Opposite regulation of the mitochondrial apoptotic pathway by C2-ceramide and PACAP through a MAP-kinase-dependent mechanism in cerebellar granule cells

Authors

  • Anthony Falluel-Morel,

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, University of Rouen, Mont-Saint-Aignan, France
    Search for more papers by this author
  • Nicolas Aubert,

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, University of Rouen, Mont-Saint-Aignan, France
    Search for more papers by this author
  • David Vaudry,

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, University of Rouen, Mont-Saint-Aignan, France
    Search for more papers by this author
  • Magali Basille,

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, University of Rouen, Mont-Saint-Aignan, France
    Search for more papers by this author
  • Marc Fontaine,

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Defense Proteins in Immune and Inflammatory Responses, INSERM U519, Faculty of Medicine and Pharmacy, Rouen, France
    Search for more papers by this author
  • Alain Fournier,

    1. INRS-Institut Armand Frappier, University of Québec, Pointe-Claire, Québec, Canada
    Search for more papers by this author
  • Hubert Vaudry,

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, University of Rouen, Mont-Saint-Aignan, France
    Search for more papers by this author
  • Bruno J. Gonzalez

    1. European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, University of Rouen, Mont-Saint-Aignan, France
    Search for more papers by this author

Address correspondence and reprint requests to Hubert Vaudry, Laboratory of Cellular and Molecular Neuroendocrinology, INSERM U413, University of Rouen, 76821 Mont-Saint-Aignan, France.
E-mail: hubert.vaudry@univ-rouen.fr

Abstract

The sphingomyelin-derived messenger ceramides provoke neuronal apoptosis through caspase-3 activation, while the neuropeptide pituitary adenylate cyclase-activating polypeptide (PACAP) promotes neuronal survival and inhibits caspase-3 activity. However, the mechanisms leading to the opposite regulation of caspase-3 by C2-ceramide and PACAP are currently unknown. Here, we show that PACAP prevents C2-ceramide-induced inhibition of mitochondrial potential and C2-ceramide-evoked cytochrome c release. C2-ceramide stimulated Bax expression, but had no effect on Bcl-2, while PACAP abrogated the action of C2-ceramide on Bax and stimulated Bcl-2 expression. The effects of C2-ceramide and PACAP on Bax and Bcl-2 were blocked, respectively, by the JNK inhibitor L-JNKI1 and the MEK inhibitor U0126. L-JNKI1 prevented the alteration of mitochondria induced by C2-ceramide while U0126 suppressed the protective effect of PACAP against the deleterious action of C2-ceramide on mitochondrial potential. Moreover, L-JNKI1 inhibited the stimulatory effect of C2-ceramide on caspase-9 and -3 and prevented C2-ceramide-induced cell death. U0126 blocked PACAP-induced Bcl-2 expression, abrogated the inhibitory effect of PACAP on ceramide-induced caspase-9 activity, and promoted granule cell death. The present study reveals that C2-ceramide and PACAP exert opposite effects on Bax and Bcl-2 through, respectively, JNK- and ERK-dependent mechanisms. These data indicate that the mitochondrial pathway plays a pivotal role in the pro- and anti-apoptotic effects of C2-ceramide and PACAP.

Abbreviations used
Bax

Bcl-2-associated protein

Bcl-2

B cell lymphoma 2

DMSO

dimethyl sulfoxide

DTT

dithiothreitiol

ERK

extracellular regulated kinase

FDA

fluorescein diacetate

JNK

jun N-terminal associated kinase

MEK

mitogen-activated protein kinase kinase

PACAP

pituitary adenylate cyclase-activating polypeptide

PBS

phosphate-buffered saline

SAPK

stress-activated protein kinase kinase

SDS

sodium dodecyl sulfate

Programmed cell death plays a critical role in a number of physiological and pathological processes including neurodevelopment (Haydar et al. 1999), neurodegenerative diseases (Pompl et al. 2003), and acute neurological injuries (Rami et al. 2003). The executioner caspase-3 is clearly implicated in the development of the central nervous system (CNS) and disruption of the gene encoding caspase-3 causes severe brain malformation (Kuida et al. 1996). Notably, in the cerebellum of caspase-3 knockout mice, the germinative neuroepithelium that gives rise to cerebellar granule cells persists longer and the number of granule neurons is significantly higher than in wild type littermates (Kuida et al. 1996). Different initiation pathways are known to activate caspase-3 (Earnshaw et al. 1999). In particular, caspase-3 activation can be induced through a caspase-8-dependent, mitochondria-independent pathway (Miyake et al. 2003) or through a caspase-9-dependent, mitochondria-dependent mechanism (Cao et al. 2002; Zhong et al. 2003). In this latter pathway, members of the Bcl-2 family control the mitochondrial integrity and play a pivotal role in the activation of caspase-9 (Sedlak et al. 1995; Rosse et al. 1998).

There is now clear evidence that the sphingomyelin breakdown products, ceramides, are important regulators of programmed cell death during development (Ariga et al. 1998; Herget et al. 2000). In particular, ceramides have been found to induce apoptosis of cortical neurons and cerebellar granule cells during brain histogenesis (Toman et al. 2002). Studies conducted in non-neuronal cells indicate that the pro-apoptotic effects of ceramides are mediated through the MAP-kinase JNK (Liu et al. 1999), the JAK/STAT pathway (Mazière et al. 2001), and the mitochondrial apoptotic pathway (Von Haefen et al. 2002). We have recently shown that in cerebellar granule cells, C2-ceramide induces apoptosis through a caspase-3-dependent mechanism (Vaudry et al. 2003a). However, the biochemical cascade mediating the effect of C2-ceramide on caspase-3 activation in neuronal cells is currently unknown.

Pituitary adenylate cyclase-activating polypeptide (PACAP) is a 38-amino acid neuropeptide that belongs to the vasoactive intestinal polypeptide (VIP) family (Miyata et al. 1989; Sherwood et al. 2000; Vaudry et al. 2000). PACAP and its receptors are actively expressed in the central nervous system during development (Basille et al. 1993; Tatsuno et al. 1994) in close association with germinative neuroepithelia, suggesting that the neuropeptide may be involved in regulation of neurogenesis (Hu et al. 2001). In agreement with this notion, in vivo studies have shown that, in the cerebellum of newborn rats, PACAP increases the number of migrating granule cells and increases the thickness of the internal granule cell layer (Vaudry et al. 1999). In vitro data confirmed that PACAP interacts with sonic hedgehog to control proliferation of cerebellar granule precursors (Nicot et al. 2002) and promotes survival and neurite outgrowth of immature granule cells (Gonzalez et al. 1997). We have recently shown that PACAP prevents apoptosis of cerebellar granule cells induced by C2-ceramide and that the anti-apoptotic effect of PACAP results from a blockage of caspase-3 activity (Vaudry et al. 2003a). However, the mechanisms leading to inhibition of C2-ceramide-activated caspase-3 by PACAP are currently unknown.

The objectives of the present study were to investigate the possible involvement of the mitochondrial apoptotic pathway in C2-ceramide-induced caspase-3 activation and to determine whether the control of the MAP-kinases JNK and ERK by PACAP and C2-ceramide contributes to the regulation of apoptosis in cerebellar granule cells.

Materials and methods

Materials

The 38-amino acid form of PACAP was synthesized by solid phase methodology as previously described (Chartrel et al. 1991). (L)-JNKI1 and LEHD-CHO were supplied by Calbiochem (San Diego, CA, USA). C2-ceramide, C8-ceramide, bovine serum albumin (BSA; fraction V), dithiothreitol (DTT), Ham's F12 medium, and the N1 supplement were purchased from Sigma Aldrich (Saint-Quentin Fallavier, France). Dulbecco's modified Eagle's medium (DMEM), Ham's F12, and the antibiotic–antimycotic solution (10 000 U/mL penicillin, 10 mg/mL Eheim 2211fungizone) were from Life Technologies (Cergy Pontoise, France).

Cell culture

Granule cell suspensions were prepared from cerebellums of 8-day-old Wistar rats, as previously described (Gonzalez et al. 1992). Briefly, neurons were cultured at a density of 4 × 103 cells per mm2 in a chemically defined medium consisting in 75% DMEM and 25% Ham's F12 supplemented with 2 mm glutamine, 1 mm sodium pyruvate, 25 mm KCl, 1% N1 supplement (× 100) and 1% antibiotic–antimycotic solution. Mitotic inhibitors were not added, as these culture conditions do not favor replication of non-neuronal cells (Gallo et al. 1987). Cells were grown at 37°C in a humidified incubator with an atmosphere of 5% CO2/95% air. For measurements of caspase activity and western blot analysis, dispersed cells were seeded in 35- and 60-mm Falcon dishes (Becton Dickinson, Plymouth, UK), respectively.

Caspase activity

Cells were incubated for 1–24 h in the absence or presence of C2-ceramide (20 µm), PACAP (10−7m), the JNK inhibitor (L)-JNKI1 (10–100 µm), the MEK-inhibitor U0126 (20 µm), or the cell-permeable caspase-9 inhibitor LEHD-CHO (20 µm). Cultured cells were washed twice with phosphate-buffered saline (PBS) at 37°C and re-suspended in PBS at 4°C. Cells were harvested by centrifugation (350 × g, 4°C, 10 min) and processed with a fluorometric caspase assay system (Promega, Charbonnières, France). Briefly, the cell pellet was resuspended in 10 µL hypotonic cell lysis buffer and centrifuged (16 000 × g, 4°C, 20 min). The supernatant was preincubated at 30°C for 30 min with 32 µL of the caspase assay buffer, 2 µL dimethyl sulfoxide (DMSO), 10 µL dithiothreitol (DTT) (100 mm) and deionized water to a final volume of 100 µL. The samples were then incubated with 2.5 mm caspase-3 or -9 substrates (30°C, 1 h). Fluorescence intensity was measured with a fluorescence microplate reader (Bio-Tek FL600, Winooski, VT, Canada).

Western blot analysis

Cerebellar granule cells were incubated for various times ranging from 2 to 24 h in the absence or presence of C2-ceramide (20 µm), C8-ceramide (80 µm), PACAP (10−7m), the JNK inhibitor (L)-JNKI1 (10–100 µm), the caspase-9 inhibitor LEHD-CHO (20 µm) or the MEK inhibitor U0126 (20 µm). Total cellular proteins were extracted by using the lysis buffer containing 1% Triton X-100, 50 mm Tris/HCl and 10 mm EDTA. The homogenate was centrifuged (20 000 × g, 4°C, 15 min) and the proteins contained in the supernatant were precipitated by addition of ice-cold 10% trichloroacetic acid. The extract was centrifuged (15 000 × g, 4°C, 15 min) and washed three times with alcohol/ether. The pellet was denatured in 50 mm Tris/HCl (pH 7.5) containing 20% glycerol, 0.7 m 2-mercaptoethanol, 0.004% (w/v) bromophenol blue and 3% (w/v) sodium dodecyl sulfate (SDS) at 100°C for 5 min, and electrophoresed on a 10% SDS–polyacrylamide gel electrophoresis (PAGE). After separation, proteins were electrically transferred onto a nitrocellulose membrane (Amersham, Les Ulis, France). The membrane was incubated with the blocking solution (1% BSA in Tris-buffered saline containing 0.05% Tween 20) at room temperature for 1 h and revealed with antibodies against Bax, Bcl-2, actin or the active and total form of JNK (Santa Cruz Biotechnology, Santa Cruz, CA, USA), cytochrome c or the active and total form of ERK (Promega), using a chemiluminescence detection kit (ECL System, Amersham). Autoradiographic films were quantified using an image analysis system (Biocom, Les Ulis, France).

Measurement of mitochondrial activity

Mitochondrial membrane potential was quantified by using the ratiometric probe JC-1 (Molecular Probes, Leiden, the Netherlands). In healthy granule cells, the intact membrane potential allows the lipophilic dye JC-1 to enter into the mitochondria where it accumulates and aggregates, producing an intense orange signal. In apoptotic cells, where the mitochondrial membrane potential collapses, the monomeric JC-1 remains cytosolic and stains the cytosol in green. Immature granule cells were treated for 12 h with graded concentrations of C2-ceramide or C8-ceramide (10–160 µm) and the effect of PACAP (10−7m) on mitochondrial activity was studied in the absence or presence of C2-ceramide (20 µm). The cells were then incubated for 30 min with 3 µL/mL JC-1 and washed twice with PBS. The proportion of aggregated versus monomeric JC-1 probe was quantified by measuring the ratio of fluorescence emissions at 590 nm (orange) over 530 nm (green) with the Bio-Tek FL600 fluorescence microplate reader.

Quantitative RT-PCR

Cells were harvested for total RNA using the RNAeasy Mini Kit (Qiagen, Valencia, CA, USA). Contaminating DNA was removed by treatment with RNase-free DNase I, and cDNA was synthesized with SuperScript First Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA, USA) by reverse transcription of 0.2–1 µg of total RNA. Quantitative-RT-PCR (Q-RT-PCR) was performed on cDNA in the presence of a 1X Mastermix (Applied Biosystems, Foster City, CA, USA) containing preset concentrations of dNTPs, MgCl2 and buffers, along with 90 nm forward and reverse primers, and either 100 nm probe or the SYBR green reporter dye, using the ABI PRISM 7000 Sequence Detection System (Applied Biosystems). RNA levels were deduced by comparison of cDNA-generated signals in samples with signals generated by a standard curve constructed with known amounts of cDNA, and internally corrected with the 18S cDNA signal for variations in amounts of input mRNA. The primers used were: Bax forward primer, 5′-TCGCTCTGTGGATGACTGAGTAC-3′; Bax reverse primer, 5′-GGGCCATATAGTTCCACAAAGG-3′; Bcl-2 forward primer, 5′-CTGGCATCTTCTCCTTCCAGC-3′; Bcl-2 reverse primer 5′-ACCTACCCAGCCTCCGTTATC-3′; 18S forward primer, 5′-GTGGAGCGATTTGTCTGGTT-3′; and 18S reverse primer, 5′-CGCTGAGCCAGTCAGTGTAG-3′.

Cell survival

Three hours after plating, granule cells were cultured for 24 h in the absence or presence of graded concentrations of C2-ceramide or C8-ceramide (10–160 µm). The caspase-9 inhibitor LEHD-CHO and the JNK inhibitor (L)-JNKI1 were added 30 min before treatment with C2-ceramide (20 µm). At the end of the 24-h period, cells were incubated for 8 min with 15 µg/mL FDA (producing green fluorescence in living neurons), rinsed once with PBS and lysed with a Tris/HCl solution. Fluorescence intensity was measured with the FL600 fluorescence microplate reader. Pilot experiments have shown that the fluorescence intensity is proportional to the cell number (in the range of 5 × 104−1 × 106 cells/mL).

Statistical analysis

Data are presented as the mean ± SEM from at least three independent experiments performed in triplicate. Statistical analyses of the data were performed using anova test followed by a Tukey's post hoc test.

Results

Effects of C2-ceramide, C8-ceramide and PACAP on mitochondrial integrity

The integrity of mitochondria was investigated by visualizing the membrane potential using the fluorescent ratiometric probe JC1. In control conditions, most granule cells possessed active mitochondria characterized by numerous fluorescent red spots in both soma and neurites (Fig. 1a). A 12-h exposure of cerebellar granule cells to C2-ceramide (20 µm) resulted in a marked decrease of the red signal; instead, green fluorescence was observed in cell bodies, indicating that the mitochondrial integrity was severely altered by C2-ceramide (Fig. 1b). Co-incubation of cerebellar granule neurons with C2-ceramide (20 µm) and PACAP (10−7m) restored the labeling of mitochondria by JC-1 (Fig. 1c).

Figure 1.

Effect of ceramides and PACAP on mitochondrial integrity in cerebellar granule cells. (a–c) Visualization of the mitochondrial potential by using the ratiometric probe JC-1 in control conditions (a), in the presence of 20 µm C2-ceramide (b), or in the presence of 20 µm C2-ceramide plus 10−7m PACAP (c). (d,e) Comparison of the effects of graded concentrations of C2-ceramide (C2-cer) and C8-ceramide (C8-cer) on cell survival (d) and mitochondrial potential (e). Each value represents the mean (± SEM) of three independent experiments performed in triplicate. *p < 0.05; **p < 0.01; ***p < 0.001 versus control. Scale bar = 20 µm.

As C2-ceramide is a ceramide analog that may not mimic the characteristics of native long-chain ceramides, we have next compared the effects of C2-ceramide to those of C8-ceramide. Both C2-ceramide and C8-ceramide increased granule cell death in a concentration-dependent manner (Fig. 1d). Quantitative analysis of the 530/590 nm fluorescent ratio in experiments similar to those illustrated in Fig. 1(a–c) also revealed that C2-ceramide and C8-ceramide both induced a concentration-dependent decrease of mitochondrial potential (Fig. 1e). As shown on Fig. 1(d,e), C2-ceramide appeared more potent and more efficacious than C8-ceramide on cell death and alteration of mitochondrial potential. Thus, C2-ceramide was used in most subsequent experiments.

Administration of graded concentrations of PACAP to C2-ceramide-treated cells induced a dose-dependent inhibition of the deleterious effects of C2-ceramide on mitochondrial integrity, the effect of PACAP being significant at concentrations from 10−8m onward (p < 0.01; Fig. 2a). Time-course experiments revealed that C2-ceramide significantly decreased mitochondrial potential after 8 h of treatment (p < 0.001) and the impairment of mitochondrial integrity was more prominent after 10 and 12 h (Fig. 2b). Administration of PACAP prevented the decrease of mitochondrial potential induced by C2-ceramide (Fig. 2b). After 12 h of treatment, PACAP reduced by 50% the impairment of mitochondrial integrity provoked by C2-ceramide.

Figure 2.

Dose–response and time-course of the effects of PACAP on mitochondrial potential in cerebellar granule cells. (a) Effects of graded concentrations of PACAP on C2-ceramide (20 µm)-induced impairment of mitochondrial potential. (b) Time-course of the effect of PACAP (10−7m) on C2-ceramide (20 µm)-induced impairment of mitochondrial potential. Each value represents the mean (± SEM) of three independent experiments performed in triplicate. *p < 0.05; ***p < 0.001 versus control; #p < 0.05; ##p < 0.01; ###p < 0.001 versus C2-ceramide.

Western blot experiments showed that a 12-h treatment of immature cerebellar granule cells with 20 µm C2-ceramide induced a marked decrease of cytochrome c level in the mitochondria and a concomitant increase of the cytosolic fraction (Fig. 3). Administration of PACAP significantly inhibited the C2-ceramide-induced release of cytochrome c (p < 0.001; Fig. 3).

Figure 3.

Western blot analysis showing the effect of C2-ceramide (20 µm) in the absence or presence of PACAP (10−7m) on the cytosolic/mitochondrial cytochrome c ratio. Each value represents the mean (± SEM) of three independent experiments performed in duplicate. Actin was used as an internal control. ***p < 0.001 versus control; ###p < 0.001 versus C2-ceramide alone.

C2-ceramide and PACAP exert opposite effects on Bax and Bcl-2 mRNA expression

The effects of C2-ceramide and PACAP on the expression of the pro- and anti-apoptotic factors Bax and Bcl-2 were investigated by conventional (Fig. 4a) and real-time RT-PCR (Fig. 4b,c). Exposure of immature cerebellar granule cells to C2-ceramide (20 µm) for various times ranging from 1 to 4 h resulted in a robust and transient increase of Bax mRNA expression (p < 0.001; Fig. 4a,b). The effect of C2-ceramide was detected after 1 h of treatment and vanished after 2 h. Administration of PACAP alone had no effect on Bax mRNA level, but totally prevented the C2-ceramide-induced increase of Bax mRNA (p < 0.001; Fig. 4b). In contrast, C2-ceramide (20 µm) did not affect the expression of Bcl-2 whatever the duration of treatment while a 2-h exposure to PACAP (10−7m) induced a threefold increase in Bcl-2 mRNA level (p < 0.001; Fig. 4a,c). Co-incubation of granule cells with PACAP and C2-ceramide significantly reduced the effect of PACAP on Bcl-2 mRNA expression (p < 0.01; Fig. 4c).

Figure 4.

Time-course of the effects of PACAP and C2-ceramide on the expression of Bax and Bcl-2 mRNAs in cerebellar granule cells. (a) Conventional RT-PCR experiments illustrating the effect of PACAP (10−7m) and C2-ceramide (20 µm) on Bax and Bcl-2 mRNA levels. 18S was used as an internal standard. (b) Quantitative RT-PCR analysis showing the effect of PACAP (10−7m), C2-ceramide (20 µm) or both on Bax mRNA level. The inset shows representative amplification curves in control conditions (dotted line) and in cells treated for 1 h with 20 µm C2-ceramide (full line). (c) Quantitative RT-PCR analysis showing the effect of PACAP (10−7m), C2-ceramide (20 µm) or both on Bcl-2 mRNA level. The inset shows representative amplification curves in control conditions (dotted line) and in cells treated for 2 h with 10−7m PACAP (full line). Each value represents the mean (± SEM) of three independent experiments performed in duplicate. ***p < 0.001 versus control; ##p < 0.01; ###p < 0.001 versus C2-ceramide.

C2-ceramide and PACAP exert opposite effects on Bax and Bcl-2 protein levels

The effect of C2-ceramide and PACAP on the concentration of the Bax and Bcl-2 proteins was investigated by western blot analysis (Fig. 5). Treatment of immature granule cells with C2-ceramide (20 µm) significantly increased Bax level within 2 h; thereafter, the concentration of Bax returned to basal level (Fig. 5a). Administration of PACAP (10−7m) had no effect on Bax protein level by its own, but abolished the stimulatory effect of C2-ceramide (Fig. 5a). Incubation of cells with PACAP (10−7m) caused a significant stimulation of Bcl-2 concentration within 4 h (p < 0.001; Fig. 5b). Addition of C2-ceramide significantly inhibited PACAP-induced Bcl-2 protein level (Fig. 5b).

Figure 5.

Time-course of the effects of PACAP and C2-ceramide on Bax and Bcl-2 proteins in cerebellar granule cells. (a) Western blot analysis showing the effect of PACAP (10−7m), C2-ceramide (20 µm) or both on Bax concentration. (b) Western blot analysis showing the effect of PACAP (10−7m), C2-ceramide (20 µm) or both on Bcl-2 concentration. Each value represents the mean (± SEM) of three independent experiments performed in duplicate. ***p < 0.001 versus control; ###p < 0.01 versus C2-ceramide.

Effects of MEK and JNK inhibitors on Bax and Bcl-2 protein levels

We have recently shown that, in immature cerebellar granule cells, C2-ceramide is a potent activator of the MAP-kinase JNK while PACAP stimulates ERK activation (Vaudry et al. 2003a). Here, we show that treatment of cerebellar granule cells with C8-ceramide also increased active JNK level (Fig. 6a). However, C8-ceramide was less potent than C2-ceramide to induce JNK phosphorylation (Fig. 6a), as already shown for cell death (Fig. 1d) and impairment of mitochondrial integrity (Fig. 1e). Pre-incubation of granule neurons with the cell permeant JNK inhibitor (L)-JNKI1 (100 µm) significantly blocked C2-ceramide-induced activation of JNK (Fig. 6b). At a concentration of 100 µm (L)-JNKI1 almost totally suppressed the effect of C2-ceramide on Bax protein level (p < 0.01; Fig. 6c). We also confirmed that the MEK inhibitor U0126 (20 µm) blocked the ERK phosphorylation induced by PACAP (10−7 m; Fig. 6d). U0126 strongly inhibited the effect of PACAP on Bcl-2 protein concentration (p < 0.001; Fig. 6e).

Figure 6.

Effect of JNK and ERK inhibitors on Bax and Bcl-2 proteins in cerebellar granule cells. (a) Western blot analysis comparing the effects of C2-ceramide (40 µm) and C8-ceramide (80 µm) on JNK phosphorylation. (b) Western blot analysis showing the effect of C2-ceramide (20 µm) in the absence or presence of the JNK inhibitor (L)-JNKI1 (10 or 100 µm) on JNK phosphorylation. (c) Western blot analysis showing the effect of C2-ceramide (20 µm) in the absence or presence of (L)-JNKI1 (10 or 100 µm) on Bax concentration. (d) Western blot analysis showing the effect of PACAP (10−7m) in the absence or presence of the MEK inhibitor U0126 (20 µm) on ERK phosphorylation. (e) Western blot analysis showing the effect of PACAP (10−7m) in the absence or presence of the MEK inhibitor U0126 (20 µm) on Bcl-2 concentration. Total JNK, total ERK, and actin were used as internal controls. Each value represents the mean (± SEM) of three independent experiments performed in duplicate. **p < 0.01; ***p < 0.001 versus control; ##p < 0.01; ###p < 0.001 versus C2-ceramide.

Effects of MEK and JNK inhibitors on mitochondrial integrity

As the actions of C2-ceramide and PACAP on Bax and Bcl-2 expression were mediated, respectively, through the JNK and ERK pathways, we investigated the effects of selective MAPK inhibitors on mitochondrial potential. Exposure of cerebellar granule cells to C2-ceramide (20 µm) resulted in a marked decrease of the 590 nm fluorescent signal reflecting accumulation of the JC-1 probe into active mitochondria (Fig. 7a,b). Treatment of granule cells with 100 µm (L)-JNKI1 alone did not affect cell morphology nor JC1 labeling (Fig. 7a,c). Pre-incubation of cells with (L)-JNKI1 prevented the deleterious action of C2-ceramide on JC1 aggregation (Fig. 7d) but did not restore granule cell morphology altered by C2-ceramide. Quantitative analysis of the 590/530 nm fluorescence ratio indicated that (L)-JNKI1 almost completely abrogated the effect of C2-ceramide on mitochondrial integrity (p < 0.001; Fig. 7e).

Figure 7.

Effect of a JNK inhibitor on mitochondrial integrity in cerebellar granule cells. (a–d) Visualization of cultured granule cells by phase contrast and mitochondrial integrity by using the ratiometric probe JC-1 in control conditions (a), or after treatment with C2-ceramide alone (20 µm; b) and (L)-JNKI1 alone (100 µm; c), or after co-treatment with C2-ceramide plus (L)-JNKI1 (d).  (e) Quantification of the effect of the JNK inhibitor (L)-JNKI1 on the deleterious effect of C2-ceramide on mitochondrial potential. Each value represents the mean (± SEM) of three independent experiments. ***p < 0.001 versus control; ###p < 0.001 versus C2-ceramide. Scale bar = 40 µm.

Incubation of granule neurons with the MEK inhibitor U0126 (20 µm) had no significant effect on either cell morphology or JC1 aggregation (Fig. 8a,d,g). Pre-incubation of cells with U0126 abolished the protective action of PACAP (10−7m) against C2-ceramide-induced decrease of mitochondrial potential (p < 0.001; Fig. 8c,e–g).

Figure 8.

Effect of a MEK inhibitor on mitochondrial integrity in cerebellar granule cells. (a–f) Visualization of cultured granule cells by phase contrast and mitochondrial integrity by using the ratiometric probe JC-1 in control conditions (a), or after treatment with PACAP alone (10−7m; b), C2-ceramide alone (20 µm; c) and U0126 alone (20 µm; d), or after co-treatment with PACAP plus C2-ceramide (e) and PACAP plus C2-ceramide plus U0126 (f).  (g) Quantification of the effect of the MEK inhibitor U0126 (20 µm) on the protective effect of PACAP against the deleterious action of C2-ceramide on mitochondrial potential. Each value represents the mean (± SEM) of three independent experiments. ***p < 0.001 versus control; ###p < 0.001 versus C2-ceramide, §§§p < 0.001 versus C2-ceramide plus PACAP. Scale bar = 40 µm.

Effects of MAPK and caspase inhibitors on caspase activation and cell survival

Treatment of cerebellar granule cells with PACAP (10−7 m) induced a significant decrease of the basal activity of caspase-9 (p < 0.05; Fig. 9a) while C2-ceramide (20 µm) significantly increased caspase-9 activity (p < 0.01; Fig. 9a). PACAP suppressed the stimulatory effect of C2-ceramide on caspase-9 activity (p < 0.001; Fig. 9a). Pre-incubation of granule cells with the MEK inhibitor U0126 (20 µm) abrogated the inhibitory effect of PACAP on caspase-9 activity (p < 0.001; Fig. 9a). Similarly, the JNK inhibitor (L)-JNKI1 (100 µm) abolished C2-ceramide-induced caspase-9 activation (Fig. 9b).

Figure 9.

Effect of C2-ceramide on caspase-9 and caspase-3 activity in cerebellar granule cells. (a) Effect of C2-ceramide (20 µm) in the absence or presence of PACAP (10−7m), U0126 (20 µm) and PACAP plus U0126 on caspase-9 activity. (b) Effect of C2-ceramide (20 µm) in the absence or presence of the JNK inhibitor (L)-JNKI1 (100 µm) on caspase-9 activity. (c) Effect of C2-ceramide (20 µm) in the absence or presence of (L)-JNKI1 (100 µm) on caspase-3 activity. (d) Effect of C2-ceramide (20 µm) in the absence or presence of the caspase-9 inhibitor DEHD-CHO (20 µm) on caspase-3 activity. Each value represents the mean (± SEM) of three independent experiments performed in duplicate. *p < 0.05; **p < 0.01; ***p < 0.001 versus control; #p < 0.05; ##p < 0.01; ###p < 0.001 versus C2-ceramide; +++p < 0.001 versus C2-cer plus PACAP.

A 3-h exposure of cerebellar granule cells to C2-ceramide (20 µm) significantly increased caspase-3 activity (p < 0.001; Fig. 9c). The C2-ceramide-induced activation of caspase-3 was blocked by (L)-JNKI1 (100 µm; p < 0.001; Fig. 9c) and significantly reduced by the caspase-9 inhibitor DEHD-CHO (p < 0.05; Fig. 9d).

As both (L)-JNKI1 and DEHD-CHO markedly attenuated C2-ceramide-induced caspase-3 activity, the effects of these two inhibitors on granule cell survival were investigated. As previously shown, C2-ceramide (20 µm), administered alone during 12 h, strongly affected cell survival (Fig. 10a). The JNK inhibitor (L)JNKI1 significantly protected cerebellar granule cells against C2-ceramide neurotoxicity (p < 0.01; Fig. 10a), and the caspase-9 inhibitor DEHD-CHO (20 µm) totally suppressed C2-ceramide-induced granule cell death (p < 0.001; Fig. 10b).

Figure 10.

Effect of the JNK inhibitor (L)-JNKI1 and the caspase-9 inhibitor DEHD-CHO on C2-ceramide-induced death of cerebellar granule cells. (a) Cell survival was assessed after treatment with C2-ceramide (20 µm) in the absence or presence of (L)-JNKI1 (100 µm). (b) Cell survival was assessed after treatment with C2-ceramide (20 µm) in the absence or presence of DEHD-CHO (20 µm). Each value represents the mean (± SEM) of three independent experiments performed in quadruplicate. ***p < 0.001 versus control; ##p < 0.01; ###p < 0.001 versus C2-ceramide.

Discussion

Elimination of excess neurons through programmed cell death is an essential process that allows harmonious development of the CNS. However, while the morphological and biochemical characteristics of apoptosis have been extensively investigated, the factors regulating programmed cell death during brain development remain largely unknown. Cerebellar granule cell is a well suited model that is widely used to investigate the basic molecular interplay of the cell death machinery during neurogenesis (Vaudry et al. 2003b). In particular, it has been recently shown that apoptosis of granule neurons is stimulated by the sphingolipid-derived messengers ceramides (Monti et al. 2001; Toman et al. 2002) and inhibited by the neuropeptide PACAP (Vaudry et al. 2003a). In the present report, we describe the molecular mechanisms underlying the opposite effects of ceramides and PACAP on apoptosis of granule neurons.

Comparison of the effects of C2-ceramide and C8-ceramide on cell survival, mitochondrial potential and JNK phosphorylation

Previous studies have shown an increased level of the naturally occurring C16-ceramide during apoptosis (Thomas et al. 1999). Because the long-chain C16-ceramide has a propensity for packing into the plasma membrane (Gidwani et al. 2003), the cell-permeable analog C2-ceramide is commonly used to investigate the effects of native ceramides on cell activity (Obeid et al. 1993; Toman et al. 2002; Willaime-Morawek et al. 2003). In order to ascertain that C2-ceramide could really mimic the characteristics of the long-chain ceramides, we have studied the effect of C8-ceramide, a cell-permeable ceramide that, unlike C2-ceramide, is metabolized within the cells to generate natural ceramides (Karasavvas and Zakeri 1999) and thus more closely reflects the effects of endogenous ceramides than C2-ceramide does. Here we show that C8-ceramide, in very much the same way as C2-ceramide, provoked granule cell death, decreased mitochondrial potential and enhanced cytochrome c release. Although C2-ceramide appeared to be more potent and more efficacious than C8-ceramide, these data provide evidence that C2-ceramide actually mimics the action of the long-chain ceramides on cell fate. In support of this hypothesis, it has been recently shown that both C2-ceramide and C16-ceramide induce cytochrome c release from mitochondria suspensions (Di Paola et al. 2004).

Control of the mitochondrial apoptotic pathway by C2-ceramide and PACAP

The activation of caspases leading to apoptosis is mediated through two distinct pathways, the extrinsic pathway that necessitates the action of a death ligand on its receptor and the intrinsic pathway that involves the participation of the mitochondria. Here, we show that exposure of cerebellar granule cells to C2-ceramide provokes a marked alteration of the mitochondrial integrity. The deleterious action of C2-ceramide was characterized by a decrease of the mitochondrial potential associated with a massive release of cytochrome c. The involvement of the intrinsic apoptotic pathway in C2-ceramide-induced apoptosis has also been reported in different tumoral cell lines including Jurkat cells (Rodriguez-Lafrasse et al. 2002), radioresistant SQ20B cells (Alphonse et al. 2002) and DU145 carcinoma cells (Von Haefen et al. 2002). Treatment of granule cells with PACAP prevented the action of C2-ceramide on both mitochondrial potential and cytochrome c release, indicating that the mitochondrial pathway plays a pivotal role in the pro-apoptotic action of C2-ceramide as well as the anti-apoptotic effect of PACAP on granule neurons.

Permeabilization of the mitochondrial membrane is under the control of pro- and anti-apoptotic factors that belong to the Bcl-2 family (Zimmerman et al. 2001). In particular, it has been previously shown that genetic deletion of the pro-apoptotic factor Bax prevents cerebellar granule cell death (Miller et al. 1997). It has also been reported that thyroid hormone deficiency, which causes massive loss of granule neurons, is associated with an increase in Bax expression and a concomitant decrease in the expression of the anti-apoptotic factor Bcl-2 (Singh et al. 2003), suggesting that Bax and Bcl-2 may play an important role in the regulation of programmed granule cell death during development. The present study reveals that C2-ceramide stimulates the expression of Bax mRNA and Bax protein in a time-dependent manner, but does not affect Bcl-2 expression. A similar effect of ceramides on the Bax-Bcl-xL ratio has recently been reported in the tumoral HL-60 cell line (Kim et al. 2001). In contrast, PACAP up-regulated Bcl-2 expression in cerebellar granule cells. Although PACAP had no effect on Bax by its own, it totally suppressed ceramide-induced Bax expression, suggesting that C2-ceramide and PACAP control the balance between the pro- and anti-apoptotic factors Bax and Bcl-2. These data provide the first evidence that ceramides and PACAP exert opposite effects on the expression of pro- and anti-apoptotic Bcl-2 members in neural cells: C2-ceramide, by increasing Bax level will lead to the formation of Bax homodimers, mitochondria permeabilization and granule cell death, whereas PACAP, by stimulating Bcl-2 expression and by abrogating C2-ceramide-induced Bax production will favor the formation of Bcl-2-Bax heterodimers and promote cell survival (Fig. 11).

Figure 11.

Schematic representation of the signaling mechanisms likely involved in the opposite effects of PACAP and C2-ceramide on cerebellar granule cell apoptosis. PACAP provokes a rapid inhibition of caspase-3 through a PKA- and PKC-dependent mechanism that does not require de novo protein synthesis. PACAP also causes an ERK-dependent induction of Bcl-2 expression. C2-ceramide stimulates Bax expression via phosphorylation of JNK. Balance between the pro- and anti-apoptotic factors Bax and Bcl-2 regulates mitochondrial potential, caspase-9 activation, and finally granule cell survival. The dotted lines denote secondary inhibitory effects. AC, adenylyl cyclase; Cyt c, cytochrome c; ERK, extracellular regulated kinase; IP3, inositol trisphosphate; H89, protein kinase A inhibitor; MEK, mitogen-activated protein kinase kinase; PAC1-R, PACAP-specific receptor; PLC, phospholipase C; PKA, protein kinase A; PKC, protein kinase C; U0126, MAP kinase kinase inhibitor. →, activation; |, inhibition; -| →, inhibition of an inhibitory effect.

Involvement of the MAP-kinases JNK and ERK in the control of the mitochondrial apoptotic pathway by C2-ceramide and PACAP

Studies conducted on different types of tumor cells, such as Jurkat cells (Caricchio et al. 2002), MRC5 fibroblasts (Mazière et al. 2001) and malignant glioma cells (Schiffer et al. 2001), indicate that ceramide-induced apoptosis is regulated by the JNK/SAPKs protein kinases. The implication of the MAP-kinase ERK in the neurotrophic effect of PACAP has been reported in cerebellar granule cells (Villalba et al. 1997; Harada and Sugimoto 1999; Le-Niculescu et al. 1999) and PC12 cells (Vaudry et al. 2002) and activation of the JNK signaling pathway seems to be a key event in granule cell apoptosis (Harris et al. 2002). We have recently demonstrated that C2-ceramide and PACAP exert opposite effects on JNK and ERK phosphorylation and that the MEK inhibitor U0126 suppresses the protective effect of PACAP against ceramide-induced cerebellar granule neurons apoptosis (Vaudry et al. 2003a). Here, we show that the JNK inhibitor (L)-JNKI1 abrogates C2-ceramide-induced expression of Bax and prevents the deleterious effect of C2-ceramide on mitochondrial potential, while the MEK inhibitor U0126 inhibits the stimulatory effect of PACAP on Bcl-2 production, suppresses the protective action of PACAP against C2-ceramide-evoked decrease of mitochondrial potential and blocks the inhibitory effect of PACAP on caspase-9 activity. These data indicate that the expression of the pro- and anti-apoptotic members of the Bcl-2 family is regulated by the MAP-kinases JNK and ERK, and that an increase in JNK phosphorylation may account for the activation of the mitochondrial apoptotic pathway in cerebellar granule neurons. In agreement with this notion, (L)-JNKI1 prevented the stimulatory effect of C2-ceramide on caspase-9 and caspase-3 activity, and significantly reduced C2-ceramide-evoked granule cell death (Fig. 11). Consistent with a relationship between JNK/Bax on the one hand, and ERK/Bcl-2 on the other hand, recent reports have shown that JNK phosphorylation induces Bax-dependent apoptosis in kidney cells (Lei and Davis 2003) while expression of the Bcl-2-related protein Bcl-Xl necessitates activation of ERK (Miranda et al. 2003).

Functional implication during brain development

Ceramides act as intracellular messengers for various pro-inflammatory cytokines. In particular, there is now clear evidence that tumor necrosis factor (TNF)-α uses ceramides as signaling molecules to induce apoptosis (Heller and Krönke 1994; Sortino et al. 1999). Recent studies conducted in transgenic mice have shown that overexpression of TNF-α impairs brain development and, notably, strongly affects the histogenesis of the cerebellar cortex (Ye et al. 2003). These observations, together with our own data, strongly suggest that ceramides may act as second messengers mediating the effect of TNF-α on the cytoarchitecture of the cerebellum. Concurrently, several studies suggest that PACAP plays a crucial role in the development of the cerebellar cortex (Vaudry et al. 2003b). For example, in vitro studies have shown that PACAP reduces proliferation of cerebellar granule cell precursors by inhibiting Shh gene transcription (Nicot et al. 2002) and in vivo experiments have revealed that PACAP increases the number of granule cells in the internal granule cell layer (Vaudry et al. 1999), indicating that the neuroprotective effect of PACAP on ceramide-induced apoptosis may play a key role during cerebellar development.

In conclusion, the present study has demonstrated that PACAP blocks the activation of the intrinsic apoptotic pathway induced by C2-ceramide in cerebellar granule cells. C2-ceramide and PACAP induce opposite effects on Bcl-2 and Bax expression through ERK- and JNK-dependent mechanisms, respectively. These pro- and anti-apoptotic mechanisms may represent a crucial process regulating the life/death fate of cerebellar neuroblasts during neurodevelopment.

Acknowledgements

This research was supported by an INSERM Grant (U413) and an IREB Grant (2001/22). AFM is recipient of a fellowship from the Conseil Régional de Haute-Normandie (LARC-Neuroscience network). NA is a recipient of a fellowship from the Conseil Régional de Haute-Normandie and the CIT-IFM Recherche. HV is Affiliated Professor at the INRS-Institut Armand Frappier (Montréal, Canada). We thank Magalie Benard for the gifts of 18S oligonucleotides and Gérard Cauchois for excellent technical assistance.

Ancillary