Oncogene (2005) 24, 178–187 & 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Sphingosine activity counteracts -mediated death in human cells: role of Bcl-2 expression

Meryem Bektas1,2, Puneet S Jolly2,3, Carola Mu¨ ller1,Ju¨ rgen Eberle1, Sarah Spiegel2 and Christoph C Geilen*,1

1Department of Dermatology, Charite´ - Universita¨tsmedizin Berlin, Campus Benjamin Franklin, Fabeckstr. 60–62, Berlin 14195, Germany; 2Department of , Virginia Commonwealth University School of Medicine, Richmond, VA 23298, USA; 3Department of Biochemistry and Molecular Biology, Georgetown University, Washington, DC 20007, USA

While most of the pharmacological therapies for melano- overexpression of members of the inhibitor of ma utilize the apoptotic machinery of the cells, the (IAP) family (Grossman et al., 1999), as well as available therapeutic options are limited due to the ability inactivation of effector molecules of the apoptotic of melanoma cells to resist . signaling cascade, such as apoptosis activating factor-1 Human melanoma cell lines A-375 and M186 are sensitive (Apaf-1) (Soengas et al., 2001). In previous studies to ceramide- and Fas-induced cell death, while Mel-2a and characterizing 11 different human melanoma cell M221 are resistant. We have now found that Mel-2a and populations, we classified melanoma cells into two M221 cells have a significantly higher ceramide/sphingo- subpopulations: one group was susceptible and another sine-1- (S1P) ratio than A-375 and M186 cells. group was resistant to apoptotic stimuli, such as As kinase (SphK) type 1 plays a critical role ceramide and Fas ligation (Raisova et al., 2000). in determining the dynamic balance between the proa- Analysis of expression of the components of the poptotic metabolite ceramide and the pro- apoptotic machinery in these two subpopulations did survival S1P, we examined its role in apoptosis of not reveal substantial differences. However, we subse- melanoma cells. Increasing SphK1 expression reduced quently noted that the resistant cells had a defect in the sensitivity of A-375 melanoma cells to Fas- and cytochrome c release as well as an increased ratio of Bcl- ceramide-mediated apoptosis. Conversely, downregulation 2 to Bax (Raisova et al., 2001). of SphK1 with small interfering RNA decreased the Numerous studies have shown that increased intra- resistance of Mel-2a cells to apoptosis. Importantly, cellular levels of the sphingolipid metabolite ceramide overexpression of the prosurvival protein Bcl-2 in A-375 inhibit cell proliferation and enhance differentiation and cells markedly stimulated SphK1 expression and activity, apoptosis (reviewed in Jayadev and Hannun, 1996; while downregulation of Bcl-2 reduced SphK1 expression. Perry and Hannun, 1998; Geilen et al., 2001). Many This link between Bcl-2 and SphK1 might be an additional anticancer drugs induce apoptosis by increasing intra- clue to resistance of malignant melanoma. cellular ceramide levels (Ogretmen and Hannun, 2001). Oncogene (2005) 24, 178–187. doi:10.1038/sj.onc.1208019 Ceramide-mediated clustering has been suggested to be required for CD95-death-inducing signaling complex Keywords: melanoma; bcl-2; sphingosine; apoptosis formation (Grassme et al., 2003). One mechanism by which cells can evade anticancer drugs is by altering their of ceramide. Blocking cer- amide generation (Bose et al., 1995) or its rapid removal (Xia et al., 1999; Strelow et al., 2000) or conversion to Introduction glucosylceramide (Liu et al., 1999) suppresses apoptosis. In contrast to ceramide, its further metabolite Melanoma is largely considered to be a chemotherapy sphingosine-1-phosphate (S1P) has been associated with refractory tumor. The underlying molecular mechan- and suppresses ceramide-mediated cell death isms for this resistance, however, have not yet been (Cuvillier et al., 1996; Osawa et al., 2001). Several characterized (Serrone and Hersey, 1999). Abundant growth and survival factors such as platelet-derived evidence indicates that melanoma cells show defects at growth factor (PDGF), nerve growth factor (NGF), multiple levels of the apoptotic program (Satyamoorthy basic fibroblast growth factor (bFGF) and phorbol et al., 2001), including altered signaling of death esters activate (SphK), the receptors (Thomas and Hersey, 1998), increased levels that phosphorylates sphingosine, leading to an increase of antiapoptotic Bcl-2 (Tang et al., 1998), in S1P levels and a concomitant decrease in ceramide levels (Edsall et al., 1997; Rius et al., 1997; Olivera et al., *Correspondence: CC Geilen; E-mail: [email protected] 1999a; Johnson et al., 2002). Moreover, overexpression Received 4 December 2003; revised 7 June 2004; accepted 29 June 2004 of SphK protects cells from apoptosis induced by serum Bcl-2 regulates sphingosine kinase M Bektas et al 179 withdrawal or ceramide (Olivera et al., 1999a, 2003). 2a and M221, than in the sensitive A-375 and M186 cells Thus, the dynamic balance between ceramide and S1P (Figure 1c), consistent with the opposite effects of can be a decisive factor in determining whether or not a ceramide and S1P on cell survival. cell undergoes apoptosis. Inhibitors of SphK1, such as N,N-dimethylsphingosine and L-threo-dihydrosphingo- sine, induce apoptosis regardless of multidrug resistance expression (Jendiroba et al., 2002), and may provide a new strategy for the treatment of anticancer drug- resistant . Indeed, SphK inhibitors reduce gastric tumor growth (Endo et al., 1991) and mammary adenocarcinoma tumor growth in mice (French et al., 2003). As SphK1 is overexpressed in a variety of solid tumors, including breast, stomach, ovary, kidney and lung, compared to normal tissues from the same patients (French et al., 2003), it was of interest to examine its role in resistance of melanoma cells to apoptosis.

Results

Extracellular-regulated kinase (ERK)1/2 is constitutively phosphorylated in melanoma cells Constitutive ERK is correlative with malignant transformation in many tumor cell lines including some melanoma cell populations (Kortylewski et al., 2001; Govindarajan et al., 2003). Thus, it was of interest to examine ERK activation in cells, which are resistant to ceramide and agonistic activation of Fas by the CH-11 antibody (Mel-2a and M221 cells) compared to Fas and ceramide sensitive (A-375 and M186 cells). All of these melanoma cell lines showed a high basal level of ERK1/2 phosphorylation. Furthermore, ERK1/ 2 phosphorylation was not increased by the presence of serum (data not shown), suggesting that apoptosis susceptibility of melanoma cells is not correlated with ERK activation.

Ceramide-resistant melanoma cells have a higher S1P/ ceramide ratio To examine the role of sphingolipid metabolites in the disparity in the apoptotic behavior between resistant and sensitive melanoma cells, we measured the intracel- lular levels of S1P and ceramide. The sensitive cell lines, A-375 and M186, had higher basal ceramide levels than the resistant cells, M221 and Mel-2a (Figure 1a). On the other hand, only M221, not Mel-2a, displayed higher levels of S1P than the sensitive cell lines, A-375 and M186 (Figure 1b). Since the ratio of S1P to ceramide within a cell may be crucial in determining the outcome of an apoptotic stimulus, the S1P/ceramide ratio was compared in the sensitive and resistant cells. The ratio was significantly higher in both resistant cell lines, Mel-

Figure 1 Increased ratio of S1P/ceramide in resistant melanoma cells. Mass levels of ceramide (a) and S1P (b) in resistant Mel-2a and M221 cells and sensitive A-375 and M186 cells were measured as described in Materials and methods and S1P/ceramide ratios were calculated (c). The values are representative of two similar experiments (n ¼ 3)

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 180 SphK activity is higher in Mel-2a and M221 than A-375 tion (G82D) in the ATP (Pitson et al., 2000) and M186 were also isolated. As expected, cells expressing wt- SphK but not catalytically inactive SphK1 had increased Increases in cellular S1P protect against apoptosis sphingosine phosphorylating activity (Figure 3b and c), et al induced by ceramide or FasL (Cuvillier ., 1996; although SphK1 protein was expressed to the same level Osawa et al., 2001). SphK, which catalyses the in both cells (Figure 3a). phosphorylation of sphingosine to S1P, is the critical Overexpression of wt-SphK1 significantly reduced cell enzyme that regulates intracellular levels of S1P. There- death induced by exogenous short-chain C -ceramide fore, we correlated S1P levels in melanoma cells with 6 (N-hexanoylsphingosine) (Figure 4a). It should be noted SphK activity. Indeed, SphK activity was nearly four- that treatment of cells with C -ceramide resulted in a fold higher in Mel-2a cells than in A-375 and M186 cells. 6 significant increase in endogenous long-chain apoptotic SphK activity in M221 cells was not as high but was still ceramide species (Ogretmen et al., 2002). Similarly, significantly greater than A-375 and M186 cells expression of SphK1 significantly reduced cell death (Figure 2a). Moreover, the higher SphK1 activity induced by agonistic Fas antibody CH-11 or by correlated with SphK1 protein expression as shown by doxorubicin, although the antiapoptotic effect of SphK1 Western blot analysis (Figure 2b). was less pronounced (Figure 4b). In sharp contrast, overexpression of inactive SphK1 did not influence the Ectopic SphK1 expression suppresses ceramide-mediated apoptotic effects of C6-ceramide (Figure 4c), CH-11 or cell death doxorubicin (data not shown). As SphK1 expression and activity were higher in ceramide-resistant Mel-2a cells than in sensitive A-375 cells and its expression suppresses ceramide-mediated cell death in other cell types (Olivera et al., 1999a, 2003), it was of interest to determine whether overexpression of SphK1 in A-375 would induce resistance to apoptosis. To this end, we prepared A-375 cells stably over- expressing wild-type (wt) SphK1. Cells stably expressing catalytically inactive SphK1 containing a point muta-

Figure 3 Ectopic expression of SphK1 in A-375 cells. Cells were Figure 2 SphK1 activity is higher in resistant melanoma cells. (a) stably transfected with GFP vector, wt-SphK1-GFP or catalyti- SphK1 activity was measured in cell lysates in the presence of cally inactive SphK1-GFP. (a) Expression was examined by 0.25% Triton X-100. Data are means7s.d. of triplicate determina- Western blot analysis with anti-GFP antibody. (b and c)In tions from three independent experiments. (b) SphK1 levels were duplicate cultures, SphK1 activity was determined in the determined in the same lysates by Western blot analysis using anti- presence of 0.25% Triton X-100 (c), and a representative TLC is SphK1 antibody shown in (b)

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 181 Sensitization of Mel-2a cells to ceramide-mediated apoptosis by downregulation of endogenous SphK1 expression To further investigate the role of endogenous SphK1 in ceramide resistance of Mel-2a cells, its expression was specifically reduced by small interfering RNA (siRNA). siRNA targeted against SphK1 markedly downregulated expression of both SphK1 mRNA and protein (Figure 5a and b). This downregu- lation was also reflected in significantly decreased SphK1 activity (Figure 5c). Furthermore, siSphK1 increased the sensitivity of Mel-2a cells to apoptosis induced by ceramide and CH-11 by about twofold as compared to cells treated with control siRNA (Figure 5d).

Correlation of Bcl-2 expression and SphK1 in melanoma cells Previously, we have demonstrated that a high Bcl-2/Bax ratio correlated with resistance of melanoma cells to apoptosis induced by ceramide or CH-11 antibody (Raisova et al., 2001). Mel-2a cells express higher levels of both Bcl-2 (Figure 6a) and SphK1 (Figure 2a) than A-375 cells. Since both proteins are antiapoptotic, it was of interest to determine whether there is a correlation between Bcl-2 and SphK1 expression. siRNA targeted to SphK1 did not significantly affect Bcl-2 expression (Figure 5b). In addition, overexpression of SphK1 in A- 375 cells did not change expression of Bcl-2 as determined by Western analysis (Figure 6b). These results suggest that SphK1 does not regulate the expression of Bcl-2. However, we noticed that stably overexpressing Bcl-2 in A-375 cells caused a twofold increase in SphK1 enzymatic activity (Figure 6c) that correlated with higher SphK1 protein expression in these cells (Figure 6d). To further substantiate the correlation between Bcl-2 expression and SphK1 activity, we used the tetracycline/ doxycycline-inducible expression system in the apoptosis-sensitive melanoma cell line Bro (Eberle et al., 2002). Bro-Tet-On cells transfected with a Bcl-2 expression vector under the control of the Tet promoter showed increased Bcl-2 expression after 8 and 24 h of doxycycline induction (Figure 7d). In agreement with other reports (Zhang et al., 1996; Raisova et al., 2002), induction of Bcl-2 in these melanoma cells Figure 4 Suppression of ceramide-induced apoptosis by SphK1 significantly decreased apoptosis induced by C6-cera- overexpression. (a and b) Stably transfected A-375 cells were mide and CH-11 as compared to vector-transfected cells treated with the indicated concentrations of short-chain C6- (Figure 7a and b). In parallel with increased Bcl-2 ceramide, CH-11 or doxorubicin for 6 h. Apoptotic cell death expression, there was an increase in SphK1 expression PLUS was measured with the Cell Death Detection ELISA assay as (Figure 7d). Moreover, this correlated with SphK1 described in Materials and methods. (c) Cells stably transfected with vector, wild-type SphK1 or catalytically inactive SphK1 were activity that was increased after 8 and 24 h of treated with the indicated concentrations of C6-ceramide for 24 h doxycycline treatment by twofold and 2.5-fold, respec- and cell death was measured. Results are means of triplicate tively (Figure 7c). Nonspecific effects of doxycycline determinations from two independent experiments and expressed were excluded as vector-transfected Bro-Tet-On as fold changes compared to untreated controls. *Po0.05 by t-test cells showed no increase in Bcl-2 (Figure 7d), no protection from apoptosis (Figure 7a), and there were no significant changes in SphK activity (Figure 7c) or expression (Figure 7d) after doxycycline treatment.

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 182

Figure 6 Bcl-2 overexpression increases SphK1. (a) Expression of endogenous Bcl-2 was determined in lysates of A-375 and Mel-2a cells by Western blotting. (b) Bcl-2 expression was determined in A- 375 cells stably transfected with vector or SphK1 by Western blotting with Bcl-2 antibody. Blots were stripped and reprobed with anti-tubulin antibody to demonstrate equal loading. (c) SphK1 activity was measured in A-375 cells stably transfected with vector or mBcl-2. Values represent triplicate determinations from three independent experiments. (d) SphK1 and mBcl-2 levels were determined by Western blotting with anti-SphK1 and anti- mBcl-2 antibodies, respectively. Blots were stripped and reprobed with anti-tubulin antibody to demonstrate equal loading

Downregulation of Bcl-2 decreases SphK1 expression in Mel-2a cells Recently, it has been reported that overexpression of Bcl-2 protects SphK1 from DNA damage-induced (Taha et al., 2004). As overexpression of Bcl-2 increases SphK1 activity (Figure 7c), it was of interest to determine whether endogenous Bcl-2 might regulate basal SphK1 expression. Therefore, endogen- ous Bcl-2 was downregulated in Mel-2a cells using specific siRNA. As shown in Figure 7e, downregulation of Bcl-2 leads to a concomitant decrease in SphK1 expression and activity.

Discussion Figure 5 Downregulation of SphK1 in Mel-2a cells sensitizes them to ceramide- and CH-11-induced apoptosis. Subconfluent To characterize differences between ceramide-resistant Mel-2a cells were transfected with siRNA against SphK1 or control siRNA. After 48 h, (a) SphK1 mRNA and GAPDH mRNA were and ceramide-sensitive melanoma cells, we previously determined by RT–PCR. (b) Protein levels were determined by investigated the expression of several proteins involved Western blotting with anti-SphK1 or anti-Bcl-2 antibodies. Blots in apoptosis. We found no differences in the expression were stripped and reprobed with anti-actin or anti-tubulin to show of FLIP, caspase-8, Fas or Fas (Raisova et al., equal loading. (c) SphK1 activity was determined in duplicate 2000). Moreover, phosphorylation of ERK1/2, which cultures. (d) Mel-2a cells transfected with siSphK1 or control siRNA were treated with vehicle or the indicated concentrations of has been shown to result in increased cell survival and short-chain C6-ceramide or CH-11. After 6 h, cells were fixed, malignant transformation (Smalley, 2003), cannot ac- stained with Hoechst dye and apoptotic cells determined by count for the differences in apoptosis sensitivity, as we fluorescence microscopy. At least two different fields were scored found that it was constitutively activated in both with a minimum of 300 cells per field. Similar results were obtained resistant and sensitive melanoma cells. in three independent experiments. *Po0.05 by t-test Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 183

Figure 7 Transient expression of Bcl-2 suppresses apoptosis and enhances SphK1 expression. (a) Bro-Tet-On vector-transfected cells and (b) Bro-Tet-On Bcl-2-transfected cells were treated without (none) or with 4 mM doxycycline together with 30 mM C6-ceramide or 1.0 mg/ml CH-11 as indicated. After 24 h, apoptosis was assessed by Hoechst staining. At least two different fields were scored with a minimum of 350 cells per field. Results are means7s.d. of triplicate determinations. Similar results were obtained in two additional experiments. (c and d) Parallel cultures were treated with 4 mM doxycycline, and at the indicated times, cells were lysed and SphK1 activity determined (c). Basal SphK1 activity in vector and Bcl-2-overexpressing cells was 22.570.9 and 23.670.1 pmol/min/mg, respectively. Values are representative of two independent experiments. (d) SphK1 and Bcl-2 expression was determined by Western blot analysis after treatment with 4 mM doxycycline for the indicated times. Blots were stripped and reprobed with anti-b-actin antibody to demonstrate equal loading. (e) Mel-2a cells were transfected with siBcl-2 or siControl. Equal amounts of protein from cell lysates were analyzsd by Western blotting with anti-Bcl-2 antibody or anti-hSphK1 antibody. Blots were stripped and reprobed with anti-b- actin antibody to demonstrate equal loading. SphK1 activity was determined in the presence of 0.25% Triton X-100. Data are means7s.d. of triplicate determinations from three independent experiments. *Po0.05 by t-test; #Po0.05 by ANOVA; NS, not significant

We next examined the relationship between ceramide rheostat model for apoptosis (Cuvillier et al., 1996), we and S1P levels in these melanoma cells. Both M221- and found significantly higher ratios of S1P to ceramide in Mel-2a-resistant cells had lower ceramide levels than the resistant M221 and Mel-2a cells compared to sensitive sensitive A-375 and M186 cell lines. Consistent with the A-375 and M186 cells. Since ceramide and S1P are

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 184 interconvertable , and S1P may suppress inducible factor 1) (Iervolino et al., 2002). VEGF- ceramide-mediated apoptosis, it is possible that resistant induced ERK activation was also shown to be mediated melanoma cells are protected from apoptosis by by SphK in T24 bladder tumor cells and in HUVEC increased SphK activity, higher S1P levels and, con- cells (Shu et al., 2002). The activation of SphK is versely, lower ceramide levels. In support of this dependent on C as shown in various cell hypothesis, we found higher SphK1 activity in both types (Buehrer et al., 1996; Johnson et al., 2002). It is Mel-2a and M221. Moreover, if S1P levels are critically therefore tempting to speculate that in melanoma cells important for survival of melanoma cells, increased Bcl-2 might activate SphK via VEGF upregulation. In expression of SphK1 should affect the apoptotic addition, the factor NF-kB has been linked sensitivity. In agreement, overexpression of wt, but not to Bcl-2 expression in MCF7ADR cells (Ricca et al., catalytically inactive mutant SphK1, increased resis- 2000). NF-kB regulates the expression of numerous tance of A-375 melanoma cells to short-chain ceramide-, important for cell survival, tumor progression and as well as to CH-11- and doxorubicin-induced apopto- invasion (Bours et al., 2000), and constitutive NF-kB sis. The cytoprotective effect of SphK1 overexpression activity has been associated with tumor development on ceramide-mediated cell death was more pronounced and progression (Yamamoto and Gaynor, 2001). The than on Fas- or doxorubicin-induced apoptosis, likely BH4 domain of Bcl-2 has been shown to be crucial for due to alternative death signaling pathways of Fas and the activation of NF-kB and, interestingly, this domain doxorubicin, which do not necessarily involve ceramide has also been described to interact with several other or ceramide metabolites (Gewirtz, 1999; Peter and signaling molecules such as calcineurin and raf-1 Krammer, 2003). In further support of the decisive role (Shibasaki et al., 1997). Further studies are needed to of SphK1, downregulation of SphK1 expression by shed light on the mechanism of Bcl-2-induced SphK treatment with specific siRNA rendered ceramide- activation. resistant Mel-2a cells sensitive to ceramide- and Fas- In conclusion, in this study, we demonstrate the induced cell death. involvement of SphK1 in the resistance of melanoma It was previously suggested that the different apopto- cells to apoptotic stimuli. Furthermore, we show that tic responses of Mel-2a and A-375 melanoma cells could inhibition of ceramide-induced apoptosis by Bcl-2 may be due to differences in their Bcl-2/Bax ratios (Raisova at least in part result from increased SphK1 expression et al., 2001), as ceramide-resistant Mel-2a cells expressed and activity. Targeting SphK1 might provide additional significantly higher levels of Bcl-2 than the sensitive A- therapeutic benefits for treatment of drug-resistant 375 cells, and Bcl-2 overexpression reverses the cera- melanoma. mide-sensitive phenotype of A-375 (Raisova et al., 2001). In the present study, we found that the disparity in apoptosis not only correlated with Bcl-2 expression but also with pronounced differences in SphK1 expres- Materials and methods sion and activity. As both Bcl-2 and SphK1 are Materials antiapoptotic, it was of interest to determine whether their expression was linked. There was a twofold The agonistic anti-CD95/Fas monoclonal antibody (clone CH- increase in SphK1 activity in A-375 cells stably over- 11) was purchased from Chemicon (Hofheim, Germany). C6- expressing Bcl-2 compared to naı¨ ve A-375 cells. To rule ceramide was purchased from Biotrend (Cologne, Germany) out possible clonal or compensatory effects, we used and dissolved in ethanol to a final concentration of 25 mM. pEGFP-neo was purchased from Clontech (Palo Alto, CA, Tet-On Bro melanoma cells in which Bcl-2 expression USA). Geneticin (G418) was from Life Technologies (Karls- was induced by doxycycline. SphK1 activity and ruhe, Germany). Doxorubicin was provided by Sigma expression were similarly increased as a consequence (Munich, Germany). Sphingosine was from Biomol (Ham- of Bcl-2 induction. On the other hand, overexpression of burg, Germany). [g-32P]ATP (3000 Ci/mmol) was purchased SphK1 in A-375 cells had no effect on Bcl-2 expression, from Amersham (Arlington Heights, IL, USA). High-perfor- indicating that Bcl-2 might be upstream of SphK1. mance TLC plates were purchased from EM Science (Gibbs- Interestingly, overexpression of Bcl-2 or failure to town, NJ, USA). Bisbenzimide hydrochloride (Hoechst 33258) upregulate p53 MOLT-4 cells following and recombinant Escherichia coli were DNA damage induced by actinomycin D inhibits the from Calbiochem (La Jolla, CA, USA). activation of proteases and, subsequently, prevents SphK1 proteolysis (Taha et al., 2004). To explore the Melanoma cell culture potential role of endogenous Bcl-2 in the regulation of SphK1 expression, Bcl-2 was specifically downregulated M186 and M221 cells were obtained from patients with with siRNA. This resulted in a reduction of endogenous histologically confirmed metastatic melanoma by surgical SphK1 activity and protein, further supporting a role intervention. A-375 and Bro cells are established melanoma for Bcl-2 in regulating SphK1 expression. cell lines originating from primary human tumors (Giard et al., 1973; Lockshin et al., 1985), whereas Mel-2a To date, not much is known about how Bcl-2 might originated from metastasis (Bruggen et al., 1981). Melanoma influence expression. Recently, it has been reported cell lines were grown in Dulbecco’s modified Eagle’s that Bcl-2 overexpression increases the expression of the medium (Sigma) supplemented with 10% fetal bovine serum angiogenesis factor vascular endothelial growth factor (Biochrom, Berlin, Germany), 100 U/ml penicillin and 100 mg/ (VEGF) in human melanoma cells via HIF-1 (hypoxia- ml streptomycin.

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 185 Transfection 1998). In some experiments, apoptotic cell death was measured by staining cell nuclei with the Hoechst dye bisbenzimide A-375 cells were transfected with pEGFP-N1 vector (Clon- (Olivera et al., 1999b) and apoptotic cells were distinguished tech) containing the cDNA for green fluorescent protein by condensed, fragmented nuclear regions. (GFP) alone or with a plasmid containing either wt-SphK1 or catalytically inactive SphK1 (Pitson et al., 2000) fused to GFP at its C-terminus. Transfection was performed using Lipofec- RT–PCR tamine (Invitrogen) according to the manufacturer’s protocol. Total RNA was isolated from cells with RNAeasy (Qiagen, After 2 days, cells were transferred into medium containing Valencia, CA, USA). RNA was reverse transcribed with 1 mg/ml G418. After 2–3 weeks, green cells were sorted with an Superscript II (Life Technologies, Gaithersburg, MD, USA). EPICS 753 argon laser flow cytometer (Beckman Coulter The following specific primers were used to amplify hSphK1: Epics Elite). The construction of the mouse Bcl-2 expression forward, 50-ATC CAG AAG CCC CTG TGT AGC CTC C-30; plasmid based on pIRES (Clontech) has been reported reverse, 50-GCA GCA AAC ATC TCA CTG CCC AGG T-30. previously and has been shown to be functional in A-375 cells (Raisova et al., 2002). Tetracycline/doxycycline-inducible Bro- Tet-On melanoma cells have been described previously (Eberle SphK activity et al., 2002). We have subcloned a full-length cDNA copy of SphK1 activity was measured as described previously (Olivera murine Bcl-2 into EcoRI and XbaI sites of pTRE1 (Clontech, and Spiegel, 1998). SphK1 activity is optimal under conditions Palo Alto, CA, USA). The construct pTRE-Bcl2 was then of low salt and 0.25% Triton X-100. Under these conditions, transfected into the teracycline-inducible Tet-On melanoma SphK2 activity is negligible (Liu et al., 2000). cell line (Eberle et al., 2002). A stably transfected cell clone (Bro-Bcl-2) was obtained by limited dilution in microtiter Measurement of S1P plates and was continuously cultured under antibiotic pressure (geneticin, hygromycin). S1P was measured as described previously with some SphK1 expression was downregulated with sequence-specific modifications (Edsall and Spiegel, 1999; Edsall et al., 2000). siRNA. siRNA for human SphK1 (sense 50-GGG CAA GGC Cells were washed with PBS and scraped in 1 ml of methanol. CUU GCA GCU C; antisense 50-GAG CUG CAA GGC were extracted by adding 2 ml chloroform/1 M NaCl CUU GCC C) and the control siRNA were obtained from (1 : 1, v/v) and 100 ml3N NaOH, and phases were separated. Xeragon-Qiagen (Birmingham, AL, USA). Specificity of the The basic aqueous phase containing S1P, devoid of sphingo- sequence was confirmed by the Blast program. siRNA for sine and the majority of , was transferred onto a human Bcl-2 (cat. #M-003307-010) and control siRNA (cat. siliconized glass tube. The organic phases were re-extracted #D-001210-01) were purchased from Dharmacon (Lafayette, with 1 ml methanol/1 M NaCl (1 : 1, v/v) plus 50 ml3N NaOH, CO, USA). Mel-2a cells were seeded into six-well dishes at and the aqueous fractions were combined. To the aqueous 4 Â 105 cells/well. After 24 h, cells were transfected for 3 h with fraction of extracts (2 ml total) was added 450 ml of the siRNA using OligofectAMINE (Invitrogen) as recommended following buffer: 200 mM Tris-HCl (pH 7.4), 75 mM MgCl2 in by the manufacturer. After 48 h, cells were treated as indicated 1.2 M glycine buffer (pH 9.0) and 50 U of alkaline . in figure legends. Samples were incubated at 371C for 30 min. To terminate the reactions, 50 ml conc. HCl was added and lipids were extracted Western blot analysis twice with 1.5 ml CHCl3. Pooled organic fractions were dried under in siliconized glass tubes and resuspended in Cells were harvested in buffer containing 50 mM HEPES (pH 180 ml of SphK buffer containing 0.25% Triton X-100 and S1P 7.4), 150 mM NaCl, 20 mM orthovanadate, 10 mM pyropho- measured as described previously (Olivera and Spiegel, 1998). sphate, 100 mM NaF, 10 mM EDTA, 1% Triton X-100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml pep- Measurement of ceramide statin A, 10 mg/ml aprotinin and 10 mg/ml leupeptin. Equal amounts of protein were separated by SDS–PAGE. Proteins Lipids were extracted and mass amounts of ceramide in were transferred onto a nitrocellulose membrane and probed cellular extracts measured by the DAG kinase enzymatic with specific antibodies as follows: mouse monoclonal anti- method as described (Toman et al., 2002). mouse Bcl-2 (100/D5, Novocastra, Dossenheim, Germany), mouse monoclonal anti-human Bcl-2 (C-2, Santa Cruz, Statistical analysis Heidelberg, Germany), rabbit polyclonal anti-SphK antibody, polyclonal phospho-ERK1/2 (New England Biolabs, Frank- Statistical significance was determined using two-tailed Stu- furt, Germany), polyclonal antibodies against ERK2 (Santa dent’s t-test or a one-way analysis of variance (ANOVA) with Cruz Biotech., Santa Cruz, CA, USA), mouse monoclonal a Tukey post hoc and Po0.05 was considered significant. anti-b-actin (Sigma, Munich, Germany) and anti-b-tubulin (Santa Cruz Biotech.). Primary antibodies were labeled with goat-anti-mouse IgG or with goat-anti-rabbit IgG conjugated Abbreviations with horseradish peroxidase (Dako, Hamburg, Germany). IAP, inhibitor of apoptosis protein; S1P, sphingosine-1- Immunocomplexes were visualized by enhanced chemilumi- phosphate; SphK, sphingosine kinase; ERK, extracellular nescence with Super Signal West Pico reagent from Pierce regulated kinase; PDGF, platelet-derived growth factor; (Weiskirchen, Germany). VEGF, vascular endothelial growth factor; NGF, nerve growth factor; bFGF, basic fibroblast growth factor; Apaf-1, Cell death assay apoptosis-activating factor-1. Cell death was measured with the Cell Death Detection ELISA PLUS kit (Roche Diagnostics), which determines cytoplasmic Acknowledgements histone-associated DNA fragments (mononucleosomes and Polyclonal anti-SphK1 antibody was a generous gift from oligonucleosomes) as described previously (Wieder et al., Dr Murate and Dr Banno. MB was a recipient of a

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 186 fellowship from the Free University of Berlin. This study from the National Cancer Institute (R01CA61774) to SS. was supported by grants from Deutsche Forschun- Cell sorting was supported in part by NIH Grant gsgemeinschaft (SFB 366 and Ge 641/5-2) to CCG and P30CA16059.

References

Bose R, Verheij M, Haimovitz-Friedman A, Scotto K, Fuks Z Ogretmen B, Pettus BJ, Rossi MJ, Wood R, Usta J, Szulc Z, and Kolesnick R. (1995). Cell, 82, 405–414. Bielawska A, Obeid LM and Hannun YA. (2002). J. Biol. Bours V, Bentires-Alj M, Hellin AC, Viatour P, Robe P, Chem., 277, 12960–12969. Delhalle S, Benoit V and Merville MP. (2000). Biochem. Olivera A and Spiegel S. (1998). Methods Mol. Biol., 105, Pharmacol., 60, 1085–1089. 233–242. Bruggen J, Fogh J and Sorg C. (1981). J. Cancer Res. Clin. Olivera A, Edsall L, Poulton S, Kazlauskas A and Spiegel S. Oncol., 102, 141–152. (1999a). FASEB J., 13, 1593–1600. Buehrer BM, Bardes ES and Bell RM. (1996). Biochim. Olivera A, Kohama T, Edsall L, Nava V, Cuvillier O, Poulton Biophys. Acta, 1303, 233–242. S and Spiegel S. (1999b). J. Cell Biol., 147, 545–558. Cuvillier O, Pirianov G, Kleuser B, Vanek PG, Coso OA, Olivera A, Rosenfeldt HM, Bektas M, Wang F, Ishii I, Chun J, Gutkind S and Spiegel S. (1996). Nature, 381, 800–803. Milstien S and Spiegel S. (2003). J. Biol. Chem., 278, Eberle J, Fecker LF, Bittner JU, Orfanos CE and Geilen CC. 46452–46460. (2002). Br. J. Cancer, 86, 1957–1962. Osawa Y, Banno Y, Nagaki M, Brenner DA, Naiki T, Edsall L, Vann L, Milstien S and Spiegel S. (2000). Methods Nozawa Y, Nakashima S and Moriwaki H. (2001). Enzymol., 312, 9–16. J. Immunol., 167, 173–180. Edsall LC and Spiegel S. (1999). Anal. Biochem., 272, Perry DK and Hannun YA. (1998). Biochim. Biophys. Acta, 80–86. 1436, 233–243. Edsall LC, Pirianov GG and Spiegel S. (1997). J. Neurosci., 17, Peter ME and Krammer PH. (2003). Cell Death Differ., 10, 6952–6960. 26–35. Endo K, Igarashi Y, Nisar M, Zhou QH and Hakomori S. Pitson SM, Moretti PA, Zebol JR, Xia P, Gamble JR, Vadas (1991). Cancer Res., 51, 1613–1618. MA, D’Andrea RJ and Wattenberg BW. (2000). J. Biol. French KJ, Schrecengost RS, Lee BD, Zhuang Y, Smith SN, Chem., 275, 33945–33950. Eberly JL, Yun JK and Smith CD. (2003). Cancer Res., 63, Raisova M, Bektas M, Wieder T, Daniel P, Eberle J, Orfanos 5962–5969. CE and Geilen CC. (2000). FEBS Lett., 473, 27–32. Geilen CC, Barz S and Bektas M. (2001). Skin Pharmacol. Raisova M, Goltz G, Bektas M, Bielawska A, Riebeling C, Appl. Skin Physiol., 14, 261–271. Hossini AM, Eberle J, Hannun YA, Orfanos CE and Geilen Gewirtz DA. (1999). Biochem. Pharmacol., 57, 727–741. CC. (2002). FEBS Lett., 516, 47–52. Giard DJ, Aaronson SA, Todaro GJ, Arnstein P, Kersey JH, Raisova M, Hossini AM, Eberle J, Riebeling C, Wieder T, Dosik H and Parks WP. (1973). J. Natl. Cancer Inst., 51, Sturm I, Daniel PT, Orfanos CE and Geilen CC. (2001). 1417–1423. J. Invest. Dermatol., 117, 333–340. Govindarajan B, Bai X, Cohen C, Zhong H, Kilroy S, Louis Ricca A, Biroccio A, Del Bufalo D, Mackay AR, Santoni A G, Moses M and Arbiser JL. (2003). J. Biol. Chem., 278, and Cippitelli M. (2000). Int. J. Cancer, 86, 188–196. 9790–9795. Rius RA, Edsall LC and Spiegel S. (1997). FEBS Lett., 417, Grassme H, Cremesti A, Kolesnick R and Gulbins E. (2003). 173–176. Oncogene, 22, 5457–5470. Satyamoorthy K, Bogenrieder T and Herlyn M. (2001). Trends Grossman D, McNiff JM, Li F and Altieri DC. (1999). Mol. Med., 7, 191–194. J. Invest. Dermatol., 113, 1076–1081. Serrone L and Hersey P. (1999). Melanoma Res., 9, 51–58. Iervolino A, Trisciuoglio D, Ribatti D, Candiloro A, Biroccio Shibasaki F, Kondo E, Akagi T and McKeon F. (1997). A, Zupi G and Del Bufalo D. (2002). FASEB J., 16, Nature, 386, 728–731. 1453–1455. Shu X, Wu W, Mosteller RD and Broek D. (2002). Mol. Cell. Jayadev S and Hannun YA. (1996). J. Lipid Mediat. Cell Biol., 22, 7758–7768. Signal., 14, 295–301. Smalley KS. (2003). Int. J. Cancer, 104, 527–532. Jendiroba DB, Klostergaard J, Keyhani A, Pagliaro L and Soengas MS, Capodieci P, Polsky D, Mora J, Esteller M, Freireich EJ. (2002). Leukemia Res., 26, 301–310. Opitz-Araya X, McCombie R, Herman JG, Gerald WL, Johnson KR, Becker KP, Facchinetti MM, Hannun YA and Lazebnik YA, Cordon-Cardo C and Lowe SW. (2001). Obeid LM. (2002). J. Biol. Chem., 277, 35257–35262. Nature, 409, 207–211. Kortylewski M, Heinrich PC, Kauffmann ME, Bohm M, Strelow A, Bernardo K, Adam-Klages S, Linke T, Sandhoff MacKiewicz A and Behrmann I. (2001). Biochem. J., 357, K, Kronke M and Adam D. (2000). J. Exp. Med., 192, 297–303. 601–612. Liu H, Sugiura M, Nava VE, Edsall LC, Kono K, Poulton S, Taha TA, Osta W, Kozhaya L, Bielawski J, Johnson KR, Milstien S, Kohama T and Spiegel S. (2000). J. Biol. Chem., Gillanders WE, Dbaibo GS, Hannun YA and Obeid LM. 275, 19513–19520. (2004). J. Biol. Chem., 279, 20546–20554. Liu YY, Han TY, Giuliano AE, Ichikawa S, Hirabayashi Y Tang L, Tron VA, Reed JC, Mah KJ, Krajewska M, Li G, and Cabot MC. (1999). Exp. Cell Res., 252, 464–470. Zhou X, Ho VC and Trotter MJ. (1998). Clin. Cancer Res., Lockshin A, Giovanella BC, De Ipolyi PD, Williams Jr LJ, 4, 1865–1871. Mendoza JT, Yim SO and Stehlin Jr JS. (1985). Cancer Res., Thomas WD and Hersey P. (1998). J. Immunol., 161, 45, 345–350. 2195–2200. Ogretmen B and Hannun YA. (2001). Drug Resist. Update, 4, Toman RE, Movsesyan V, Murthy SK, Milstien S, Spiegel S 368–377. and Faden AI. (2002). J. Neurosci. Res., 68, 323–330.

Oncogene Bcl-2 regulates sphingosine kinase M Bektas et al 187 Wieder T, Orfanos CE and Geilen CC. (1998). J. Biol. Chem., Yamamoto Y and Gaynor RB. (2001). Curr. Mol. Med., 1, 273, 11025–11031. 287–296. Xia P, Wang L, Gamble JR and Vadas MA. (1999). J. Biol. Zhang J, Alter N, Reed JC, Borner C, Obeid LM and Hannun Chem., 274, 34499–34505. YA. (1996). Proc. Natl. Acad. Sci. USA, 93, 5325–5328.

Oncogene