Nội dung

Chronic high-dose morphine treatment promotes SH-SY5Y cell apoptosis via c-Jun N-terminal kinase-mediated activation of mitochondria-dependent pathway Xin Lin1,*, Yu-Jun Wang2,*, Qing Li2, Yuan-Yuan Hou1, Min-Hua Hong1, Ying-Lin Cao2, Zhi-Qiang Chi1 and Jing-Gen Liu1 1 State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, China 2 School of Life Science and Biopharmaceutics, Shenyang Pharmaceutical University, China Keywords apoptosis; JNK signaling; mitochondria; morphine; ROS Correspondence J.-G. Liu, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 201203, China Fax: +86 21 50807088 Tel: +86 21 50807588 E-mail: jgliu@mail.shcnc.ac.cn *These authors contributed equally to this work (Received 4 December 2008, revised 22 January 2009, accepted 28 January 2009) doi:10.1111/j.1742-4658.2009.06938.x Chronic high doses of morphine inhibit the growth of various human cancer cell lines. However, the mechanisms by which such high-dose morphine inhibits cell proliferation and induces cell death are not fully understood. Here we show that c-Jun N-terminal kinase (JNK) plays a pivotal role in high-dose morphine-induced apoptosis of SH-SY5Y cells in a mitochondria-dependent manner. Activation of JNK by morphine led to reactive oxygen species (ROS) generation via the mitochondrial permeability transition pore, because the mPTP inhibitor cyclosporin A significantly inhibited ROS generation. ROS in turn exerted feedback regulation on JNK activation, as shown by the observations that cyclosporin A and the antioxidant N-acetylcysteine significantly inhibited the phosphorylation of JNK induced by morphine. ROS-amplified JNK induced cytochrome c release and caspase-9 ⁄ 3 activation through enhancement of expression of the proapoptotic protein Bim and reduction of expression of the antiapoptotic protein Bcl-2. All of these effects of morphine could be suppressed by the JNK inhibitor SP600125 and N-acetylcysteine. The key role of the JNK pathway in morphine-induced apoptosis was further confirmed by the observation that decreased levels of JNK in cells transfected with specific small interfering RNA resulted in resistance to the proapoptotic effect of morphine. Thus, the present study clearly shows that morphine-induced apoptosis in SH-SY5Y cells involves JNK-dependent activation of the mitochondrial death pathway, and that ROS signaling exerts positive feedback regulation of JNK activity. Opioids, in addition to their well-recognized analgesic effects, may act as modulators of cell proliferation and cell death. It has been shown that opioids can protect astrocytes from apoptosis triggered by apoptosispromoting agents [1], delay neuronal death in the avian ciliary ganglion [2], and promote the growth of tumor cells [3–5]. On the other hand, opioids have also been demonstrated to induce apoptosis of immunocytes [6,7], cancer cells [8,9], neuroblastoma cells such as SK-N-SH, NG108-15 and PC12 cells [10–12], and neuronal cells [13,14], as well as human microglia [15]. The effects of opioid-mediated cell proliferation and death appear to be dependent on the concentrations and durations employed for treatment. Growth- Abbreviations ALP, allopurionol; CsA, cyclosporin A; DCFH2-DA, 2,7-dichlorodihydrofluorescein diacetate; DPI, diphenylene iodonium; FITC, fluorescein isothiocyanate; HE, hydroethidine; JNK, c-Jun N-terminal kinase; mPFC, medial prefrontal cortex; mPTP, mitochondrial permeability transition pore; NAC, N-acetylcysteine; PI, propidium iodide; PTX, pertussis toxin; ROS, reactive oxygen species; SEM, standard error of the mean; siRNA, small interfering RNA; SRB, sulforhodamine B; VTA, ventral tegmental area. 2022 FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS X. Lin et al. promoting effects occur at low concentrations or single doses of opioids, whereas growth-inhibitory effects occur with chronic opioid treatment or relatively high in vitro concentrations [16]. For example, it was reported by previous studies that morphine and its derivatives inhibited the growth of various human cancer cells, including neuroblastoma cells, with IC50 values over the millimolar level [17,18]. For humans, high plasma concentration of opioids occur under two circumstances. One is the application of high doses of opioids for pain treatment in cancer patients. Chronic high-dose morphine therapy has been widely used for severe cancer pain in palliative care [19]. The other is the abuse of opioids. In support of this, it was reported that, in animal models of addiction, the plasma concentration of morphine is as high as 2.5 mm [20]. It is possible that such high plasma concentrations of opiate could result in neuronal toxicity and death. Indeed, accumulating evidence demonstrates that chronic exposure to morphine or other opiates leads to alterations in the morphology, structure and function of neurons in certain brain regions associated with the development of opioid dependence, such as the ventral tegmental area (VTA), nucleus accumbens, and medial prefrontal cortex (mPFC). For example, chronic morphine exposure has been found to reduce the size of neurons and spine density in the VTA [21], and decrease the number of dendritic spines and alter the complexity of dendritic branches in the nucleus accumbens and mPFC [22]. Also, reduction of the immunodensity of neurofilament proteins, the major intermediate filaments of the neuronal cytoskeleton, has been observed in the VTA of brains from chronic morphine-treated rats [23], and in the mPFC from chronic opioid abusers [24]. In addition, a significant loss of ventricular and cortical volume was found in the brains of human opioid addicts in a clinical study [25]. All of these findings suggest that the abuse of opioids may induce neuronal toxicity and affect neuronal survival. Although opioids have been clearly shown to modulate cell proliferation and cell death, the molecular mechanisms have not been fully elucidated. We recently found that morphine at relatively low concentrations was able to inhibit doxorubicin-induced apoptosis through inhibition of reactive oxygen species (ROS) accumulation and mitochondrial cytochrome c release, and blockade of nuclear factor-jB transcriptional activation, in SH-SY5Y cells [26]. The present study was undertaken to study the mechanisms by which chronic high doses of morphine inhibit cell proliferation and induce cell death in SH-SY5Y cells, which constitute a subclone derived from the parent Role of JNK signaling in morphine-induced apoptosis SK-N-SH cell line, a human neuroblastoma cell line that possesses the growth, biochemical and cytogenetic properties of neurons [27] and expresses both d-opioid and l-opioid receptors [28]. SH-SY5Y cells have been used extensively in the study of neuronal death [29–31]. Results Morphine inhibited SH-SY5Y cell proliferation and induced cell apoptosis Previous studies reported that morphine inhibited the growth of various human cancer cells, including neuroblastoma cells, with IC50 values of 2.7–8.8 mm [17,18]. To assess the effect of morphine on SH-SY5Y cell proliferation, equal numbers of cells were treated with various concentrations of morphine (0.5–4 mm) for 48 h. Cell viability was detected by the sulforhodamine B (SRB) assay. As shown in Fig. 1A, morphine (0.5–4 mm) caused dose-dependent inhibition of cell proliferation, with a significant reduction at 0.5 mm and an almost 80% reduction at 4 mm. The cell death caused by morphine could be mediated by several different mechanisms. To determine whether cell death was caused by apoptosis, an extreme consequence of neurotoxicity, we examined the apoptotic percentage of morphine-treated cells by flow cytometric analysis of permeabilized cells double stained with annexin V–fluorescein isothiocyanate (FITC) ⁄ propidium iodide (PI). SH-SY5Y cells were treated with morphine (1–4 mm) or left untreated (control) for 48 h. As shown in Fig. 1B, morphine-treated cells exhibited significantly greater apoptosis than control cells. In the presence of 4 mm morphine, the percentage of apoptotic cells reached 60%, which confirmed the results obtained using the SRB assay. To determine whether morphine mediates its effect on inhibition of cell proliferation and induction of apoptosis via an opioid receptor-related mechanism, SH-SY5Y cells were treated for 48 h with morphine in the presence of the opioid receptor antagonist naloxone (0.1 mm) or the Gi ⁄ o protein inhibitor pertussis toxin (PTX, 0.1 lgÆmL)1), and cell viability and the apoptotic percentage of morphine-treated cells were then assessed. The inhibitory effects of morphine on cell growth were not antagonized by either 0.1 mm naloxone or 0.1 lgÆmL)1 PTX (Fig. 1C). Even naloxone up to 1 mm and PTX up to 0.4 lgÆmL)1 had no effect on morphine-induced inhibition of cell proliferation (data not shown). Furthermore, the results from flow cytometric analysis and SRB assay agreed well with each other, showing that naloxone and PTX pretreat- FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS 2023 Role of JNK signaling in morphine-induced apoptosis X. Lin et al. C 140 Cell viability (%) Cell viability (%) A 120 100 80 60 40 20 0 Con 0.5 1 2 3 100 80 60 40 20 0 4 0 B 80 D 80 Apoptosis (%) Apoptosis (%) 60 40 20 Con 1 2 3 2 Concentration of Mor (mM) Concentration (mM) 0 Mor 0.1 µg·mL–1 PTX + Mor 0.1 mM Nx + Mor 120 4 Concentration (mM) 60 Mor 0.1 µg·mL–1 PTX + Mor 0.1 mM Nx + Mor 40 20 0 0 4 Concentration of Mor (mM) Fig. 1. Morphine-induced cell death in a naloxone-irreversible and PTX-irreversible manner in SH-SY5Y cells. (A) Cells were treated with various concentrations of morphine (Mor) for 48 h, and cell viability was determined using the SRB assay as described in Experimental procedures. (B) Cells were treated with various concentrations of morphine for 48 h, and then permeabilized, double stained with annexin V–FITC ⁄ PI, and detected by flow cytometry as described in Experimental procedures. (C) Cells were treated with 2 mM morphine in the absence or presence of 0.1 mM naloxone or 0.1 lgÆmL)1 PTX, as indicated, for 48 h. (D) Cells were treated with 4 mM morphine in the absence or presence of 0.1 mM naloxone or 0.1 lgÆmL)1 PTX, as indicated, for 48 h. Data are expressed as a percentage of the untreated control cell samples, and represent means ± SEMs for at least three independent experiments performed in triplicate. ment were not able to block morphine-induced apoptosis in SH-SY5Y cells (Fig. 1D), suggesting that a typical opioid receptor-related mechanism was not involved, consistent with previous findings that an opioid receptor-related mechanism is not involved in opioid-induced apoptosis of tumor cells [9,32]. Morphine treatment induced release of cytochrome c and activation of caspase-9 and caspase-3 Previous studies have shown that a mitochondriadependent pathway is implicated in morphine-induced apoptosis in neuroblastoma cells and neurons [12,14]. To determine the molecular pathway for morphineinduced apoptosis, we examined the effects of morphine treatment on the release of cytochrome c and the activation of caspase-9 and caspase-3 in SH-SY5Y cells by western blot analysis. Because activation of caspase-9 and caspase-3 is required for mitochondriadependent apoptosis [33], we first detected activation of caspase-9 and caspase-3 upon morphine treatment. Activation of caspases can be manifested either by a 2024 reduction in the amount of procaspase (inactive form) or by an increase in the amount of cleaved caspase (active form). As shown in Fig. 2A,B, a concentrationdependent decrease in procaspase-3 or procaspase-9 expression and an increase in cleaved caspase-3 expression were observed in SH-SY5Y cells treated with morphine (1–4 mm) for 48 h, indicative of caspase-9 and caspase-3 activation. Release of cytochrome c from mitochondria to the cytosol is essential for caspase-9 activation [34]. Next, we examined the effect of morphine treatment on cytochrome c release. The cytosolic fractions from cells were isolated, and the presence of cytochrome c was detected by antibody against cytochrome c. Treatment of cells with morphine (1–4 mm) for 48 h led to large amounts of cytochrome c release into the cytosol, as compared with control cells (Fig. 2C). Cytochrome c release is a common event in the cell death pathway, initiated by diverse apoptosis-inducing agents. Cytochrome c can be either a trigger or a consequence of caspase activation. Although cytochrome c release occurs mostly upstream of caspase activation, in some models of apoptosis, such as death receptor- FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS X. Lin et al. Role of JNK signaling in morphine-induced apoptosis A D B C Fig. 2. Morphine treatment induced cytochrome c release and caspase-3 and caspase-9 activation. (A–C) Morphine (Mor) dose-dependently activated caspase-3 and caspase-9, and released cytochrome c from mitochondria. Cells were treated with increasing concentrations of morphine, as indicated, for 48 h. Extracts from whole cells or the cytosol were subjected to 12% SDS ⁄ PAGE, and immunoblotted with antibodies against procaspase-3, procaspase-9, cleaved caspase-3, and cytochrome c. (D) zVAD-fmk was able to inhibit morphine-induced caspase-3 activation, but unable to suppress cytochrome c release. Cells were treated with 4 mM morphine for 24 h in the absence or presence of 12 lM zVAD-fmk, and then harvested for detection of cleaved caspase-3 and cytochrome c. (a) A representative image of immunoblots for cleaved caspase-3 and cytochrome c. (b) Densitometric analysis of changes in levels of cleaved caspase-3 and cytosolic cytochrome c. All images are representative of three independent experiments yielding similar results. Data are means ± SEMs for three independent experiments. *P < 0.05 as compared with morphine. dependent apoptosis, caspase activation is upstream of cytochrome c release [35,36]. To examine the sequence of the process of cytochrome c release and caspase activation, we examined the effect of zVAD-fmk, a broad-spectrum caspase inhibitor, on morphineinduced cytochrome c release. SH-SY5Y cells were treated with morphine alone or concomitantly with morphine and zVAD-fmk for 24 h. As shown in Fig. 2D, cotreatment of cells with zVAD-fmk led to a significant decrease in cleaved caspase-3 levels in morphine-treated cells, but had no effect on the cytosolic accumulation of cytochrome c, suggesting that active caspases are not required for cytochrome c release in morphine-induced cell death, and that cytochrome c release is upstream of caspase activation during morphine-induced apoptosis in SH-SY5Y cells. Thus, mitochondria may be direct targets of death signals initiated by morphine. The results are consistent with previous studies showing that SH-SY5Y cells do not express caspase-8 [37], an essential mediator of CD95triggered apoptosis [38]. ROS were generated in SH-SY5Y cells following morphine treatment It has been reported that ROS are implicated in the mediation of caspase-dependent cell apoptosis by promoting cytochrome c release [39]. Previous studies have also demonstrated that morphine-induced apoptosis requires the generation of ROS [40]. To investigate whether ROS generation is one of the molecular events upstream of release of cytochrome c, we detected ROS generation in response to morphine treatment for varying times. Using flow cytometry to assess ROS generation with the fluorescent indicators hydroethidine (HE) and 2,7-dichlorodihydrofluorescein diacetate (DCFH2DA), to detect O2·) and H2O2, respectively, we found that treatment of cells with 4 mm morphine led to increases in O2·) and H2O2 levels. An increase in ROS generation was detected as early as after 6 h of morphine treatment, and the maximal enhancements of O2·) and H2O2 levels were detected after 24 h of morphine treatment (Fig. 3A). N-Acetylcysteine (NAC) FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS 2025 Role of JNK signaling in morphine-induced apoptosis A C X. Lin et al. D B Fig. 3. Morphine treatment increased intercellular O2·) and H2O2 levels in SH-SY5Y cells. (A) Morphine (Mor) time-dependently induced ROS generation. Cells were treated with 4 mM morphine for the indicated time periods, and intracellular O2·) and H2O2 levels were determined by flow cytometry using 5 lM HE and 10 lM DCFH2-DA as fluorescent probes, as described in Experimental procedures. The figure is representative of four independent experiments yielding similar results. (B, C) NAC, but not DPI and ALP, suppressed morphine-induced ROS generation in SH-SY5Y cells. Cells were pretreated with 5 mM NAC, 100 lM ALP and 2 lM DPI for 30 min prior to incubation with 4 mM morphine for 6 h, or pretreated with 5 mM NAC prior to incubation with 4 mM morphine for 24 h, and intracellular H2O2 and O2·) levels were then detected. (a) A representative image of five independent experiments yielding similar results. (b) Quantification of O2·) and H2O2 generation. (D) CsA decreased morphine-induced enhancement of O2·) levels. Cells were treated with either 4 mM morphine alone or with 4 mM morphine in combination with 1 lM CsA for 6 h, and intracellular O2·) levels were then determined. Values are means ± SEMs for at least three independent experiments performed in triplicate. **P < 0.01 as compared with morphine. (5 mm), a well-characterized antioxidant, significantly inhibited the generation of O2·) in SH-SY5Y cells treated with morphine for 6 h. However, this effect of morphine could not be suppressed by diphenylene iodonium (DPI) (2 lm), an inhibitor of NADPH oxidase, and allopurionol (ALP) (100 lm), an inhibitor of xanthine oxidase (Fig. 3B), suggesting that mitochondria may be the major source of ROS. In addition, NAC (5 mm) also clearly inhibited the elevation of intracellular O2·) and H2O2 levels after 24 h of morphine treatment (Fig. 3C). It has been suggested that mitochondrial permeability transition pore (mPTP) opening increases ROS production in vivo [41] and in isolated mitochondria [42]. To determine whether ROS came from the mitochondria, we examined the effect of the mPTP inhibitor cyclosporin A (CsA) on morphineinduced ROS generation. Treatment with CsA (1 lm) alone had no effect on O2) generation, but it significantly inhibited morphine-induced O2) generation (Fig. 3D), supporting the idea that the mitochondria were the source of ROS generation. 2026 c-Jun N-terminal kinase (JNK) activation was upstream of ROS generation and was in turn regulated by ROS Studies of JNK-induced neuronal apoptosis suggest that JNK-induced phosphorylation of the transcription factor c-Jun and the consequent expression of c-Juninduced genes mediate JNK-induced apoptosis [43,44]. As ROS activate the JNK cascade [45,46], apoptosis induced by morphine may depend on activation of the JNK pathway. To determine whether morphine-induced ROS generation leads to activation of JNK and its substrate target (c-Jun) in SH-SY5Y cells, we examined the effect of morphine treatment on the phosphorylation of JNK and c-Jun by western blot analysis, using phosphospecific antibodies. As shown in Fig. 4A, the phosphorylation of JNK and the phosphorylation of c-Jun were both increased by morphine treatment (4 mm), starting at 3 h and peaking at 24 h. However, no differences in the phosphorylation of p38 were observed between controls and morphine-treated cells. FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS X. Lin et al. Role of JNK signaling in morphine-induced apoptosis A B C D E Fig. 4. JNK activation induced by morphine (Mor) was upstream of ROS generation and was in turn regulated by ROS. (A) Morphine activated the JNK ⁄ c-Jun, but not the p38, signaling pathway. Cells were treated with 4 mM morphine for increasing time periods, as indicated. Extracts from whole cells were subjected to 12% SDS ⁄ PAGE and immunoblotted with antibodies against JNK, phospho-JNK, p38, phosphop38, c-Jun, phospho-c-Jun and actin. (B) SP600125, but not SB203580, blocked morphine-induced O2) generation. Cells were treated with 4 mM morphine alone, or treated with 4 mM morphine in the presence of 20 lM SP600125 or 20 lM SB203580 for 6 h, and intracellular O2·) levels were then determined by flow cytometry. Values are means ± SEMs for at least three independent experiments performed in triplicate. **P < 0.01 as compared with morphine. (C) NAC failed to suppress phosphorylation of JNK induced by 3 h of morphine treatment. Cells were either treated with 4 mM morphine alone or treated with 4 mM morphine in combination with 5 mM NAC for 3 h. (D) NAC and SP600125 were able to inhibit phosphorylation of JNK and c-Jun induced by 12 h of morphine treatment. Cells were treated with 4 mM morphine for 12 h in the absence or presence of 5 mM NAC or 20 lM SP600125. (E) CsA abolished phosphorylation of JNK induced by 12 h of morphine treatment. Cells were treated with 4 mM morphine for 12 h in the absence or presence of 1 lM CsA. The results showed that JNK was activated as early as after 3 h of morphine treatment; this preceded ROS generation, which was detected after 6 h of morphine treatment, indicating that JNK activation was located upstream of ROS generation (Fig. 3A). To determine whether JNK activation is involved in ROS generation, we detected the effect of SP600125, a selective inhibitor of JNK that acts by binding to the ATPbinding site, on O2) generation induced by morphine treatment. Pretreatment of cells with 20 lm SP600125 effectively attenuated O2) generation (Fig. 4B), indicating that JNK activation is essential for ROS generation. Consistent with the absence of activation of p38 by morphine, the p38 mitogen-activated protein kinase inhibitor SB203580 had no effect on morphine-induced ROS generation (Fig. 4B). To obtain further evidence that JNK activation precedes ROS generation, we determined the effect of NAC on JNK activation induced by 3 h of morphine treatment, because JNK activation started at 3 h. NAC had no significant effect on the phosphorylation of JNK induced by 3 h of morphine treatment (Fig. 4C). These results support the idea that JNK activation precedes ROS generation, consistent with recent studies showing that ROS generation was regulated by the JNK signaling pathway [47,48]. To determine whether ROS exert feedback regulation on JNK activation, we next examined the effects of the antioxidant NAC and the mPTP inhibitor CsA on the phosphorylation of JNK and on its substrate transcription factor c-Jun following prolonged morphine treatment. Figure 4D shows that, in the presence of 5 mm NAC, morphine treatment for 12 h failed to induce increases in the phosphorylation of JNK and c-Jun. Additionally, in the presence of 1 lm CsA, morphine treatment for 12 h also failed to induce JNK FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS 2027 Role of JNK signaling in morphine-induced apoptosis X. Lin et al. and c-Jun phosphorylation (Fig. 4E). The effect of morphine on the phosphorylation of JNK and c-Jun could also be inhibited by SP600125, an inhibitor of JNK activation. These results support the idea that ROS exert a positive feedback effect on JNK activation after prolonged morphine treatment. Altogether, the findings suggest that morphine induces ROS accumulation by activation of the JNK pathway, and that ROS can, in turn, activate JNK in a positive feedback fashion. tion of Bim in SH-SY5Y cells. As was expected, in parallel with JNK activation, treatment with morphine resulted in enhancement of Bim expression in a timedependent and dose-dependent manner (Fig. 5A,B). Bcl-2 prevents release of cytochrome c by heterodimerizing with Bax [51,52]. Therefore, the effects of treatment with morphine on the protein levels of Bcl-2 and Bax were assessed. As shown in Fig. 5C, treatment of cells with morphine induced a robust decrease in the protein level of Bcl-2, but failed to produce a significant change in the protein level of Bax. As a result, the Bcl-2 ⁄ Bax ratio was decreased, which is indicative of an increase in mitochondrial permeability. Morphine treatment differentially regulated the expression of Bcl-2 family members Bcl-2 family members are major regulators of mitochondrial integrity and mitochondria-initiated cytochrome c release and caspase activation. The Bcl-2 family includes antiapoptotic members such as Bcl-2 and Bcl-XL, and proapoptotic members such as Bax, Bak, and Bim. Bax and Bak are potent regulators of cytochrome c release from mitochondria under a variety of stress conditions. JNK ⁄ c-Jun signaling has been implicated in the induction of the BH3-only Bcl-2 family member Bim (Bcl-interacting mediator of cell death), a key mediator of Bax-dependent cytochrome c release during neuronal apoptosis [49,50]. Next, we investigated whether morphine treatment caused induc- A Inhibitors of JNK and antioxidant suppressed morphine-induced changes in Bim and Bcl-2 expression, cytochrome c release, and caspase-3 activation The data presented above show that a positive feedback cycle operates in morphine-treated SH-SY5Y cells, in which JNK activation causes ROS generation, which, in turn, leads to further activation of JNK. We thus hypothesized that JNK activation might play a crucial role in morphine-induced cytochrome c release and consequent caspase-3 activation by increasing the expression of the BH3-only protein Bim and decreasing C B Fig. 5. Morphine (Mor) treatment differentially regulated the expression of Bcl-2 family members. (A, B) Morphine dose-dependently and time-dependently increased the expression of Bim. Cells were exposed to various concentrations of morphine for 24 h or 4 mM morphine for various times, as indicated. Extracts from cells were subjected to 12% SDS ⁄ PAGE and immunoblotted with antibody against Bim. (C) Morphine decreased Bcl-2 levels in a dose-dependent manner. Cells were treated with various concentrations of morphine for 48 h, and extracts from cells were subjected to 12% SDS ⁄ PAGE and immunoblotted with antibodies against Bax and Bcl-2. (a) Representative immunoblots for Bcl-2 and Bax. (b) Densitometric analysis of changes in levels of Bcl-2 ⁄ Bax. All images are representative of at least three independent experiments yielding similar results. Values are means ± SEMs for at least three independent experiments. 2028 FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS X. Lin et al. A Role of JNK signaling in morphine-induced apoptosis B Fig. 6. SP600125 and NAC suppressed morphine-induced changes in Bim and Bcl-2 expression, cytochrome c release and caspase-3 activation in SH-SY5Y cells. Cells were either left untreated or treated with 4 mM morphine for 24 h, in the presence or absence of 20 lM SP600125 or 5 mM NAC. Extracts from whole cells or the cytosol were subjected to 12% SDS ⁄ PAGE and immunoblotted with antibodies against Bim, Bcl-2, cytochrome c, and cleaved caspase-3. (A) A representative image of immunoblots for Bim, Bcl-2, cytochrome c, and cleaved caspase-3. (B) Densitometric analysis of changes in levels of Bim, Bcl-2, cytochrome c, and cleaved caspase-3. All images are representative of three independent experiments yielding similar results. Data are means ± SEMs for three independent experiments. *P < 0.05 as compared with morphine. Bcl-2 levels. To test this hypothesis, we determined whether the selective inhibitor of JNK SP600125 and the antioxidant NAC suppressed morphine-induced changes in Bim and Bcl-2 expression, cytochrome c release, and caspase-3 activation. As shown in Fig. 6, pretreatment of cells with 20 lm SP600125 or 5 mm NAC alone had no effects on Bim and Bcl-2 expression, cytochrome c release, and caspase-3 activation. However, in the presence of SP600125 or NAC, treatment of cells with 4 mm morphine for 24 h failed to increase Bim expression, decrease Bcl-2 levels, increase cytochrome c release, or activate caspase-3, indicating that the positive feedback amplification of JNK activation by ROS and the consequent changes in the expression of Bim and Bcl-2 may contribute to morphine-induced cytochrome c release and caspase-3 activation. SP600125, small interfering RNA (siRNA) against JNK and NAC attenuated morphine-induced apoptosis Accumulating evidence shows that activation of the JNK ⁄ c-Jun pathway with subsequent enhancement of cytochrome c release via induction of Bim expression and reduction of Bcl-2 expression are key events required for neuronal apoptosis [50,53]. In this study, we found that SP600125 could inhibit or prevent the morphine-induced occurrence of these molecular events. Therefore, the last set of experiments addressed the question of whether JNK-mediated cytochrome c release and caspase-3 activation via alteration of Bim and Bcl-2 was correlated with morphine-induced apoptosis in SH-SY5Y cells. SH-SY5Y cells were treated with 4 mm morphine for 48 h in the absence or presence of 20 lm SP600125 or SB203580, which were added 30 min before morphine. Apoptotic cells were measured by flow cytometry and phase contrast microscopy. Morphine at 4 mm caused drastic apoptosis, and this apoptosis was substantially inhibited by preincubation of the cells with the JNK inhibitor SP600125 (20 lm; Fig. 7A). The involvement of JNK in morphine-induced apoptosis was further investigated in SH-SY5Y cells that had been transiently transfected with siRNA against JNK mRNA. Transfection of JNK siRNAs resulted in a decrease in the basal protein level of JNK and markedly suppressed the apoptosis of cells in response to morphine treatment (Fig. 7B), indicating a pivotal role of the JNK ⁄ c-Jun pathway in morphine-induced apoptosis. Given the role of ROS in feedback amplification of JNK activation, we next investigated the effect of the antioxidant NAC on morphine-induced apoptosis. As shown in Fig. 7C, treatment of cells with 4 mm morphine for FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS 2029 Role of JNK signaling in morphine-induced apoptosis A X. Lin et al. B (a) C (a) Mock SiRNA – – + – Apoptosis (%) JNK Ratio JNK/β-action Action (b) 1.2 1.0 0.6 40 40 20 0 * 0.4 0.2 0.0 Con NAC Mor NAC + Mor Mock SiRNA Con 80 60 60 0.8 * (b) Mor Mor + Mor SiRNA JNK + Mor * 60 20 0 4 0 Concentration (mM) Apoptosis (%) Apoptosis (%) 80 Mor SP600125 + Mor SP203580 + Mor ** 80 – + 40 20 0 0 4 Concentration (mM) Fig. 7. SP600125, siRNA against JNK and NAC attenuated morphine-induced apoptosis. (A) SP600125, but not SB203580, attenuated morphine-induced apoptosis. Cells were treated with 4 mM morphine for 48 h in the absence or presence of 20 lM SP600125 or 20 lM SB203580. (a) Representative phase contrast microscopy images. (b) The apoptotic cells were detected by flow cytometry. (B) JNK siRNA markedly decreased morphine-induced apoptosis. (a) Cells were transfected with 200 nM JNK1 ⁄ 2 siRNA or Mock, and the protein levels of JNK were determined by western blot after transfection for 24 h, and densitometric analysis of changes in the basal protein levels of JNK, with *P < 0.05 in comparison with both untransfected and mock-transfected cells. (b) Cells were transfected with 200 nM JNK1 ⁄ 2 siRNA or Mock for 24 h, and then treated with 4 mM morphine for 48 h. The apoptotic cells were detected by flow cytometry as described in Experimental procedures. (C) NAC inhibited morphine-induced apoptosis. Cells were treated with 4 mM morphine for 48 h in the absence or presence of 5 mM NAC, and the apoptotic cells were detected by flow cytometry. All images are representative of at least three independent experiments yielding similar results. Values are means ± SEMs for at least three independent experiments performed in triplicate. *P < 0.05, **P < 0.01 as compared with morphine. 48 h in the presence of 5 mm NAC greatly reduced morphine-induced apoptosis, by 35% (from 57% in the absence of NAC to 22% in the presence of NAC), supporting its role in the amplification of JNK activation. Discussion In the present study, we demonstrated that chronic high-dose morphine treatment was able to cause apoptotic cell death of SH-SY5Y cells in an opioid receptor-independent manner, consistent with previous studies [9,32]. Moreover, we found that JNK signaling 2030 played a central role in the regulation of SH-SY5Y cell apoptosis via activation of the mitochondria-dependent pathway. We also demonstrated that a positive feedback cycle operates in SH-SY5Y cells, in which morphine-mediated JNK activation caused ROS generation, which in turn led to further activation of JNK. The activation of JNK by morphine (starting after 3 h of treatment) preceded ROS generation (starting after 6 h of treatment), indicating that the JNK activation occurred upstream of the ROS generation. SP600125 almost completely inhibited the accumulation of morphine-induced intracellular O2), further FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS X. Lin et al. indicating that ROS generation was JNK-dependent. Additional evidence to support JNK activation prior to ROS generation was that NAC, a well-characterized antioxidant, failed to prevent JNK activation induced by 3 h of morphine treatment. Regulation of ROS generation by JNK has also been observed recently in the fibroblast and PC3 prostate carcinoma cell lines [47,48]. Our study further demonstrated that ROS were released from mitochondria through opening of the mPTP because the mPTP inhibitor CsA, but not DPI and ALP, robustly attenuated morphine-induced ROS generation. These results are consistent with the findings reported by previous studies, showing that mPTP opening increases ROS production in vivo [41] and in isolated mitochondria [42,54], and that mPTP is the target downstream of JNK activation [55], and is inhibited by SP600125 [56]. Moreover, we also demonstrated that ROS could in turn activate JNK in a positive feedback fashion, because NAC significantly attenuated prolonged JNK activation induced by 12 h of morphine treatment. In addition, CsA also attenuated prolonged JNK activation induced by 12 h of morphine treatment. Therefore, this study clearly shows a route from JNK activation by morphine to initiation of ROS release from mitochondria, the ROS, in turn, activating JNK in a positive feedback fashion. This novel route is not consistent with that reported by previous studies, in which ROS generation initiated JNK activation [45,57,58]. JNK can be activated by various stress stimuli. In the present study, we showed that morphine treatment resulted in sustained activation of JNK, consistent with previous studies showing that prolonged exposure to morphine activates JNK in neurons and other cell lines [59,60]. We further demonstrated that activation of JNK was necessary for the morphine-induced apoptosis by using the pharmacological inhibitor SP600125. Inhibition of JNK activity by SP600125 led to decrease of apoptotic cell death in SH-SY5Y cells. Additionally, downregulation of endogenous JNK with transfected specific siRNA resulted in resistance to the proapoptotic effect of morphine. Moreover, NAC, which was shown to suppress ROS-amplified JNK activation, also robustly attenuated morphine-induced apoptosis in SH-SY5Y cells. These results strongly suggest that JNK signaling plays a pivotal role in morphineinduced apoptosis. JNK is thought to induce apoptosis via transcriptiondependent and transcription-independent mechanisms. We found that JNK-mediated morphine-induced apoptosis appeared to be transcription-dependent. Activation of JNK by morphine led to a robust increase in the expression of the BH3-only protein Bim via Role of JNK signaling in morphine-induced apoptosis activation of the transcription factor c-Jun, and this effect could be suppressed by SP600125. In addition, NAC attenuated morphine-induced Bim expression, whereas it inhibited morphine-induced phosphorylation of JNK and c-Jun. It has been shown that, in sympathetic neurons, activation of the JNK ⁄ c-Jun pathway and increased expression of Bim are key events required for cytochrome c release and apoptosis following nerve growth factor withdrawal [50]. There is strong evidence that Bax and Bak are potent regulators of cytochrome c release from mitochondria under a variety of stress conditions. Recent evidence has shown that BH3-only proteins (e.g. Bim, Bad, HRK, and Bid) are essential for Bax and Bak activation. Bim has been shown to potentiate the proapoptotic effects of Bax and Bak while concomitantly suppressing the prosurvival function of Bcl-2 [61,62]. In agreement with these findings, we also demonstrated that activation of the JNK ⁄ c-Jun pathway by morphine caused dramatic enhancement of cytochrome c release and caspase-3 activation, both of which were inhibited by SP600125. NAC also suppressed morphine-induced cytochrome c release and caspase-3 activation, through its inhibition of the morphineactivated JNK ⁄ c-Jun pathway. Our findings thus indicate that Bim may be involved in the release of cytochrome c and initiation of the intrinsic death pathway in response to prolonged morphine treatment. In addition to enhancement of the expression of the proapoptotic protein Bim, prolonged morphine treatment also dose-dependently decreased the expression of the antiapoptotic protein Bcl-2 in a JNK-dependent manner. Activation of JNK has been reported to induce the phosphorylation of Bcl-2, leading to Bcl-2 degradation through the proteasome pathway [63,64]. SP60025 could inhibit the morphine-induced reduction of Bcl-2 expression. Decreases in Bcl-2 expression by prolonged morphine treatment have been observed in cell lines and in rat brain in previous in vitro and in vivo studies [13,14,65]. However, Bax expression was not significantly altered following prolonged morphine treatment. As a result, the Bcl-2 ⁄ Bax ratio was decreased. Bcl-2 can block cytochrome c release by heterodimerizing with Bax [51,52]. Conversely, Bim promotes cytochrome c release by freeing Bax to be incorporated into the mitochondrial membrane [62]. Also, Bim can interact with Bax and induce a conformational change that facilitates the formation of channels for release of cytochrome c [61]. Thus, morphine seems to promote the proapoptotic effect of Bax by both enhancement of Bim expression and downregulation of Bcl-2 expression, leading to cytochrome c release, caspase-3 activation, and apoptosis. FEBS Journal 276 (2009) 2022–2036 ª 2009 Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Journal compilation ª 2009 FEBS 2031