δ‐Tocotrienol sensitizes and re‐sensitizes ovarian cancer cells to cisplatin via induction of G1 phase cell cycle arrest and ROS/MAPK‐mediated apoptosis

Abstract Objectives Among gynaecologic malignancies, ovarian cancer (OC) represents the leading cause of death for women worldwide. Current OC treatment involves cytoreductive surgery followed by platinum‐based chemotherapy, which is associated with severe side effects and development of drug resistance. Therefore, new therapeutic strategies are urgently needed. Herein, we evaluated the anti‐tumour effects of Vitamin E‐derived δ‐tocotrienol (δ‐TT) in two human OC cell lines, IGROV‐1 and SKOV‐3 cells. Materials and Methods MTT and Trypan blue exclusion assays were used to assess δ‐TT cytotoxicity, alone or in combination with other molecules. δ‐TT effects on cell cycle, apoptosis, ROS generation and MAPK phosphorylation were investigated by flow cytometry, Western blot and immunofluorescence analyses. The synergism between δ‐TT and chemotherapy was evaluated by isobologram analysis. Results We demonstrated that δ‐TT could induce cell cycle block at G1‐S phase and mitochondrial apoptosis in OC cell lines. In particular, we found that the proapoptotic activity of δ‐TT correlated with mitochondrial ROS production and subsequent JNK and p38 activation. Finally, we observed that the compound was able to synergize with cisplatin, not only enhancing its cytotoxicity in IGROV‐1 and SKOV‐3 cells but also re‐sensitizing IGROV‐1/Pt1 cell line to its anti‐tumour effects. Conclusions δ‐TT triggers G1 phase cell cycle arrest and ROS/MAPK‐mediated apoptosis in OC cells and sensitizes them to platinum treatment, thus representing an interesting option for novel chemopreventive/therapeutic strategies for OC.

experience tumour relapse a few months after treatment and eventually develop resistance to therapy. Chemotherapy resistance, whether intrinsic or acquired, represents the major challenge in OC management. 5,6 Additionally, platinum-based chemotherapy is associated with multiple severe side effects, including nausea, vomiting, myelosuppression, neurotoxicity, nephrotoxicity, hepatotoxicity and ototoxicity. 7 For all these reasons, more effective and better-tolerated therapeutic options are urgently needed.
Tocotrienols (TTs) are Vitamin E derivatives endowed with several anti-tumour properties. 8,9 We previously demonstrated that δ-TT could induce ROS-mediated cell death in prostate cancer and melanoma. [10][11][12] Moreover, δ-TT has been recently reported to synergize with bevacizumab in a phase II trial conducted on chemotherapyrefractory OC. 13 However, the molecular mechanisms underlying δ-TT anti-tumour activity in OC cells have not been elucidated yet.
The present study is aimed at exploring the antiproliferative and proapoptotic effects of δ-TT on human OC cells, with a focus on its pro-oxidant action and on its synergistic combination with cisplatin.

| Chemicals
δ-TT (average final purity >98%) was isolated from a commercial extract of Annatto seeds (Bixa Orellana L.) (kindly provided by American River Nutrition Inc), as previously described. 14  Horseradish-peroxidase-conjugated secondary antibody and enhanced chemiluminescence reagents were from Cyanagen Srl.
Alexa Fluor 488 secondary antibody was from Thermo Fisher Scientific.

| Cell lines and cell culture
Cisplatin-sensitive (IGROV-1) and cisplatin-resistant (SKOV- 3) human OC cells were from American Type Culture Collection (ATCC), while IGROV-1/Pt1 cell line, exhibiting a stable drug-tolerant phenotype, was obtained by continuous exposure of parental cells to platinum-based chemotherapy. 15 They were all cultured in RPMI medium supplemented with 10% FBS, glutamine and antibiotics, in humidified atmosphere of 5% CO 2 /95% air at 37°C. Original stocks of cells were stored frozen in liquid nitrogen; after resuscitation, cells were kept in culture for no more than 10-12 weeks. Cells were detached through trypsin-EDTA solution and passaged once/week.

| MTT viability assay
Cells were seeded at a density of 3 × 10 4 cells/well in 24-well plates for 24 hours and then exposed to the specific compounds. The medium was then changed with MTT solution (0.5 mg/ml) in RPMI without phenol red and FBS; cells were incubated at 37°C for 30 minutes, and violet precipitate was dissolved with isopropanol. Absorbance at 550 nm was measured through an EnSpire Multimode Plate reader (PerkinElmer).

| Cell cycle analysis
Cells were plated (1.5 × 10 5 cells/dish) in 6 cm dishes. After 24 hours, cells were treated with δ-TT (15 μg/ml, 24 hours). Adherent (viable) and floating (dead) cells were harvested, washed in PBS, fixed in methanol and resuspended in Invitrogen™ FxCycle™ PI/RNase Staining Solution, according to the manufacturer's protocol. The flow cytometry analyses were performed with a Novocyte3000 instrument (ACEA Biosciences). Data were analysed with Novoexpress software.

| Annexin V/PI apoptosis assay
Cells were plated (1.5 × 10 5 cells/dish) in 6 cm dishes. After 24 hours, cells were treated with δ-TT (15 μg/ml, 24 hours). Adherent (viable) and floating (dead) cells were harvested, washed in PBS and incubated with Annexin V and PI, using the eBioscience™ Annexin V-FITC Apoptosis Detection Kit. The flow cytometry analyses were performed with a Novocyte3000 instrument (ACEA Biosciences).
Data were analysed with Novoexpress software.

| Measurement of mitochondrial membrane potential (ΔΨm)
in PBS and incubated with MitoTracker Orange CMTMRos (Thermo Fisher Scientific) 10 nmol/L for 30 minutes. The flow cytometry analyses were performed with a Novocyte3000 instrument (ACEA Biosciences). Data were analysed with Novoexpress software.

| Detection of mitochondrial cytochrome c release
Cells were seeded at 3 × 10 4 cells/well in 24-well plates on polylysine- Cytochrome c release was quantified as the percentage of cells with no cytochrome c/mitochondria co-localization relative to the total number of cells; at least 200 randomly selected cells in multiple fields were assessed.
To confirm the above data, cytochrome c release was also evaluated by Western blot analysis. In particular, the cytosolic fraction was separated by the overall cellular protein extract by using a Cell Fractionation Kit (Abcam), and cytochrome c expression was analysed in both total cellular and cytoplasmic compartments.

| Western blot analysis
Cells were seeded at 5 × 10 5 cells/dish in 10 cm dishes. After each treatment, cells were lysed in RIPA buffer; protein preparations (25 μg) were resolved on SDS-PAGE and transferred to nitrocellulose membranes. Membranes were incubated with the specific primary antibodies. Detection was done using horseradish peroxidaseconjugated secondary antibodies and enhanced chemiluminescence (Westar Eta C Ultra 2.0). α-Tubulin was utilized as a loading control.

| Isobologram analysis
Cells were treated for 24 hours using six different concentrations of cisplatin and δ-TT 15 μg/ml. Viable cells were quantitated by MTT assay as described above, and CalcuSyn software was used to generate an isobologram.
F I G U R E 1 δ-TT reduces OC cell viability and proliferation. (A), IGROV-1 and SKOV-3 cells were treated with δ-TT (5-20 μg/ml) for 24 and 48 h. Cell viability was then evaluated by MTT assay. Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by Dunnett's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle). (B), Cells were treated with δ-TT (5-20 μg/ml) for 24 and 48 h. Cell proliferation was then evaluated by Trypan blue exclusion assay. Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by Dunnett's test after one-way analysis of variance. *P < .05 vs (C), controls (vehicle) F I G U R E 2 δ-TT promotes G1 phase cell cycle arrest in OC cells. (A), IGROV-1 and SKOV-3 cells were treated with δ-TT (15 μg/ml, 1-24 h); cell cycle distribution was then evaluated by cytofluorimetric analysis after staining with Invitrogen™ FxCycle™ PI/RNase Staining Solution (according to the manufacturer's protocol). Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by t test. *P < .05 vs C, controls (vehicle). **P < .01 vs C, controls (vehicle). ***P < .001 vs C, controls (vehicle). (B), After δ-TT treatment (15 μg/ml, 1-24 h), Western blot analysis was performed to investigate the expression levels of cyclin D1, cyclin D3, CDK4, CDK6, p21 and p27. Tubulin expression was evaluated as a loading control. One representative of three experiments performed is shown. Data represent mean values ± SEM and were analysed by Dunnett's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle); **P < .01 vs C, controls (vehicle); ***P < .001 vs C, controls (vehicle)

| Statistical analysis
Statistical analysis was performed with a statistic package (GraphPad Prism5, GraphPad Software). Data are represented as the mean ± SEM of three-four independent experiments. Differences between groups were assessed by t test or one-way analysis of variance (ANOVA) followed by Dunnett's or Bonferroni's test. A P value <0.05 was considered statistically significant.

| δ-TT promotes G1 phase cell cycle arrest in OC cells
To determine how δ-TT treatment could affect OC cell growth, cell cycle distribution was assessed by flow cytometry. Our results show that the compound (15 μg/ml, 24 hours) significantly increased the number of both IGROV-1 and SKOV-3 cells in G1 phase as compared to control group (Figure 2A). Supporting these results, the expression levels of cell cycle regulation proteins involved in G1/S transition were evaluated by Western blot. As evidenced in Figure 2B, the protein levels of cell cycle promoters cyclin D1, cyclin D3, CDK4 and CDK6 were time-dependently reduced after δ-TT treatment, while the expression of cell cycle inhibitors p21 and p27 was parallelly enhanced.

| δ-TT triggers mitochondrial apoptosis in OC cells
Since inhibition of cell proliferation is strictly associated with apoptosis induction, we further investigated whether δ-TT could trigger apoptotic OC cell death. Our data from Annexin V/PI staining indicate that the compound (15 μg/ml, 24 hours) induced apoptosis in both IGROV-1 and SKOV-3 cells, with apoptotic rates being around 22% in both cell lines ( Figure 3A). Additionally, after 12 hours of treatment,

| ROS generation is involved in the apoptosis induced by δ-TT in OC cells
We have previously reported that δ-TT is able to generate massive oxidative stress in prostate cancer and melanoma cells. 10-12 Therefore, we measured ROS formation in both IGROV-1 and SKOV-3 cells treated with the compound. As described in Figure 5A

| MAPK activation is implicated in the ROSrelated apoptotic OC cell death caused by δ-TT
It is now well established that oxidative stress can mediate the activation of growth-suppressing MAPK pathways able to promote Bax phosphorylation and translocation to mitochondria and to parallelly attenuate Bcl-2 expression, eventually culminating in the initiation of the caspase cascade. [22][23][24][25][26][27] Thus, we analysed

F I G U R E 4 δ-TT activates the intrinsic caspase cascade in OC cells. (A)
, After δ-TT treatment (15 μg/ml, 1-24 h), Western blot analysis was performed to investigate the expression levels of Bax, Bcl-2, cleaved caspase 9, cleaved caspase 3 and PARP. Tubulin expression was evaluated as a loading control. One representative of three experiments performed is shown. Data represent mean values ± SEM and were analysed by Dunnett's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle); **P < .01 vs C, controls (vehicle); ***P < .001 vs C, controls (vehicle). (B), Cells were pretreated with the pan-caspase inhibitor Z-VAD-FMK (50 μM, 4 h) and then with δ-TT (15 μg/ml, 24 h). Cell viability was assessed by MTT assay. Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by Bonferroni's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle). #P < .05 vs δ-TT-treated cells successfully rescued OC cell viability ( Figure 6B) and prevented caspase 3 and PARP activation ( Figure 6C). In addition, pretreatment of tumour cells with NAC (4 mmol/L, 2 hours) markedly blocked δ-TT-related MAPK activation (15 μg/ml, 24 hours), suggesting that JNK and p38 upregulation is critically implicated in the ROS-mediated apoptotic OC cell death caused by the compound (Figure 7).

| δ-TT synergizes with cisplatin in reducing OC cell viability
To investigate whether δ-TT could enhance cisplatin cytotoxicity in OC cell lines, we treated both IGROV-1 (platinum-sensitive) and To further validate the above results, we tested δ-TT efficacy on chemotherapy-adapted IGROV-1/Pt1 cells, in the absence or presence of cisplatin. As shown in Figure 9A

| D ISCUSS I ON
It is now widely accepted that TTs possess potent anti-tumour properties. 8,9 Although most of the studies so far reported were performed with γ-TT, the δ isomer has recently shown clinical promise in the treatment of recurrent OC when given in combination with bevacizumab. 13 However, its mechanism of action is still poorly understood.
Here, we dissected the molecular mechanisms underlying the anti-tumour activity of δ-TT in two different OC cell lines, IGROV-1 and SKOV-3 cells, known to be cisplatin-sensitive and -resistant,

respectively.
We found that δ-TT can exert significant anti-tumour effects on OC cells, by decreasing both cell viability and proliferation.
Specifically, we demonstrated that the compound can trigger To get further insights into the molecular targets of δ-TT, we focused our attention on the analysis of intracellular oxidative stress.
Indeed, several natural compounds have been shown to induce ROSrelated cell death in different types of tumour, including OC. [28][29][30][31] Regarding TTs, the γ isoform has been reported to trigger mitochondrial dysfunction-and oxidative imbalance-associated apoptosis in gastric adenocarcinoma cells, as well as to sensitize colorectal cancer cells to TRAIL proapoptotic activity via ROS overproduction. 32,33 Similarly, δ-TT can cause ROS-mediated cytotoxicity in various models of breast and prostate cancer, 11,34 while α-tocopheryl succinate, a redox-silent Vitamin E analogue, has been demonstrated to stimulate cytochrome c release in neuroblastoma through a severe alteration of redox homeostasis. 35 In OC cells, we observed that δ-TT promotes mitochondrial ROS generation and that NAC can successfully rescue cell viability by preventing caspase 3 and PARP cleavage, thus highlighting the involvement of oxidative stress in the apoptosis induced by the compound.
Since ROS generation is often linked to MAPK activation, 22 we analysed p-JNK and p-p38 expression in δ-TT-treated OC cells. A significant increase in the levels of both these proteins was found, and it was shown to be significantly counteracted by NAC-mediated ROS scavenging. Interestingly, by using SP600125 and SB203580, upregulation of both p38 and JNK was demonstrated to correlate with δ-TT-related cytotoxicity and apoptosis. Notably, numerous phytochemicals have been shown to specifically target the MAPK signalling in a variety of cancers. 36,37 In particular, TTs are known to trigger JNK-and p38-dependent cell death in several types of tumour, such as melanoma, lymphoma and breast and prostate cancer. 10,11,[38][39][40] In this setting, our data not only support previous findings about the MAPK-targeting ability of TTs but also indicate that the induction of a ROS/MAPK cascade is deeply implicated in the anti-OC effects of δ-TT.
Platinum-based chemotherapy represents the standard treatment for OC. 4 However, it is generally associated with high toxicity and relapse rates. [5][6][7] In order to overcome these issues, we conducted a combination study with δ-TT, evidencing F I G U R E 5 ROS generation is involved in the apoptosis induced by δ-TT in OC cells. (A), IGROV-1 and SKOV-3 cells were treated with δ-TT (15 μg/ml, 12 h); cellular ROS production was then evaluated by cytofluorimetric analysis after staining with 2',7'-dichlorofluorescin diacetate (DCFDA, 10 µmol/L, 30 min). Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by t test. ***P < .001 vs C, controls (vehicle). (B), Cells were treated with δ-TT (15 μg/ml, 12 h); mitochondrial ROS production was then evaluated by cytofluorimetric analysis after staining with MitoSOX Red (5 μmol/L, 10 min). Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by t test. ***P < .001 vs C, controls (vehicle). (C), Cells were pretreated with the ROS scavenger N-acetyl-L-cysteine (NAC, 4 mmol/L, 2 h) and then with δ-TT (15 μg/ml, 24 h). Cell viability was assessed by MTT assay. Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by Bonferroni's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle). #P < .05 vs δ-TT-treated cells. (D), Cells were pretreated with NAC (4 mmol/L, 2 h) and then with δ-TT (15 μg/ml, 24 h). Caspase 3 and PARP cleavage was evaluated by Western blot analysis. Tubulin expression was evaluated as a loading control. One representative of three experiments performed is shown. Data represent mean values ± SEM and were analysed by Bonferroni's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle); ***P < .001 vs C, controls (vehicle). #P < .05 vs δ-TTtreated cells; ##P < .01 vs C, controls (vehicle) a strong synergistic in vitro interaction between the two anticancer agents not only in platinum-sensitive (IGROV-1) OC cells but also in two models of drug resistance (wild type SKOV-3 and chemotherapy-adapted IGROV-1/Pt1 cell lines). Remarkably, several natural compounds are able to sensitize cancer cells to cisplatin 41 ; among them, berberine, β-elemene, genistein, F I G U R E 6 MAPK activation is implicated in the ROS-related apoptotic OC cell death caused by δ-TT. (A), After δ-TT treatment (15 μg/ml, 1-24 h), Western blot analysis was performed to investigate the expression levels of p-JNK and p-p38. Tubulin expression was evaluated as a loading control. One representative of three experiments performed is shown. Data represent mean values ± SEM and were analysed by Dunnett's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle); **P < .01 vs C, controls (vehicle); ***P < .001 vs C, controls (vehicle). (B), Cells were pretreated with JNK (SP600125, 20 μmol/L, 2 h) or p38 (SB203580, 20 μmol/L, 2 h) inhibitors and then with δ-TT (15 μg/ml, 24 h). Cell viability was assessed by MTT assay. Each experiment was repeated three times. Data represent mean values ± SEM and were analysed by Bonferroni's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle). #P < .05 vs δ-TT-treated cells.
(C), Cells were pretreated with SP600125 (20 μmol/L, 2 h) or SB203580 (20 μmol/L, 2 h) and then with δ-TT (15 μg/ml, 24 h). Caspase 3 and PARP cleavage was evaluated by Western blot analysis. Tubulin expression was evaluated as a loading control. One representative of three experiments performed is shown. Data represent mean values ± SEM and were analysed by Bonferroni's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle); **P < .01 vs C, controls (vehicle); ***P < .001 vs C, controls (vehicle). #P < .05 vs δ-TT-treated cells; ##P < .01 vs C, controls (vehicle) hirsutenone, morin, saikosaponin-d, cardamonin, theaflavin-3,3'-digallate and withaferin A have been found to specifically enhance the chemosensitivity of OC cell lines. [42][43][44][45][46][47][48][49][50] Furthermore, δ-TT itself has been reported to synergize with a wide range of anti-tumour drugs, including common anti-OC chemotherapeutics (i.e. taxanes) 8,9 ; in particular, its ability to overcome chemoresistance in tumours seems to be associated with the induction of multiple growth-suppressive and pro-death pathways, as well as with a specific targeting of the mechanisms responsible for treatment escape, such as growth factor receptor hyperexpression, ATP-binding cassette (ABC) transporter activation and cancer stem cell (CSC) propagation. 8,9,51 To our knowledge, this is the first study highlighting the synergism between δ-TT and cisplatin in OC.  Figure 6A were converted to Fraction Affected (FA) and plotted against Combination Index (CI). Straight line on the graph designates a CI equal to 1. Combination Index interpretation was as follows: CI value of 1 indicates additivity; CI<1 indicates synergism; and CI >1 indicates antagonism F I G U R E 7 The ROS/MAPK axis mediates the apoptotic OC cell death triggered by δ-TT. Cells were pretreated with NAC (4 mmol/L, 2 h) and then with δ-TT (15 μg/ml, 24 h). JNK and p38 phosphorylation was evaluated by Western blot analysis. Tubulin expression was evaluated as a loading control. One representative of three experiments performed is shown. Data represent mean values ± SEM and were analysed by Bonferroni's test after one-way analysis of variance. *P < .05 vs C, controls (vehicle); ***P < .001 vs C, controls (vehicle). #P < .05 vs δ-TTtreated cells; ##P < .01 vs C, controls (vehicle) In conclusion, these results demonstrate that δ-TT exerts a potent anti-tumour activity in OC cell lines by triggering both G1 phase cell cycle arrest and mitochondrial apoptosis. In particular, we showed that it can induce MAPK/ROS-mediated OC cell death and enhance cisplatin anti-OC efficacy, providing a deeper understanding of its anti-tumour properties.

ACK N OWLED G EM ENTS
This research was funded by MIUR Progetto di Eccellenza (Department of Pharmacological and Biomolecular Sciences, Università degli Studi di Milano). FF was supported by an AIRC fellowship for Italy.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
FF designed the study. FF, MM, MR, VZ and NZ collected and analysed the data. FF, MM, MR, VZ, NZ and PL prepared the manuscript.
PL was involved in funding acquisition.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.