced apoptosis.

There are many possible targets of the JNK signaling pathway that may affect the mitochondria, including members of the Bcl2 group of apoptotic regulatory proteins (22). Indeed, the antiapoptotic protein Bcl2 has been reported to be phosphorylated and inactivated by JNK (23). However, an electrophoretic mobility shift, characteristic of phosphorylated Bcl2, was not detected in WT MEF exposed to UV radiation (Fig. 3B), which suggests that Bcl2 phosphorylation may not mediate the proapoptotic effects of UV-activated JNK. A second potential target of JNK signaling is Bid, a proapoptotic BH3-only member of the Bcl2 group, which is proteolytically activated to generate a fragment that translocates to the mitochondria and induces cytochrome c release (24). Fas caused Bid cleavage in both WT andJnk1−/− Jnk2−/− MEF, but UV caused Bid cleavage only in WT MEF (Fig. 4B). The caspase inhibitor zVAD inhibited Bid cleavage and cytochrome c release caused by Fas, but not UV, indicating that Bid cleavage and cytochrome c release in response to UV may be caspase-independent (5). These data indicate that a BH3-only molecule (like Bid) is a possible mediator of proapoptotic signaling by JNK.

In conclusion, our study indicates that UV-induced apoptosis in fibroblasts requires JNK for cytochrome c release from the mitochondria. Other proteins required for UV-induced apoptosis include the cytochrome c effectors Apaf-1, caspase 9, and caspase 3 (18–20). Together with JNK, these proteins function as a signaling pathway that mediates UV-induced death.

  • * Present address: Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN 46285, USA.

  • To whom correspondence should be addressed. E-mail: Roger.Davis{at}umassmed.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article