|Home | About | Journals | Submit | Contact Us | Français|
Ceramide levels are elevated in Mantle Cell Lymphoma cells following treatment with cannabinoids. Here, we investigated the pathways of ceramide accumulation in the MCL cell line Rec-1 using the stable endocannabinoid analogue R(+)-methanandamide (R-MA). We further interfered with the conversion of ceramide into sphingolipids that promote cell growth. Treatment with R-MA led to increased levels of ceramide species C16, C18, C24 and C24:1 and transcriptional induction of ceramide synthases (CerSs) 3 and 6. The effects were attenuated using SR141716A, which has high affinity to cannabinoid receptor 1 (CB1). The CB1-mediated induction of CerS3 and CerS6 mRNA was confirmed using Win-55,212-2. Simultaneous silencing of CerS3 and CerS6 using siRNA abrogated the R-MA-induced accumulation of C16 and C24. Inhibition of either of the enzymes serine palmiotyl transferase, ceramide synthase, and dihydroceramide desaturase within the de novo ceramide pathway reversed ceramide accumulation and cell death induced by R-MA treatment. In order to enhance the cytotoxic effect R-MA, sphingosine kinase-1 (SK-1) and glucosylceramide synthase (GCS), enzymes that convert ceramide to the pro-proliferative sphingolipids sphingosine-1-phospate and glucosylceramide, respectively, were inhibited. Suppression of either enzyme using inhibitors or siRNA potentiated the decreased viability, induction of cell death and ceramide accumulation induced by R-MA treatment. Our findings suggest that R-MA induces cell death in MCL via CB1-mediated upregulation of the de novo ceramide synthesis pathway. Furthermore, inhibition of SK-1 and GCS potentiated ceramide accumulation and cell death induced by R-MA. This is the first study were the cytotoxic effect of a cannabinoid is enhanced by modulation of ceramide metabolism.
Ceramide accumulation is a widely described event in cancers after various treatments . C16-Ceramide is described as one of the major ceramide sub-species whose levels are elevated during apoptosis induced by various agents . For instance, C16 ceramide, generated de novo, was accumulated during androgen ablation in the prostate cell line LNCaP . Both C16 and C24 ceramide accumulated during BcR crosslinking in Ramos cells [4, 5]. When ceramide species C18 was specifically induced in UM-SCC-22A cells (sqaumous cell carcinoma of hypopharynx) by overexpression of CerS1, cell growth was inhibited .
In previous publications [7–9] we and others observed that induction of ceramide accumulation by cannabinoids leads to apoptosis in MCL, glioma and pancreatic cancer. The signaling leading to cell death was blocked by inhibition of ceramide synthase with Fumonisin B1 (FB1). In MCL, the accumulation of ceramide was mediated through the cannabinoid receptors type 1 (CB1) and type 2 (CB2) which are overexpressed on MCL cells, while control cells lacking the receptors remained unaffected.
There are six species of ceramide synthases (CerSs), and several iso-forms have been described . CerSs 1–6 synthesize ceramides of varying chain length . When CerS3 was overexpressed in HEK-293T cells, an increased production of C18 and C24 ceramide species was observed , while overexpression of CerS6 showed that the enzyme preferably synthesized the long chain ceramide species C14 and C16 ceramide .
CerSs can act through two different pathways, as they are involved in both de novo synthesis of (dihydro)ceramide as well as regeneration of ceramide from sphingosine in the salvage/recycling pathway, see Fig 1. Several enzymes are involved in the de novo synthesis of ceramide which starts with the precursors L-serine and palmitoyl-CoA. Their conversion into 3-ketosphinganine is catalyzed by Serine Palmitoyl Transferase (SPT) . Further downstream, sphinganine is acylated to dihydroceramide by ceramide synthase (CerS). The dihydroceramide is desaturated by dihydroceramide desaturase (DEGS) to ceramide . On the other hand, in the salvage/recycling pathway, CerSs act on sphingosine that is generated from the breakdown of complex sphingolipids. Since FB1 inhibits CerS, its actions do not distinguish between the activation of the de novo pathway vs. the operation of the salvage pathway. Thus, it became important to determine the specific pathway activated by cannabinoids.
Once ceramide is synthesized, it can be rapidly metabolized into sphingomyelin, glucosylceramide or sphingosine, see Figure 1, and the latter two can be further converted to complex glycosphingolipids or sphingosine-1-phosphate (S-1-P), respectively. Metabolism of active ceramide into such species by glucosylceramide synthase (GCS) or sphinogsine kinase-1 (SK-1) is the limiting factor in the cell death response to ceramide-inducing stimuli . It has been shown in multiple cell types  that manipulating ceramide metabolism by blocking enzymes leads to a potentiation of cell death. Also, the balance between ceramide and S-1-P is vital to the cell death decision in many cancer types  . Safingol, an inhibitor of SK-1, has been shown to synergistically increase the efficacy of the cytotoxic drug fenretinide in neuroblastoma cells . Down regulation of SK-1 by ActD in Molt-4 cells has been shown to decrease viability and induce cell death . Resistant melanoma cells Mel-2a showed increased rate of apoptosis after treatment with siRNA against SK-1 together with Fas antibody CH-11 or C6-ceramide . Several studies have shown that overexpression of GCS in cancers can generate multidrug resistance caused by subsequent upregulation of the multi drug resistance 1 (MDR1) gene [20, 21]. There are multiple publications stating that GCS inhibitors e.g. PDMP, PPMP and PPPP can enhance the effect of chemotherapeutic drugs in resistant cells , . Using antisense to downregulate GCS in resistant breast cancer cells, MCF-7 Adr, Gouaze et al  showed a decrease in MDR1 expression leading to an increased cell death by vinblastine.
In our previous publications we have induced cell death by treatment of lymphoma cells with different cannabinoids in vitro [7, 25], and observed a 40% reduction of tumor burden in NOD/SCID mice xenotransplanted with human MCL by treatment with the stable endocannabinoid analogue R(+)-methanandamide (R-MA) in vivo . These results together with those implicating ceramide in the action of cannabinoids raised the possibility that preventing the transformation of ceramide into other forms of sphingoplipids could enhance the cell death response in MCL. Further, the Nordic lymphoma Network reported that adding the chemotherapeutic agents doxorubicine and Ara-C, both inducers of ceramide accumulation, to MCL treatment has improved the event free survival for MCL patients. Thus, ceramide accumulation appears to contribute to the reduction of malignant MCL cells in vivo.
In this study, we investigated the mechanisms and specificity of the ceramide response to R-MA. We further exploited this understanding in order to determine if modulating ceramide levels could potentiate the cytotoxic response to R-MA. The obtained data show that R-MA treatment leads to increased expression of ceramide synthases, de novo synthesis of specific ceramide species and apoptosis in the MCL cell line Rec-1. Modulation of ceramide metabolism using inhibitors or RNA interference potentiates the apoptosis-inducing effect of R-MA.
R(+)-methanandamide (R-MA), Win-55,212-2 (Win55), Fumonisin B1, Myriocin, SKI II, DMS and PDMP were purchased from Sigma-Aldrich Sweden AB (Stockholm, Sweden). C8-Cyclopropenylceramide (C8CPPC) was obtained from Matreya LLC (PA, USA). 3[H]-palmitate was purchased from Amersham Biosciences (Buckinghamshire, England). SiRNAduplexes against CerS3, CerS6, GCS and SK1 were purchased from Ambion and diluted to 200nM in siRNA dilution buffer (Qiagen, Valencia, CA). AIM-V medium was purchased from Invitrogen Corporation (Carlsbad, CA).
The MCL cell line Rec-1 was a kind gift from Dr. Christian Bastard, Ronan, France. The plasma cell line SK-MM-2 was obtained from Deutsche Sammlung von Microorganismen und Zellkulturen (DSMZ) (Braunschweig, Germany). Cell lines were maintained in RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 2 mM L-glutamine and 10% fetal calf serum (FCS) and 50 μg/ml gentamicin (Invitrogen) under standard conditions (humidified atmosphere, 95% air, 5% CO2, 37°C).
After treatment, samples were frozen at −80 °C and transferred to the Lipidomics Core facility at MUSC. There lipids were measured by high performance liquid chromatography tandem mass spectrometry (MS/MS) as described earlier .
Phospholipids were extracted according to Bligh and Dyer . The samples and phosphate standards made of NaH2 PO4 were ashed in ashing buffer (10 N H2SO4: 70% HClO4: H2O) at 160 °C overnight. Thereafter, 900 μl water, 500 μl 0, 9% ammounium molybdate and 200 μl 9% ascorbic acid was added to each sample followed by incubation at 45 °C for 30 minutes. The amount of lipid phosphate was determined by measuring absorption at 590 nm.
Rec-1 cells were resuspended in AIM-V media containing 2μCi [3H]-palmitic acid to a concentration of 2 million cells/ml. After 12 hrs, cells were washed in PBS and treated with 10μM R-MA with or without pre-treatment with inhibitors in fresh AIM-V media.
Cells were harvested and washed in cold PBS three times. Subsequently, cells were dissolved in 50 μl PBS and the suspension was added to 1 ml of MeOH;CHCl3(1:2). To extract the lipids, 1 ml water was added to the samples that were centrifuged 20 min at 4°C to attain two distinct phases. The lower phase containing the lipids was dried by SpeedVac for 45 minutes. Lipids were resuspended in 50 μl Chloroform/methanol (2:1). The lipid samples were loaded onto a 60 Ångström silica TLC plate (Scleicher&Schnell, Germany) pre-washed in acetone. The TLC plate was run for 45 minutes in a solvent system for ceramide (90 ml Ethylacetate:50 ml Ocanoic Acid:20 ml Acetic Acid), then dried followed by treatment 3 times with [3H]-enhancer spray. The plate was then developed for 48 hrs at −80°C. To quantify [3H]-ceramide, the area of interest was scraped into 3,5 ml scintillation fluid and analyzed by liquid scintillation. [3H]-ceramide was normalized to total [3H]-labeled cells loaded per lane. Treated samples were compared to vehicle treated control.
Total RNA was prepared using Qiagen RNA purification system as directed by the supplier (Qiagen, Valencia, CA). The samples were treated with Turbo DNase Kit to eliminate genomic DNA (Ambion, Austin, TX).
First strand cDNA synthesis was carried out according to the protocol for Omniscript Reverse Transcription (Qiagen, GmbH, Hilden, Germany). One microgram of RNA was used in the reactions for RT-PCR. The Beacon Designer 3 program (Biosoft International, Palo Alto, CA) was employed for design of primers (See Table 1). Primers were synthesized by Integrated DNA Technologies (Leuven, Belgium). The quantification of CerS compared to β-actin was carried out with an iCycler iQ (Bio-Rad Laboratories, Sundbyberg, Sweden). The iCycler iQ reaction detection system software from the same company was used for data analysis. cDNA was amplified using the qPCR Kit Platinum SYBR Green qPCR SuperMix-UGD with FITC (Invitrogen) according to the manufacturer’s instructions. The samples were divided into triplicates in a 96-well PCR plate (Abgene, Hamburg, Germany) and run at 95°C for 10 min followed by 40 cycles, each cycle consisting of 15 sec at 95°C and 1 min at 55°C. Threshold (Ct) cycle numbers were obtained from amplification of primers (Table 1). ΔCt values were calculated by subtracting the Ct value of β-actin from the Ct value of primers for the genes of interest. The relative fold increase (RFI) of the genes of interest was calculated as follows. The ΔCt for controls and treated samples was first determined. The ΔCt value was calculated by subtracting the Ct value for housekeeping control from the Ct value for the gene of interest. The RFI of genes of interest was calculated by the equation: RFI = 2-ΔCt. To use this calculation, the PCR efficiencies of the target and control assays must be similar. This was achieved by adjusting primer concentrations. The criterion for using the ΔCt method was fulfilled because by graphing serial dilutions of input cDNA of a random sample against ΔCt values (genes of interest-β-actin), the slope of the line as 0.1 (data not shown).
Cell Death ELISA (Roche, Mannhein, Germany) is a quantitative sandwich ELISA that detects histone and intra-nucleosomal DNA fragmentation by binding to two different monoclonal antibodies. It allows specific determination of mono- and oligonucleosomes in the cytoplasmatic fraction in cell lysates. The anti-histone-biotin antibody binds to histones H1, H2A, H2B, H3 and H4. The Anti-DNA-POD antibody reacts with double or single stranded DNA in the cytoplasm. In brief, cells were washed and resuspended in AIM-V medium. After treatment (see individual experiments in Results) cells were harvested and lysed with lysis buffer (Roche). The cell lysate was allowed to bind to the enzyme immunoassay plate for two hours together with immunoreagent containing anti-DNA-POD and anti-histone-biotin and incubation buffer. Thereafter, ABTS substrate (Roche, Mannheim, Germany) was added for 10 min. Adding stop solution terminated the reaction, and the absorption was determined at 405 nm.
The cells were split 2 days before the experiment in order to assure that they are in logarithmic phase during electroporation. On the day of electroporation, 5 million cells were resuspended in 100 μl of Nucleofector solution C (Amaxa Biosystems, Gaithersburg, MD)/sample. To the cell suspension, 20nM of duplex siRNA was added. The sample was then electroporated using Amaxa Nucleofector Systems, program X-001 (Amaxa Biosystems, Gaithersburg, MD). The electroporated cells were then resuspended in 2,5 ml pre-heated RPMI media supplemented with 10% FBS. Transfection efficiency was evaluated by quantitative PCR.
Viability of cells treated with R-MA and various modulators of ceramide metabolism was determined by using the XTT kit (Roche Diagnostics, Germany) according to manufacturer’s instructions. Absorbance was measured at 490 nm and reported as optical density.
Quantitative Cell Death ELISA was evaluated using the Kruskal-Wallis test comparing control and treated cells. P-values are presented in figure legends. The software Statistica (Statsoft AB, Tulsa, OK) was used.
We have previously shown that treatment of MCL cells with cannabinoids leads to an accumulation of ceramide . The effect was mediated by the cannabinoid receptors CB1 and CB2. To further study the time-course of ceramide accumulation, the MCL cell line Rec-1 was treated with 10 μM R(+)-methanandamide (R-MA), and total [3H]-ceramide was analysed using tritium labeling and liquid scintillation. After 30 min of treatment, only a slight increase in [3H]-ceramide accumulation was observed compared to the mock-treated control. After four hrs, there was a 30% increase in the accumulation of [3H]-ceramide, which was even more pronounced after 12 or 24 hrs (Fig 2).
The induction of ceramide in MCL is dependent of signaling via the CB1 receptor . To investigate the accumulation of different ceramide species, Rec-1 cells, which express the CB1 receptor, were treated with increasing doses of R-MA. As a control, the cell line SKMM, which has higher endogenous levels of ceramide C16 but lacks CB1 expression, was used. The cell lines were treated for 12 hrs with R-MA, and the levels of ceramide species were measured by HPLC/MSMS. A 2–3.5 fold increase of ceramide species C16, C18, C24 and C24:1 was observed when the MCL cell line Rec-1 was treated with 5 or 10 μM R-MA, while lower doses had no effect (Fig 3a). The levels of ceramide species C14, C20, C22, C22:1, C26 and C26:1 remained unaltered (data not shown). No accumulation of any of the ceramide species was observed after stimulation of the control cell line SKMM with R-MA (Fig 3a). These results show that R-MA induced a subset of ceramide species specifically in MCL. Pre-treatment with SR141716 (SR1) an antagonist with high affinity for CB1, significantly prevented the accumulation of ceramide species C16, C18 and C24, while SR144528 (SR2), which binds to CB2, partially inhibited the accumulation of C16 and C18 (Fig 3b). These results indicate that R-MA causes accumulation of specific ceramide species mainly via the CB1 receptor.
The relatively late increase in total ceramide levels following treatment with methanandamide raised the possibility that transcriptional upregulation of the CerSs could contribute to the accumulation of ceramide. Therefore, Rec-1 cells and SKMM cells were treated with 10μM R-MA (Fig 4a upper panel) for 12 hrs, and the expression of CerSs was investigated by quantitative real time PCR. In Rec-1, CerS3 and CerS6 showed a 2 and 2.6-fold increase of expression, respectively, after R-MA treatment. In contrast, CerS1 variant 2 showed decreased expression, while the expressions of CerS1 variant 1, CerS2, CerS4 and CerS5 were only marginally altered (Fig 4a upper panel). There was no substantial increase in the expression of CerSs in the control cell line SKMM. These results were confirmed using 10 μM of the synthetic cannabinoid Win-55,212-1 (Win55) (Fig 4a lower panel). When Rec-1 cells were pre-treated with 10nM of SR1, the induction of CerS3 and CerS6 was inhibited to a large extent (Fig 4b). SR2 caused a partial inhibition of the R-MA-induced upregulation of CerS3, while the attenuation of the induction of CerS6 in response to R-MA and of both CerSs in response to Win-55 was not significant (Fig 4b).
To investigate if the upregulation of CerSs following treatment with R-MA gave rise to increased accumulation of ceramides, Rec-1 cells were transfected with siRNA against CerS3 and CerS6. No change in accumulation of ceramide subspecies was observed when either enzyme was silenced. However, when both CerS3 and CerS6 were knocked down simultaneously, the accumulation of ceramide species C16 and C24 was abrogated (Fig 4c). Thus, CerS3 and CerS6 together appear to contribute to part of the earlier observed synthesis of ceramide species (Fig. 3) in response to treatment with R-MA. Similarly, suppression of both CerS3 and CerS6 caused a reproducible, but non-significant, decrease in the cell death induced by 10 μM R-MA (Fig. 4d).
To delineate the pathway leading to synthesis of ceramide after stimulation with R-MA, inhibitors targeting enzymes in the de novo pathway were used. Cells were labeled with radioactive tritium and pretreated with Myriocin, Fumonisin B1 (FB1) or C8CPPC, inhibitors to SPT, CerSs and DEGS, respectively [28–30] (Fig. 1), prior to treatment with 10μM R-MA. The formation of ceramide was disrupted following pre-treatment with each of these inhibitors (Fig 5a), strongly suggesting that ceramide was synthesized through the de novo pathway and not from the salvage pathway. The study was extended by analyzing different ceramide species using HPLC/MS/MS. After treatment with inhibitors and R-MA as described above, accumulation of ceramide species C16, C18, C24 and C24:1 was abrogated (Fig 5b).
To examine the role of de novo ceramide synthesis in the induction of apoptosis, cell death was estimated using Cell Death ELISA and AnnexinV PI staining combined with flow cytometry. Rec-1 cells were pre-incubated with Myriocin, FB1 or C8CPPC prior to treatment with 10μM R-MA. When de novo synthesis was disrupted, apoptosis induced by R-MA was abrogated (Fig 6a, b). Since these inhibitors act at different stages of the de novo pathway, the results show that ceramide (or downstream metabolites) and not dihydroceramide or other upstream metabolites in the de novo pathway is the primary mediator. This confirms and extends our previous observations were FB1, an inhibitor of CerS, abrogated the induction of cell death.
It has been shown in earlier studies  that increasing ceramide levels by manipulating ceramide metabolism can sensitize cells to cytotoxic treatment. In order to enhance the targeted cell death induced by cannabinoids in MCL , we combined the R-MA treatment with inhibition of two ceramide-metabolizing enzymes; SK-1 and GCS. Rec-1 cells were co-treated with 10μM R-MA and the SK-1 inhibitor SKI II or DMS, and after 72 hrs viability was examined by XTT. Co-treatment with SKI II (Fig 7a upper left panel) or DMS (Fig 7a upper right panel) potentiated the effect of R-MA on its own. Co-treatment with the GCS inhibitors C9DGJ (Fig 7b lower left panel) and PDMP (Fig 7b lower right panel) had a similar effect. To confirm that the effects observed using the inhibitors of ceramide metabolism in MCL cells were specific, SK-1 and GCS were silenced using siRNA. Rec-1 cells were transfected with 20nM siRNA duplexes binding either to either enzyme prior to treatment with 10 μM R-MA. SiRNA against SK-1 or GCS potentiated the decrease in viability induced by R-MA alone (Fig 7c).
To determine the effect of inhibitors to ceramide metabolism in combination with 10 μM R-MA on induction of apoptosis, Cell Death ELISA was performed. C9DGJ was used in this experiment since it is more specific to GCS than PDMP . SKI II was used since French et al  showed that it specifically inhibits the formation of S-1-P. Rec-1 cells were co-treated as above and analyzed after 24 hrs. In accordance with the decrease in viability, there was a potentiation of cell death when SKI II or C9DGJ was combined with 10μM R-MA (Fig 8a). The potentiation was confirmed using siRNA against SK-1 and GCS (Fig 8b). Taken together, these results demonstrate that inhibition of the transformation of ceramide into S-1-P or glucosylceramide potentiates the viability-suppressing and cell death-inducing effects of R-MA.
To investigate which ceramide species are converted by SK-1 and GCS in Rec-1, siRNA against each enzyme was used in combination with R-MA. Significantly increased levels of ceramide specie C16 were detected when SK-1 was inhibited by siRNA against SK-1 prior to treatment with R-MA (Fig 9a left panel), while the levels of C18 remained unaffected (Fig 9a right panel). Co-treatment with R-MA and siRNA against GCS induced significantly increased levels of both ceramide C16 and C18 (Fig 9b). No change in the levels of C24 or C24:1 was observed (data not shown). Thus, inhibition of SK-1 or GCS is likely to potentiate cell death in Rec-1 via R-MA induced accumulation of C16 or C16 and C18, respectively.
Ceramide is known to function as a second messenger in different cellular processes, e.g. induction of apoptosis and differentiation, and its accumulation can be induced by a variety of stimuli . Cannabinoids have been shown to induce ceramide accumulation in pancreatic cancer-, glioma- and leukemia cell lines [8, 9, 34] as well as in MCL . The production of ceramide following treatment with cannabinoids can be caused by either hydrolysis of sphingomyelin or de novo ceramide synthesis [35, 36]. In the current study, we have shown accumulation of ceramide species C16, C18, C24 and C24:1 in the MCL cell line Rec-1 after long-term stimulation with the stable endocannabinoid analogue R-MA. Ceramides exhibit a tissue-dependent bias for amide-linked fatty acids, characterized by chain length, degree of saturation and degree of hydroxylation. C16 ceramide is known to be most abundant in fibroblasts, endothelial cells and immune cells [37, 38]. In accordance with the present study using a cell line of B-cell origin, C16 and C24 ceramide species were accumulated after BcR crosslinking in the B-cell line Ramos [4, 5].
It has been shown previously that cannabinoids can induce apoptosis via de novo synthesis in C6 glioma cells . In the same study, increased activity of SPT was the rate-limiting step in ceramide synthesis de novo . Here, we observed upregulation of messages for ceramide synthases within the same pathway. The six known forms of CerSs are active in two processes, acylating either dihydrosphingosine to form dihydroceramide or sphingosine to form ceramide. After stimulation with R-MA, the Rec-1 cells overexpressed CerS3 and CerS6 two and 2,6 times respectively. CerS6 is a major CerS that is expressed at high levels in a variety of tissues , and is associated with synthesis of ceramide species C14 and C16. CerS3 is regarded as a minor CerS that has mainly been described in skin and testis [10, 40], and has been shown to synthesize the longer ceramide species C18, C20 and C24 [10, 11]. The induction of CerS3 after stimulation with cannabinoids is intriguing, since database searches show that CerS3 expression is rarely present in lymphoma tissue, while CerS6 expression is present in a majority of the lymphomas (Omnibus GEO database).
Silencing of either CerS3 or CerS6 using siRNA had no effect, while simultaneous knockdown abrogated the accumulation of C16, C18 and C24. Our results suggest that CerS3 and CerS6 may have redundant functions in the production of ceramide species, and indicate that CerS6 and CerS3 are not restricted to the synthesis of C14 and C16 or longer ceramide species, respectively. Similarly to the effects on ceramide accumulation, simultaneous suppression of CerS3 and CerS6 had a reproducible, but non-significant inhibitory effect on R-MA-induced cell death. Generally, the induction of ceramide accumulation and cell death following R-MA treatment was greater in untransfected compared to transfected cells, as observed when comparing Fig 3 and and4d4d or Fig 8a and 8b, respectively. Thus, electroporation itself is likely to affect the background levels of ceramides and DNA fragmentation, rendering the effects of R-MA less pronounced.
The CB1 antagonist SR1 completely inhibited the R-MA-induced accumulation of ceramide species C16, C18 and C24 and the upregulation of CerS3 and CerS6 following treatment with R-MA or Win55. SR2, an antagonist to CB2, partially inhibited the increase in C16, C18 and CerS3, while the inhibitory effects on the accumulation of other ceramide species and on the upregulation of CerS6 were not significant. Given the selectivity of R-MA towards CB1, it may seem surprising that SR2 counteracted the induction of ceramide accumulation and the upregulation of CerSs. These results are in accordance with our earlier studies showing that antagonists to either CB1 or CB2 attenuated cell death induced by R-MA in MCL and other lymphomas expressing both receptors [7, 41]. At the doses used, it is possible that R-MA acts as an agonist also to CB2, despite its much higher affinity to CB1. Alternatively, the binding of SR2 to CB2 induces changes downstream of the receptor that affect the signaling via CB1.
We have previously blocked cell death in Rec-1 cells by inhibiting CerSs with FB1 . The same inhibitor has been used to prevent induction of ceramide accumulation by cannabinoids in glioma cells . However, CerSs can act in two pathways; both in regeneration of ceramide from sphingosine and in de novo synthesis, see Fig 1. To exclude the involvement of SL degradation, we here added inhibitors to enzymes that are active only in the de novo pathway; SPT and DEGS. Dbaibo et al.  have shown that inhibition of SPT with myriocin abrogates cell death induced by arsenic trioxide in T-cell leukemia and lymphoma. In the present study, inhibition of either of three different enzymes, SPT, CerS and DEGS, led to an abrogation of ceramide accumulation and suppression of cell death in MCL (Fig 1). This fortifies the role of de novo synthesis in the responses to R-MA.
It has previously been suggested that ceramide is the active mediator of apoptosis, while dihydroceramide is merely an inactive precursor to ceramide . In these studies, exogenous ceramide and dihydroceramide have been added to cells or mitochondria  . In contrast, high levels of dihydroceramide were observed in HL-60 leukemia cells prior to cell death,  and SMS-KCNR cells were cell cycle arrested when DEGS was inhibited . In our MCL cells, the inhibition of DEGS using C8-CPPC caused accumulation of dihydroceramide species (data not shown) and inhibited the induction of cell death by R-MA. This supports the theory that dihydroceramide species cannot induce apoptosis when the transformation to ceramide is inhibited.
The above conclusion on the central role of ceramide suggested that the cell death-promoting effects of R-MA could be enhanced by inhibiting the transformation of ceramide into species with opposing effects. Inhibition of both SK-1 and GCS potentiated the effects of R-MA on viability and cell death. French et al  have shown that SKI II is a specific inhibitor to S-1-P formation, which also inhibits tumor growth in vivo. In view of the significant potentiation of the R-MA-induced effects observed using SKI II in vitro, future in vivo studies using a combination of R-MA and SKI II in mice xenotransplanted with MCL cells are warranted. To be assured that the effects observed in our experiments were enzyme-specific, GCS and SK1 were silenced using siRNA. In addition to potentiation of the effects induced by R-MA, silencing of SK-1 led to a decrease in cell viability by itself. Taha et al  and Sarkar et al  have observed that down-regulation of SK-1 by siRNA in MCF-7 cells resulted in a reduction of cell viability. Interference with GCS RNA has also been shown to affect growth of neuroepitelioma  and sensitize breast carcinoma cells to cytotoxic drugs . However, the viability of an astrocytoma cell line following treatment with the cannabinoid Δ-9 tetrahydrocannabinol (THC) was not affected by inhibitors or siRNA against GCS . In our cells, knockdown of SK-1 led to significant potentiation of R-MA-induced C16 accumulation, whereas the accumulation of both C16 and C18 was potentiated using siRNA against GCS. Instead, the levels of C24 and C24:1 remained unaltered. It is possible that these ceramide species are metabolized by other ceramide-converting enzymes, e.g. sphingomyelin synthase or ceramide kinase .
In conclusion, we have shown that induction of CerSs regulate de novo ceramide synthesis in response to the stable endocannabinoid analogue R-MA in MCL cells. The effect of inhibition of DEGS supports earlier studies showing that ceramide, and not dihydroceramide, is the active mediator of apoptosis. Moreover, inhibition of enzymes that convert ceramide to growth-promoting sphingolipid species potentiated the ceramide accumulation and cell death induced by R-MA in MCL. Cannabinoids have been suggested as a new non-toxic therapeutic option for cancer treatment . This is the first study showing that the cytotoxic effect of a cannabinoid can be enhanced by modulation of ceramide metabolism. The results suggest that interference with ceramide conversion may provide a tool to enhance the targeted cell death-promoting effects of cannabinoids in MCL and other malignant lymphomas overexpressing the CB1 receptor.
We thank Dr Alicja Bielawska for her helpful technical advice and comments regarding the manuscript. This study was supported by grants from The Swedish Cancer Society, The Swedish Research Council, The Gunnar Nilsson Cancer Foundation, The Cancer Society in Stockholm, The Magnus Bergvall Foundation, Karolinska Institutet Funds, Stockholm County Council funds, The Fernström funds, The Swedish Society of Medicine, Boehringer-Ingelheim travel fund and NIH (CA87584).