BEO was kindly provided by CAPUA s.r.l. (Reggio Calabria, Italy; www.webcapua.com). BEO contained 39.76% limonene, 29.59% linalyl acetate, 8.09% γ-terpinene, 7.32% ß-pinene, 6.71% linalool, 1.28% α-pinene, 1.23% sabinene, 1.00% myrcene, 0.45% ß-bisabolene, 0.35% terpinolene, 0.34% neryl acetate, 0.33% α-thujene, 0.32% geranyl acetate, 0.31% ß-caryophyllene, 0.31% trans-α-bergamotene, 0.25% geranial, 0.17% phellandrene, 0.16% neral, 0.13% p-cimene, 0.12% decanal, 0.03% camphene, 0.02% o-cimene as determined by GC-FID analysis, and 2.331 g/kg bergapten (HPLC/UV). BEO was diluted 1∶1 in a 1∶9 water/ethanol solution and then further diluted in culture medium to obtain final concentrations of 0.005–0.03%; ethanol was added to the medium of parallel control cultures to obtain concentrations ranging from 0.0045 to 0.027% (vehicle-treated cells).

d-Limonene (CAS No 5989-27-5) and linalyl acetate (CAS No 115-95-7) were obtained from Sigma-Aldrich (Milan, Italy). Stock solution (10%) of limonene in dimethyl sulfoxide (DMSO; Sigma-Aldrich) and linalyl acetate in 90% ethanol were further diluted in culture medium to obtain final concentrations of 125–750 µM limonene and 62.5–375 µM linalyl acetate. DMSO or ethanol were added to the medium of parallel control cultures (vehicle-treated cultures) to obtain concentrations equal to those present in drug-treated cultures and these never exceeded 0.108% DMSO or 0.066% ethanol.

Adherent SH-SY5Y human neuroblastoma cells were obtained from ICLC-IST (Genoa, Italy); MCF7 human breast cancer cells were purchased from American Type Culture Collection (Manassas, VA, USA). Cell lines were routinely grown in RPMI medium (Invitrogen, Carlsband, CA) cointaining 10% foetal bovine serum (FBS: Invitrogen, Carlsband, CA). Cells were seeded onto 6-well plates; 24 h after plating, the growth medium was replaced with fresh normal medium (control cells) or medium supplemented with vehicle or different compounds. Cells were exposed to BEO (0.005–0.03%), d-limonene (Limo; 125–750 µM) or linalyl acetate (LinAc; 62.5–375 µM) for the indicated period of time. Stock solution of bafilomycin A1 (BafA1, 100 µM; Sigma-Aldrich) or rapamycin (2.74 mM; Sigma Aldrich) in DMSO were further diluted in culture medium. BafA1 or rapamycin were applied to SH-SY5Y cells 2 h and 48 h respectively before exposure to BEO or Limo for 1 h. Where indicated cells were subjected to starvation in serum-free media for 24 h and exposed to BEO (0.01% and 0.02%) for 1 h.

After treatment, cells were collected, co-labelled with 2 µg/ml propidium iodide (PI) and 0.1 µM fluorescein diacetate (FDA) and analyzed by a FACSCalibur flow cytometer (Becton Dickinson, CA, USA) to detect the proportion of viable (FDA⁺/PI^-), apoptotic (FDA^-/PI^-) and necrotic cells (FDA^-/PI⁺) by quadrant analysis as previously described [22], [23]. Ten thousands events/sample were acquired. Data acquisition and analysis were performed with CellQuest Software (Becton Dickinson).

SH-SY5Y cells were plated on coverslips, treated with BEO for the indicated time and fixed with formalin solution, containing 4% paraformaldehyde (Sigma-Aldrich). After blocking with 10% donkey serum for 30 min, coverslips were incubated with primary antibodies against microtubule-associated protein 1 light chain 3 (LC3; code PD036, MBL International Corporation, Nagoya, Japan; 1∶500 dilution) and/or lysosomal-associated membrane protein 1 (LAMP1; Developmental Studies Hybridoma Bank, Iowa City, IA, USA; 1∶400 dilution) overnight at 4°C followed by 1h incubation with Alexa Fluor secondary antibodies (Alexa Fluor 488 donkey anti-mouse and Alexa Fluor 594 donkey anti-rabbit, Molecular Probes, Life Technologies Italia, Monza, Italy; 1∶500 dilution). Coverslips were mounted with Vectashield solution containing 1.5 µg/ml 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, CA, USA) to visualize nuclei. Images were acquired under a 63X objective using a confocal microscope (Leica TC-SP2 Confocal System; Leica Microsystems Srl, Milan, Italy).

SH-SY5Y cells were lysed in ice-cold RIPA buffer (50 mM Tris-HCl, pH 8, 150 mM NaCl, 1 mM EDTA, 0.1% SDS, 1% IGEPAL and 0.5% sodium deoxycholate) containing a protease inhibitor cocktail (code P8349; Sigma-Aldrich). For detection of phosphoproteins, SH-SY5Y cells were lysed in ice-cold buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 1% Triton, 1 nM okadaic acid, protease (code P8349, Sigma-Aldrich) and phosphatase (code 524625, Calbiochem) inhibitor cocktails.

Lysates were centrifuged for 15 min at 20.000 g at 4°C and supernatants were assayed for protein content by Bio-Rad DC Protein Assay Kit (Bio-Rad Laboratories, Milan, Italy). Equal amount of total proteins was resolved by 12% SDS-polyacrylamide gel electrophoresis and transferred onto PVDF membranes (Immobilon-P, Sigma-Aldrich). Membranes were blocked with 5% non-fat milk in Tris-buffered saline containing 0.05% Tween 20 for 1 h at room temperature. Primary antibodies were incubated overnight at 4°C, followed by a horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at room temperature. Protein bands were visualized with the ECL Western Blotting Detection kit (ECL, Amersham Biosciences, GE Healthcare, Milan, Italy), and the chemiluminescence signal detected using X-ray films (Hyperfilm ECL, Amersham Biosciences). Exposure time to obtain the optimal density of the images varied from 5 s to 20 min. Autoradiographic films were scanned, digitalized at 300 dpi and band quantification was performed using ImageJ software (NIH, Bethesda, MD, USA).

The following primary antibodies and dilutions were used: anti-LC3 1∶1000 (code PD036; MBL), anti-p62/SQSTM1 1∶2000 (code PM045; MBL), anti-Beclin-1 1∶2000 (code PD017; MBL), anti-phospho p70^S6K (Thr389) 1∶1000 (code 9234; Cell Signalling Technology), anti-p70^S6K 1∶2000 (code 2708; Cell Signalling Technology), anti-phospho ULK (Ser757) 1∶1000 (code 6888; Cell Signalling Technology), anti-ULK 1∶1000 (code 8054, Cell Signalling Technology), anti-actin 1∶1000 (clone AC-40; Sigma-Aldrich), anti-β-tubulin 1∶30000 (clone B-5-1-2; Sigma-Aldrich), anti-GAPDH 1∶40000 (clone 6C5; Applied Biosystems, Carlsbad, CA, USA). HRP-conjugated goat anti-rabbit or anti-mouse IgG (Pierce Biotechnology, Rockford, IL, U.S.A) were used as secondary antibodies.

Cells were transfected with small interfering RNA (siRNA) targeting Beclin-1 or scramble control (Dharmacon RNAi Technologies; Chicago, IL, USA) at a final concentration of 50 nM, 48 h before BEO exposure, by using Lipofectamine 2000 transfection reagent (Invitrogen, Life Technologies) according to the manufacturer's instruction.

Data are expressed as the mean ± s.e.m. of the indicated number of independent experiments and evaluated statistically for difference by ANOVA followed by Tukey-Kramer test for multiple comparisons. Where indicated, Student's t test was used to evaluate differences between two means. A value of P less than 0.05 was considered to be significant.

Our previous data show that a significant percentage of apoptotic and necrotic cell death occurs within 1 h exposure to 0.02% BEO and this dramatically increases in SH-SY5Y cells incubated for the same time period with 0.03% BEO; conversely, no cytotoxic effects are observed following incubation with lower concentrations (0.005–0.01%) of BEO for 1 h and up to 24 h [22]–[24]. Accordingly, immunofluorescence analysis here revealed DNA fragmentation, chromatin marginalization, and nuclear shrinkage and condensation in a significant proportion of cells exposed to 0.02% BEO for 1 h (Figure 1); signs of nuclear alterations, such as nuclear condensation, were also detected at an earlier time (30 min; Figure 1), though less pronounced. Nuclear morphological alterations were absent in cells treated for up to 1 h with lower concentrations (0.005–0.01%) of BEO (Figure 1). Based on these and previous observations, biochemical assessment of autophagy was initially performed following 1 h exposure to 0.005–0.03% BEO, i.e. a dilution range encompassing both non cytotoxic and cytotoxic concentrations. As shown in Figure 2A, treatment with BEO resulted in a concentration-dependent conversion of the non-lipidated form of LC3, LC3I, to the lipidated form, LC3II, that specifically associates with the membrane of expanding autophagosomes [25]. As compared to vehicle-treated cells, enhanced LC3I to LC3II conversion, measured as the LC3II/LC3I ratio, was detected in cells exposed to 0.01–0.03% BEO but not to a lower concentration (0.005%) (Figure 2A). Changes in LC3II levels were paralleled by a concentration-dependent reduction of the selective autophagy substrate p62 [26], [27] (Figure 2A), indicative of an increased autophagy. To analyze the autophagic flux, the effects of BEO were studied in the presence of the lysosomal inhibitor bafilomycin A1 (BafA1), which blocks autophagosome-lysosome fusion [28], thus preventing LC3II degradation [29]. In SH-SY5Y cells pretreated with BafA1, BEO further enhanced LC3II levels as compared with BafA1 alone (Figure 2 B and C) which indicates that BEO induces autophagosome formation. In fact, immunofluorescence analysis of intracellular distribution of endogenous LC3 protein showed that the latter was evenly distributed in vehicle-treated cells, with scattered LC3 punta (autophagosomes) reflecting ongoing (basal) autophagy, whereas a gradual increase in LC3 punctal pattern occurred in cells treated 30 min before with increasing concentrations of BEO (Figure 3). To determine whether autophagy modulation precedes the occurrence of nuclear alterations on treatment with 0.02% BEO, biochemical and morphological assessment of autophagy was carried out 15 min after BEO exposure. The results show that, at this time, BEO has stimulated the formation of autophagosomes, yet nuclei maintain a normal morphology; further, most of the LC3 puncta colocalize with the lysosomal protein LAMP-1, indicating that autophagosome-lysosome fusion occurred (Figure 4A). Coincident with this, enhanced LC3II/LC3I ratio and reduced levels of the autophagy substrate p62 were, indeed, revealed by western blotting analysis (Figure 4B). Altogether, these results indicate that 0.02% BEO rapidly stimulates autophagic flux in SH-SY5Y cells and this temporally precedes the profound alterations of nuclear morphology detected at later time points (Figure 1). To ascertain whether autophagy might have a causative role in cell death triggered by BEO, the latter was evaluated in cells pretreated with the autophagy inhibitor, BafA1. We found that BafA1 given alone did not affected neuroblastoma cell viability nor it protected from cell death induced by BEO; rather, BafA1 caused a trend toward an increase in the percentages of both necrotic and apoptotic cells induced by 0.02% BEO (Figure 2D).

The ability of BEO to modulate autophagic markers is not cell-line specific because enhanced LC3-II expression and reduced p62 levels were also observed following 1 h exposure of human breast cancer MCF7 cells to increasing concentrations of BEO (Figure 5).

The above reported results demonstrated that BEO enhances basal autophagy. Further experiments were then performed to investigate the effect of BEO on stimulated autophagy. To this end, autophagy was induced by serum starvation [29]. SH-SY5Y cells exposed to serum deprivation for 24 h showed an increase in LC3I/LC3II conversion and a reduction of p62 levels when compared to not starved cells (Figure 6A). Treatment of serum deprived cells with BEO induced further accumulation of LC3II and reduction of p62 levels (Figure 6A). These results indicate that BEO enhances basal autophagy and autophagy induced by serum starvation. Furthermore, they suggest that signalling events in BEO-induced autophagy might be partially distinct from those induced by starvation known to stimulate autophagy through inhibition of mTOR kinase [2]. To establish whether or not BEO activates autophagy by repressessing mTOR, the phosphorylation level of p70^S6K (Thr389) and ULK1 (Ser 757), two downstream targets of mTOR kinase [30], [31], was examined at 5 and 15 min after addition of 0.01 and 0.02% BEO. As shown in Figure 7A and B, no change in the level of phosphorylation of p70^S6K (Thr389) and ULK1 (Ser 757) was observed indicating that induction of autophagy by BEO is mTOR independent. Accordingly, BEO further enhanced LC3 lipidation and reduction of p62 triggered by the mTOR inhibitor, rapamycin [32], [33]; as expected, the latter caused loss of p70^S6K phosphorylation at Thr389 (Figure 7C). Under these same experimental conditions, no change in p70^S6K phosphorylation accompanied modulation of autophagic markers caused by 1 hour exposure to BEO (Figure 7C).

Autophagy modulation by BEO is associated to a trend toward a reduction of beclin-1 (Figure 6A), a protein involved in vesicle nucleation step of autophagosome formation [34]. There are evidence that certain pro-apoptotic stimuli may induce a non-canonical type of autophagosomal formation which is independent from beclin-1 [35]. To study the relevance of the upstream protein beclin-1 in LC3II and p62 modulation induced by BEO, SH-SY5Y cells were transiently transfected with beclin-1 siRNA; as control, cells were transfected with scramble non-targeting sequence. Western blot analyses confirmed the effective reduction of beclin-1 expression and the tendency of BEO to decrease the expression of beclin-1 (Figure 6B); however, silencing of beclin-1 failed to prevent the increase in LC3II and the reduction of p62 levels induced by BEO (Figure 6B) thus suggesting that BEO-induced autophagy can bypass the beclin-1-dependent nucleation of autophagosomal precursor.

To identify the active constituent responsible for the modulation of autophagic markers induced by BEO, we focussed on the monoterpene hydrocarbon d-limonene and the monoterpene ester linalyl acetate, which altogether account for about 70% of the whole oil. SH-SY5Y cells were treated with d-limonene (125–750 µM) or linalyl acetate (62.5–375 µM) at concentrations comparable with those found in 0.005–0.03% BEO. As compared to vehicle-treated cells, a significant (P<0.05), concentration-dependent increase in LC3I to LC3II conversion occurred in neuroblastoma cells exposed for 30 min to 250–750 µM limonene, whereas a lower concentration was ineffective (Figure 8A). Enhanced LC3II expression was paralleled by a graded reduction of p62 levels, which resulted statistically significant (P<0.05) in cells treated with 750 µM d-limonene (Figure 8A). Changes in LC3II expression were time-related and indicative of a rapid autophagic flux induction by d-limonene (Figure 6C). Moreover, in cells pretreated with BafA1, d-limonene further enhanced LC3II levels as compared with BafA1 alone (P<0.001; Figure 8E), thus suggesting that it induces autophagosome formation.

Exposure of SH-SY5Y cells for 30 min to linalyl acetate, given at concentrations (62.5–375 µM) found in 0.005–0.03% BEO, did not elicit LC3II accumulation (Figure 8B). Further, no changes in LC3II could be noticed on treatment with linalyl acetate (250 µM) for 15 min to 3 hours (Figure 8D). Collectively, these findings indicate that d-limonene, but not linalyl acetate, plays a key role in modulation of autophagic markers triggered by BEO; interestingly, however, linalyl acetate significantly (P<0.05) enhanced LC3 lipidation triggered by d-limonene (Figure 8F).