N-palmitoylethanolamine and palmitic acid were obtained from Tocris Bioscience. The enzyme inhibitors URB597, CAY10402 and CAY10499 were bought from Cayman Europe and MAFP from Tocris Bioscience. CCP (N-cyclohexanecarbonylpentadecylamine) was kindly synthesized in our lab by Coco N. Kapanda (Université catholique de Louvain, Belgium) according to the synthetic procedure described by Vandevoorde [32] and Tsuboi [33]. All the receptor antagonists (AM251, capsazepine, GW6471, T0070907 and (-)-cannabidiol) were purchased from Tocris Bioscience. All drugs were prepared as 20 mM stock solutions in DMSO and extemporaneously diluted in media for the experiments conducted on cells. The final concentration of DMSO was kept below 0.2%. [³H]-anandamide (60 Ci/mmol), [³H]-2-oleoylglycerol (40 Ci/mmol) and [³H]-PEA (20 Ci/mmol) were purchased from American Radiolabeled Chemicals (St Louis, MO, USA).

The murine melanoma cell line B16 was obtained from the American Type Culture Collection and routinely cultured in Minimum Essential Medium (MEM) α medium supplemented with 10% fetal bovine serum, 100 UI/ml penicillin, 100 mg/ml streptomycin and MEM Vitamins Solution. The human melanoma cell line MZ2-MEL.43 was kindly given by Pierre Coulie (Université catholique de Louvain, Belgium) and cultured in Iscove's medium supplemented with 10% fetal calf serum. Cells were maintained at 37°C in a humidified atmosphere of 5% CO₂.

Tumors were induced in 5 week-old male C57BL/6 mice (Elevage Janvier, France) by s.c. flank injection of 10⁶ B16 melanoma cells. When the tumor reached a volume of 20-40 mm³, mice were randomly divided in groups and injected i.p. for six days with vehicle or treatment (either PEA and/or URB597 at 10 mg/kg/day, daily). Tumor size was calculated according to the following formula: (4π/3) × (width/2)² × (length/2). The procedure was approved by a local ethical review committee according to national animal care regulations.

Total RNA was extracted from the cultured cells with the TriPure Isolation reagent (Roche). To measure mRNA expression, reverse transcription was performed using the Reverse Transcription System (Promega) and the generated cDNA was amplified by PCR using the primers mentioned in the Table 1. Polymerase chain reactions were performed according to the following parameters: 95°C for 10 min, 95°C for 3 s, 60°C for 26 s, and 72°C for 10s (45 cycles). After amplification, agarose gel electrophoresis was used to detect the expression of the genes.

Quantitative PCR (qPCR) was performed to study the quantitative mRNA expression of the FAAH and the NAAA. RPL19 was used as house keeping gene. The samples were run in a 96-well reaction plate and data were analyzed according to the 2^-ΔCT method.

The effect on cell viability of the different treatments was measured using MTT assay, which is based on the transformation of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) in formazan crystals by the mitochondrial succinate dehydrogenase of viable cells. Cells were plated in 96-well plates at a density of 2000 cells/well in medium supplemented with 10% serum. After 5 h of incubation at 37°C in a 5% CO₂ humidified atmosphere, test compounds diluted in culture medium were added in each well for 24 h, 48 h or 72 h. The medium was then removed and 100 μl of MTT solution (0.3 mg/ml) were added for a 2 h incubation. The MTT solution was removed, replaced by 100 μl DMSO to dissolve the crystalline formazan product and the absorbance was read at 570 nm (with a reading at 650 nm as reference) using a microplate spectrophotometer.

For the treatments with the receptor antagonists, only the 72 h time point was considered and the antagonists were added 1 h before the beginning of the cytotoxic treatment.

Detection and quantification of apoptosis was performed by the analysis of phosphatidylserine on the outer leaflet of apoptotic cell membranes using Annexin-V-Fluorescein. Propidium iodide was used for the differentiation from necrotic cells. Cells were incubated for 24 h with the cytotoxic treatment before being stained with the Roche Annexin-V-FLUOS Staining kit (Mannheim, Germany) following the manufacturer's instructions. Cells were examined using a fluorescence microscope from Optika (Ponteranica, Italy). Pictures were taken with a Moticam 2300 from Motic (Hong Kong, China).

Cells (10⁷ cells/condition) were seeded in 10% FBS media for 12 h prior to the incubation (8 h) with drugs, or vehicle, in 1% FBS media. Cells and media were then recovered and the lipids extracted, in the presence of deuterated standards (200 pmol), by adding 14 ml of chloroform and 7 ml of methanol. Following vigorous mixing and sonication, the samples were centrifuged and the organic layer was recovered and then dried under a stream of N₂.

Endocannabinoid levels in B16 tumors were quantified by directly homogenizing the tissue in chloroform (14 ml) before adding the deuterated standards, methanol (7 ml) and water (3.5 ml). After mixing and phase separation by centrifugation, the organic layer was recovered and dried under a stream of N₂.

Both for cells and tumors, the resulting lipid extracts were purified by solid-phase extraction using silica and elution with an EtOAc-Acetone (1:1) solution [34,35]. The resulting lipid fraction was analysed by HPLC-MS using an LTQ Orbitrap mass spectrometer (ThermoFisher Scientific) coupled to an Accela HPLC system (ThermoFisher Scientific) [36]. Analyte separation was achieved using a C-18 Supelguard pre-column and a Supelcosil LC-18 column (3 μM, 4.6 × 150 mm) (Sigma-Aldrich). Mobile phases A and B were composed of MeOH-H₂O-acetic acid 75:25:0.1 (v/v/v) and MeOH-acetic acid 100:0.1 (v/v), respectively. The gradient (0.5 ml/min) was designed as follows: transition from 100% A to 100% B linearly over 15 min, followed by 10 min at 100% B and subsequent re-equilibration at 100% A. We performed MS analysis in the positive mode with an APCI ionisation source. The capillary and APCI vaporiser temperatures were set at 250°C and 400°C, respectively. N-acylethanolamines were quantified by isotope dilution using their respective deuterated standards with identical retention [34]. The data were normalized by cell number in the in vitro experiments and by tumor sample weight in the in vivo testing.

Tumors were excised after five or six days of treatment with tested compounds or vehicle and either fixed in 4% paraformaldehyde or embedded in O.C.T. compound for standard paraffin sections or for cryosectioning, respectively. Tissue samples were sliced in 5 μm sections for paraffin samples and 10 μm sections for frozen samples. The paraffin sections were stained with Haematoxylin & Eosin and photographed on a Zeiss MIRAX microscope to allow a global overview of tumor necrosis. Induction of apoptosis was assessed by TUNEL assay using an in situ cell death detection kit (Roche Diagnostics, Vilvoorde, Belgium) on frozen slices. Vascularization was evaluated by immunostaining of tumor cryosections using an antibody directed against CD31 (BD Biosciences, San Jose, CA). Nuclei were also counterstained with 4,6-diamidino-2-phenylindole (DAPI). Necrosis, apoptosis and blood vessel area were quantified using Frida software and expressed as a percentage of the whole tumor area.

The antiangiogenic effect of the different treatments was determined by seeding 15 × 10³ endothelial cells (HUVEC) in 96-well plates containing Matrigel (BD Biosciences). Formation of capillary-like tubular structures was quantified by counting the number of tube intersections in each well.

Values were expressed as mean ± SEM. Statistical analysis was performed by ANOVA or by unpaired Student's t test.

In order to evaluate the hydrolysis of AEA, 2-AG and PEA, we used [³H]-AEA, [³H]-2-OG and [³H]-PEA and found that B16 cell homogenates significantly hydrolyzed endocannabinoids (Figure 1A). When looking for the expression of enzymes responsible for endocannabinoid hydrolysis, we detected mRNA expression of the major endocannabinoid degrading enzymes, i.e. FAAH, NAAA and MAGL (Figure 1B). Since the aim of this work was to increase endocannabinoid cytotoxicity by inhibiting their hydrolysis, we ensured that AEA, 2-AG and PEA reduced cell viability in our B16 model using a MTT test. We observed that at 10 μM and already after 24 h of treatment, AEA, 2-AG and PEA decreased cell viability, in comparison to vehicle (Figure 1C). This effect was amplified after 48 h and 72 h of incubation and was slightly more pronounced for PEA. Therefore, we further investigated the cytotoxic effect of PEA and obtained a dose-response when this molecule was tested at 1, 10 and 20 μM for 72 h (Figure 1D). Because endocannabinoids' fatty acid metabolites are known to possess numerous functions we tested whether they affect cell viability of B16 cells. We found that neither palmitic acid (Figure 1E) nor arachidonic acid (data not shown) affect cell viability. Taken together these data support our hypothesis that increasing endocannabinoid levels, by blocking their degradation, would reduce tumor cell viability.

Five inhibitors that were reported to decrease N-palmitoylethanolamine hydrolysis either by inhibiting FAAH (URB597, CAY10402, MAFP and CAY10499) or NAAA (CCP) (see Additional file 1) were tested at 1 μM and 10 μM on total cell homogenates. The inhibition assays were also performed on intact cells in culture to confirm that the inhibitors do reach their targets in culture conditions (Table 2).

We observed that URB597, CAY10402, MAFP and CAY10499 all inhibit PEA hydrolysis in homogenates and cultured cells even though the inhibition is slightly less pronounced in the latter case. Note that the use of CCP did not reduce PEA hydrolysis in homogenates and only decrease it by 26% ± 7.0 in intact cells at 10 μM. The absence of inhibition observed in homogenates compared to cells in culture could be explained by a NAAA activity known to be the highest at acidic pH while the assay was performed on homogenates at physiological pH [12]. Quantitative measurements of FAAH and NAAA mRNA expression were also performed in order to investigate the possibility that a high level of FAAH, in comparison to NAAA, could lead to the lack of efficacy of CCP on PEA hydrolysis in our system. The results indicated that the relative mRNA levels of the two enzymes were in the same order of magnitude (Figure 2) and thus the lack of NAAA expression at the mRNA levels does not account for the inefficiency of CCP.

Next, PEA (10 μM) was co-incubated for 72 h with the FAAH inhibitors - URB597, CAY10402, MAFP, CAY10499 - and the NAAA inhibitor - CCP - at 10 μM, with the objective of enhancing individual effect on cell viability (Figure 3A). All four inhibitors enhanced PEA-induced cytotoxicity although the effect was more pronounced with the selective inhibitors (URB597, CAY10402) as compared to the dual FAAH/MAGL inhibitors (MAFP, CAY10499). Of note, the co-incubation of PEA and URB597 dramatically reduced cell viability which was no more than 11% of the vehicle control after 72 h of treatment. In line with the results obtained in the enzymatic inhibition assay, CCP had no potentiating effect on cytotoxicity when added to PEA.

The effects on cell viability of the inhibitors alone were also evaluated at 10 μM and after 72 h of incubation. We did not use CCP anymore because it was poor at inhibiting PEA hydrolysis in our cellular model and did not induce supplemental cytotoxicity when co-incubated with PEA. Only the three FAAH inhibitors URB597, MAFP and CAY10499 provoked a significant increase in cytotoxicity while the reversible FAAH inhibitor CAY10402 did not influence cell viability (Figure 3B).

Because the PEA-URB597 combination produced the highest cytotoxicity (Figure 3A), we focused on these compounds for the next experiments. We wondered if URB597 could exert its cytotoxicity through inhibition of FAAH and subsequent increase of PEA concentrations in our conditions. For this purpose, PEA levels were measured in B16 cells using an isotope-dilution HPLC-MS method and we observed that incubating B16 cells with URB597 (1 μM, 8 h) resulted in increased PEA levels (163 ± 13% of the control) (Figure 3C).

We next looked at the influence of PEA and URB597 on cell death by measuring annexin-V positive cells (A+/PI-) and double stained cells (A+/PI+) representing apoptotic and necrotic (or late apoptotic) cells, respectively. Translocation of phosphatidylserine is an early event in apoptosis and its measurement allows the detection of cells undergoing caspase-dependent or independent apoptosis. Treatment of B16 cells with PEA or URB597 (10 μM, 24 h) did not result in an increased number of dead cells as compared to vehicle, even though PEA tended to enhance cell death (Figure 4). On the contrary, co-incubation with PEA and URB597 increased the percentage of cells dying by both apoptosis and necrosis.

We then asked if the cytotoxic effects of the PEA-URB597 combination could be mediated by the classical molecular targets of endocannabinoids. Thus, we co-incubated PEA, URB597 or PEA-URB597 with AM251 (0.1 and 1 μM), capsazepine (0.1 and 1 μM), GW6471 (1 and 5 μM), T0070107 (1 and 5 μM) and cannabidiol (1 and 10 μM). These compounds are selective antagonists of the CB1, TRPV1, PPARα, PPARγ and GPR55 receptors respectively, the mRNA of which were found in B16 melanoma cells (see Additional file 2). None of the antagonists reduced the effect of PEA (10 μM) or URB597 (10 μM) alone or in combination (see Additional file 3). We would like to point out that antagonist concentrations were selected according to the literature and that we tested their own cytotoxicity to exclude the possibility that they could affect B16 cell viability by themselves (see Additional file 4). At 10 μM, cannabidiol has enhanced the cytotoxic effect of PEA and URB597 when each of these drugs was used alone. One explanation is that this compound is a weak agonist of the TRPV1 and was shown to activate this receptor in this range of concentrations [37].

Since PEA and URB597 were able to induce cell death when used in co-incubation in vitro, we wanted to confirm that this was also the case in vivo. Therefore, we evaluated their effect on malignant melanoma growth in C57BL/6 mice. Tumors were generated by subcutaneous injection of B16 cells, and when tumors reached a volume of approximately 20-40 mm³, mice were treated by intraperitoneal injections of vehicle, PEA and/or URB597 for up to six days. The tumor size of mice treated with PEA or URB597 alone were not significantly different from those injected with vehicle. However, co-administration of PEA and URB597 resulted in a significantly reduction of tumor growth and of their size at the end of the experiment (Figure 5A). We also weighted tumors at the end of the experiment and observed a significant difference between normal and PEA/URB597 treated tumors (Figure 5B).

With the aim to correlate the anticancerous effect of PEA and URB597 with endocannabinoid levels inside the tumor, we measured AEA, 2-AG and PEA amounts in the excised tumors by HPLC-MS. On the one hand, AEA and 2-AG levels were not significantly affected by PEA, URB597 or both molecules injection, even though AEA concentration tended to increase after URB597 or PEA/URB597 treatments (Figure 6A, B). On the other hand, PEA levels were increased after co-treatment with PEA and URB597 but not if these compounds were injected alone, even though a trend towards increased levels was observed following URB597 administration (Figure 6C).

Necrosis and apoptosis were quantified on tumor slices after Haematoxylin & Eosin (H&E) and TUNEL staining respectively. Tumors were excised after six days of co-treatment with PEA and URB597. Vehicle-treated tumors were excised after five or six days of injection in order to be able to compare either tumor treated during the same period of time, or tumors having the same volume at the end of the experiment. Indeed, we observed that size of six days drug-treated tumors and five days vehicle-treated tumors did not significantly differ (Figure 5A, B). It is known that more voluminous tumors may present larger necrotic regions and we wanted to exclude this artifact. Here we show that tumors co-treated with PEA and URB597 present enlarged necrotic regions (53 ± 6%) as compared with both tumors treated during five or six days with vehicle (28 ± 3% and 38 ± 2% respectively) (Figure 7A). These results were consistent with observations we made during in vitro assays and demonstrate that drug treatment delays tumor progression by provoking cell death. Figure 7B displays representative tumor slices after H&E staining. TUNEL assay on tumors slices did not show more positively stained cells in treated tumors than in control tumors (Figure 7C).

Finally, we wondered if the increase in the initial level of necrosis produced by co-injection of PEA and URB597 could be the result of vascular events. Thus we performed an endothelial tube formation assay using HUVEC. Here, PEA (10 μM) and URB597 (10 μM), incubated alone or in co-incubation, did not alter the capacity of endothelial cells to form tubes when cultured on Matrigel (Figure 7D). Additionally, evaluation of tumor vascularization by immunostaining did not reveal any significant change in blood vessel area between treated and untreated mice (Figure 7E).

In a view to strengthen the potential interest of using PEA and URB597-based treatment for melanoma growth management, we measured the effect of these compounds on a human melanoma cell line. URB597 (10 μM) slightly decreased cell viability of MZ2-MEL.43 melanoma, which was 90% of the vehicle control after 72 h of treatment. The cytotoxicity produced by PEA (10 μM) led to a reduction of cell viability to 66%, while it was potentiated by URB597 to reach 48% of residual viable cells (Figure 8).